Scoping systematic review of treatments for eczema

When is eczema ‘atopic’?

Although the word ‘atopic’ is often used when describing eczema, up to two-thirds of people with eczema do not have measurable levels of circulating allergen-specific immunoglobulin E (IgE) antibodies, which are a necessary criterion to denote a person as ‘atopic’. 4 The relationship between atopy and eczema is unclear. Atopy is more common in more severe disease in hospital populations than in community populations and atopy is also more common in affluent than in non-affluent countries. It has been suggested that the relationship between eczema and atopy might be apparent only because of shared causes for both conditions. 5

The nomenclature for allergy6 has been revised and is now based on the mechanisms that initiate and mediate allergic reactions. The term ‘atopic eczema’ should be used only when IgE sensitisation has been confirmed by allergen-specific IgE antibodies in the blood or by a positive skin-prick test to common allergens such as house dust mite. 7 As this review includes cases of eczema based purely on a clinical diagnosis, the term ‘eczema’ is used throughout to refer to what is more often described as ‘atopic dermatitis’ or ‘atopic eczema’.

How is eczema defined in clinical studies?

Quite often no definition of eczema is given in clinical studies, which leaves the reader guessing as to what sort of people were studied. Eczema is a difficult disease to define as the clinical features are highly variable. This variability can relate to the skin rash morphology (e.g. it can be dry and thickened or weeping and eroded), location (e.g. it commonly affects the cheeks in infants and skin creases in older children) and time (it can be bright red one day and apparently gone a couple of days later) (Figure 1). There is no specific diagnostic test for all people with typical eczema to serve as a reference standard; diagnosis is therefore a clinical one.

FIGURE 1.

Facial eczema. Severe eczema with signs of weeping, crusting and scratching. Image reproduced with permission.

Until the late 1970s at least 12 synonyms for eczema-like conditions were in common usage in the dermatological literature and it is not certain whether physicians were all referring to the same disease when using these terms. A major milestone in describing the main clinical features of eczema was the Hanifin and Rajka diagnostic criteria of 1980. 8 These consensus criteria are frequently cited in clinical trial articles, thereby providing a degree of confidence that researchers are referring to a similar disease. Scientifically developed refinements of the Hanifin and Rajka diagnostic criteria have been developed by a UK Working Party9 and these criteria have been validated widely10 and used throughout the world and in National Institute for Health and Care Excellence (NICE) guidance. To qualify as a case of eczema, the person must have an itchy skin condition plus three or more of the following:

• past involvement of the skin creases, such as the bends of elbows or behind the knees

• personal or immediate family history of asthma or hay fever

• tendency towards a generally dry skin

• onset under the age of 2 years (except when aged ≤ 4 years)

• visible flexural dermatitis as defined by a photographic protocol.

Binary or continuous disease?

It is unclear whether eczema is an ‘entity’ in itself or whether it is part of a continuum when considered at a population level. 11 Although it may be appropriate to ask the question, ‘How much atopic eczema does he/she have?’ as opposed to ‘Does he/she have atopic eczema – yes or no?’, most population and clinical studies require a categorical cut-off.

Is it all one disease?

It is quite possible that there are distinct subsets of eczema, for example those with filaggrin mutations,12,13 which may lead to more persistent and severe disease. When assessing which treatments are going to be effective for eczema it is still sensible to consider the clinical disease as one condition. As more evidence about the different phenotypes of eczema is collected (e.g. those who are definitely atopic with raised circulating IgE levels to allergens or those with severe disease and associated asthma),11 sensitivity analyses can be carried out to evaluate whether such subdivisions are useful for predicting treatment response.

Age

Eczema is more common in childhood, particularly in the first 5 years of life. 17 The prognosis of eczema shows a mixed picture, with one study18 showing a 90% clearance rate for children within 10 years, and other studies finding the rate to be around 60% by age 16 years. These figures may still not reflect the true level of eczema clearance, as many people relapse at some point in their life. Eczema prognosis may differ between the community setting (where the majority of cases are mild) and the hospital setting (where cases tend to be more severe). One study found that 10% of hospital eczema patients still suffer as adults. Adults constitute around one-third of all those with eczema in a general practice community. 17 Adults also tend to represent a more persistent and severe subset of cases.

Severity distribution

Most cases of childhood eczema are mild. A study from 1998 by Emerson and colleagues19 found that 84% of 1760 children aged 1–5 years from four urban and semi-urban general practices in Nottingham were mild cases (as defined globally by the examining physician), with 14% of cases in the moderate category and 2% in the severe category. However, there has been little research regarding the severity distribution of eczema, even though this would provide useful information for allocating health resources (as the costs of managing severe eczema are disproportionally large in comparison to the costs of managing mild eczema).

Economic costs

In financial terms, the cost of atopic eczema is potentially very large. The costs associated with the management of eczema are largely indirect, such as workdays lost by parents and travel costs for heath-care appointments, with much of this expense being met by the family of the person with eczema. There have been a number of studies in the UK, the Netherlands and the USA that have shown the costs to vary between country and according to severity. 16,24,25

Genetics

There is strong evidence to suggest that genetic factors are important in the predisposition to eczema. In addition to family studies, twin studies have shown a much higher concordance for monozygotic (identical) (85%) than for dizygotic (non-identical) (21%) twins. 26 Mutations that occur in the filaggrin (filament-aggregating-protein) gene give rise to a faulty filaggrin protein that results in dry skin and an increased risk of developing eczema, as well as more severe and persistent disease and associated asthma. 13

Environment

Although genetic factors are probably a very important factor for disease predisposition, there are numerous general and specific clues that point strongly to the crucial role of the environment in disease expression. 27 The recent, large increases in the prevalence of atopic eczema are difficult to explain purely in genetic terms. 15 It has been shown that atopic eczema is more common in wealthier families. 28,29 It is unclear whether this positive social class gradient is a reflection of indoor allergen exposures or whether it reflects a whole constellation of other factors associated with ‘development’. The observation that many cases of atopic eczema improve spontaneously around puberty is also difficult to explain in genetic terms alone. Recent large birth cohort studies14 have not found any protective effect of elder siblings, conflicting with earlier studies. 30 The original observation that increasing family size was associated with decreased eczema prevalence led to the ‘hygiene hypothesis’, which proposed that children in larger families were protected from expressing atopy because of frequent exposure to infections. 31 The role of exposure to microbes and allergens in the environment is still being debated.

Migrant studies also point strongly to the role of environmental factors in eczema development. It has been shown that 14.9% of black Caribbean children living in London develop atopic eczema (according to the UK diagnostic criteria) compared with only 5.6% of similar children living in Kingston, Jamaica. 25 Other migrant studies reviewed elsewhere have consistently recorded large differences in ethnic groups migrating from warmer climates to more prosperous cooler countries. 32

Further work has suggested that the tendency to develop eczema could be programmed at birth and could be related to factors such as maternal tobacco exposure. 33 Specific risk factors for eczema expression in the environment include furry pets; however, there is evidence that these can also have protective effects. 34 Allergic factors such as exposure to the house dust mite could be important, but non-allergic factors such as exposure to nylon clothing, dust or shampoo may also be important. 35

National Institute for Health and Care Excellence guidance

The NICE guidance41 covers diagnosis, assessment and management of eczema and information on eczema. The evidence was systematically reviewed and opinions from clinicians, researchers and consumers were used to develop the clinical guidelines.

The guidelines recommend a holistic approach at each consultation, taking into account the severity of the atopic eczema and the child’s quality of life, including everyday activities, sleep, and psychosocial well-being.

The guidelines give a clear list of the first-line (emollients and topical corticosteroids), second-line (topical calcineurin inhibitors) and third-line (systematic treatments and phototherapy) treatments based on need and severity.

The mainstay of treatment for all severities of eczema is emollients. The guidelines emphasise the importance of using emollients even when the skin is clear. It also recommends allowing the patient a choice of emollients for washing, bathing and moisturising.

Topical corticosteroids should be tailored to the severity and area of eczema and should be used once or twice daily for an appropriate length of time. Topical steroids are an important tool for eczema treatment, although care regarding the duration of treatment, site and age of the person treated was emphasised. Potent steroids are not recommended for use in children aged < 12 months without specialist dermatologist supervision.

Oral antihistamines should not be used routinely in the management of atopic eczema in children. However, if sleep disturbance (for the child or their parents/carers) becomes significant during an exacerbation of eczema (flare), health-care professionals should offer a 7- to 14-day trial of an age-appropriate sedating antihistamine (for children aged ≥ 6 months). This treatment can be repeated during subsequent flares if successful.

Occlusive medicated dressings and dry bandages should not be used to treat infected atopic eczema in children.

Phototherapy or systemic treatments should be initiated in children with atopic eczema only after assessment and documentation of the severity of atopic eczema and quality of life.

Health-care professionals should spend time educating children with atopic eczema and their parents or carers about atopic eczema and its treatment.

Scottish Intercollegiate Guidelines Network guidance

The SIGN guidance42 covers diagnosis, referral, management and patient education. The evidence for the SIGN guidelines came from a systematic review of the literature from 2004 to 2009, involving five bibliographic databases. The analysis of the evidence was conducted by the SIGN committee using standard forms, with any additional evidence submitted by members of the committee, including key reviews outside the search period. A review of studies on the issues facing patients was conducted and presented to the committee.

The guidelines give the following key recommendations:

• Emollients should be given on a continuous basis, even when topical corticosteroids are being used.

• Topical corticosteroids should be applied once a day, and twice-weekly treatment should be considered when those with moderate to severe eczema have frequent relapses.

• Topical calcineurin inhibitors are recommended for use in children aged ≥ 2 years if topical corticosteroids are not controlling the eczema or the level of use of topical corticosteroid could lead to adverse effects.

• Topical antibiotics are not recommended for the treatment of non-infected eczema.

Other recommendations included:

• As a precaution topical calcineurin inhibitors should not be applied to skin that appears actively infected.

• Patients with non-infected moderate to severe eczema should be advised to cover affected areas with dry wrap dressings to provide a physical barrier to scratching and improve retention of emollient.

• Swabs of potential Staphylococcus aureus carriage sites (of both the patient and family members) should be considered in patients with recurrent infection.

• In patients with atypical features, or when there is concern about possible streptococcal infection, skin swabs of affected areas should be considered.

• Short-term use of sedating antihistamines at night-time should be considered in patients with atopic eczema when there is debilitating sleep disturbance.

• When an irritant effect is suspected, patients should be advised to avoid biological washing powders, fabric conditioners and fragranced products such as soaps and shower gels.

• Dietary exclusion is not recommended for management of atopic eczema in patients without confirmed food allergy.

• When there is suspicion of food allergy in infants or children with atopic eczema, general practitioners (GPs) should refer to an allergist or paediatrician with a special interest in allergy.

• Exercise caution when using herbal medicines and be wary of any herbal product that is not labelled in English or does not come with information about safe usage.

The guidelines also detail what information can reasonably be expected by a patient during diagnosis and treatment.

How is care organised in the UK?

Most children with atopic eczema in the UK are managed by a primary care team, with around 4–10% of children with atopic eczema referred to a dermatologist for further advice. 19

The quality of service provided by secondary care has been audited by the British Association of Dermatologists. Although most departments provided a high-quality service, some aspects of care, such as the administration of simple standardised record forms, could be improved. 43 The audit found that the outcomes may not be as good as some doctors believe, with the improvements in quality of life and numbers of adults returning to work not meeting the working standards. 43

Adherence (or, more correctly, concordance) seems to be a major cause of apparent treatment failure. A survey conducted in Nottingham found that most parents worry that topical steroids cause adverse effects, although many were not able to distinguish between weak and strong ones. 44

The National Eczema Society [see www.eczema.org (accessed 12 October 2015)] is the UK’s largest self-help organisation for people with eczema. It has a well-organised information service and national network of activities geared to help those with eczema and their families. Sources of alternative care abound in the community, especially with the increased access to the internet, ranging from the highly professional to elaborate, expensive diagnostic and therapeutic measures of dubious value.

A question- or data-driven review?

The very broad-ranging scoping nature of this review implies that it cannot be hypothesis driven. Trying to answer similar questions for each of the 92 or so interventions used for the treatment of atopic eczema would be impossible in one short report.

This updated review is still unashamedly data driven. It is a review that aims to map out what has been done in terms of RCTs in atopic eczema to date and to reflect and comment on the coverage of already-researched areas in relation to questions that are commonly asked by physicians and their patients.

The authors are aware that there is a danger that a data-driven review can serve to amplify and perpetuate current trends in evaluating minor differences between a profusion of similar pharmacological products. The authors have mitigated against this inevitable hazard by drawing attention to gaps that have not been addressed when summarising the reported studies and also by including a section on unanswered questions in the discussion and conclusions section of this report, based on a recently completed James Lind Alliance Priority Setting Partnership. 57

Electronic searching

To retrieve all RCTs on atopic eczema treatments in accordance with the inclusion criteria, a systematic and mainly electronic search was carried out. The Cochrane Handbook for Systematic Reviews of Interventions58 was used as a template.

The following electronic databases were searched from the end of 2000 up to and including August 2013:

• MEDLINE

• EMBASE with its higher yield of non-English reports

• Cochrane Central Register of Controlled Trials

• The Cochrane Skin Group Specialised Trials Register

• Latin American and Caribbean Health Sciences (LILACS) database

• Allied and Complementary Medicine Database (AMED)

• Cumulative Index to Nursing and Allied Health Literature (CINAHL).

Disease terms for atopic eczema [as a text word and medical subject heading (MeSH) term] are shown in Appendix 1. Possible trials were identified from each of the six databases as follows:

• MEDLINE (OvidSP) – the Cochrane Collaboration ‘highly sensitive electronic search string’ for RCTs was used (see Appendix 1). Publications from 2000 to 31 August 2013 were searched and yielded > 4000 references using the disease search terms in Appendix 1.

• EMBASE (OvidSP) – because of the different format of this database an alternative search strategy was employed (see Appendix 1). Publications from 2000 to 31 August 2013 were searched and yielded > 2000 references using the same eczema terms as for MEDLINE.

• Cochrane Central Register of Controlled Trials – The Cochrane Library 2013 was searched for controlled trials within the Cochrane Central Register of Controlled Trials section by exploding the disease-specific search terms separated by the Boolean ‘AND’ with the advanced search option. These included clinical controlled trials (quasi randomisation) and RCTs (randomisation).

• Cochrane Skin Group Specialised Register – this was searched with the disease-specific terms and the kind help of the Cochrane Skin Group Trials Search Coordinator.

• www.controlled-trials.com was searched for completed and ongoing RCTs using the terms ‘atopic dermatitis’, ‘atopic eczema’ and ‘eczema’. This meta-register of trials contains the entries from the following databases:

  • ISRCTN register (international) – copy of the ISRCTN register

  • Action Medical Research (UK) – subset of the ISRCTN register

  • Medical Research Council (UK) – subset of the ISRCTN register

  • US National Institutes of Health ClinicalTrials.gov register (international) – subset of randomised trial records.

Handsearching

As there are > 200 specialist dermatology journals and none specific to atopic eczema, separate handsearching was not carried out for this report. Some trials published in journals not listed in the bibliographic databases searched, or published within the body of a letter to the editor, might therefore have been missed. However, specialist dermatology journals are handsearched by the Cochrane Skin Group and these are included in the Cochrane Skin Group Specialised Register of trials, which was searched. This includes the results of handsearching the following dermatology journals over varying dates: Cutis, Acta Dermato-Venereologica Supplementum, Archives of Dermatology, British Journal of Dermatology, Clinical and Experimental Dermatology, International Journal of Dermatology, Journal of Investigative Dermatology and Journal of the American Academy of Dermatology.

In addition, conference proceedings of previous symposia such as the Atopic Dermatitis Symposia and all meeting abstracts for the annual meetings of the Society for Investigative Dermatology, European Academy of Dermatology and British Association of Dermatologists were handsearched by one of the authors and the results made available to the Cochrane Skin Group Specialised Register. Furthermore, one of the authors has been prospectively handsearching five dermatology journals (Clinical Experimental Dermatology, British Journal of Dermatology, Journal of the American Academy of Dermatology, Journal of Investigative Dermatology and Paediatric Dermatology) since January 1998 for atopic eczema treatment trials.

Other trial sources

In addition to checking citations in retrieved RCTs, additional trials were sought by personal contact with atopic eczema researchers and by writing to 15 pharmaceutical companies with a product (commercially available or in development) related to atopic eczema treatment.

Filtering

The initial search yielded 7168 references. A manual filtering process was conducted to assess whether each reference fitted the preliminary labels of ‘trial’ and ‘eczema’.

Not all references had abstracts and therefore studies were included if their title related to an eczema trial, to avoid premature exclusion. In cases of uncertainty, the full paper was requested and scrutinised by one of the authors and discussed with a second senior author as necessary. Papers excluded were categorised by reason for exclusion by one reviewer; this was checked by a second reviewer in cases of uncertainty. When a trial fitted the inclusion criteria but was found to be published only as an abstract, the trial was excluded from the review. All papers were catalogued on a specialised referencing database (EndNote X7; Thomson Reuters, CA, USA).

Non-English-language studies

Non-English-language studies were screened by international colleagues (see Acknowledgements) and full data extraction performed for those meeting the inclusion criteria.

Identifying treatments with no randomised controlled trials

A data-driven review can identify only treatments for which some evidence exists and therefore there may be other treatments that are currently used throughout the world for atopic eczema but which are not necessarily supported by RCTs.

Study quality

The methodological quality of each study was assessed using the Cochrane Collaboration risk-of-bias assessment tool,58 with potential sources of bias evaluated:

• method of generating the randomisation sequence

• concealment of the allocation sequence

• blinding of participants, study personnel and outcome assessments

• other issues including incomplete outcome data and selective outcome reporting and the extent to which the primary analysis included all participants initially randomised (i.e. an intention-to-treat analysis).

The method of generation of the randomisation sequence, allocation concealment and blinding have been consistently shown to predict bias in effect estimates67 and so these have been tabulated throughout the review to assist readers in making comparisons between studies. When assessing the risk of bias for blinding, a trial was assessed as low risk if the primary outcome assessor was reported as blinded even if other parties, such as the participants, were not. When the trials are summarised, standard statements about the collective risk of bias have been used (Table 3). This allows readers to see which aspects of the study reporting were deficient for each treatment. Because of the sheer size of this scoping review, the report authors were not blinded to the identity of the RCT authors when assessing quality or carrying out data extraction.

Quantitative data synthesis

Pooling of the data did not make sense clinically for any of the interventions and so no pooling of quantitative data was undertaken.

Methods of presenting results

Summarising the evidence for treatments and harms from 287 RCTs covering at least 92 different interventions in a way that would be helpful to health-care commissioners, providers and physicians is challenging. There is always a conflict in such a situation of providing too much information, resulting in loss of the general picture, or of omitting important details in some specific areas.

Readers are encouraged to read the original studies when doubt occurs over the reported data or the report authors’ conclusions. One of the report authors has led the development of a database of eczema RCTs [the Global Resource for EczemA Trials (GREAT) database68] in which information about and links to the included and excluded studies in this review can be found.

In many of the studies it would have been impractical to document every outcome and therefore the report authors have highlighted:

• patient-rated global improvement or itch or sleep loss

• global severity score based on several skin signs, or individual skin sign scores.

In some studies evaluating multiple outcomes, statistically significant post hoc tests were highlighted in the paper’s conclusions or abstract. The report authors have mitigated against this bias by stating whether the results were from a post hoc test.

If pre-existing systematic reviews were identified for any of the interventions, these were highlighted at the beginning of the relevant sections.

Separating trial data from authors’ opinions

The report authors have been careful to make a clear distinction between the facts abstracted from individual studies and the respective authors’ interpretations of what those results or lack of results mean. Thus, actual data on efficacy and possible harms have been clearly separated from the authors’ ‘overall implications for research and practice’ sections. In the comments sections of the risk-of-bias tables, the report authors have also commented on possible sources of bias.

Existing systematic reviews

Since 2000, 15 systematic reviews79–93 and three guidelines41,42,94 have been published on the use of topical corticosteroids and/or topical calcineurin inhibitors. Two reviews have assessed occlusive therapy. 95,96

Topical corticosteroids compared with placebo

Topical corticosteroids compared with active treatments (except topical immunomodulators)

Once-daily compared with twice-daily application

There were three trials published before 200055 (see Appendix 3). No new studies have been published since 2000.

Overall implications for research and practice

The publication of new trials comparing topical corticosteroids has dramatically slowed in the last 10 years, possibly because of emerging competitors such as the topical calcineurin inhibitors. The few trials reported, mostly with manufacturer involvement, still do not seem to directly address the knowledge gaps about the best choices of topical corticosteroids. There is a small amount of evidence that combining a topical corticosteroid with a penetration-enhancing chemical shows an increased benefit,118 but the trial was not clear enough about the methodology used to assess the risk of bias and so this must be treated with caution until stronger, long-term treatment evidence becomes available. There is some very weak evidence that pulsed treatment is more effective than continuous treatment. 122

As all of the trials comparing an active treatment that is not another topical corticosteroid or topical immunomodulator use the topical corticosteroid as the ‘standard treatment comparator’, they have been discussed in the chapters of this report that cover the treatments they have been compared against. Three of the trials123,124,126 are discussed in Chapter 5; the other is discussed in Chapter 10. 125

Tacrolimus compared with placebo

Tacrolimus compared with active treatments

Pimecrolimus compared with placebo

Pimecrolimus compared with active treatments

Tacrolimus compared with pimecrolimus

Studies

One new trial by Meurer and colleagues181 compared a combination of fluticasone propionate (with hydrocortisone acetate for the face neck and hands) and pimecrolimus with fluticasone propionate with hydrocortisone acetate and vehicle cream. This multinational trial randomised 376 children aged between 2 and 17 years to treatments applied twice a day for 4 weeks. Those children who were clear or almost clear after 4 weeks of treatment were observed for a further 12 weeks to assess time to relapse.

Another new within-person multicentre trial by Hebert and colleagues182 compared treatment with tacrolimus (0.1%) ointment used concurrently with desoximetasone (0.25%) ointment with tacrolimus ointment used concurrently with the vehicle of desoximetasone, twice daily until clear or for 21 days. Eighty-two adults with eczema in the target lesions of ≥ 8/15 for total symptom score had their left and right sides randomised to treatment. There was no follow-up of participants beyond 21 days.

Assessment of risk of bias for the new studies

Table 9 provides the risk-of-bias assessment for the new studies.

Benefits

The trial by Meurer and colleagues,181 which was primarily concerned with safety, combined pimecrolimus and fluticasone and compared this treatment with fluticasone only. The trial found no significant differences between the treatments except for an increased time to relapse in what appeared to be a post hoc subgroup analysis of only those participants who were ‘clear’ of eczema as assessed by the IGA at the end of treatment.

In the trial by Hebert and colleagues182 the reduction in participant-assessed pruritus over 21 days was significantly better in the combined treatment group than with tacrolimus alone (p = 0.006); however, the scores were not reported and only 69 out of 82 participants were included in the pruritus assessment because of missed visits. The percentage of participants with a score of 0 (no pruritus) at baseline was 58% for combination treatment and 61% for tacrolimus treatment; this compared with 84% for combination treatment and 71% for tacrolimus treatment after 3 days. This was reported as statistically significant in favour of combination treatment (p = 0.04). A difference of 0.3 (95% Cl 0.1 to 0.5) between the groups in the physician’s global assessment score at day 21 was reported, which was significantly significant in favour of the combination treatment. For the total symptom score for the target lesions, the difference between the reduction in scores after 21 days was 0.8 (95% Cl 0.4 to 1.2) in favour of combination treatment (p = 0.0002).

Harms

The trial by Meurer and colleagues181 found a relatively high rate of infections and infestations (25–30% of participants) in both treatment groups. Bronchitis was far more common in the pimecrolimus and topical corticosteroid group (2.7% vs. 0%). The proportion of participants experiencing adverse events suspected to be related to the study drug was slightly higher on combination treatment (6.3%) than with the topical corticosteroid-only treatment (4.4%). For adrenal suppression, 54 participants (28.4%) on combination treatment and 44 participants (24.0%) on the topical corticosteroid-only treatment were suspected cases, but the majority of patients were confirmed not to be suppressed.

Hebert and colleagues182 reported that five participants withdrew from the trial because of ‘non-compliance’, ‘protocol violation’ or an ‘adverse event’, but the number of participants effected by adverse events and the nature of the events was not reported.

Overall implications for research and practice

There is not enough convincing evidence that combining topical corticosteroids and topical calcineurin inhibitors confers any short-term beneficial effects compared with topical corticosteroids alone. One large trial in children181 and a smaller trial in adults182 based their positive conclusions on the results of post hoc subgroup analysis or complete case analysis that excluded many participants.

Studies

No trials involving wet wrap bandages for the treatment of eczema were reported before 2000. Five new trials involving the use of wet wrap bandages were reported after 2000. 183–187

A left–right within-person comparison trial by Foelster-Holst and colleagues183 compared prednicarbate ointment and tubular bandages soaked in warm water and then covered with dry dressings with prednicarbate ointment alone for 48–72 hours. All of the 24 adults and children had experienced an acute episode of atopic eczema and also used an emollient.

A single-centre UK trial by Hindley and colleagues184 compared conventional treatment [hydrocortisone (1%) or stronger steroids] under wet wraps with hydrocortisone (1%) and emollients. Fifty children aged 4–27 months (eligibility was from 3 months to 5 years) with eczema that scored ≥ 15 using SCORAD scores were randomised to treatment. The children were assessed over a 4-week period but it was not clear whether they were using the study treatment for the whole 4 weeks.

A four-arm trial conducted in Hong Kong, China, by Pei and colleagues185 compared a one-tenth dilution of mometasone furoate ointment (0.1%) only for 4 weeks (first group) and the same treatment with wet wraps under dry wraps for the second 2 out of 4 weeks (second group) with a one-tenth dilution of fluticasone propionate (0.005%) only for 4 weeks (third group) and the same treatment with wet wraps under dry wet wraps for the second 2 out of 4 weeks (fourth group). The 40 children randomised were aged between 1 and 15 years with active eczema with a severity of at least 40–144 using a composite scale despite being treated with UK class II or stronger topical corticosteroids plus soap substitutes and emollients.

The single-centre trial by Schnopp and colleagues186 compared mometasone furoate (0.1%) with wet wraps against the vehicle for mometasone furoate with wet wraps. The wet wraps were applied only to a test area of eczema twice daily for 5 days, but not at night, and the mometasone furoate was applied morning and evening. Basic adjuvant treatment was allowed, although there were no details of what was deemed acceptable. Twenty children aged 2–17 years with exacerbated atopic eczema were randomised into the trial.

A small pilot trial conducted in the UK by Beattie and Lewis-Jones187 compared treatment with hydrocortisone and wet wraps with treatment with hydrocortisone without wet wraps. The treatment was applied twice a day for the first week and once a day for the second week. Participants could use emollients freely as long as a 20-minute gap between application of the hydrocortisone and application of the emollient was observed. All participants used emollients only for a third week. The 19 participants were aged ≤ 5 years and had atopic eczema with > 30% body surface area affected and no clinical evidence of infection.

Assessment of risk of bias for the new studies

Table 10 provides the risk-of-bias assessment for the new studies.

Benefits

In the trial of prednicarbate ointment and wet wraps by Foelster-Holst and colleagues183 only the local severity of eczema using SCORAD scores was reported. The prednicarbate and wet wrap group improved by 4.4 SCORAD points whereas the prednicarbate-only group improved by 3.0 SCORAD points after 48–72 hours. This was reported to be a statistically significant difference (p = 0.011) in favour of prednicarbate with wet wraps. The baseline local SCORAD score was an average of 12.0 (SD 1.04). The participants had not been allowed any topical corticosteroid treatment in the 2 days before the treatment.

The trial by Hindley and colleagues184 found no significant difference between conventional treatment with wet wraps and conventional treatment only after 4 weeks for the primary outcome of severity of eczema measured by SCORAD. ‘Tolerability’, ‘ease of application’ and ‘efficacy’, each on a 5-point Likert scale, were stated to be outcomes, but no results were reported.

The trial by Pei and colleagues185 reported the severity of eczema using a composite scale; however, the groups were not compared against each other. The number of participants analysed for each group was not reported after the second randomisation. The trial reports that the two groups that used wet wraps continued to improve compared with baseline whereas the two groups that did not use wet wraps plateaued.

The trial by Schnopp and colleagues186 measured the severity of eczema using local SCORAD scores. Over 5 days the severity of eczema decreased from 10.6 to 2.5 in the mometasone group and from 11.1 to 4.0 in the vehicle group. This was a significantly better improvement in severity in the mometasone furoate and wet wraps group compared with the vehicle and wet wraps group (p < 0.01). The nurses and carers rated the wet wraps more difficult to use in a questionnaire.

The pilot trial by Beattie and Lewis-Jones187 found that there was a greater mean reduction in Six Area, Six Sign Atopic Dermatitis (SASSAD) score without wet wraps (8 more SASSAD points than with wet wraps) after the 2-week treatment period (95% CI for difference –18 to + 2; p = 0.11). For quality of life measured using the IDLQI, the group without wet wraps had a greater median decrease (5 more IDLQI points) than the group with wet wraps (95% CI for difference –10 to + 3; p = 0.24). The Dermatitis Family Impact score showed no significant difference between the treatments.

Harms

Two trials185,186 did not report adverse events, one trial187 reported that two participants withdrew because of folliculitis and one trial183 reported that there were no withdrawals and no adverse effects. The trial by Hindley and colleagues184 reported that recruitment was stopped early because of clinically significant adverse event differences between the treatment groups, although it was not clear what these were. It was also reported that 5 out of 28 of the participants using the wet wraps needed antibiotic treatment for infected eczema compared with none in the group not using wet wraps.

Overall implications for research and practice

These five small, short-term and poorly reported trials do not currently provide evidence of a beneficial effect of combining wet wrap treatment with topical corticosteroid treatment, but this may be because of their small size and design flaws. The between-group results were not properly compared in one trial,185 making interpretation of the evidence difficult.

Topical corticosteroids compared with placebo

• There were 13 trials reported before 2000 and the trials that did report the magnitude of benefit suggested a large treatment effect of topical corticosteroids compared with placebo.

• Nine trials reported from 2000 onwards that gave a continuous short- to medium-term course of treatment provide further evidence of the large beneficial effect of topical corticosteroid treatment compared with placebo.

• There is evidence from four trials that used an initial 4-week continuous treatment regimen to ‘get control’ followed by a 16-week period of twice-weekly treatment on consecutive or evenly spaced days to ‘keep control’ of a large beneficial effect of topical corticosteroids compared with placebo.

Topical corticosteroids compared with active treatments (except topical immunomodulators)

• There were 40 trials reported before 2000 comparing topical corticosteroids against other topical corticosteroids. The trial evidence was mixed and it was difficult to provide a summary as none of the trials compared all of the main topical corticosteroids together. Fluticasone propionate and mometasone furoate were found to be reasonably equivalent to older topical corticosteroids when used once daily.

• Eleven new trials reported from 2000 onwards compared a new topical corticosteroid against another commonly used topical corticosteroid. These trials still do not compare more than two treatments and only add to the previous mixed results seen before 2000. Of these:

  • one trial provided some very weak evidence that pulsed treatment is more effective than continuous treatment with the same topical corticosteroid, which needs further research

  • one trial provided some evidence that adding a penetration-enhancing chemical to topical corticosteroid treatment was more beneficial than topical corticosteroid alone

  • one trial provided some evidence that mometasone furoate in multilamellar emulsion was more beneficial than methylprednisolone aceponate.

• Four trials reported after 2000 compared a topical corticosteroid against a different active comparator. All of these trials used the topical corticosteroid as the standard comparator and these trials are discussed in the relevant sections of this review.

Topical immunomodulators

Topical calcineurin inhibitors combined with topical corticosteroid

• There were no trials of topical calcineurin inhibitors combined with topical corticosteroids reported before 2000.

• One large, methodologically unclear trial in children that was primarily concerned with safety did not provide any evidence of a short-term significant difference in benefit when pimecrolimus and a topical corticosteroid were combined over 4 weeks of treatment. There was a significant difference in relapse rates over 12 weeks in favour of the combined treatment from a, probably post hoc, subgroup of all those clear at the end of treatment.

• One small, methodologically unclear trial in adults provides some weak evidence of benefit for the combination of tacrolimus (0.1%) and a topical corticosteroid.

Topical corticosteroids with occlusive therapy

• There were no trials involving occlusive therapy in combination with topical corticosteroids reported before 2000.

• There were five new RCTs reported after 2000 but they were all small, very short term and methodologically unclear. The results do not provide clear evidence of a significant benefit from the addition of occlusive therapies to topical corticosteroids.

Existing systematic reviews

There is one systematic review focusing on emollients by Tarr and Iheanacho. 188 The SIGN,42 NICE41 and American Academy of Dermatology (AAD)94 guidelines have all covered emollients. A review of reducing pruritus for eczema93 covers emollients, topical doxepin and sodium chromoglycate and a Cochrane review189 covers evening primrose oil and borage oil for eczema treatment.

Studies

Five trials were reported before 2000 that tested emollients for eczema55 (see Appendix 3).

Most of the 15 new trials123,126,190–202 looking at emollients reported after 2000 compared twice-daily treatment with an emollient (often applied concurrently with topical corticosteroid treatment) with treatment with topical corticosteroids alone, other emollients or, in one case, no treatment. Most trials had a treatment length of around 4–6 weeks and none of the trials gave treatment for > 2 months. Nearly all of the trials included participants with mild to moderate eczema, with only three trials123,126,191 including participants with severe eczema.

Albolene^®

One ‘equivalence’ trial of 60 patients with mild eczema sponsored by the manufacturer of the over-the-counter emollient Albolene^® (DSE Healthcare Solutions) compared Albolene against a prescription device emollient (MimyX^™, Stiefel Laboratories), with concurrent use of topical triamcinolone (0.1%) cream in both groups, twice daily for 4 weeks. 190

Exomega milk

A 6-week unblinded trial funded by the manufacturers of Exomega milk^® (Pierre Fabre Limited) randomised 173 infants with moderate to severe eczema to Exomega milk applied twice daily on non-inflammatory areas of skin or no equivalent treatment. 191 All of the participants were allowed to use topical corticosteroids during the trial to treat inflammatory lesions. The primary outcome was steroid sparing (decrease in topical corticosteroid use).

A smaller trial,192 also involving the Pierre Fabre Laboratories, included 76 infants and children aged from 6 months to 2 years with mild to moderate eczema. Participants were randomised to use either Exomega milk or a cleaning bar (A-Derma^®) all over the body, twice a day, for 2 months. All participants were allowed to use topical corticosteroids during the trial.

Urea and glycerine emollients

Two virtually identical trials by Loden and colleagues,193,194 each lasting for 30 days, compared a glycerine cream (20%) with its vehicle (with glycerine substituted with water) and a cream containing urea (4%) and sodium chloride (4%). For the earlier trial193 the treatment was applied only twice a day to a patch of dry skin identified by the dermatologist. For the second trial194 participants were allowed to use the treatments as much as necessary and at least once a day. The first trial was primarily concerned with physical markers of efficacy but also measured skin dryness. The second trial was more concerned with efficacy and measured both participant- and investigator-assessed skin dryness, as well as participant-assessed degree of stinging, smarting, itching and dryness/irritation.

A trial by Bissonnette and colleagues195 compared a urea moisturiser (5%) against a urea lotion (10%) but did not include a control arm. The trial included 100 adults aged > 18 years with mild eczema (SCORAD score of < 30) and treatments were applied twice daily for 42 days.

A trial by Amichai and Grunwald196 compared the liquid soap Axera^™ (Perrigo-Pharma), containing 12% ammonium lactate and 20% urea, with a commercially available liquid soap for showering over a 3-week period. No other emollients or soaps were permitted during the trial but participants could continue to use their current eczema treatments. The study included 36 adults and children aged 3–40 years with mild to moderate eczema, diagnosed according to the UK Working Party’s criteria. 9

Lipid emollient

One trial by Wiren and colleagues197 conducted in Sweden compared an emollient with 20% lipid content (Canoderm^® cream, ACO Hud) with no treatment until relapse or 6 months. All 55 adults with eczema who were recruited into the trial initially used the topical corticosteroid betamethasone (0.01%) for 3 weeks to induce remission. Only those participants who had ‘cleared eczema’ according to an assessment by a dermatologist (n = 44) were then randomised to the maintenance period of emollient or no treatment. The aim of the trial was to assess whether emollient use prolonged the time spent in remission from eczema.

Sunflower oleodistillate emollient

Two industry-sponsored trials123,198 from the same group compared a sunflower oleodistillate (2%)-containing emollient (Stelatopia^®; Mustela DermoPediatrie, Laboratoires Expanscience) with a topical corticosteroid for 3 weeks.

In the first trial, by Msika and colleagues,123 86 infants and young children aged from 4 months to 48 months with mild to moderate eczema were randomised to five groups. Each group received a treatment based on a moderate-potency topical corticosteroid (Tridesonit^®; CS Dermatologie) (0.05%) once or twice daily or every other day combined or not with application of the sunflower emollient twice daily for 21 days. Overall, two groups received topical corticosteroids alone and three groups received topical corticosteroids plus emollient. The second trial, by De Belilovsky and colleagues,198 compared twice-daily application of Stelatopia with the mild-potency topical corticosteroid hydrocortisone butyro-propionate cream (1 mg/g) in infants and young children aged from 4 months to 4 years with mild to moderate eczema.

Emollients containing ceramide

An industry-sponsored multicentre trial by Sugarman and Parish126 compared twice-daily application of a ceramide-dominant barrier repair formulation (EpiCeram) against the topical corticosteroid fluticasone propionate (0.05%) (Cutivate^™; GlaxoSmithKline) on affected areas in body folds. The trial included 121 infants and children aged from 6 months to 18 years. All of the participants used the emollient lotion Cetaphil^™ (Galderma Laboratories) on unaffected areas of skin.

A trial by Berardesca and colleagues200 compared a lipid mixture containing ceramide-3, cholesterol, palmitic acid and oleic acid in water-in-oil with nanoparticles with the same lipid mixture in combination with topical corticosteroids. Out of a trial population of 508 participants, 91 participants had eczema. All participants applied the treatment once or twice a day until healing had occurred or for a maximum of 8 weeks.

A trial by Draelos201 compared a ceramide-based emollient against a hyaluronic acid-based foam, the details of which are discussed later in this chapter.

A within-person trial by Simpson and Dutronc202 compared a body wash and moisturiser containing ceramide (Restoraderm^®; Galderma (UK) Ltd) in addition to standard eczema treatment with standard eczema treatment alone. The trial included 127 adults and children aged > 3 years with mild to moderate eczema according to IGA, randomised to emollient treatment twice daily on one side of the body and no emollient treatment on the other side, for an unreported length of time.

Hyaluronic acid-based emollient

A within-person trial by Draelos201 compared a hyaluronic acid-based foam emollient (Hyaltopic^™; Onset Therapeutics) against a ceramide-based emollient (EpiCeram) used twice daily for 4 weeks. The participants used the treatment on either their arms or legs, one limb per treatment, and used randomised barrier treatment on the rest of each side of their body. The 20 adults randomised had mild to moderate eczema as assessed by IGA and symmetrically distributed lesions

Overall implications for research and practice

Although it is pleasing to note an increase in the number of emollient trials since the 2000 report,55 the lack of reporting of methodological detail is disappointing. Some studies have tried to demonstrate sparing of use of topical corticosteroids, presumably on the rationale that the latter may exhibit side effects such as skin atrophy if used inappropriately. Although skin atrophy and systemic effects have been associated with prolonged use of potent topical corticosteroids in the past, these adverse effects are probably rare nowadays with appropriate use in the context of eczema management. Furthermore, reducing cutaneous inflammation by using topical corticosteroids is probably a way to improve cutaneous barrier functionality. None of the studies involving topical corticosteroids to date have shown an increase in skin atrophy caused by topical corticosteroids. It is important that studies aiming to show topical steroid sparing demonstrate that such sparing is not achieved at the expense of lack of eczema control. One new study has suggested that regular emollients used once disease has been controlled by topical corticosteroids may reduce the frequency of flares and time to next flare, an important finding that needs to be replicated in larger studies that include those with mild, moderate and severe eczema. Despite a lot of interest in emollients that have been designed to contain specific ‘barrier repair’ ingredients, there is no clear evidence to date that any of these more expensive preparations are superior to simple cheaper emollients. There is some evidence, however, that some emollients such as aqueous cream, which contains sodium lauryl sulfate, may harm the skin barrier203 and more refined mechanistic studies to identify which emollients are helpful and which are not are needed before large-scale comparative trials are carried out. Research into the effectiveness of increasing compliance with regard to the use of emollients, such as allowing patients to choose their own emollient from a range of consistencies, and how long should be left after applying an emollient before applying a topical corticosteroid is needed. Emollient application from birth is now being considered as an intervention to prevent or at least delay the onset of eczema. 204

Assessment of risk of bias

Table 18 provides the risk-of-bias assessment for the new study.

Benefits

There was no significant difference between the treatment groups for the investigator assessment of improvement. A significant difference was reported in the number of participants assessed as having an improvement in their itching scores (six out seven in the diamide derivative group compared with two out of seven in the control group). This was a per-protocol analysis that included only 7 out of 13 participants in the diamide derivative group and seven out of eight in the control group. All of the withdrawn participants were withdrawn because they did not use the treatment four or more times a week.

Harms

There was no information on adverse events included in this report.

Overall implications for research and practice

It is unclear from this small trial206 whether adding a diamide derivative to bath emollients improves eczema. The study was not powered appropriately to pick up anything but very large changes. The significant difference in rates of improvement in pruritus reported for the per-protocol population of 14 participants is highly susceptible to attrition bias given the differential dropout rate. Much larger, pragmatic and long-term trials are needed to evaluate the possible additional benefit of bath emollients in eczema. It is possible that the emollient effects from adding moisturisers to the bath are minimal compared with leave-on emollients applied after bathing. 188

Studies

No RCTs on furfuryl palmitate for eczema were reported before 2000.

One new trial was reported after 2000. 207 This new trial by Tripodi and colleagues207 compared an emollient with the antioxidants furfuryl palmitate, superoxide dismutase, 18-beta-glycyrrhetinic acid, vitamin E and alpha-bisabolol against the same treatment without furfuryl palmitate. The trial randomised 117 children aged 3 months to 14 years with eczema according to the UK Working Party’s criteria9 to treatment twice a day for 2 weeks, using one fingertip unit for every patch of skin the area of two adult hands. In total, 109 participants completed the trial, 21 of whom took medications not permitted by the trial protocol. Participants did not change their lifestyle and diet during the trial. Participants were not allowed to have used any topical or systemic treatments for eczema for 1 week before randomisation and during the trial.

Assessment of risk of bias

Table 19 provides the risk-of-bias assessment for the new study.

Benefits

This trial surprisingly reported that the emollient base without furfuryl palmitate showed better efficacy than the furfuryl palmitate emollient, with 70% (38/54) of the base emollient group compared with 29% (14/48) of the furfuryl palmitate group having a ≥ 20% reduction in baseline SCORAD scores at the end of the 2-week study. The investigator- and participant-assessed efficacy results supported this finding, with the emollient base cream being significantly more effective than the furfuryl palmitate cream.

Harms

No formal data were provided in the trial report about adverse events. The authors report in the discussion that some of the participants using the furfuryl palmitate emollient cream reported itching and burning sensations after application. There was no particular difference in the number of withdrawals from the trial; however, 14 out of 53 of the participants in the furfuryl palmitate group used other topical or systemic treatments compared with seven out of 56 in the control group.

Overall implications for research and practice

This methodologically robust trial207 based on a formal sample size calculation clearly shows that the addition of furfuryl palmitate in an emollient base seems to confer a negative effect for eczema. It is a credit to the trial authors and journal editors for publishing this negative study, which might save money from being spent on the further pursuit of this treatment.

Studies

No trials involving pill masks for eczema were reported before 2000.

One new trial was reported after 2000. 208 This manufacturer-funded, open, three-arm trial by Palombo and colleagues208 compared a pill mask containing a chitosan-derived anti-inflammatory compound ATOBIOL, tocotrienols and hyaluronic acid against a lamellar active emulsion containing ATOBIOL or petroleum ointment only. Thirty-five children applied the study treatments after using bath oils or used petroleum ointment only, twice a day for the first 8 days. After this treatment period, all participants used triamcinolone (0.1%) either once a day or twice a week (the abstract and main text do not agree) and the lamellar gel only. It is not entirely clear from the trial report exactly what the ‘pill mask’ is in this trial, except that it is a topical method of delivery for medication.

Assessment of risk of bias

Table 20 provides the risk-of-bias assessment for the new study.

Benefits

Erythema and pruritus, scaling and crusting were assessed on a VAS and both of these scores were also combined to make the ‘clinical score’. The improvement in clinical score after 4 weeks compared with baseline was reported as 58% in the pill mask group compared with the petroleum ointment, 82% in the pill mask group compared with the lamellar gel group and 64% in the petroleum group compared with the lamellar gel group. The erythema and pruritus, scaling and crusting scores were reported in graphs, but the scores were not compared between groups.

Harms

No information about adverse events was reported.

Overall implications for research and practice

The lack of methodological and treatment detail, the low number of participants and the unclear analysis of the results between groups mean that the potential benefits or harms of this topical pill mask cannot be adequately assessed from this one trial. 208

Studies

There were no RCTs for black seed oil topical treatment before 2000.

One new study was reported after 2000. 209 This within-person study involving 20 people with eczema was conducted by Stern and Bayerl209 in Germany. An ointment containing 15% black seed oil was applied on one arm and the base treatment was applied on the other arm twice a day for 28 days. The application of the two treatments was randomised to the participants’ left or right side. No other creams or ointments were allowed on the arms during the study.

Assessment of risk of bias

Table 21 provides the risk-of-bias assessment for the new study.

Benefits

A modified SCORAD score was used to record the overall severity of eczema; there was only a slight decrease in severity in both treatment groups, with no significant difference between the groups. The intensity of pruritus was measured on a VAS; there was hardly any change in either group compared with baseline and no significant difference between the groups at the end of treatment.

Harms

It was reported that there were no adverse events during the trial.

Overall implications for research and practice

It is likely that many of the participants, and possibly the outcome assessors, in this small trial209 will have known which treatment they were using. Neither the black seed oil treatment nor the base ointment changed the severity of eczema or pruritus levels. The smell of the black seed oil was reported as unacceptable. It seems that the addition of black seed oil to topical eczema treatments is not a good potential candidate for eczema treatment.

Studies

No trials on rosmarinic acid for eczema were reported before 2000.

One new trial was reported after 2000. 210 This study from Korea by Lee and colleagues210 compared a cream containing rosmarinic acid (0.3%) against the vehicle cream applied twice daily for 8 weeks in 21 participants. It is not clear whether this was a within-person trial or a parallel-group trial as the trial report describes the area being treated as ‘half an elbow flexure’.

Assessment of risk of bias

Table 22 provides the risk-of-bias assessment for the new study.

Benefits

After treatment for 8 weeks the severity of eczema decreased from a mean SCORAD score of 7.37 to 3.27 in the rosmarinic acid group and from a mean SCORAD score of 6.49 to 5.63 in the placebo group. The individual symptom scores for erythema, oedema/papulation, oozing/crusting, lichenification and local pruritus were all reported to have significantly decreased after 8 weeks of treatment compared with baseline in the rosmarinic acid group. As no between-group comparisons were reported, it is not known whether rosmarinic acid showed a beneficial effect in comparison to the vehicle cream; however, nearly all severity and symptom scores for both groups were almost the same at 8 weeks.

Harms

The trial report states that there were no reactions to rosmarinic acid at 4 or 8 weeks and that no adverse reactions such as erythema, burning and pruritus were observed.

Overall implications for research and practice

The lack of methodological clarity and appropriate analyses of the results in this trial210 mean that the trial does not provide any evidence of a beneficial effect of rosmarinic acid.

Studies

One trial involving H. rhamnoides for eczema was reported before 2000.

One new trial was reported after 2000. 211 This trial by Thumm and colleagues,211 conducted in Germany, included 58 Caucasian participants who were randomised to one of three creams for 28 days: H. rhamnoides (20%), H. rhamnoides (10%) or Miglyol^® (CREMER OLEO) (placebo). The vehicle cream for all three treatments was based on beeswax, glycerine and paraffin.

Assessment of risk of bias

Table 23 provides the risk-of-bias assessment for the new study.

Benefits

The severity of eczema significantly improved in all three groups, but the degree of improvement appeared comparable (improvement of 10.98 points for H. rhamnoides (20%), 9.52 points for H. rhamnoides (10%) and 13.76 points for Miglyol cream), although this was not statistically analysed. These results were closely mirrored by the quality-of-life scores as well as participant-assessed signs and symptoms of eczema (redness, itching, dryness and general skin condition).

Harms

No information was given in the trial report about adverse events.

Overall implications for research and practice

This short-term trial211 did not show any significant benefit of adding H. rhamnoides to an emollient base.

Studies

No RCTs on shale oil for eczema were reported before 2000.

One new trial was reported after 2000. 212 This manufacturer-sponsored trial conducted in Germany by Korting and colleagues212 compared a cream containing sodium bituminosulfonate (4%) (Ichthosin^® cream; Ichthyol-Gesellschaft) (pale sulfonated shale oil) with the vehicle cream in 99 children aged 0–12 years with mild to moderate eczema. All participants applied the treatments three times a day for 4 weeks and no concomitant medications were allowed during the trial. Skincare products were allowed on unaffected skin.

Assessment of risk of bias

Table 24 provides the risk-of-bias assessment for the new study.

Benefits

The severity of eczema was assessed using the authors’ own severity assessment of erythema, crusts, excoriations, scales, lichenification and itch, and the percentage area affected in each body site, combined into one score, with a maximum possible score of 45, as for other measures such as EASI or SCORAD. A statistically significant reduction in severity of 8.9 (SD ± 7.4) was found after 4 weeks of using the pale sulfonated shale oil cream compared with a reduction of 1.3 (SD ± 8.3) using the vehicle cream, which was also present after 1 week. Participant-assessed tolerability also showed a significant difference in favour of pale sulfonated shale oil, with 73 participants in the pale sulfonated shale oil group compared with 42 participants in the vehicle group rating the treatment as ‘very good’. All individual signs and symptoms assessed, including itch, were reported as significantly better in the pale sulfonated shale oil group from 2 weeks onwards.

Harms

There were no serious adverse events reported during this trial. The other adverse events reported resulted in the participants withdrawing from the study: two in the pale sulfonated shale oil group, because of pruritus, erythema and spreading of eczema, and four in the vehicle group, one because of bacterial superinfection and three who experienced itch plus worsening/spreading of eczema or erythema.

Overall implications for research and practice

Even though the two treatments were matched for colour, the slight odour of the shale oil cream means that there is a risk that participants and possibly the investigators who were outcome assessors were unblinded. Nevertheless, this trial212 demonstrated a reasonable beneficial effect of adding pale sulfonated shale oil to a topical cream for the treatment of mild to moderate eczema in Caucasian children, with no evidence of any particular adverse events. It is important that this trial is followed up with further long-term, large-scale research that incorporates a range of skin tones and pays attention to blinding the outcome assessors of the objective outcomes by making sure that the cream is not applied close to an assessment.

Studies

No RCTs on V. filiformis for eczema were reported before 2000.

Two new trials were reported after 2000. 213,214 These two manufacturer-funded trials by Guéniche and colleagues213,214 compared a cream with V. filiformis lysate (5%) against vehicle cream. The first trial213 involved 13 participants with mild to moderate eczema, who used the treatments on either side of the body twice daily for 4 weeks. The second trial214 involved 75 participants (aged 6–70 years) with mild eczema and a history of atopy.

Assessment of risk of bias

Table 25 provides the risk-of-bias assessment for the new studies.

Benefits

In the first study213 a statistically significant but relatively small difference in the severity of eczema of around half a point using the EASI score (p = 0.012) and the modified EASI score (p = 0.008), which included itch scores, was reported after 28 days of treatment.

In the second, larger, study,214 the V. filiformis lysate cream showed a significant beneficial effect on eczema severity and pruritus after 30 days. Mean SCORAD scores at the end of treatment were 24.9 in the placebo group and 15.1 in the V. filiformis group (p = 0.004). Baseline scores were 29 (SD ±9.7) in the placebo group and 31 (SD ±11.9) in the V. filiformis group. The severity scores were analysed over time and also gave a significant result in favour of the bacterial lysate cream.

Harms

The first study213 reported the most common adverse event as pricking and burning sensations (at the same rate for both treatments). The authors suggest that this may have been caused by the composition of the vehicle cream, which was not designed specifically for eczema. The trial also reported that few participants reported dryness for both sides of the body.

Although it was reported in the second study214 that adverse events were monitored, no information was given on whether any occurred or not.

Overall implications for research and practice

These trials213,214 suggest a possible beneficial effect of V. filiformis on eczema severity. It is not clear if participants were allowed any other eczema treatments during the trials and so we are left guessing whether the beneficial treatment effect seen is the result of use of the V. filiformis lysate cream or an increase in the use of standard eczema treatments. Frustratingly, information on adverse event monitoring is not provided at all for the second, larger, trial, making it impossible to weigh up benefits compared with any potential harms. With so little practical information in the context of potential use in a clinical setting, more independent research with a longer treatment period, full information on any adverse events and a clear report of any concomitant eczema treatments used are needed before the true potential of this treatment can be assessed.

Background

Miltefosine is a phospholipid analogue that was originally developed as a chemotherapy treatment. It has also been found to have antiprotozoal activity and it is effective against both visceral and cutaneous leishmaniasis.

Studies

No RCTs on miltefosine for eczema were reported before 2000.

One new trial was reported after 2000. 215 This part industry-sponsored within-person trial by Dolle and colleagues215 compared a solution containing miltefosine against a hydrocortisone solution for 3 weeks. The dose for both treatments was gradually increased from two drops per lesion once a day for the first week to two drops per lesion twice a day for the second and third weeks. Sixteen adults with moderate to severe eczema according to the criteria of the Hanifin and Rajka8 and having at least two comparable lesions were recruited.

Assessment of risk of bias

Table 26 provides the risk-of-bias assessment for the new study.

Benefits

The severity of the eczema lesions treated with hydrocortisone improved slightly more than the severity of the miltefosine-treated lesions after 3 weeks of treatment. The primary outcome of ‘improvement’ in eczema severity, defined as a > 1.5-point drop in Three-Item Severity (TIS)63 score, occurred in 10 out of 16 participants for the miltefosine-treated lesion. The results of the primary outcome for hydrocortisone treatment were reported only in a graphical form, although this appears to show a greater reduction in severity for the hydrocortisone group. Four weeks after treatment was stopped, the hydrocortisone-treated lesions increased in severity by a median of 0.5 TIS points to 2 (minimum 1.5, maximum 3), whereas the miltefosine-treated lesions decreased by a median of 2 TIS points to exactly the same TIS score as the hydrocortisone-treated lesions.

Harms

There were a relatively high number of local topical adverse events related to the treatments, with miltefosine treatment producing more of the adverse events (10/16 participants affected) than hydrocortisone treatment (7/16 participants affected). These adverse events included pruritus, burning, tingling and dry skin. Dry skin was seen only with miltefosine treatment. There were no withdrawals because of adverse events and no systemic adverse events.

Overall implications for research and practice

Although the trial authors focused on the deterioration of the hydrocortisone-treated lesions after treatment was stopped, the rate of deterioration was slow and the lesions only reached the same severity as those lesions treated with miltefosine. It is interesting that miltefosine seemed to show a perpetuating beneficial effect on the severity of eczema, although it was slower than the hydrocortisone treatment to take effect. With the lack of methodological clarity and wide CIs, the evidence for treatment of eczema with miltefosine is poor. This early trial has not provided a promising signal to justify further trials.

Studies

No RCTs assessing opiate receptor antagonists for eczema were reported before 2000.

One new trial was reported after 2000. 216 This multicentre crossover trial by Bigliardi and colleagues216 in Germany compared naltrexone cream against vehicle cream to be applied for up to 28 days when the participants were experiencing intense symptoms of itching. Forty-five adults with eczema and experiencing bouts of itching of > 50 mm on a 100-mm VAS were randomised. All participants were allowed to use topical eczema treatments as rescue medication, except in an itching intensity monitoring period.

Assessment of risk of bias

Table 27 provides the risk-of-bias assessment for the new study.

Benefits

A 52.6-mm (26.7%) difference between the decrease in pruritus intensity score for the two treatments in favour of naltrexone was reported for the 40 out of 45 participants who made up the full analysis set. This was reported as a statistically significant treatment effect between the treatment groups of p = 0.047. No confidence limits were given for any of the results in the trial. The results of the per-protocol analysis of 39 participants were similar. This trial did not record the severity of eczema.

Harms

Information on adverse events was not reported.

Overall implications for research and practice

Although a statistically significant effect on pruritus appears to be a positive result, this poorly reported trial216 does not provide enough data to support the adoption of naltrexone. The severity of the eczema in the two treatment groups and the amount of topical corticosteroid used is not reported and, although pruritus levels tend to decrease as the severity of eczema decreases, this trial on its own is not robust evidence of a beneficial effect of naltrexone. Trials that measure the severity of eczema as well as pruritus and document any concurrent eczema treatment are needed.

Studies

No RCTs looking at topical vitamin B₁₂ were reported before 2000.

Two new trials have been reported since 2000. 217,218 These two within-person trials,217,218 both industry funded, compared cyanocobalamin (0.07%) cream with a base cream. In the trial by Stücker and colleagues,217 48 adults aged 18–70 years applied the treatments twice a day for 8 weeks. In the trial by Januchowski,218 22 children aged from 6 months to 18 years (no severity inclusion criteria stated) were randomised to treatment for 4 weeks, but it was not clear how often they were instructed to use the treatment.

Assessment of risk of bias

Table 28 provides the risk-of-bias assessment for the new studies.

Benefits

The trial by Stücker and colleagues217 found that there was a significant decrease in eczema severity after 8 weeks of topical vitamin B₁₂ treatment compared with the base cream, as measured using a modified SASSAD score (maximum of 240 points). The mean decrease for topical vitamin B₁₂ was 55.34 (standard error of the mean 5.74) whereas the mean decrease for the base cream was 28.87 (standard error of the mean 4.86; p < 0.0002). This significant difference was seen from week 4 onwards. Participant- and investigator-assessed efficacy both showed a significant difference in favour of topical vitamin B₁₂ treatment after 8 weeks when the responses were grouped as ‘effective’ or ‘non-effective’ (p < 0.005).

The trial by Januchowski218 also reported a significant decrease in eczema severity, measured using modified SCORAD scores (maximum of 27 points), for topical vitamin B₁₂ cream compared with the base cream after 4 weeks of treatment. Taking data from a graph, the decrease in total SCORAD scores for topical B₁₂ treatment after 4 weeks was 4.5 whereas that for the placebo treatment was 1.7 (p = 0.011). The objective SCORAD score mirrored this result and the subjective SCORAD score also showed a significant difference, although not as pronounced, by week 4.

Harms

In the trial by Stücker and colleagues,217 33 cutaneous events were reported, which were all mild except for one, which involved a moderate reaction after applying the placebo cream. Of these cutaneous events, four cases were considered ‘probably related’ and two cases were ‘possibly related’ to the application of vitamin B₁₂.

In the trial by Januchowski,218 only one adverse event was reported, which resulted in withdrawal from the trial. The participant had a reaction to both the active and placebo treatments but the nature of the reaction was not reported.

Overall implications for research and practice

The two trials217,218 appear to show a significant beneficial effect of topical vitamin B₁₂ cream on the severity of eczema. The lack of CIs for the results presented in both trials means that there is no way to judge the statistical robustness of the results. There were a number of adverse events relating to skin irritation, probably caused by topical vitamin B₁₂. Whether or not to try this topical treatment will probably remain a decision based on individual circumstance and preference.

Studies

No RCTs looking at WBI-1001 were reported before 2000.

Two new trials were reported after 2000. 219,220 The trial by Bissonnette and colleagues in 2010,219 reported as a research letter, compared the synthetic compound 2-isopropyl-5-[(E)–2-phenylethenyl] benzene-1,3-diol (WBI-1001; Welichem Biotech Inc.) at both 0.5% and 1.0% concentrations against vehicle cream. The 37 participants had eczema scores of ≤ 12 on the EASI scale and scored 2 (mild) or 3 (moderate) for IGA; they were randomised to one of the three treatment groups and applied the treatment twice daily for 4 weeks.

Bissonnette and colleagues220 published a second trial in 2012, again comparing the same topical treatment WBI-1001 at 0.5% and 1% concentrations with vehicle as the placebo. This was a larger and slightly longer trial. In total, 148 adults aged 18–65 years with ‘chronic’ eczema for ≥ 6 months, diagnosed according to the criteria of Hanifin and Rajka,8 a body surface area of 3–20% and an eczema severity of mild to severe (IGA 2–4) were included.

Assessment of risk of bias

Table 29 provides the risk-of-bias assessment for the new studies.

Benefits

The earlier trial by Bissonnette and colleagues219 stated that all efficacy analyses were not planned before database lock and so should be considered post hoc. Change in the severity of eczema from baseline measured by IGA, SCORAD and EASI showed significant differences in favour of WBI-1001. For the IGA after 4 weeks’ treatment, the percentage decreases from baseline were 5.6% for placebo, 38.9% for WBI-1001 0.5% and 45.8% for WBI-1001 1%. Active treatment (WBI-1001 0.5% and 1.0%) was significantly beneficial compared with placebo (p = 0.003). The severity of eczema measured by EASI and SCORAD closely mirrored these results. Body surface area also showed almost the same significant percentage decreases by week 4. Pruritus scores showed the same pattern but no between-group analyses were reported.

The larger trial by Bissonnette and colleagues in 2012220 reported a significant reduction in pruritus at day 42 from baseline for the WBI-1001 0.5% group (29.8%) and the WBI-1001 1.0% group (66.9%) compared with placebo (9.5%) (p < 0.001 for both WBI-1001 treatments). There was also a significant reduction in eczema severity at day 42 from baseline, measured using IGA, for both of the WBI-1001 treatments (0.5% and 1.0%) compared with placebo (WBI-1001 0.5%: –1.3 SD ± 0.97, 95% CI –1.2 to –0.5; WBI-1001 1.0%: –1.8 SD ± 1.02, 95% CI –1.6 to –0.9; placebo: –0.5 SD ± 0.89, 95% CI was not reported) (p < 0.001 for both WBI-1001 treatments). Significant reductions in eczema severity as measured by EASI and SCORAD were also reported for the WBI-1001 creams compared with placebo.

Harms

In the earlier trial by Bissonnette and colleagues219 there were no serious adverse events and no withdrawals as a result of adverse events. One participant in the WBI-1001 group had a T-wave anomaly, although it is not reported if this was related to treatment. Two participants in the placebo group and one in the WBI-1001 group had mild papules and two participants in the placebo group had pruritus.

Two serious adverse events were reported in the trial by Bissonnette and colleagues220 published in 2012: one case of cellulitis in the WBI-1001 0.5% group and one case of acute cholecystitis in the WBI-1001 1.0% group. Neither of these events was reported as being related to study treatment. Nine participants stopped treatment because of adverse events. In the placebo group four events were eczema and one was worsening eczema. Two events, one of eczema and one of contact dermatitis, were reported in the WBI-1001 0.5% group and two events of contact dermatitis were reported in the WBI-1001 1.0% group.

Overall implications for research and practice

The small Phase 2 trial,219 which did not prespecify the efficacy outcomes reported, offered a hint that it may be worth carrying out a larger, long-term Phase 3 trial of WBI-1001. The trial was not designed to determine clinical efficacy and included a small number of participants. The same group conducted a larger, slightly longer-term trial aimed at assessing the clinical benefits and harms of WBI-1001 and found clear evidence of benefit compared with placebo for clinician-assessed severity of eczema. The trial did not use any patient-reported outcome measures such as itching as separate outcomes. Longer trials that use a pragmatic comparator such as topical corticosteroids or topical immunomodulators and measure patient-reported outcomes are now needed to be clearer about whether this treatment could be useful in routine clinical care. Adverse events such as contact dermatitis should also be assessed in more detail.

Studies

No RCTs of carbohydrate-derived fulvic acid treatments for eczema were reported before 2000.

Gandy and colleagues221 compared carbohydrate-derived fulvic acid in an emollient base with the emollient base only in 36 participants aged > 2 years. The participants applied the emollient twice daily for 4 weeks on the affected area. All participants were allowed to use Epizone (VanDyk Pharmaceutical Products) (an emollient buffered with acetic acid) as needed.

Assessment of risk of bias

Table 30 provides the risk-of-bias assessment for the new study.

Benefits

Although the report states that there were statistically significant decreases in the carbohydrate-derived fulvic acid group for investigator global response to treatment, investigator-assessed severity of disease and participant-assessed severity of disease, there were also significant decreases in the emollient placebo group. The trial does not report between-group differences but it seems unlikely from the data that there were any statistically significant differences between the treatments.

Harms

The only adverse event reported was a short-lived burning sensation after treatment, although the report does not state which treatment was involved.

Overall implications for research and practice

This short trial without a clear methodology seems to show no additional benefit of adding carbohydrate-derived fulvic acid to emollient treatment. The trial did not include two participants in the final analyses as they used concomitant treatments. The extent to which participants and investigators, both of whom were the outcome assessors, were blinded is not known. Perhaps reassuringly, this trial adds to evidence that regular emollient use improves the severity of eczema.

Studies

No RCTs looking at the topical matrix metalloproteinase and aureolysin inhibitor SRD441 for eczema were reported before 2000.

One new trial was reported after 2000. This industry-funded study by Foelster-Holst and colleagues222 compared SRD441 against vehicle in 93 adults with mild to moderate atopic eczema confirmed by a dermatologist. The participants used SRD441 cream (1 mg/g) or vehicle cream on all affected and commonly affected areas twice a day for 28 days. The trial was run in 13 centres in Germany, Bulgaria and Finland.

Assessment of risk of bias

Table 31 provides the risk-of-bias assessment for the new study.

Benefits

There was no significant difference between the treatments for the primary outcome of rate of ‘success’ (defined as an IGA score of 0 or 1) at day 21. There were also no significant differences in any of the secondary outcomes, which included time to resolution of the primary exacerbation (IGA score of 0 or 1), IGA score (all visits), participant-assessed total pruritus over the previous 24 hours, number of participants requiring rescue medication and quality of life measured using the Dermatology Life Quality Index (DLQI).

Harms

In total, 60.0% (n = 27) of the SRD441 group and 70.8% (n = 34) of the vehicle group experienced adverse events. Of these, 40.0% (n = 18) in the SRD441 group and 58.3% (n = 28) in the vehicle group were possibly or probably related to study treatment. The adverse events related to treatment were mostly application site reactions and occurred at roughly equal rates in both groups. Seven participants in the SRD441 group and 11 participants in the vehicle group withdrew because of adverse events. The main reason for withdrawal from treatment was application site reactions.

Overall implications for research and practice

This is a clearly reported and methodologically sound trial. 222 Both the vehicle control and the study treatment were poorly tolerated, precipitating a higher than expected withdrawal rate because of adverse events. Although problems with the vehicle control may have masked any potential beneficial effects elicited by SRD441, this trial shows no evidence of benefit for SRD441 in the treatment of eczema.

Studies

No RCTs looking at the oligosaccharide raffinose for eczema were reported before 2000.

One new trial was reported after 2000. 223 This trial, conducted in France by Misery and colleagues,223 compared a lipiderm cream with raffinose (1%) added (Tefirax^®; Laboratoire G-pharm) against the lipiderm cream alone. Participants were instructed to apply as much as necessary for 3 days and as needed for persistent symptoms of pruritus. The 11 adults in the trial all had eczema and current pruritus.

Assessment of risk of bias

Table 32 provides the risk-of-bias assessment for the new study.

Benefits

The short trial report223 does not give any detailed data apart from a graph of pruritus intensity on the first application and a graph of the mean pruritus intensity for all applications of treatment (67 applications between 11 participants). After treatment, six out of 11 participants reported a benefit from the cream containing raffinose, four participants reported a benefit from both treatments and one participant reported no benefit from either treatment. The report states that the study is too small to analyse whether the treatment showed significant benefit and no between-group analyses were reported. Although the primary outcome results (intensity of pruritus) were presented, the results of other specified outcomes were not reported.

Harms

The report states that the cream containing raffinose produced application site burning.

Overall implications for research and practice

It is debatable whether a crossover trial223 of 11 participants, which the trial authors admit was not appropriately sized to assess the treatment’s potential benefits, just serves to confuse more than aid those looking for evidence of treatment benefits. If the relative merits or otherwise of topically applied raffinose are to be considered at all, an appropriately designed trial will need to be conducted.

Studies

There were no trials on Atopiclair reported before 2000.

Five new trials were reported after 2000. 199,224–227

Assessment of risk of bias

Table 33 provides the risk-of-bias assessment for the new studies.

Benefits

Harms

Overall implications for research and practice

Three of the five trials show a significant improvement in IGA for participants treated with Atopiclair compared with those treated with vehicle, with two also reporting an improvement in EASI score. 225–227 No conclusions can be drawn from the trial by Belloni and colleagues224 as no between-treatment comparison was performed. One trial showed no difference between Atopiclair and two other emollients for IGA and EASI. 199

Overall, there is reasonable evidence of benefit for Atopiclair compared with vehicle. Further trials comparing Atopiclair against other active treatments are required and these should ideally be independent from the manufacturers of any interventions involved.

Studies

No RCTs looking at farnesol or xylitol for eczema were reported before 2000.

One new trial was reported after 2000. 229 This within-person trial by Katsuyama and colleagues229 compared an oil-in-water 17% moisturiser cream containing 0.02% farnesol and 5% xylitol against the oil-in-water 17% moisturiser cream only. The treatments were applied for 7 days by 17 participants to their left and right forearms.

Assessment of risk of bias

Table 34 provides the risk-of-bias assessment for the new study.

Benefits

Although this trial was focused on physiological indicators of efficacy and numbers of S. aureus, changes in dryness, redness, excoriation, scaling and papules were assessed by a dermatologist. The trial reported no significant differences between the treatments; however, no detailed data were presented in the report.

Harms

It was reported that no adverse events occurred during the trial.

Overall implications for research and practice

The changes in dermatologist-assessed skin condition were not significantly different between the two treatments.

Studies

No RCTs that assessed a combination of levomenol and heparin for eczema were reported before 2000.

One new trial was reported after 2000. This trial by Arenberger and colleagues, published separately in both German230 and English,231 compared a cream containing a combination of levomenol and heparin with a cream containing levomenol alone, a cream containing heparin alone and a cream without any active ingredient.

Assessment of risk of bias

Table 35 provides the risk-of-bias assessment for the new study.

Benefits

The primary outcome of intensity of itching (on a VAS) was significantly reduced in the combined levomenol and heparin group compared with the levomenol, heparin and vehicle groups. The most significant difference (mean ± SD) was between the vehicle group and the combined levomenol and heparin group at week 8 (24.3 mm ± 2.1 mm, 95% CI 20.2 mm to 28.5 mm), although the two separate levomenol and heparin groups were not compared against vehicle. The severity of eczema measured by SCORAD mirrored the pruritus results, with the most significant difference (mean ± SD) being between the combined levomenol and heparin group and the vehicle group after 8 weeks of treatment (14.6 ± 1.3, 95% CI 12.8 to 17.1). The participants were asked to rate the efficacy of the treatment on a 4-point scale and the combined treatment had the highest percentage of assessments rated as ‘good’ or ‘very good’ (97%). This assessment was not analysed across treatment groups, except in a subgroup of children aged 0–12 years, with 100% in the combined group and 42.9% in the vehicle group assessing the treatment as ‘good’ or very good’ (p = 0.002). It is not clear whether this subgroup was prespecified or not from the trial report.

Harms

One participant in the heparin group reported a transient increase in itching, which the investigators assessed as most likely being caused by the participant’s eczema. No other adverse events were reported.

Overall implications for research and practice

There is reasonable evidence from this one trial230,231 of benefit from combined heparin and levomenol treatment compared with vehicle and some evidence that the combination is significantly more effective than each of the treatments given separately. The subgroup analysis of children aged ≤ 12 years must be treated with caution as the numbers of participants were low and it is not clear whether this was a post hoc analysis or not. It may be worth comparing this treatment against a more pragmatic comparator of ‘standard care’ with emollients and topical corticosteroids to obtain a clearer picture of its potential benefit.

Studies

No RCTs that assessed a topical bacterial antigen suspension for eczema were reported before 2000.

One new trial was reported after 2000. 232 This trial by Mora and colleagues232 compared a topical suspension containing Lantigen B against a placebo solution used twice a day for 3 months on the eczema lesions, using one drop per year of age. Eighty children aged between 2 and 6 years with external auditory eczema lesions were randomised.

Assessment of risk of bias

Table 36 provides the risk-of-bias assessment for the new study.

Benefits

The trial report states that the clinical efficacy scores were lower for the antigen suspension than for the placebo in the second study period, which is possibly referring to the period of 1 year after finishing treatment, but no values were reported. For the 3-month treatment period, the clinical efficacy score decreased from 7.1 to 3.4 in the antigen group and from 7.3 to 6.4 in the placebo group; however, the difference between the groups was not statistically analysed.

Harms

The authors reported that no participants experienced side effects.

Overall implications for research and practice

The trial report232 does not give enough detail about the trial methodology and results to be able to gather any good evidence of benefits or harms of this topical bacterial antigen treatment.

Studies

No trials of chamomile extract for eczema were published before 2000.

One new trial involving camomile extract was reported after 2000. 233 This within-person trial by Patzelt-Wenczler and Ponce-Poschl233 randomised 72 participants with eczema whose severity was described as ‘at least moderate’. Participants applied Kamillosan^® cream (Dales Pharmaceuticals Ltd) (containing 2% ethanolic extract of chamomile flowers), vehicle or hydrocortisone twice daily to a specified arm. The severity of eczema was recorded as the sum of the pruritus, erythema and desquamation scores. The IGA scores for each arm separately were also recorded. Five individual signs of oedema, papules/pustules, lichenification, excoriation and fissures were each assessed on a 4-point scale and combined into one score. Adverse events were also measured. It is not clear how long the participants used the interventions for; however, the participants were followed up for 2 weeks from baseline.

Assessment of risk of bias

Table 37 provides the risk-of-bias assessment for the new study.

Benefits

The trial report233 does not compare the interventions against each other for any of the outcomes measured. The assessment of Kamillosan, hydrocortisone or the vehicle cream, which has emollient properties in its own right, did not show any big differences in effect apart from hydrocortisone not performing quite as well as the other two comparators. This is not surprising as the potency of the hydrocortisone used (0.5%) falls well below the potency required to effectively treat nearly all cases of eczema.

Harms

The only information given on adverse events in this trial was that three participants in the combined Kamillosan/placebo group withdrew early because of intolerability.

Overall implications for research and practice

This trial fails to compare the interventions and hence does not provide any evidence regarding the use of Kamillosan for eczema. A methodologically robust trial of Kamillosan compared with other topical treatments that pays greater attention to recording adverse events is needed to better inform the many people with eczema who buy this relatively expensive treatment over the counter and the clinicians who are asked to give advice on this treatment.

Studies

No trials involving camellia oil were published before 2000.

One new trial involving camellia oil was published after 2000. 234 This small crossover study by Hamada and colleagues234 investigated the use of a spray containing camellia oil against a spray containing purified water. Forty-two participants with eczema described as ‘less than moderate’ used the spray in addition to their usual care, as desired, for 2 weeks. After 2 weeks, participants switched to the other spray.

Assessment of risk of bias

Table 38 provides the risk-of-bias assessment for the new study.

Benefits

The camellia oil spray showed significant benefits for itching (p < 0.01) and moisturising (p < 0.01) compared with the purified water spray. The amount of ointment being used was significantly decreased when using camellia oil.

Harms

There were no adverse events reported in this trial. 234

Overall implications for research and practice

This one small trial234 hints at the potential benefit of this treatment and has not reported any adverse events. A large trial that addresses the issue of blinding an intervention that has a distinctive aroma is needed to assess this treatment further.

Studies

There were no trials of topical cipamfylline cream reported before 2000.

One new trial of cipamfylline cream has been reported since 2000. 124 This trial by Griffiths and colleagues124 compared cipamfylline cream (1.5 mg of cipamfylline per gram of cream) used up to a maximum of 2 g of cream per day against hydrocortisone 17-butyrate cream (0.1%) or vehicle of cipamfylline. The 103 adults with eczema according to the Hanifin and Rajka8 criteria, who had stable symmetrical lesions on the arms, applied up to 2 g of study treatment per day for 14 days.

Assessment of risk of bias

Table 39 provides the risk-of-bias assessment for the new study.

Benefits

For the primary outcome of eczema severity measured using a total severity score, cipamfylline cream was significantly more effective than vehicle after 14 days (mean difference 1.67, 95% CI 1.06 to 2.28; p < 0.001) and hydrocortisone 17-butyrate was significantly more effective than cipamfylline cream (mean difference –2.10, 95% CI –2.93 to –1.27; p < 0.001). For both the investigator- and the participant-assessed overall response, the cipamfylline cream was significantly more effective than the vehicle and hydrocortisone 17-butyrate cream was significantly more effective than the cipamfylline cream. For participant-assessed pruritus after 14 days, the cipamfylline cream was significantly more effective than the vehicle and hydrocortisone 17-butyrate cream was significantly more effective than the cipamfylline cream. The participants found the hydrocortisone 17-butyrate cream to be significantly more cosmetically acceptable than the cipamfylline cream and the cipamfylline cream significantly more acceptable than the vehicle. For those who only needed emollient on treated areas after 14 days of study treatment, there was no significant difference in the relapse rate after 7 days for cipamfylline cream compared with vehicle. The relapse rate was significantly lower after hydrocortisone 17-butyrate cream compared with cipamfylline cream (p = 0.022).

Harms

There was no difference in cutaneous adverse events assessed as possibly or probably related to trial treatment on the treatment sites in either group (p = 0.13 for both treatment comparison groups). In the cipamfylline/vehicle comparison group, 29 (55.8%) participants reported 63 adverse events in total. In the hydrocortisone/cipamfylline group, 20 (40.8%) participants reported 41 adverse events in total. The adverse events were mostly application site reactions, including itching, stinging or burning, and drug reactions.

Overall implications for research and practice

This trial has clearly placed cipamfylline cream as less effective than hydrocortisone cream but more effective than vehicle for both participant-assessed and objective outcomes. The methodology of the trial was fairly clear and robust enough to exclude further testing of topical cipamfylline cream. Topical cipamfylline may have some limited short-term benefits for those with difficult to manage eczema because of steroid phobia, steroid resistance or contraindications for steroids, but it should be used with caution as there are no data on the safety of long-term treatment with this cream.

Studies

There were no trials of lipoxin A₄ reported before 2000.

One new trial conducted in China by Wu and colleagues235 compared 15(R/S)-methyl-lipoxin A₄ 0.1% cream against mometasone furoate 0.1% cream and also a placebo of distilled water in 1% dimethyl sulfoxide mixed with the identical cream base as used for the 15(R/S)-methyl-lipoxin A₄. All treatments were applied to the face twice a day for 10 days using cotton sticks. Sixty participants with infantile acute or subacute facial eczema were randomised, 20 to each group.

Assessment of risk of bias

Table 40 provides the risk-of-bias assessment for the new study.

Benefits

The efficacy of 15(R/S)-methyl-lipoxin A₄ cream was not directly compared with that of mometasone furoate cream or placebo in the trial report, making it impossible to assess the results of the trial. It was reported that of the six components of the Severity Scale Score, the 15(R/S)-methyl-lipoxin A₄ cream significantly reduced erythema and pruritus/scratching at day 3, papulation, vesiculation and scaling at day 5 and lichenification at day 10 compared with baseline. Mometasone furoate cream significantly reduced erythema, papulation, vesiculation, scaling and pruritus/scratching at day 3. Placebo significantly reduced scaling at day 5.

Harms

No clinical adverse events were reported and none of the safety tests, including full blood count and kidney and liver function tests as well as an electrocardiogram, showed any significant differences compared with baseline for all three treatment groups.

Overall implications for research and practice

There is no attempt to compare the results of one treatment against another in this trial report. An appropriate analysis of between-group differences is needed before being able to assess this potential treatment.

Studies

There were no trials of N-acetyl-L-hydroxyproline reported before 2000.

One new trial of N-acetyl-L-hydroxyproline has been reported since 2000. 236 This within-person trial conducted in Japan randomised 15 adults with slight eczema as assessed by a dermatologist according to the guidelines of the Japanese Dermatological Association to N-acetyl-L-hydroxyproline 1% cream and the vehicle cream for 4 weeks, with the treatments being applied twice daily to each forearm.

Assessment of risk of bias

Table 41 provides the risk-of-bias assessment for the new study.

Benefits

The only clinically relevant efficacy outcome assessed was the reduction in pruritus, assessed using a 100-mm VAS; however, the change in pruritus was not compared between the two treatments. The control treatment resulted in a reduction in pruritus from 27.1 SD ± 5.9 mm at baseline to 19.6 SD ± 6.0 mm at 4 weeks whereas the N-acetyl-L-hydroxyproline treatment resulted in a reduction in pruritus from 27.8 SD ± 5.9 mm at baseline to 16.4 SD ± 5.0 mm at week 4. The only scores to be statistically compared between groups were the pruritus scores at 4 weeks (p = 0.07).

Harms

It was reported that there were no adverse events, defined as no inflammation, no irritation and no allergic reactions, during the trial.

Overall implications for research and practice

This trial report does not appropriately compare the results of the two treatment groups and so does not provide any evidence about the comparative effectiveness of this potential treatment.

Studies

There were no trials of nalmefene hydrochloride monohydrate (SRD174) reported before 2000.

A new crossover trial compared SRD174 cream against vehicle cream. 237 The participants had to have active and pruritic eczema covering a body surface area of ≤ 20% and at least three episodes of moderate to severe pruritus defined as ≥ 40 on a 101-point VAS in the 7 days prior to randomisation. The study treatments were applied when a participant experienced an itch of ≥ 40 on a 101-point VAS during two treatment periods of 7 days each. The 62 participants randomised had to identify a target area of highest intensity and treat both the target area and other areas of bothersome itch. Participants could treat more than one episode of itch in a day provided that the episodes were > 8 hours apart and the total amount of study drug used in a day was less than one 10-g tube.

Assessment of risk of bias

Table 42 provides the risk-of-bias assessment for the new study.

Benefits

The primary outcome of the sum of pruritus intensity difference between 0 and 4 hours was 210.7 (SD ± 20.4) in the SRD174 group and 212 (SD ± 20.2) in the vehicle group, a difference of –1.3 (95% CI –25.9 to 23.3; p = 0.91).

None of the secondary efficacy end points tested (sum of pruritus intensity difference between 0 and 8 hours, pruritus intensity difference at each assessed time point) demonstrated a statistically significant or clinically important difference between the test product and the vehicle. Change in EASI score during each time period, IGA during each treatment period, quality of sleep recorded during each treatment period, pruritus relief, time to achieve > 30%, > 50% and 80% reduction in itch sensation, time to achieve a reduction in itch sensation to below a VAS score of 40 and use of rescue medication for pruritus were also not significantly different between the groups.

Harms

There was a higher incidence of adverse events in the SRD174 group than in the vehicle group: 22 (36.7%) participants reported a treatment-emergent adverse event in the SRD174 group and 14 (23.3%) participants reported a treatment-emergent adverse event in the vehicle group.

Overall implications for research and practice

This trial, although lacking some methodological clarity, does not provide any supportive evidence of a potential benefit for this treatment for any of the large number of outcomes. The trial authors are also clear in the report that there is no evidence of benefit for this treatment. It is probably not worth pursuing this treatment further.

Studies

There were no trials of licochalcone A reported before 2000.

A new within-person trial by Udompataikul and Srisatwaja238 compared a cream containing licochalcone A (0.025%) in a ceramide and linoleic acid lipid base (12% omega-6-fatty acids, 0.05% ceramide and 10% glycerine) for 6 weeks against hydrocortisone 1% lotion for 4 weeks followed by 2 weeks of ‘cream base’, which was not described further. Thirty children aged 2–15 years with mild to moderate eczema (SCORAD 1–40) that was present in the flexures on both sides of the body applied each treatment to one side of the body twice daily. Before starting treatment, those taking oral treatments had a washout period of 4 weeks and those using topical treatments had a washout period of 2 weeks.

Assessment of risk of bias

Table 43 provides the risk-of-bias assessment for the new study.

Benefits

The trial report stated that there were no significant differences between treatments in the proportion of participants who rated their satisfaction as ‘good’ or ‘excellent’ (licochalcone A cream: 10/26 excellent, 12/26 good, total 84.6%; hydrocortisone/base cream: 12/26 excellent, 10/26 good, total 84.6%). It is clear that there were some differences if those rating their satisfaction as ‘good’ or ‘excellent’ were assessed separately. There was no significant difference between groups in the reduction of the severity of eczema, measured using SCORAD (p = 0.199), although it was unclear which data this statistical comparison referred to. The last 2 weeks of the trial are described as a follow-up phase to evaluate relapse, but the participants were still using the active licochalcone A treatment on one side of their body. There was no significant difference between the treatments in relapse rate in the follow-up period using a Kaplan–Meier survival analysis (p = 0.240).

Harms

The trial report stated that there were no side effects of either treatment.

Overall implications for research and practice

Although the trial authors appear to suggest that the trial provides evidence that treatment with licochalcone A cream is beneficial compared with hydrocortisone, there was no evidence of superiority from this trial.

Studies

There were no trials of AR-GG27^® (Giuliana SpA) reported before 2000.

A new single-centre trial by Patrizi and colleagues239 compared AR-GG27 cream, which contains many different ingredients, with placebo cream, which contained 10 ingredients that are in the AR-GG27 cream and citric acid. Sixty children aged from 2 months to 15 years with pityriasis alba on the face and/or limbs and/or trunk, and eczema diagnosed using the Hanifin and Rajka8 criteria, with xerosis and pruritus present, were randomised to either AR-GG27 cream or placebo cream applied twice daily about 12 hours apart on affected and perilesional areas for 30 days. No other topical or systemic treatments were allowed during the trial, including sun exposure.

Assessment of risk of bias

Table 44 provides the risk-of-bias assessment for the new study.

Benefits

The intensity of itching (mean ± SD) in the intention-to-treat population was significantly reduced in the AR-GG27 group (–2.048 mm ± 2.330 mm) compared with the placebo group (–0.388 mm ± 2.22 mm) after 15 days of treatment (p = 0.011). The significant reduction in itch after 15 days was also reported in the population of participants who began the trial with itching (AR-GG27 group n = 18, placebo group n = 11). The severity of eczema (mean ± SD) using the IGA was significantly reduced in the AR-GG27 group (–6.30 ± 3.27) compared with the placebo group (–2.80 ± 3.19) after 15 days of treatment (p = 0.0007). This was also significant after 30 days of treatment.

Harms

There were no serious adverse events reported. There were seven adverse events in six participants in the placebo group, with one case of urticaria and one case of worsening eczema and pityriasis alba. These two events, which were reported as being possibly correlated with treatment, caused discontinuation. It is not reported whether these two events occurred in the same participant or not. Five of the events were not related to study treatment and were mild.

Overall implications for research and practice

This trial provides some evidence of short-term clinical benefit from using AR-GG27 cream compared with placebo. The trial methodology is unclear and it is not known how this treatment compares with any of the current standard treatments for eczema. Also, as the participants had both pityriasis alba and eczema, it is unclear how much beneficial effect would be seen when applied to those with only eczema. If this treatment is trialled again for eczema, particular attention needs to be paid to the adverse effects of the treatment.

Emollients

• Five trials of emollients were reported before 2000, which found no significant difference between using one emollient and using another, some evidence of benefit for using an emollient in addition to a topical corticosteroid compared with a topical corticosteroid alone and evidence of benefit for using an emollient containing urea (10%) compared with the vehicle base, but no evidence of any difference in beneficial effect for different concentrations of urea.

• There were 12 new trials of emollients reported from 2000 onwards:

  • Four trials, three large and one small, with an overall unclear risk of bias, did not provide any evidence of benefit from using one emollient compared with another emollient. Participants were allowed to use topical corticosteroid treatment as well as the emollients in two of the trials.

  • Four trials, three small and one medium sized, with a mostly unclear risk of bias, did not provide evidence of benefit or equivalence for emollient treatment compared with topical corticosteroid treatment.

  • One small trial, with a mostly unclear risk of bias, did not provide any evidence of benefit for a combination of emollient and topical corticosteroid compared with emollient alone.

  • One medium-sized trial, with a mostly unclear risk of bias, provided evidence of benefit for an emollient compared with no treatment. All participants were allowed to use topical corticosteroids.

  • One moderately sized trial, with an overall unclear risk of bias, provided evidence of benefit for an emollient compared with a cleansing wash.

  • One small proactive therapy trial, with an overall unclear risk of bias, provided evidence of benefit for an emollient compared with no emollient after induction of remission with topical corticosteroids.

Bath additives

• One trial of bath additives published before 2000 compared Oilatum^® with Oilatum^® Plus (Stiefel Laboratories) (which contained an added antiseptic) and found some evidence of a beneficial effect of the emollient/antiseptic combination compared with Oilatum alone.

• There were two trials of bath additives reported since 2000:

  • One trial, with a mostly unclear risk of bias, compared a dilute bleach bath and a once-a-month mupirocin treatment of the nares against placebo in children with infected eczema and found a significant beneficial effect for the bleach and mupirocin treatment.

  • Another trial, with a mostly unclear risk of bias, did not find any significant benefit from using a bath additive containing a diamide derivative compared with the same bath additive without the diamide derivative.

• We did not find any trials comparing a bath additive in which there is no antimicrobial component against no bath additive, nor did we find any trials comparing bath emollients against direct application of emollients to the skin after bathing.

Cipamfylline cream

• One trial, with a mostly low risk of bias, provided evidence of a modest benefit from using cipamfylline cream compared with vehicle. However, it also provided evidence that hydrocortisone 17-butyrate was more beneficial than cipamfylline cream.

Camellia oil (Camellia japonica extract)

• One small trial, with a mostly unclear risk of bias, provided evidence of benefit for camellia oil spray for 2 weeks compared with placebo spray.

Furfuryl palmitate (antioxidant)

• One small trial, with an overall low risk of bias, provided evidence of benefit for an emollient when furfuryl palmitate was removed (vehicle) compared with the emollient when furfuryl palmitate was added.

Atopiclair

• There were five trials overall, four funded by the makers of Atopiclair, with a mixed risk of bias. Three of the five trials showed improvements for the Atopiclair group compared with vehicle, one showed no difference between Atopiclair and other available emollients and one trial failed to compare groups.

SRD441 (protease inhibitor)

• One small industry-funded trial, with a mostly low risk of bias, did not provide evidence of benefit for the protease inhibitor SRD441 compared with vehicle.

Vitamin B₁₂

• Two small trials, one with a mostly low risk of bias and one with a mostly unclear risk of bias, provided evidence of benefit for topical vitamin B₁₂ cream compared with vehicle.

WBI-1001 (an inhibitor of T-cell inflammatory cytokine secretion)

• Two small trials, with a mostly unclear risk of bias, provided evidence of benefit for WBI-1001 (0.5%) and (1.0%) compared with vehicle, although the smaller trial did not prespecify outcomes before data lock.

Other topical treatments [Hippophae rhamnoides, black seed oil, pill mask, rosmarinic acid, Vitreoscilla filiformis, shale oil, miltefosine, opiate receptor antagonist, carbohydrate-derived fulvic acid, raffinose, farnesol and xylitol, bacterial antigens, chamomile extract, heparin and levomenol, 15(R/S)-methyl-lipoxin A₄, N-acetyl-L-hydroxyproline, nalmefene hydrochloride monohydrate (SRD174)]

• Each of these treatments were tested in one trial reported from 2000 onwards. None of the trials found any evidence of benefit for the treatment tested compared with placebo or, in the case of licochalcone A, compared with hydrocortisone.

Rationale

The relationship between secondary infection or skin colonisation with the bacterium S. aureus and atopic eczema disease activity has been debated for many years. People with atopic eczema carry S. aureus in about 90% of clinically involved areas and about 75% of clinically uninvolved areas. S. aureus represents about 90% of the total aerobic bacterial flora of such individuals compared with 30% in normal skin. The density of S. aureus tends to increase with the clinical severity of the atopic eczema lesions. It has been suggested that the dry skin of atopic eczema is deficient in certain inhibitory fatty acids, which may encourage growth of the organism. S. aureus may also show enhanced adherence properties to skin cells, which has been shown when comparing atopic eczema sufferers with normal control subjects. 240,241 Other studies reviewed elsewhere242,243 have suggested that the balance of pathogenic and synbiotic bacterial species on the skin is altered in atopic eczema, resulting in an agitated skin microbiome.

Few clinicians would dispute that grossly infected atopic eczema with oozing and sore pus spots requires treatment with some form of antibiotic or antiseptic, and that the bacteria are contributing at least in part to that particular flare-up. However, the role of S. aureus in non-clinically infected atopic eczema skin or for borderline infection (e.g. with just redness and oozing) is far from clear and the definition of what constitutes ‘clinically infected atopic eczema’ among physicians is also not clear. Skin swabs taken for bacteriological culture are of little use because of the almost universal colonisation of atopic eczema skin with S. aureus, although such swabs may reveal additional bacteria such as streptococci species.

If S. aureus does play a pathogenic role in atopic eczema, then this could be mediated in a number of ways including direct chemical irritation, a non-specific reaction of the protein A component of the bacterium to immune cells and by the production of specific exotoxins called superantigens. Superantigens are capable of activating large populations of T lymphocytes distant from the site of colonisation, giving rise to widespread eczematous inflammation.

Although in many cases of non-clinically infected atopic eczema, the presence of S. aureus could be considered as an ‘innocent bystander’, which has simply colonised a dry and broken skin surface, there is at least some rationale for considering the role of S. aureus in more acute forms of atopic eczema. This has led to the use of many antimicrobial compounds, such as oral antibiotics that are active against S. aureus given in short or prolonged courses, topically applied antibiotics and antiseptic agents applied directly or by mixing with emollients applied directly to the skin or within bath additives.

Existing systematic reviews

The efficacy of antimicrobials and antiseptics for eczema has been reviewed in a Cochrane review that was published in 2008 by Birnie and colleagues. 244 This has since been updated in 2010, although not as a Cochrane review, with a new search ending in March 2009. 245 A systematic review in 2007, with a search end date in September 2005, assessed the safety of topical therapies for atopic dermatitis and this included topical antibiotics and antiseptic treatments. 80 All three of the current eczema guidelines from the AAD,94 SIGN42 and NICE41 cover antimicrobials and antiseptics.

Fusidic acid

Fusidic acid (Fucidin^®; Leo Laboratories Ltd) is a bacteriostatic agent that inhibits bacterial protein synthesis. Its biological action is attributed to its effect on Gram-positive bacteria such as staphylococcus and streptococcus species.

Mupirocin

Mupirocin (Bactroban^®; GlaxoSmithKline) is an antibiotic that is bacteriostatic at low concentrations and bactericidal at high concentrations. It is effective against Gram-positive bacteria, including methicillin-resistant Staphylococcus aureus (MRSA).

Tetracycline

Tetracycline is a broad-spectrum antibiotic produced by Streptomyces species of Actinobacteria.

Clarithromycin

One trial by Capella and colleagues251 compared oral montelukast against a ‘standard’ treatment combination of clarithromycin, cetirizine and mometasone furoate. No data were presented for clarithromycin treatment alone and so it is impossible to draw any conclusions about the use of clarithromycin for eczema on its own from this one trial.

Triclosan

This antiseptic and disinfectant is widely used in everyday items such as toothpaste, chopping boards and rubbish bags.

Bleach baths

Common household bleach should not be applied to the skin as it can cause burns on contact and is toxic if ingested. However, research has investigated the use of extremely small amounts of a certain form of bleach. The amount of bleach added to the bath makes the concentration very dilute.

Rationale for antifungal agents

The role of fungi and yeasts in eczema is not clear. Although fungal infections such as athlete’s foot (tinea pedis) can result in secondary eczematisation, secondary fungal infection co-existing or superimposed on atopic eczema lesions is apparently uncommon. However, it has been suggested that allergic sensitisation to Malassezia yeast species, which is common on the scalp and head, may contribute to some patterns of atopic eczema affecting the head and neck in adults. Although the role of fungi and yeasts in atopic eczema is tenuous, some have used antifungal agents combined with topical corticosteroids and some have used antifungals in shampoo or tablet form in an attempt to improve atopic eczema, and these will be discussed at the end of this chapter.

Ketoconazole (oral and topical)

Ketoconazole is an antifungal agent available in oral and topical forms. The topical form (Nizoral^® shampoo; Janssen-Cilag) is used in shampoos for dandruff and scalp psoriasis but is not currently licensed for the treatment of eczema.

In July 2013, the European Medicines Agency’s Committee for Medicinal Products for Human Use issued a statement that oral medicines containing ketoconazole should no longer be used for the treatment of fungal infections because of safety concerns regarding the risk of developing hepatotoxicity and adrenal insufficiency, and the potential for fatal drug interactions. 254 As such, the use of ketoconazole in developed countries has been largely superseded by newer azoles, such as itraconazole (Sporanox^®; Janssen-Cilag), because of a lower risk of liver toxicity and drug interactions.

Miconazole (topical)

Miconazole is an imidazole agent that is used topically to treat fungal infections. When taken orally it has also been shown to be effective against some forms of leishmaniasis and it also has some antibacterial properties.

Itraconazole (oral)

Itraconazole belongs to the triazole group of antifungal medications. It has a broad spectrum of action against fungi including yeasts and dermatophytes. It is similar to fluconazole but also treats Aspergillus infections.

Ciclopirox olamine (topical)

Ciclopirox olamine is a hydroxypyridine antifungal agent. It has been shown to be highly effective against Malassezia species, which have been implicated in difficult-to-treat atopic eczema in the head and neck area.

Antibiotics

Antiseptics

Antifungals

Existing systematic reviews

The NICE,41 SIGN42 and AAD94 guidelines and associated evidence reviews cover antihistamines. A systematic review covering the safety of eczema treatments80 also covers topical doxepin. A systematic review of interventions to reduce itching for eczema93 also covers many of the antihistamines in this chapter.

Cetirizine (oral) (less sedating)

Cetirizine (Zirtek^™; UCB Pharma) is a potent antihistamine and is used in adults and children.

Loratidine (oral) (less sedating)

Loratidine is a non-sedating second-generation H1 antihistamine used to treat allergies. In the UK it is sold over the counter (Clarityn^®; Merck Sharp & Dohme Ltd) as well as being available on prescription.

Fexofenadine (oral) (less sedating)

Fexofenadine (Telfast^®; Sanofi) is used for hay fever and other allergic conditions with similar symptoms. It is not as sedating as some other H1 antihistamines.

Ketotifen fumarate and epinastine hydrochloride (oral) (less sedating)

Ketotifen fumarate (Zaditen^®; Swedish Orphan Biovitrum) functions as a mast cell stabiliser and has been used as a treatment for chronic idiopathic urticaria because of its antipruritic properties. Similarly, epinastine hydrochloride is both an antihistamine and a mast cell stabiliser.

Olopatadine hydrochloride (oral) (non-sedating)

Chlorpheniramine (oral) (sedating)

Chlorpheniramine (Piriton^®; GlaxoSmithKline) is a relatively weak sedative antihistamine that is used for the prevention of rhinitis and urticaria.

Doxepin (topical) (sedating)

Doxepin (a tricyclic antidepressant) has powerful antihistamine properties by antagonising the H1 and H4 receptors. It is available in oral and topical formulations (Xepin^®; Cambridge Healthcare Supplies).

Sodium chromoglycate (topical)

Sodium chromoglycate has been used as an inhaled powder for the treatment of asthma for over 30 years and has a very strong safety profile. The mechanism of action is partly the result of the drug inhibiting the release of inflammatory mediators from mast cells. It is now being investigated as a treatment for diseases such as eczema, for which it is added to topical preparations.

Cetirizine (oral) (less sedating)

• Four trials, two small and two medium sized, involving cetirizine for eczema were reported before 2000. The largest of these trials provided evidence of benefit for cetirizine but only at four times the normal dose. The other three trials did not provide any evidence of benefit.

• One very large trial published in 2005, with an overall unclear risk of bias, provided evidence of no benefit of long-term twice-daily cetirizine (0.25 mg/kg) treatment.

Loratadine (oral) (less sedating)

• Three trials involving loratadine were reported before 2000. Two trials, one small and one very small, with an overall unclear risk of bias, provided evidence of benefit for loratadine compared with placebo. The largest trial did not provide any evidence of benefit for loratadine compared with cetirizine.

• Only one new trial, published in 2008, used loratadine for the treatment of eczema, comparing loratadine with and without the addition of modified Jiawei Danggui Decection (a type of Chinese medicine). As loratadine was used in both groups, this trial cannot be used to assess the effectiveness of loratadine.

Fexofenadine (oral) (less sedating)

• No trials of fexofenadine were reported before 2000.

• One large trial reported in 2006, with an overall unclear risk of bias, provided evidence of benefit for fexofenadine (60 mg twice daily) compared with placebo.

• A further medium-sized trial reported in 2006, with an overall unclear risk of bias, did not provide any evidence of benefit for fexofenadine (30 or 60 mg twice daily) compared with ketotifen (1 mg once daily).

Ketotifen and epinastine (oral) (less sedating)

• Two small trials on ketotifen compared with placebo reported pre 2000 gave conflicting results.

• There were no trials involving epinastine for eczema reported before 2000.

• One medium-sized trial reported in 2003, with an overall unclear risk of bias, provided evidence of the non-inferiority of ketotifen to epinastine.

Olopatadine hydrochloride (oral) (non-sedating)

• There were no trials involving olopatadine hydrochloride for eczema treatment reported before 2000.

• One large trial reported in 2011, with an overall unclear risk of bias, did not show a difference for children treated with olopatadine hydrochloride compared with ketotifen fumarate in terms of itch or eczema severity.

Chlorpheniramine (oral) (sedating)

• One very small trial involving chlorpheniramine, with missing baseline data, was reported pre 2000; however, this trial did not compare the chlorpheniramine-only group with the placebo group and so did not provide any information on the possible benefit of chlorpheniramine compared with placebo.

• One medium-sized trial reported in 2002, with an overall unclear risk of bias, did not provide any evidence of benefit for chlorpheniramine compared with placebo.

Doxepin (topical) (sedating)

• Four fairly well-reported manufacturer-sponsored trials involving topical doxepin, two large and two small, were reported before 2000. Two of the trials provided evidence of benefit for topical doxepin compared with placebo and two did not.

• One small trial reported in 2006, with an overall unclear risk of bias, provided evidence of benefit for doxepin compared with placebo.

Sodium chromoglycate (topical)

• Ten trials of sodium chromoglycate were reported before 2000.

• One new medium-sized trial, with a mostly low risk of bias, provided evidence of benefit for sodium chromoglycate (4%) compared with vehicle.

Existing systematic reviews

There have been 20 systematic reviews of dietary interventions for established eczema. Five reviews79,272–275 specifically examine probiotic supplementation, a Cochrane review276 and subsequent non-Cochrane updated review277 specifically evaluate dietary exclusions, another Cochrane review278 specifically explores dietary supplementation and other reviews35,41,42,93,94,279–283 cover more than one dietary intervention. A Cochrane review of evening primrose oil and borage oil supplementation has also recently been published. 189

Mixed-strain probiotics

Single-strain probiotics

Overall implications for research and practice for mixed-strain and single-strain probiotics

There is reasonably strong evidence from three trials284–286 that mixtures of probiotic strains do not show any benefit over placebo for children with eczema, whether or not they have symptoms of cow’s milk allergy or are sensitised to one or more allergens. All three trials undertook subgroup analyses that were not prespecified and reported a significant benefit for those children with raised IgE levels. Only one small trial comparing a mixture of four different probiotic strains compared with skimmed milk powder in children reported a significant difference in change in disease severity between the probiotic group and the placebo group. 287 However, although not compared statistically, the severity of eczema in the active treatment group was noticeably higher than in the placebo group at baseline and it was unclear whether this imbalance had been adjusted for in the statistical comparisons. In the absence of any trials specifically designed and powered to test mixtures of probiotic strains in people with eczema who are proven as ‘atopic’ with raised circulating IgE antibodies, it is impossible to assess whether probiotics are of true benefit over placebo. In reality, routine tests for atopy are not conducted, especially in primary care, and with up to two-thirds of those with eczema not ‘atopic’ there is only limited potential for mixtures of probiotic strains for the treatment of established eczema.

Probiotic strains of lactobacilli mostly do not show any good evidence of significant benefit over placebo in addition to standard treatments for eczema over a period of 8–12 weeks. Most of the trials do not give enough information about the methodology used, especially for allocation concealment, blinding and which outcome and statistical analyses were prespecified. The trials have mostly been small and often raise questions about the generalisability of the results, as selective trial populations have been studied. All five trials on L. rhamnosis have not shown any evidence of clinical benefit for eczema compared with placebo and probably now equate to evidence of no benefit, given the overall weight of evidence. One moderately sized trial involving L. paracasei in unselected infants provided no evidence of benefit of this treatment compared with placebo. Another trial involving L. plantarum provided some weak evidence of benefit compared with placebo. The only potential evidence of benefit may be for people with truly ‘atopic’ eczema, as suggested by a few reported subgroup analyses; however, this requires further specific research as none of these subgroups appears to have been prespecified. Even though there is a large amount of ‘noise’ in these trials because of a lack of clarity of reporting, it seems likely that, with the majority of trials on lactobacilli probiotics not showing any significant benefit for the treatment of eczema, there is no beneficial effect to be found.

Studies

No RCTs involving prebiotics were reported before 2000.

Two new trials have been published since 2000. A trial conducted in Japan by Shibata and colleagues305 included children under 3 years of age with atopic eczema defined according to the criteria of the Japanese Dermatological Association. 26 The trial compared oral kestose, an oligosaccharide that encourages the growth of bifidobacteria, every day for 12 weeks against no treatment. The participants were able to use topical corticosteroids during the trial.

A trial by Ghanei and colleagues306 carried out in Iran compared fructo-oligosaccharides and inulin in powder added to milk against a placebo of dextrin powder. The trial included 90 children aged between 7 and 24 months who were full term and a normal birth weight. The children had been delivered by caesarean section and diagnosed with eczema by a physician and had started solid food by the age of 6 months. The treatment was given every day for 90 days (5 g for those aged 7–12 months, 7.5 g for those aged 13–18 months and 10 g for those aged 19–24 months).

Assessment of risk of bias

Table 63 provides the risk-of-bias assessment for the new studies.

Benefits

The trial by Shibata and colleagues305 showed a significant reduction in eczema severity for the participants using the kestose probiotic compared with those using the placebo after 12 weeks, with a median SCORAD score in the kestose group of 19.5 (range 25.6–22.0) and in the placebo group of 37.5 (range 25.5–43.5; p < 0.001). The baseline SCORAD values were 41.3 (range 36.2–46.4) in the kestose group and 38.3 (range 26.3–41.8) in the placebo group and, although the change in severity of each group was not compared, it does seem to show a greater decrease in the kestose group; however, no analysis for this was performed. This apparent benefit was reflected in the measurement of eczema severity using the Intensity of Atopic Dermatitis (maximum score 12 points), although, again, no analysis of the difference between the groups in change in severity was carried out.

The trial by Ghanei and colleagues306 reported a significant difference between the groups in eczema severity at the end of treatment; however, again, no comparison was made of the change in eczema severity between the groups over the treatment period. This change was reported as significant for both groups and so it is unlikely that there is a significant benefit from the fructo-oligosaccharide and inulin powder compared with placebo for the treatment of eczema.

Harms

The trial by Shibata and colleagues305 did not report information on adverse events. In the trial by Ghanei and colleagues306 there was one report of diarrhoea lasting for > 1 week in the placebo group; no other information was reported.

Overall implications for research and practice

The lack of information on the relative use of other eczema treatments by participants and the appropriate comparisons of eczema severity, coupled with a large difference in baseline eczema severity in one trial, make interpretation of the clinical significance of these trials very difficult. The relatively large number of withdrawals of those not taking the treatment for > 75 out of 90 days in the trial by Ghanei and colleagues306 may be masking potential problems with the treatment, such as unpalatability or even adverse events. Trials that pay detailed attention to randomisation, allocation concealment and blinding of all stakeholders, especially the outcome assessors and the participants, are needed. The age range of the participants in these trials also poses a problem as there is a large change in the prevalence of eczema over the age range from 6 months to 3 years.

Studies

No trials involving synbiotics were reported up to the year 2000.

Eight trials307–314 involving synbiotics were reported from 2000. The eight trials on synbiotics all tested different combinations of probiotics and prebiotics. All but two trials309,313 compared the synbiotics against an inactive placebo or a placebo of the hydrolysed formula given to both groups. These other two trials compared the synbiotic against the probiotic or the prebiotic only.

Assessment of risk of bias

Table 64 provides the risk-of-bias assessment for the new studies.

Benefits

Farid and colleagues312 found a significant difference in change in eczema severity according to SCORAD scores from baseline to 4 weeks [treatment group mean –29.51 (SD 19.09) vs. control group mean –11.06 (SD 10.96); p = 0.001] and 8 weeks [treatment group mean –39.2 (SD 24.22) vs. control group mean –20.10 (SD 8.63); p = 0.005], but the difference was not significant between week 4 and week 8.

Gerasimov and colleagues307 found a significant difference in change in eczema severity from baseline after 8 weeks of treatment [L. acidophilus DDS-1 and B. lactis UABLA-12 combined with fructo-oligosaccharide group mean –14.2 (SD 9.9) vs. placebo group mean –7.8 (SD 7.7); p = 0.001]. It is not clear whether the analysis was carried out on an intention-to-treat basis including the five participants in the treatment group and one in the placebo group who withdrew during the trial.

Van der Aa and colleagues308 compared a mixture of B. breve M-16V and galacto-/fructo-oligosaccharide against placebo and found no significant difference in change in eczema severity after 12 weeks in the whole trial population. There was, however, a significant difference in change in severity within the IgE-associated eczema subgroup of 45 participants (synbiotic group –18.1 SD ± 1.6 vs. control group –13.5 SD ± 1.6; p = 0.04), although it was not clear whether this was a preplanned analysis.

Hattori and colleagues310 also found a significant difference in change in eczema severity from baseline, measured using cutaneous signs, between B. breve M-16V with raffinose and placebo (p = 0.0344). This analysis involved only 15 participants and the median change in score from baseline was –3.5 (range 0 to –6) for the synbiotics group compared with –1 (range 0 to –4) for the placebo group.

Murosaki and colleagues311 and Passeron and colleagues309 reported no significant difference in change in severity of eczema, itching or quality of life for synbiotic treatment compared with placebo311 or probiotic. 309 Shafiei and colleagues314 reported no significant difference in change in eczema severity between synbiotic and placebo treatment for 2 months.

Wu and colleagues313 compared a mixture of L. salivarius plus prebiotic (fructo-oligosaccharide) for 8 weeks against prebiotic alone in 60 children with moderate to severe eczema. They found a significant difference in eczema severity measured by the SCORAD index between the experimental group and the control group at 8 weeks (27.4 vs. 36.3; p = 0.022). No significant differences were found between groups for quality of life assessed by an unvalidated scale or for medication use.

Harms

These trials reported a range of levels of adverse events. Only the trial by Hattori and colleagues310 did not report any information about adverse events. Serious adverse events in the probiotics group consisted of one burn and one croup and those in the placebo group consisted of one head injury and two food poisoning cases. 307 Two serious adverse events that resulted in hospitalisation were a case of respiratory syncytial virus and a severe allergic reaction to cow’s milk. 308 Farid and colleagues312 reported no significant adverse events but did not give any further data on adverse events. Passeron and colleagues309 reported no serious adverse events and three cases of abdominal pain, two in the synbiotics group and one in the probiotics group. Murosaki and colleagues311 reported one case of nausea in the placebo group. Van der Aa and colleagues308 reported that 91.1% of participants in the synbiotic group and 84.1% of participants in the placebo group experienced adverse events but that none of these was considered to be related to treatment; one participant used antibiotics in the synbiotics group and five participants used antibiotics in the placebo group, but this was not significantly different between the groups. Gerasimov and colleagues307 reported that 60.5% (n = 26) in the synbiotics group and 51.5% (n = 24) in the placebo group experienced adverse events. The trial by Shafiei and colleagues314 did not specifically report adverse events; however, two children were withdrawn because of diarrhoea in a twin. Wu and colleagues313 reported mild diarrhoea in two patients in the synbiotics group.

Overall implications for research and practice

A feature of these trials appears to be the level to which ‘data mining’ is taking place, that is, trying to find a significant result using outcomes and subgroups that were not originally planned. The evidence is conflicting over whether synbiotics are beneficial for eczema. This may be because of the differences in the combinations of probiotics and prebiotics tested; however, the lack of methodological detail and small trial populations also mean that the evidence is not convincing.

Borage oil

Borage oil is extracted from the seeds of the borage plant (Borago officionalis) and has high levels of omega-6 fatty acids, higher than evening primrose oil and blackcurrant seed oil.

Evening primrose oil (oral)

Fish oil (omega-3)/soybean oil (omega-6)

Fish oils contain the omega-3 fatty acids EPA and DHA. EPA and DHA are precursors to a number of substances that have been shown to reduce inflammation. Soybean oil is one of the most widely consumed cooking oils and contains high levels of omega-6 fatty acids.

Docosahexaenoic acid

This acid belongs to the group of n-3 polyunsaturated fatty acids and inhibits T-cell activation and proliferation and reduces the numbers of granulocytes circulating in the blood.

Hempseed oil

Up to 35% of the weight of hempseed is an edible oil with a high EFA content (approximately 80%), including omega-6 (55%) and omega-3 (22%) fatty acids. This oil is extracted and often used as a dietary supplement.

Overall implications for research and practice

The large fairly well-reported trial on borage oil315 showing no evidence of benefit for objective and participant-assessed outcomes, similar to the results found in a previous large trial,320 adds to the weight of evidence showing no beneficial effect of borage oil for eczema. The questions around whether there is a subgroup of the population with eczema who may respond to borage oil and whether increased adherence may show beneficial effects have not been addressed.

The additional small study on oral evening primrose oil316 showed a significant benefit of taking evening primrose oil compared with a placebo of sunflower oil. The potential bias introduced by analysing only the first 25 participants from each treatment group, and the lack of clarity about whether the outcome assessors were blinded to treatment allocation, means that this result must be treated with caution. This trial does not change the overall body of evidence on evening primrose oil for eczema, with the largest and best-reported trials not providing any evidence of benefit and other trials providing conflicting results (ranging from no hint of benefit to a 10–20% improvement).

A small pilot trial317 showed a short-term but marked improvement in eczema in a small number of participants treated with fish oil and a slower to develop, but longer-lasting improvement in participants treated with soybean oil. It is not easy to interpret whether these results are clinically significant because of the lack of a control group of standard care and a placebo. A concern with this trial is the addition of new randomised participants when participants were withdrawn as it is not clear which participants were included in the final analysis. Further trials on these treatments should evaluate whether they can be taken orally as infusions may not be as readily acceptable to patients and take more resources to administer.

The trial on DHA318 showed a hint of a modest beneficial effect of DHA compared with EPA on eczema severity after the treatment had been stopped. Larger, longer-term trials are needed that concentrate on participant-assessed outcomes such as quality of life and acceptability of treatment, as well as on objective and subjective measures of severity.

The very small pilot trial on hempseed oil,319 in which the participants would have very easily been able to determine which treatment they were taking, did not show any significant benefit from hempseed oil for skin dryness and itchiness as assessed by the participants.

Studies

No trials reported before 2000 involved oral vitamin D as a treatment for eczema.

Three trials were reported after 2000, with one trial examining oral ergocalciferol (vitamin D₂),326 one examining cholecalciferol (vitamin D₃) and one examining vitamin D₃ and vitamin E supplementation.

The small pilot trial by Sidbury and colleagues,326 reported in a letter, compared 1000 IU per day of ergocalciferol with placebo for 1 month in children aged 2–13 years. The eczema must have had a winter onset or exacerbation and the trial was conducted in the winter. All but one child had mild eczema (EASI score between 10 and 18.6). The 11 children randomised into the trial were allowed to continue all current treatments but were not allowed to start any new treatments and were not allowed to travel to temperate climates during the trial.

The trial by Amestejani and colleagues327 compared 1600 IU per day of cholecalciferol for 2 months with placebo in adults and adolescents aged ≥ 14 years. In total, 60 participants diagnosed with eczema using the Hanifin and Rajka8 criteria and with a mean SCORAD score of 25 were randomised; however; only the 54 participants who completed the study were included in the analysis. Usual treatments (emollients, topical corticosteroids, oral antihistamines) were permitted during the study period in both treatment groups.

The trial by Javanbakht and colleagues328 randomised 52 adult participants with eczema according to the Hanifin and Rajka8 criteria into four treatment groups: 1600 IU per day of vitamin D₃ taken concurrently with placebo; 600 IU per day of vitamin E taken concurrently with placebo; 1600 IU per day of vitamin D₃ taken concurrently with 600 IU per day of vitamin E; and placebo. The participants took the treatments once daily, separately, with a meal (vitamin D or placebo as one capsule, vitamin E or placebo as two soft gels). All participants could use standard prescription eczema treatments during the trial.

Assessment of risk of bias

Table 70 provides the risk-of-bias assessment for the new studies.

Benefits

In the study by Sidbury and colleagues,326 four out of five children who took vitamin D₂ and one out of six children in the placebo group improved by one category on the IGA scale compared with baseline. One out of five children who took vitamin D₂ and four out of six children in the placebo group did not change score on the IGA scale compared with baseline. One out of six children in the placebo group worsened by one category on the IGA scale.

In the trial by Amestejani and colleagues,327 the efficacy of vitamin D₃ was not statistically compared with that of placebo in the trial report, making it impossible to assess the results of the trial. Disease severity in the vitamin D₃ group measured by the SCORAD index showed a significant improvement between baseline and the end of the study (from 24.8 to 15.3; p = 0.01). Likewise, disease severity in the vitamin D₃ group measured by the TIS score also showed a significant improvement. No significant reduction in eczema severity was observed in the placebo group.

In the study by Javanbakht and colleagues,328 four small groups were included (n = 13) yet the trial failed to compare the changes in disease severity between the groups. The rates of SCORAD score improvement from baseline were 34.8% for the vitamin D₃ plus placebo group, 35.7% for the vitamin E plus placebo group, 64.3% for the vitamin D₃ plus vitamin E group (p = 0.004) and 28.9% for the placebo group. The changes in the pruritus VAS in the four groups were from 5.7 to 5, from 6.2 to 3.4, from 5.5 to 2 and from 7 to 4, respectively. The changes in the sleeplessness VAS were from 2.3 to 1.4, from 4.3 to 1.3, from 3.2 to 1 and from 1.1 to 1, respectively. The percentage decrease in topical corticosteroid use was 66.8%, 70.2%, 88.7% and 37.5%, respectively (p = 0.05).

Harms

Information about adverse events was not reported for these trials.

Overall implications for research and practice

These trials are too small and have not been analysed appropriately to provide any good evidence about whether oral vitamin D or vitamin E supplementation is beneficial or not for eczema. However, there is enough of a signal that a full trial of oral vitamin D or vitamin E supplementation may be worth pursuing. Without additional research evidence, the role of vitamin D or vitamin E supplementation remains unclear.

Studies

Goat’s milk and ass’s milk had not previously been tested as treatments for eczema before 2000.

One new trial involving goat’s and ass’s milk was reported after 2000. This crossover trial by Vita and colleagues329 randomised 28 children aged from 6 months to 3 years with active (SCORAD score of > 20) atopic eczema and a clinical history of cow’s milk allergy. The children were randomised to goat’s milk or ass’s milk for 6 months and then switched to the other milk for a further 3 months.

Assessment of risk of bias

Table 71 provides the risk-of-bias assessment for the new study.

Benefits

This small trial329 found a large magnitude of effect for ass’s milk compared with goat’s milk, with a median SCORAD score reduction of 34 points for ass’s milk and 11 points for goat’s milk. Despite the participants being aware of the milk that they were drinking, their assessment of skin symptoms also showed the same result. Treatment with ass’s milk resulted in a 4.5-point decrease on a VAS whereas treatment with goat’s milk resulted in a 0.5-point decrease. Treatment with goat’s milk also resulted in 23 out of 26 participants having a positive reaction to a double-blind placebo-controlled food challenge whereas only one participant out of 26 had a positive reaction to ass’s milk.

Harms

No specific details about adverse events were reported; however, it was stated that one participant had to withdraw because of a systemic reaction to goat’s milk involving shortness of breath, sneezing and severe generalised urticaria.

Overall implications for research and practice

This one small crossover trial329 in a very specific eczema population provides some interesting but fairly weak evidence against the use of goat’s milk for children with eczema and a cow’s milk allergy because of a lack of a beneficial effect and high levels of development of sensitivity to the goat’s milk. Ass’s milk appears to show a large beneficial effect in this population but this needs further confirmation in trials comparing ass’s milk with the commonly used cow’s milk substitute of soy milk.

Studies

Hypoallergenic milk formula had not previously been tested as a treatment for eczema before 2000.

Two new studies were reported after 2000. 330,331 A crossover trial by Leung and colleagues330 involving hypoallergenic formula was reported in 2004. This trial randomised 15 children aged < 3 years and taking at least 500 ml of soy or cow’s milk-based formula to a hypoallergenic formula that was lactose and sucrose free (Neocate^®, SHS International) or the child’s pre-existing milk formula. Each formula milk was given for 6 weeks in a randomised order.

Jin and colleagues331 conducted a 12-week study comparing a partially hydrolysed whey and casein formula milk with conventional cow’s milk formula in 113 babies with mild to moderate eczema according to the Hanifin and Rajka8 criteria. Twenty-seven infants dropped out of the study (16 infants in the experimental group because of the taste of the formula being unacceptable to them, the introduction of solid food, or suffering from other severe diseases that required additional therapy and 11 infants in the control group because of their eczema not improving).

Assessment of risk of bias

Table 72 provides the risk-of-bias assessment for the new studies.

Benefits

In the Leung study,330 eczema severity, measured using the SCORAD index, showed a significant treatment by period interaction of 7.23 points (p = 0.012); however, it did not did not show a statistically significant treatment difference (3.97 points; p = 0.274). For the individual components of the SCORAD index, intensity and pruritus showed a statistically significant treatment by period interaction (p = 0.036 for both). The caregiver global health assessment did not show a statistically significant difference for either the treatment by period interaction or the treatment difference.

In the Jin study,331 there was a significant difference between the groups in eczema severity measured by the SCORAD index and the Japanese Dermatological Association Scoring System after 12 weeks. Eczema severity, measured using the SCORAD index, was 13 in the experimental group and 20 in the placebo group, resulting in a significant difference (p = 0.001). The number of flare-ups was significantly decreased in the experimental group but not in the control group (but results were not compared between the groups). The trial report did not state clearly the disease severity at baseline in the two groups.

Harms

These trials did not report whether or not there were any adverse events. In the trial by Jin and colleagues331 there were no differences in nutritional status between groups after 12 weeks.

Overall implications for research and practice

These two trials have given some evidence of the benefit of hypoallergenic formulas for the reduction of eczema severity, especially in very young children. More independent studies are needed to confirm these results.

Probiotics

• There were no trials involving probiotics for eczema treatment reported up to 2000.

• Twenty-one trials were conducted after 2000 and showed varied results depending on the strains of probiotics examined:

  • Five small trials, with a mostly unclear risk of bias, provided evidence of benefit for mixed-strain probiotics compared with placebo, mostly in post hoc subgroups of participants with raised IgE levels.

  • Five trials, three small, one medium and one large, with a mostly unclear risk of bias, did not provide any evidence of benefit for L. rhamnosis GG compared with placebo or vehicle.

  • One small trial, with an overall unclear risk of bias, provided evidence of benefit for L. sakei KCTC 10755BP compared with placebo.

  • One small trial, with an overall unclear risk of bias, did not provide any evidence of benefit for L. fermentum VRI-033 PCC compared with placebo.

  • One small trial, with an overall unclear risk of bias, did not provide any evidence of benefit for L. acidophilus NCFM (ATCC 700396) compared with placebo.

  • One small trial, with an overall unclear risk of bias, did not provide any evidence of benefit for heat-killed L. paracasei K71 compared with placebo.

  • One large trial, with an overall unclear risk of bias, did not provide evidence of benefit for L. paracasei CNCM I-2116 compared with placebo in infants.

  • One small trial, with a mostly unclear risk of bias (low risk of bias for the method of randomisation), provided evidence of benefit for L. acidophilus L-92 compared with placebo.

  • One medium trial, with a mostly low risk of bias, provided evidence of benefit for L. plantarum CJLP133 compared with placebo.

  • One small trial, with an overall unclear risk of bias, did not provide any evidence of benefit for L. salivarius LS01 compared with placebo.

  • One small trial with an overall unclear risk of bias, did not provide any clear evidence for live lysophilised B. breve YY versus placebo in adults.

  • One small trial, with an overall unclear risk of bias, provided evidence of benefit for B. lactis Bb-12 supplemented formula or L. rhamnosis GG supplemented formula compared with unsupplemented formula in previously exclusively breastfed infants.

  • One small trial, with an overall unclear risk of bias, did not provide any evidence for B. breve M-16 supplemented formula compared with unsupplemented formula in participants with a proven allergy to cow’s milk.

Prebiotics

• There were no trials involving prebiotics (substances that promote the growth of beneficial micro-organisms in the gut) for treating established eczema reported up to 2000.

• Two small trials reported after 2000, one with a mostly low risk of bias and the other with an overall unclear risk of bias, provided evidence of benefit for prebiotic treatment compared with placebo.

Synbiotics

• There were no trials involving synbiotics (a combination of probiotics and prebiotics) for eczema reported up to 2000.

• Eight trials were reported after 2000 and provided mixed evidence of benefit for synbiotics:

  • Six small trials, with a mostly unclear or low risk of bias, provide mixed evidence for synbiotics compared with placebo. Three trials provided evidence of benefit and three trials did not. The two largest trials, with a mostly low risk of bias, gave conflicting results.

  • One small trial, with an overall unclear risk of bias, did not provide any evidence of benefit for synbiotic treatment compared with probiotic treatment.

  • One small trial, with a low risk of bias, provided evidence of benefit for synbiotic treatment compared with prebiotic treatment alone in children.

Essential fatty acid supplementation

Oral vitamins D and E

• There were no trials involving oral vitamin D (ergocalciferol or cholecalciferol) for eczema reported up to 2000.

• Three small trials were reported after 2000 and did not provide evidence of benefit of oral vitamin D or E:

  • One very small pilot trial, with a mostly unclear risk of bias, did not provide any evidence of benefit for ergocalciferol compared with placebo.

  • Two small trials, with a low or unclear risk of bias, did not provide sufficient evidence of benefit for cholecalciferol compared with placebo or vitamin E or cholecalciferol plus vitamin E in adolescents and adults.

Goat’s/ass’s milk

• There were no trials involving goat’s or ass’s milk for eczema reported up to 2000.

• One small trial reported in 2007, with a high risk of bias for blinding, provided evidence of benefit for ass’s milk compared with goat’s milk. The participants had eczema and a clinically relevant cow’s milk allergy.

Hypoallergenic formula

• There were no trials involving hypoallergenic formula for established eczema reported up to 2000.

• One very small trial reported in 2004, with a high risk of bias for blinding, provided evidence of benefit for hypoallergenic formula compared with infants’ previous soy- or cow’s milk-based formula.

• One moderately sized trial reported in 2011, with an overall unclear risk of bias, provided evidence of benefit for hypoallergenic formula compared with conventional cow’s milk formula in infants.

Existing systematic reviews

There have been nine systematic reviews35,244,245,332–337 and three guidelines (NICE,41 SIGN42 and AAD94) that include non-pharmacological interventions. Eight of these41,42,94,332–336 assessed psychological and/or educational interventions for eczema; five41,42,94,245,337 assessed textiles and/or specialised clothing for eczema; and four35,41,42,94 assessed the reduction of allergens, mostly aeroallergens.

Studies

Three trials involving specialised clothing were reported before 200055 (see Appendix 3).

Silk

Three new trials assessing the efficacy of DermaSilk^® fabric (Espère Healthcare Ltd) were identified. 338–340

In 2007, Koller and colleagues338 published a trial comparing DermaSilk arm tubes against ordinary silk arm tubes. Twenty-two children with mild to moderate eczema were included in the trial. For the initial 2 weeks, arms were covered in either silk fabric or DermaSilk fabric; for the remaining 10 weeks, one arm was covered with a cotton arm tube and the other with a DermaSilk tube. The fabric tubes were randomised to be worn on the left or right arm so that the interventions were compared within the same person. The fabric tubes were worn all day and washed daily. The severity of the eczema in the area covered by the fabric tubes was measured using local SCORAD scores by an assessor who was not aware of the allocation of the interventions. Participants in the study were not allowed to use any antimicrobial or anti-inflammatory treatments during the trial.

A double-blind, within-person randomised trial by Stinco and colleagues339 published in 2008 compared DermaSilk sleeves against the same fabric without the impregnated antimicrobial for 28 days in 30 children and adults with eczema. The ages of the participants ranged from 3 to 31 years. The fabric arm tubes were identical apart from the seam colour. No one with an acute infection was included in this trial. Local SCORAD scores for the arm and participant-assessed pruritus were measured up to 28 days.

A trial by Fontanini and colleagues340 compared a DermaSilk long-sleeve top and trousers against a cotton long-sleeve top and trousers. Twenty-two infants who were aged < 18 months and who had eczema were randomised to wear the garments every day for 2 years, except in the summer and on other very hot days. All of the families taking part were given mite-impermeable mattress and pillow encasings for both the parents’ and the children’s beds. All of the children were treated with mometasone furoate as needed during the trial.

Ethylene vinyl alcohol fibre

Two new small studies341,342 from Japan reported since 2000 looked at ethylene vinyl alcohol fibre fabric as underwear and compared it with cotton underwear.

Thirty participants who were being well maintained on standard eczema treatment participated in the 8-week crossover study by Ozawa and colleagues. 341 Each participant wore each fabric (ethylene vinyl alcohol fibre and cotton) for 4 weeks without a washout period in between the different fabrics. It is not known how much of the day and night the participants were instructed to wear the underwear and how often they were asked to wash it.

A small parallel-group trial342 of 24 children aged 3–9 years also compared ethylene vinyl alcohol fibre fabric with cotton underwear. The children were not allowed to use topical corticosteroids throughout the study.

Textiles with added silver

Garments with added tourmaline

A trial by Kim and colleagues345 compared undergarments made of anion textile, pure polyester fibres containing nanoparticles of crushed tourmaline, with undergarments made of pure cotton. Fifty-two adults and children aged between 2 and 30 years with mild to severe eczema were randomised to wear the undergarments all of the time for 4 weeks. The use of topical or systemic treatments for eczema during the trial was not permitted.

Overall implications for research and practice

The study of specialised clothing is challenging because of the inability to blind participants. Interpreting the clinical significance of participant-rated outcomes in the presence of a lack of blinding raises concerns about information/detection bias as patients often have high expectations of benefit from these garments. The studies on specialised clothing collated for this review provide some evidence of potential benefits without any harms. Unfortunately, when potential bias in the subjective results, because of most participants being aware of their treatment allocation, is taken into account, the results have to be interpreted with caution. The objective measures of severity assessed by blinded outcome assessors can be given more weight. DermaSilk showed statistically significant benefits for the severity of eczema in comparison to ordinary silk and cotton and DermaSilk fabric without antimicrobial, but the time period of the trials was very short for such a chronic, long-term condition. The numbers of participants included in the trials were low and the absolute point differences were so small that they may not be clinically meaningful. The X-STATIC silver textile trial is difficult to interpret because of confounding from topical corticosteroid use. One participant in this trial had silver in their urine, which had disappeared by the end of the trial. The ethylene vinyl alcohol fibre did not show even a hint of efficacy for eczema, although this has been assessed only in a Japanese population to date. The silver textile Padycare and the anion textile with tourmaline both resulted in a significant difference in participant-rated improvement in pruritus, a measure that is prone to detection bias for an unblinded intervention. The anion textile with tourmaline also provided evidence of significant benefit for the severity of eczema compared with cotton, but this was not significant when only the objective severity measures were assessed.

Specialised clothing is a relatively expensive treatment both in direct monetary terms and in terms of hidden indirect costs (for increased washing of clothing, having to have multiple changes of garments and rapidly growing children needing bigger sizes). For clinicians, guideline writers and people with eczema and their families to be confident that this clothing has a favourable cost–benefit ratio, more rigorous long-term trials that compare these textiles against standard care in a pragmatic setting are needed. The CLOTHing for the relief of Eczema Symptoms (CLOTHES) trial (ref. HTA 11/65/01)346 is currently running in the UK, comparing the addition of silk clothing to normal eczema treatment against normal eczema treatment alone.

Adults

Children with eczema and their parents and carers

Dermatology nurse consultations

The role of nurses in eczema, as well as for other dermatological conditions, has evolved in the last 10 years, which has resulted in specialist nurses taking on a much more autonomous role within clinics. For some specialist nurses, such as nurse consultants, this has meant having their own list of patients to manage through consultations.

Support groups

Support groups of people who have or who care for someone with a common condition are often formed both through the NHS and outside it. These groups offer a good diversity of methods of support and some of the larger groups offer helplines and written or online educational information. In the UK, the National Eczema Society is the largest national support group for those with eczema.

E-health portal

As access to the internet increases, more health services are exploring ways of offering all or part of their service online. Direct access to care via the internet, termed ‘e-health’ or ‘telemedicine’, has been shown to have small to moderate benefits on health outcomes for the management of chronic illnesses.

Studies

A trial by Schut and colleagues361 conducted in Germany compared a standardised cognitive behavioural therapy stress management programme with a waiting list control group. Twenty-eight participants with eczema diagnosed according to the Hanifin and Rajka8 criteria took part in the trial for 10–14 weeks. The age of the participants was not reported. The stress management programme consisted of four 3-hour sessions over 2 weeks, in groups of six to eight participants, which covered cognitive restructuring and enhancing problem-solving skills. A booster session was then given 3 weeks after the last session. The trial was primarily concerned with endocrine stress levels but measured clinical outcomes as well.

Assessment of risk of bias

Table 83 provides the risk-of-bias assessment for the new study.

Benefits

The trial reported that there was no significant difference between the treatment groups in the severity of eczema, measured using the SCORAD index (p = 0.179). Although the data for baseline were not reported, the values for the severity of eczema appear to have been significantly more severe in the stress management programme group at baseline according to the graph of the severity of eczema during the trial.

Harms

The authors did not report any information about adverse events.

Overall implications for research and practice

This very small, short-term trial does not provide any evidence of benefit in terms of improving the severity of eczema using cognitive behavioural therapy. Future trials in this area need to be larger with longer-term follow-up and assess patient-reported outcomes such as quality of life and sleep loss to gain a true picture of the potential of this psychological technique.

Studies

No trials involving ion-exchange water softeners were published before 2000.

One new trial involving ion-exchange water softeners was reported after 2000. This study, conducted in the UK by Thomas and colleagues,364 included 336 children with moderate to severe eczema who were living in hard water areas. The trial compared usual eczema care with usual eczema care plus an ion-exchange water softener to soften the water for bathing and washing. The trial used a parallel-group design and the primary outcome was analysed at 12 weeks. At the end of the study, participants were ‘crossed over’ for 4 weeks to allow the participants in the control group to experience the intervention and to monitor the potential decline in beneficial effects in the intervention group. The nurse who recorded these observations was blinded; however, the participants were not.

Assessment of risk of bias

Table 84 provides the risk-of-bias assessment for the new study.

Benefits

The severity of eczema after 12 weeks was not significantly different in the water softener group compared with the control group. The CIs for this blinded objective outcome were very tight, making a robust argument for the lack of additional benefit of softening the water using an ion-exchange water softener for eczema severity (mean difference in SASSAD scores at 12 weeks –0.66, 95% Cl –1.37 to 2.69). In contrast to the blinded primary outcome, unblinded participant- or carer-assessed secondary outcomes did show small statistically significant beneficial effects of the water softeners. It is likely that these positive effects were a result of the bias introduced by the participants’ awareness of treatment allocation as expectation in the effectiveness of the water softeners was high.

Harms

Adverse events were not formally recorded; however, the parents of three participants reported an exacerbation of eczema, which they thought may have been due to the softened water. Two water softener units were removed early for this reason.

Overall implications for research and practice

This large, methodologically robust trial364 of ion-exchange water softeners for eczema clearly shows no benefit of the softeners for the severity of eczema. The significant differences in favour of the water softeners in three of the participant- or carer-assessed outcomes were small and are unlikely to be clinically relevant given the detection bias resulting from participants and carers being aware of the intervention that they were allocated to. This trial gives a robust message that ion-exchange water softeners cannot be recommended as an effective treatment for eczema.

Studies

No trials involving staying in another climate were published before 2000.

One new trial has been published since 2000. This open trial365 involving 61 participants in Norway investigated the effect of sending school children to Gran Canaria compared with remaining in Norway for a duration of 1 month. The children who went to Gran Canaria had to go to school, which included 1 hour of gymnastics a day, and also had to bathe in seawater for 1–2 hours a day. The children in the control group continued attending school as usual, with only 2–4 hours of gymnastics a week. No requirement to bathe in seawater was reported for this group. The control group were offered a trip to Gran Canaria at the end of the study. The severity of eczema and quality of life were measured in all of the children when the Gran Canaria group returned to Norway after 1 month and again 3 months later.

Assessment of risk of bias

Table 85 provides the risk-of-bias assessment for the new study.

Benefits

A significant improvement in the severity of eczema after 1 month, measured using the SCORAD index, was reported in the Gran Canaria group, with the score decreasing from a mean of 37.15 (95% CI 29.40 to 44.90) to 21.18 (95% CI 17.24 to 25.13) in the Gran Canaria group and from a mean of 36.84 (95% CI 30.00 to 43.69) to 30.62 (95% CI 24.13 to 37.11) in the control group (p = 0.045). The improvement was sustained for 3 months after the children’s return. Although quality of life improved significantly in the Gran Canaria group, it is not reported whether this was significant in comparison with the control group.

Harms

It was reported that some of the participants in the Gran Canaria group had mild sunburn. No adverse event data for the control group were provided.

Overall implications for research and practice

It is very difficult to decide how much clinical significance this treatment approach has because of the many confounding factors that could have affected the severity of eczema that were different between the two groups, such as bathing in seawater and the amount of exercise undertaken. The application of sun cream in the Gran Canaria group could potentially have had an effect on eczema because of its emollient effect. Both the inclusion criteria and the method of randomisation were not clear and so the generalisability of the treatment effects to the wider eczema population is hard to gauge from this trial. Even if genuinely effective, these benefits could be attributed to the psychological effects of going to a different country, the effect of ultraviolet light on skin inflammation and vitamin D synthesis, saltwater bathing, a changed diet or altered allergen exposure. In the light of the economic and social implications of removing a person with eczema and possibly other family members from their normal life, choosing this treatment over other options must be very carefully considered.

Studies

Four trials involving house dust mite reduction were reported before 200055 (see Appendix 3).

Three new trials367–369 have been reported since 2000. Ricci and colleagues369 randomised a group of 41 children aged from 2 to 10 years and sensitised to food or inhalant allergens to either ‘recommended’ house dust mite reduction measures (mattress and pillow encasings, hot wash of bedding at least once a week, vacuuming the living room and bedroom at least twice a week, carpets vacuumed at least once a week or removed and no pets) or no recommendations (normal cleaning patterns) for 2 months. After this all participants followed the house dust mite reduction recommendations for a further 10 months. The severity of eczema was assessed using SCORAD scores and total dust mite load and the specific load of Dermatophagoides pteronyssinus and D. farinae was measured.

A larger Dutch mite avoidance study by Oosting and colleagues368 was reported in which 86 participants with eczema and sensitised to house dust mites were randomised to either mite-impermeable mattress, duvet and pillow encasings made of GORE-TEX^® material or cotton ‘placebo’ encasings for 1 year on all beds in the participants’ bedroom. The clinical severity and extent of eczema were assessed using the Leicester Sign Score. 61 Sensitivity to house dust mite was measured by intradermal and patch testing alongside total and specific IgE. It is not clear whether the participants were adults, children or both.

The third trial by Gutgesell and colleagues367 randomised only 20 participants with eczema and sensitised to house dust mite to either allergen-impermeable polyurethane encasings and acaricide (mite-killing) spray made up of tannic acid and benzyl benzoate or cotton ‘placebo’ encasings and ‘placebo’ acaricide spray (water with traces of ethanol) for a year. The severity of eczema was assessed using the SCORAD index; daytime pruritus and pruritus-induced sleeplessness were assessed using a VAS, and participant-assessed skin status and the amount of topical corticosteroids used were both recorded. It was unclear whether the participants were children, adults or both.

Assessment of risk of bias

Table 86 provides the risk-of-bias assessment for the new studies.

Benefits

All three trials367–369 failed to show any benefit of house dust mite reduction interventions for eczema compared with placebo or normal cleaning practices over a period of up to 12 months. The trial by Ricci and colleagues369 showed a mean decrease in SCORAD scores from 33 to 24 in the first month for the mite avoidance group compared with a mean decrease from 27 to 22 in the placebo group. After the additional 10 months of the mite avoidance intervention for all participants, the groups had SCORAD values of 16 and 17, respectively. In the trial by Gutgesell and colleagues,367 the SCORAD values in both groups fluctuated throughout the year, with no overall trend, and were almost identical, with a difference of only 3 points between the groups (p = 0.901). Pruritus-induced sleeplessness (p = 0.399), daytime pruritus (p = 0.799) and the participant-assessed skin status (p = 0.583) also showed no marked differences between the groups. The Dutch mite avoidance study by Oosting and colleagues368 reported a significant reduction in the house dust mite load in the GORE-TEX encasing group whereas the placebo group load did not change significantly. This effect does not seem to translate into clinical benefit, however, as sleeplessness, itching, disease activity and extent all decreased by only a few points in both groups, with no significant difference between the groups.

The three trials367–369 gave only scant details of the trial methodology, with none reporting the method of randomisation and allocation concealment. It is unlikely that 2 months of intervention in the trial by Ricci and colleagues369 is long enough to be of clinical relevance when testing house dust mite reduction for a long-term condition such as eczema. On the other hand, two other trials367,368 with a duration of 1 year saw no significant effects in favour of house dust mite reduction interventions.

Harms

None of these trials specifically reported adverse events, although one participant withdrew from the trial because of sweating-induced exacerbation of eczema, attributed to the allergen-impermeable encasings.

Overall implications for research and practice

Applying encasings to a bed and having to wash or vacuum more frequently with a high-quality vacuum cleaner may be fairly achievable in many cases and, if proven to be of significant benefit, could have great potential for the treatment of eczema. However, a complex intervention such as in the trial by Ricci and colleagues,369 which could potentially add a large physical and mental burden to members of the family (because of extra housework, strict regimens and even loss of treasured soft toys), raises questions about the balance between the effectiveness of the intervention and quality of life for the participant and his or her family.

No long-lasting, significant clinical benefits of any of these interventions have yet been shown, but the trials have so far lacked methodological clarity and the important question of the impact on quality of life must be addressed in any future trials on reduction of house dust mites. There is a need for simple, pragmatic, long-term clinical trials of individual house dust mite interventions with blinded outcome assessments. Given that increased exposure to allergens can sometimes induce tolerance, it is also important to explore whether reducing allergen levels from different baseline levels actually induces more harm than good by periodically increasing sensitisation.

Studies

There were no trials involving additional clinic visits reported before 2000.

One new trial was reported after 2000. This trial, by Sagransky and colleagues,373 reported as a short communication, involved 30 children treated with topical tacrolimus 0.03% daily for 4 weeks. The participants were randomised in an open manner to have either one extra visit to the clinician 1 week after starting treatment or no extra visit. All participants visited the clinician after 4 weeks of treatment.

Assessment of risk of bias

Table 87 provides the risk-of-bias assessment for the new study.

Benefits

This pilot trial373 found a significant improvement in eczema severity assessed by EASI scores, IGA score and participant-assessed itch in both groups, but there were no significant differences between those who had an extra visit to the clinician at 4 weeks and those who did not. The mean difference in percentage improvement in severity measured using EASI scores (extra visit group 76% vs. control group 45%; p = 0.06) showed some difference; however, only 26 out of the 30 participants randomised were included in the analysis.

Harms

The extra visit group reported three adverse events and the control group reported two adverse events, none of which was assessed as being related to treatment.

Overall implications for research and practice

This small, unblinded pilot trial,373 powered to detect large differences in treatment effect, did not find any significant beneficial effect of one extra visit to the doctor. The trial also found no correlation between adherence and reduction in severity and no correlation between baseline severity and adherence, but the study was underpowered to exclude even moderate differences. Given that visits to the doctor are expensive, future trials of adherence enhanced by extra clinic visits should include an economic evaluation. Although the unblinded nature of the intervention may raise concerns about performance bias, in the context of this intervention blinding is not desirable as the intervention is clearly intended to increase adherence/performance. Measuring treatment adherence is challenging, especially in the case of emollients, where the amount applied will depend on severity and the current level of control. Developing new methods for accurately measuring treatment adherence will be a difficult but necessary move to further investigate the influence of the number of doctor visits on eczema severity.

Studies

There were no studies involving vaccines for eczema reported before 2000.

Four new trials involving vaccines for the treatment of established eczema were reported after 2000. The trial by Arkwright and David,375 reported in 2001, followed 41 participants aged 5–18 years with moderate to severe eczema for 3 months following a single vaccination or placebo injection. The severity of eczema and potency of topical corticosteroids used were recorded at 1 and 3 months after vaccination.

Another similar trial376 by the same group compared a vaccination with heat-inactivated Mycobacterium vaccae against a placebo vaccination. The 56 participants were aged between 2 and 6 years and had moderate to severe eczema. The severity of eczema scale included the extent of involvement and severity (dermatitis score), with a maximum score of 300 points. The potency of topical corticosteroids used was recorded at 1, 3 and 6 months after the vaccination.

A multicentre, parallel-group trial in the UK and Croatia by Berth-Jones and colleagues377 was reported in 2006. This trial involved 166 children aged between 5 and 16 years with moderate to severe eczema, diagnosed according to the UK Working Party’s criteria. 9 Children were randomised to receive a single injection containing either heat-inactivated M. vaccae or placebo. The participants were followed up for 24 weeks with assessments at 4, 8, 12 and 24 weeks. The participants could not make any changes to their eczema medication in the 4 weeks prior to randomisation and could not use any topical or systemic immunomodulatory treatments, Chinese traditional medicine or phototherapy during the trial.

A trial by Brothers and colleagues378 conducted in New Zealand also compared heat-inactivated M. vaccae immunisation (three injections at 2-week intervals) with placebo in 129 children aged 5–16 years with moderate to severe eczema but otherwise good general health. The participants could keep using their other treatments for eczema or discontinue them, as long as they were not tacrolimus, ciclosporin, methotrexate, pimecrolimus, ultraviolet A (UVA) or ultraviolet B (UVB) phototherapy, systemic corticosteroids, high-dose inhaled steroids or traditional Chinese medicines. The participants were followed up for 6 months after immunisation.

Assessment of risk of bias

Table 88 provides the risk-of-bias assessment for the new studies.

Benefits

The first trial in children and adolescents aged 5–18 years375 showed a greater reduction in eczema severity and surface area affected in the M. vaccae vaccination group than in the placebo group, with the dermatitis score significantly improving from month 1 and at month 3 (total mean change from baseline 41 points vs. 10 points; p < 0.01) and the surface area affected also significantly improved by month 3 (total mean change from baseline 17% vs. 0%; p < 0.01).

The trial in children aged 2–6 years376 failed to show any significant difference between M. vaccae and placebo, although it must be noted that the M. vaccae group had significantly more severe disease than the placebo group at baseline (p = 0.05; difference in mean dermatitis score of 9 points). Neither of these trials found a significant decrease or increase in the potency of topical corticosteroids used in the trial.

The largest trial, carried out with children and adolescents aged 5–16 years in the UK and Croatia,377 reported a mean decrease in eczema severity after 12 weeks, measured by SASSAD, of 9.4 in the 1-mg M. vaccae group, 7.0 in the 0.1-mg M. vaccae group and 8.8 in the placebo group. The decrease in the 1.0-mg group was statistically significant compared with the placebo group (95% CI –4.3 to 5.4; p > 0.05). There were significant differences in favour of the 0.1-mg M. vaccae group compared with placebo and also both M. vaccae groups combined compared with placebo for sleep disturbance after 8 weeks. The trial report states that there were no significant differences in all of the other outcomes of change in severity of eczema at 24 weeks, change in body surface area affected, participant-assessed global assessment of response, pruritus severity and frequency of use of topical corticosteroids.

The trial by Brothers and colleagues378 reported that there was no significant change in eczema severity, measured using SASSAD, between the placebo group and the immunisation group at 3 months (p = 0.77) and 6 months (p = 0.70) after immunisation. There was also no significant difference in the extent of eczema (p = 1.0) at 3 months. There was also no significant difference in the change in quality of life, sleep disturbance and frequency and potency of topical corticosteroid use, but the data for these outcomes were not reported.

Harms

Both trials by Arkwright and David375,376 showed a very similar pattern of adverse events. There were no systemic adverse events reported in either trial; however, approximately half of the participants (13/21 and 13/29, respectively) who were vaccinated with M. vaccae developed a temporary localised red lump at the injection site.

In the trial by Berth-Jones and colleagues,377 103 participants reported 260 adverse events. The most common adverse event was reported as eczema (53 participants) followed by infected eczema (24 participants). Forty-one participants (25%) reported one or more adverse events assessed as at least being possibly related to treatment. There were five serious adverse events assessed as not being related to study treatment.

The trial by Brothers and colleagues378 reported that 47% of the participants had a local injection site reaction and that 75% of these had received the M. vaccae vaccine.

Overall implications for research and practice

The evidence from these four trials, comprising almost 400 participants, with a low risk of bias, indicates that there is insufficient evidence of benefit for vaccination with M. vaccae for atopic eczema treatment. Additional research evidence on M. vaccae vaccination for treating eczema is not needed.

Specialised clothing

• There were three small trials involving specialised clothing for eczema reported up to 2000. One of these trials found evidence of benefit for clothing made from cotton compared with clothing made from two other fibres and another trial found evidence of benefit for warp knits compared with jersey knits. The third trial found evidence of benefit for gel-filled absorbent core nappies compared with cellulose absorbent core nappies for nappy rash but not for eczema.

• Eight trials reported after 2000 covered four different types of specialised clothing:

  • Three trials, with a mostly low and unclear risk of bias, found evidence of benefit for silk clothing (DermaSilk).

  • One trial, with a high risk of bias for blinding, provided evidence of benefit for clothing containing silver (Padycare) and one trial, with a mostly low risk of bias, was difficult to interpret and therefore did not provide evidence of benefit for clothing containing silver (X-STATIC).

  • Two trials, with a mostly high risk of bias, did not provide any evidence of benefit for clothing made from ethylene vinyl alcohol fibre.

  • One trial, with an overall unclear risk of bias, provided evidence of benefit for anion textile with added tourmaline.

Education

• One small unblinded trial was published before 2000 and provided evidence of benefit for an educational intervention given by a nurse compared with no education.

• Twelve trials reported after 2000, with an overall mixed risk of bias, covered different educational approaches and provided some evidence of benefit for educational approaches.

E-Health portal

• One small trial, with a mixed risk of bias, did not provide any clear evidence of benefit for an online health-care portal compared with standard care.

Stress management

• There were no trials of stress management for eczema reported up to 2000.

• One very small trial reported in 2013, with a mostly low risk of bias, did not provide any evidence for the effectiveness of a stress management treatment for eczema based on cognitive behavioural therapy compared with no treatment.

Ion-exchange water softeners

• There were no trials of ion-exchange water softeners for eczema reported up to 2000.

• One large trial reported in 2011, with an overall low risk of bias, provided evidence of no benefit from the use of an ion-exchange water softener in the home compared with no water softener.

Living in a different climate

• There were no trials of living in a different climate for eczema reported up to 2000.

• One small trial, with a mostly high risk of bias, provided evidence of possible benefit of living in a warmer climate compared with staying at home.

House dust mite reduction

• There were four small trials, with a mostly unclear risk of bias, reported up to 2000. The results were conflicting, with only one of the trials providing evidence of benefit for a house dust mite reduction intervention (mite-impermeable encasings, intensive/high-filtration vacuuming, acaricide spray) compared with placebo (mite-permeable encasings, normal cleaning patterns, vacuuming with reduced suction).

• Three small trials, with an overall unclear risk of bias, did not provide any evidence of benefit for interventions to reduce house dust mites (more frequent vacuuming, hot washing of bedding and soft toys, removing pets, mite-impermeable mattress and bedding encasings, acaricide spray) compared with placebo (normal patterns of cleaning and washing, allowing pets, mite-permeable encasings, placebo spray).

Additional visits to a doctor

• There were no trials of additional visits to a doctor for eczema reported up to 2000.

• One very small trial published in 2010, with a high risk of bias for blinding, provided no evidence of benefit of one extra visit to the doctor compared with no extra visit.

Vaccines

• There were no trials of vaccination for treating eczema reported up to 2000.

• Three small trials published after 2000, with an overall low risk of bias, did not provide any evidence of benefit for vaccination with M. vaccae for treating established eczema.

Existing systematic reviews

Since 2000, three systematic reviews have been published. 83,379,380 The AAD94 and NICE41 guidelines also cover phototherapy.

Studies

A left/right within-person trial by Selvaag and colleagues,381 reported in 2005, compared standard UVB fixed-dose increments against UVB using skin reflectance guided dosing. The trial included 20 adults aged 16–38 years with mild to moderate eczema. Treatment was given for up to 6 weeks and was stopped early if a participant’s SCORAD score fell to < 10 on either side of the body. The whole of the face was given standard UVB treatment. Emollients and topical corticosteroids were allowed during the trial as long as they were used symmetrically. It was not reported how many treatments were given per week.

Assessment of risk of bias

Table 89 provides the risk-of-bias assessment for the new study.

Benefits

The severity of eczema was recorded using the SCORAD index. It was reported that no significant difference was found in the reduction of eczema severity between the two treatment regimens, but no detailed data were reported. The times taken for the SCORAD scores to reduce to < 10 were reported. The standard treatment regimen had a median time (5th and 95th percentiles) of 3.5 weeks (1.5, 6.0) and the skin reflectance regimen had a median time (5th and 95th percentiles) of 3.0 weeks (2.0, 5.5). The cumulative UVB dose was significantly lower for the skin reflectance regimen (median 39 × 10 mJ/cm²; p < 0.01) than for the standard regimen (median 124 × 10 mJ/cm²). The initial UVB dose was reported as higher for the skin reflectance regimen, with a median of 3.4 standard erythemal dose (SED) (×10 mJ/cm² at 298 nm using the CIE erythema action spectrum) compared with 2.6 SED (×10 mJ/cm² at 298 nm using the CIE erythema action spectrum) for the standard regimen, but this was not statistically significant.

Harms

Data on adverse events were not reported for this trial.

Overall implications for research and practice

The potential to administer low-dose UV radiation and achieve the same clinically beneficial effects as reported with high-dose UV radiation is encouraging but requires confirmation in larger pragmatic studies. This one small trial381 indicates that this could be achievable; however, the results must be treated with caution. Problems with the study include the lack of blinding, the claim of equivalence of non-inferiority being based on a very small sample size and no methodological details about the study design being reported. Such an important question should be examined in detail in an appropriately designed and powered trial to give patients and clinicians clear guidelines on the best treatment regimen for UV phototherapy.