Alitretinoin versus phototherapy as the first-line treatment in adults with severe chronic hand eczema: the ALPHA RCT

Article history

The research reported in this issue of the journal was funded by the HTA programme as award number 12/186/01. The contractual start date was in January 2015. The draft manuscript began editorial review in January 2023 and was accepted for publication in June 2023. The authors have been wholly responsible for all data collection, analysis and interpretation, and for writing up their work. The HTA editors and publisher have tried to ensure the accuracy of the authors’ manuscript and would like to thank the reviewers for their constructive comments on the draft document. However, they do not accept liability for damages or losses arising from material published in this article.

Copyright statement

Copyright © 2024 Wittmann et al. This work was produced by Wittmann et al. under the terms of a commissioning contract issued by the Secretary of State for Health and Social Care. This is an Open Access publication distributed under the terms of the Creative Commons Attribution CC BY 4.0 licence, which permits unrestricted use, distribution, reproduction and adaptation in any medium and for any purpose provided that it is properly attributed. See: https://creativecommons.org/licenses/by/4.0/. For attribution the title, original author(s), the publication source – NIHR Journals Library, and the DOI of the publication must be cited.

2024 Wittmann et al.

Risk and trigger factors

A number of factors have been described that influence the onset or flare-up of HE. As mentioned, HE in the general population affects women about twice as often as men. 21–23 AD and/or childhood eczema, which is almost always associated with dry skin and skin barrier impairment, seems to be one of the most important risk factors. 24,25 Exposure to wet work, frequent hand washing and irritants are recognised relevant trigger factors. A high HE incidence rate is thus associated with contact allergy, AD, wet work and also female sex. 23

Important genetic factors are those linked to AD, and the most recognised ones are loss of function (LOF) mutations in the filaggrin gene. Filaggrin is an important constituent of the skin barrier. 26 Filaggrin significantly contributes to water retention in the skin, while LOF mutations result in increased transepidermal water loss. There are four common filaggrin mutations prevalent among Europeans and more frequently found in patients with AD. 27–29 Deletions in the late cornified envelope (LCE) genes LCE3B and LCE3C have been described to be associated with CHE with allergic contact dermatitis. 30 The LCE complex is also a key component of the skin barrier.

To date, it is not fully elucidated whether the association of HE with AD is primarily due to filaggrin and skin barrier changes or due to atopy resulting in a type 2 shift in the immune response. The ALPHA trial collected information on LOF filaggrin mutation, LCE variants and atopy status.

Lifestyle factors discussed to a great extent include tobacco smoking. However, despite a wealth of available observational, retrospective studies, the data are still not clear, and overall the association seems weak; it may be stronger for foot eczema. 31–37 No study has found an association between alcohol consumption and HE. 35 Predictive factors for HE persistence include (1) age of onset before 20 years, (2) positive history of childhood eczema (3) and moderate to severe extension of HE – thus, HE severity. 23

As there is no clear subclassification of HE, there is also a lack of clear treatment guidelines. 38 Mostly, CHE therapy is delivered in escalating steps. When patient education, allergen/irritant avoidance and topical treatment are not sufficient to control the disease, ultraviolet (UV) therapy or systemic immune-modifying drugs are used. 17,39 Alitretinoin (AL) is licensed for severe CHE unresponsive to treatment with potent topical corticosteroids. However, currently available clinical evidence for the treatment of CHE is not compelling enough to guide clinical practice. 39,40 The lack of clear evidence-based data has been outlined by different national and international expert groups as well as by a recent Cochrane review. 38,41 Given the high socioeconomic impact of the disease, there is a pressing need for comparative studies on available first-line treatments and on the long-term outcome of currently used therapies. The National Institute for Health and Care Research (NIHR) has recognised this need and launched a commissioned call aiming to compare AL with other treatment options. Our choice of comparator for AL was based on published clinical trials and on feedback from UK dermatologists and the United Kingdom Dermatology Clinical Trials Network (UKDCTN). 41 As research and audit data on treatment choices for CHE were unavailable, we performed a survey among 194 UK dermatologists; the most frequent first-choice approaches for CHE were 8-methoxypsoralen combined with ultraviolet A (PUVA), oral steroids and AL. 42 When asked which strategy was thought to be most efficient for long-term outcome, 43% of clinicians reported AL, 30% PUVA and a further 20% of clinicians indicated that they did not know.

The following immunosuppressive treatments have been described as being used for the treatment of CHE: azathioprine, mycophenolate mofetil, methotrexate, ciclosporin A and systemic corticosteroids. Recently described drugs include the interleukin (IL)-4/IL-13R blocker dupilumab and Jak inhibitors including baricitinib and upadicitinib, all of which are licensed for the treatment of AD. 43,44 Clinical trials of topical Jak inhibitors are being conducted, and phase IIb results for delgocitinib have recently been reported. 45

The current literature suggests that dyshidrotic/vesicular subtypes with atopic diathesis respond preferentially to immunosuppressants and dupilumab, whereas hyperkeratotic rhagadiform subtypes may show a better response pattern to AL. However, there is insufficient clarity on which subtypes respond to which therapy approaches.

A generally unfavourable therapeutic response is believed to be linked to smoking, long disease duration and barrier defects such as LOF mutations affecting filaggrin. 5,11,39 There are two important shortcomings affecting clinical trials in CHE. Firstly, there is no consensus regarding the best outcome measures of extent and severity of the disease, disease impact on quality of life (QoL), or assessment of comorbidities, and there is a lack of measures to highlight impairment in functional use of the hands. Numerous disease severity assessments have been proposed, including photoguide, Physician’s Global Assessment (PGA), Hand Eczema Severity Index (HECSI), Hand Eczema Extent Score (HEES) and modified Total Lesion Symptom Score (mTLSS). 46,47 Although none has been fully recognised as ‘gold standard’ yet, the HECSI is predominantly used in clinical studies and has advantages regarding accurate assessment of morphology features and extent. 48,49 ALPHA has used the HECSI as the primary outcome but included other outcome measures frequently used in previous trials. The second shortcoming relates to the short duration of clinical trials, most of which have been limited to 12 weeks. This indeed seems inadequate for a disease characterised by its tendency for chronicity and which often has a long disease duration.

Summary

There is a lack of controlled clinical trials that directly compare treatments or demonstrate effectiveness under daily practice conditions. Moreover, the lack of clear evidence-based data has been outlined by a number of national and international expert groups. 38,41 Given the high socioeconomic impact of the disease, there was a pressing need for comparative studies on available first-line treatments and on long-term outcomes of currently used therapies. Therefore, we conducted a randomised controlled trial (RCT) comparing AL with Immersion PUVA for patients with severe CHE, who were followed up for a maximum period of 52 weeks to understand the longer-term impact of these treatments as first-line therapies.

Aims and objectives

The primary aim was to determine the clinical effectiveness and cost-effectiveness of AL compared with 8-methoxypsoralen combined with UV-A (Immersion PUVA) in conjunction with concomitant topical corticosteroids, emollients and patient education as the first-line treatment in patients with severe CHE.

Overview of methods

This chapter outlines the main methods for the trial, including all data collection for the trial and substudy work and the analysis methods for the primary and secondary clinical results, which are detailed in Chapter 3. Analysis methods and results for the main trial health economics are detailed in Chapter 4. The discussion chapter draws the work together in Chapter 5.

Trial design

The trial was a multicentre, Phase III, open, prospective, adaptive, two-arm parallel group RCT, with one planned interim analysis.

Participants were randomised on a 1 : 1 basis to receive either AL at a dose of 30 mg/day (with the option to reduce to 10 mg if participants suffered with headaches and restored to 30 mg dose once headaches ceased) or Immersion PUVA (3 mg/l Meladinine® with UV A, twice weekly) for a 12-week interventional phase. Randomised treatment was given in conjunction with concomitant topical corticosteroids, emollients and patient education. Partial responders in both arms continued their randomised treatment for up to a further 12 weeks. Treatment could be discontinued between 12 and 24 weeks if participants achieved a clear/almost clear assessment (responder) or had a severe assessment (non-responder). All participants stopped treatment by 24 weeks. Participants who relapsed and non-responders continued with ‘standard clinical practice’, as determined by the attending clinical team.

The study protocol for this trial has already been published. 50 Summary details of the methods are given below.

Ethical approval

Ethical approval for the study was given by Leeds West Research Ethics Committee [REC reference: 14/YH/1259, Integrated Research Application System (IRAS) project ID: 163195]. All trial activity took place according to the ethically approved protocol, principles of Good Clinical Practice (GCP) and Declaration of Helsinki 1996.

Eligibility and informed consent

Patients were screened in secondary care dermatology outpatient, community hospital and General Practice settings. Formal eligibility assessment and recruitment was undertaken in secondary care dermatology outpatient clinics. Patients were eligible if they fulfilled the following criteria:

• suffering from uncontrolled, severe CHE, defined as the presence of both of the following criteria:

  1. PGA score of severe

  2. resistance to treatment with potent topical corticosteroids for ≥ 4 weeks prior to the point of eligibility screening

• aged ≥ 18 years

• able to provide written informed consent

• expected to comply with treatment and protocol schedule.

Patients were excluded if they fulfilled any of the following criteria:

Skin-related:50

• had a clinically suspected infection (fungal, bacterial or viral) as cause for dermatitis of the hands

• known clinically relevant allergic contact dermatitis of the hands, unless they had made a reasonable effort to avoid the contact allergen

• atopic eczema covering more than 10% of body surface (excluding hands)

• had skin conditions worsened by the sun, that is, do not tolerate UV-light (for example lupus erythematosus, porphyria).

Treatment-related:

• received systemic vitamin A derivatives or systemic immunosuppressants, for example, methotrexate or biologics treatment for HE, or received phototherapy/photochemotherapy, in the 3 months prior to randomisation

• received ciclosporin A or systemic glucocorticoid steroid treatment for HE or topical calcineurin antagonist treatment within 1 week prior to randomisation

• receiving concomitant treatment with tetracyclines, medication with potential for drug–drug interaction with AL (e.g. CYP3A4 inhibitor ketoconazole), or concomitant treatment with relevant photosensitisers, that cannot be suspended or switched to an acceptable alternative

• history of melanoma skin cancer, or patients with a history of non-melanoma skin cancer depending on history, location and ‘severity’ of the non-melanoma skin cancer based on experience from routine practice

• received prior treatment with arsenic agents or ionising radiation in the treatment area (e.g. hands).

General:50

• if female:

  • lactating

  • of child bearing potential [Woman of Child Bearing Potential (WCBP)] and:

    1. - with positive pregnancy test (absence of pregnancy confirmed with a negative pregnancy test before randomisation)

    2. - unwilling to follow pregnancy prevention programme measures (rigorous contraception for women of childbearing potential, unless exempt according to standard of care practice, was required 1 month before treatment, during the treatment period and 1 month after cessation of treatment as per usual standard practice) while receiving treatment and after the last dose of protocol treatment as indicated in the relevant summary of product characteristics (SmPC)

• hepatic insufficiency (alanine aminotransferase and/or aspartate aminotransferase > 2.5 times the upper limit of normal), known severe renal insufficiency, uncontrolled hyperlipidaemia [for all of the following: triglycerides, cholesterol and/or low-density lipoprotein (LDL) cholesterol] or uncontrolled hypothyroidism in the 12-week period prior to randomisation

• known hypersensitivity to peanut, soya or vitamin A derivatives or with rare hereditary fructose intolerance as determined by patient history

• suffering from hypervitaminosis A as directed by clinical symptoms or patient history

• previous participation in the ALPHA trial.

Blood samples were required to confirm eligibility and atopy status prior to randomisation. Research sites could choose to use either a one-stage process to obtain full informed consent or a two-stage consent process involving obtaining consent for blood sampling to confirm eligibility and atopy status, and then full informed consent for trial participation. If a period of > 12 weeks elapsed prior to the baseline visit, the eligibility blood test was redone.

If available and within the 12 weeks prior to baseline visit, existing blood results (taken for other reasons) in participants’ medical notes could have been used to confirm eligibility. Similarly, if existing atopy results {bon presence/absence of specific immune globulin E [IgE] to inhalant or other relevant allergens as appropriate [e.g. via prick test, Radioallergosorbent Test (RAST) blood test]}, were available in participants’ hospital notes, a blood sample to confirm atopy status was not required.

A log was completed of all patients screened but not recruited, either because they were ineligible or because they were eligible but declined participation. The following anonymised information was included:

• age

• gender

• ethnicity

• date screened

• how the patient first heard about the trial

• the reason why they were not eligible for study participation, OR

• the reason for declining participation.

Interventions

Participants were randomised to either AL or Immersion PUVA for a minimum of 12 weeks and a maximum of 24 weeks. In both arms of the study, it was expected that patients would follow good self-care practices in the use of emollients, irritant avoidance and – if applicable – diligent continuation of contact allergen avoidance. Education on HE in using emollients, avoiding irritants and relevant contact allergens was delivered face to face in a standardised way based on study-specific educational information by the research nurse involved in the trial. The same education material was handed to participants. The information material for participants was based on sources used in clinical practice [British Association of Dermatology (BAD), National Eczema Society, Eczema Society patient information leaflets].

According to standard clinical practice and National Institute for Health and Care Excellence (NICE) guidelines, AL was administered at a starting dose of 30 mg, to be taken once daily with the main meal. Participants self-administered the treatment at home. Dose adjustment of AL down to 10 mg or temporary cessation (i.e. dose interruption) was permitted according to standard practice in participants who suffered from AL-related headaches.

For participants allocated to Immersion PUVA, hands were immersed for 15 minutes in a Meladinine® 0.75% solution diluted to 3 mg/l for 15 minutes followed by up to a 30-minute delay before exposure to UV-A radiation according to standard practice at the participating site; all sites operated based on ‘British Photodermatology Group’ guidelines. The exact dose of UV-A radiation that participants received was individually tailored to the participant depending on phenotype (as per BAD guidelines) and the erythematous response of the skin following treatment.

The treatment was performed in outpatient phototherapy departments within secondary care units and was administered and supervised by the specialised nurses/dermatologists. Treatments were carried out twice weekly.

As directed by NICE guidelines for AL, the PGA score directed treatment pathway decisions as determined by the treating clinician.

Criteria of response

Criteria are shown in the protocol schedule (Figure 1).

FIGURE 1.

Protocol schedule.

Responders: defined as a PGA score of clear/almost clear at 12 weeks post planned start of treatment, discontinued randomised treatment and continued to receive ‘standard clinical practice’ and follow-up monitoring.

Partial responders: defined as a PGA score of mild/moderate, continued with randomised treatment for up to 24 weeks post planned start of treatment. During this 12- to 24-week treatment period, the patients were monitored at 4-weekly intervals; randomised treatment could be stopped at any time if the patient responded (PGA score of clear/almost clear) or if the symptoms worsened (PGA score of severe) and in the opinion of the attending clinical team there was no clinical benefit to continuation. All randomised treatment was discontinued at the maximum 24 weeks’ treatment period, and patients continued to receive ‘standard clinical practice’ and follow-up monitoring.

Non-responders: defined as a PGA score of severe at 12 weeks post planned start of treatment, discontinued randomised treatment and continued to receive ‘standard clinical practice’ and follow-up monitoring.

At the end of trial treatment and during follow up, treatment was at the discretion of the attending clinical team as per ‘standard clinical practice’. Follow-up monitoring continued until 52 weeks post planned start of treatment, with the exception of participants randomised from 1 October 2019, who completed follow-up assessments for 24 weeks post planned start of treatment.

Randomisation

Recruitment of participants to the ALPHA trial required a blood sample taken within 12 weeks of the baseline visit to confirm eligibility and atopy status, and at least 1 month’s duration of the contraception prevention programme prior to randomisation (if applicable) to confirm eligibility. Patients were registered by an authorised member of the attending clinical or research team and were issued a unique trial number. Registration was performed using either the 24-hour automated registration telephone system or via a web address based at the Clinical Trials Research Unit (CTRU).

Participants were randomised once eligibility was confirmed and the baseline assessments and questionnaires were completed. Participants were randomised in a 1 : 1 allocation ratio to receive either AL or Immersion PUVA in conjunction with concomitant topical corticosteroids, emollients and patient education. A computer-generated minimisation programme that incorporated a random element equal to 0.8 was used to ensure intervention groups were well balanced for the following factors:

• randomising site

• disease duration (< 6 months/6–24 months/> 24 months)

• clinical phenotype (predominantly hyperkeratotic/predominantly vesicular/fingertip dermatitis)

• atopy status as determined by presence of specific IgE to inhalant allergens (prick test or RAST blood test to detect specific IgE to inhalant or other suspected allergens as appropriate)

• DLQI (< 15, ≥ 15)

• skin type (white/fair/dark).

A total of 100 participants who consented to take part in a tape stripping (non-invasive epidermal sampling) substudy were selected at their baseline visit. A quota sampling approach was taken so that the first 25 participants allocated to Immersion PUVA with predominantly hyperkeratotic CHE, the first 25 participants allocated to Immersion PUVA with predominantly vesicular CHE, the first 25 participants allocated to AL with predominantly hyperkeratotic CHE, and the first 25 participants allocated to AL with predominantly vesicular CHE were selected. Participants presenting exclusively with fingertip dermatitis were excluded from this substudy. Epidermal samples in the form of tape strip samples for randomised participants were obtained from both lesional skin and non-lesional, healthy-looking skin of the forearm. Full details of the sample technique are provided in Appendix 1.

Blinding

The trial was openlabel, as participants and investigators could not be blinded to treatment allocation due to the nature of the Immersion PUVA intervention. However, the assessment of the HE severity scores (HECSI, mTLSS and PGA) was undertaken by an assessor who was blinded to the randomised treatment. Participants were reminded not to reveal which treatment they had received to the blinded assessors in order to preserve blinding.

In addition, photographs were taken, for 20% randomly identified, consenting participants of white ethnicity and all consenting participants from minority ethnic groups from each centre, at baseline and 12 weeks post planned start of treatment. A blinded central review of the photographs was conducted to assess intercentre differences in severity scoring.

Outcome measures

Data collection schedule

Clinical data and patient-reported data were collected at baseline, every 4 weeks to week 36 and 8-weekly thereafter to week 52. For participants randomised after 1 October 2019, data collection was stopped at week 24. Trial activity was paused due to the COVID-19 pandemic from March 2020 in all centres. Some centres restarted research activity from July 2020. When trial activity restarted, visits were permitted to be conducted in person or virtually (telephone or video call), and QoL questionnaires were posted to participants where possible. The primary end point was only permitted in person, with visits at 12 weeks and 24 weeks prioritised.

End points

Trial organisational structure

The Trial Sponsor was the University of Leeds; responsibilities were delegated to the (CTRU) as detailed in the trial contract.

Trial oversight and management were conducted by the Trial Management Group (TMG), the Trial Steering Committee (TSC) and the Data Monitoring and Ethics Committee (DMEC). All groups met regularly throughout the trial.

Data quality and monitoring were performed by the CTRU.

Statistical methods50

Summary of main changes to the protocol

Centres opened to protocol v3.0 and patient information sheet and consent form v3.0 on 5 November 2015. The following substantial changes were made during the lifetime of the trial:

• Addition of a nail assessment substudy at the Chief Investigator’s centre, as there are currently no validated tools available for nail assessment in eczema. Note that the results of this are not included in this report because it was outwith the grant application.

• Following TSC and DMEC feedback, addition of a new stratification factor informed by the Fitzpatrick score for skin colour because erythema may be underestimated in darker skin.

• As part of the original Health Technology Assessment (HTA) funding envelope, a recruitment time extension was requested and approved, with a reduction in follow up to 6 months for all patients recruited from October 2019 to maximise accruals within the trial extension period. The trial stopped recruitment in accordance with the timelines agreed with the funder.

Participant flow

In total, 1557 participants were assessed for trial eligibility, 582 (37.4%) registrations took place to enable trial-specific tests and pregnancy prevention programmes to be implemented as required, and 441 (75.8% of those registered) randomisations took place between October 2015 and June 2021. Of those randomised, 59 (13.4%) were recruited after 1 October 2019 with a follow-up period of 6 months. Thirty-five NHS Secondary Care hospitals opened, of which 31 registered and randomised participants. The recruiting hospitals were located in England in the Yorkshire and Humber, North West, West Midlands, East Midlands, Eastern, London and South West regions, and in Scotland and Wales. The number of participants registered by each recruiting hospital ranged from 1 to 153 with a median of 10 (see Appendix 2 – Clinical results supplementary tables and figures, Figure 13). The mean number of registrations per centre was 0.30 participants per month. The number of participants randomised by each hospital ranged from 1 to 51, with a median of 4 (Figure 2). The mean number of participants randomised per centre was 0.23 per month. Follow-up was completed by 31 December 2021, after the last participant reached 24 weeks post planned start of treatment.

FIGURE 2.

Randomisations by anonymised centres.

A consolidated standards of reporting trials (CONSORT) flow diagram of trial progress is presented in Figure 3. Of the 1557 patients who were screened, 642 (41.2%) were ineligible, with over half of the patients not having a diagnosis of severe CHE [N = 347 (54.0%)]. Of the 915 eligible patients, 582 (63.6%) were consented and registered. Regarding the reasons for not consenting to the trial, 74 (22.2%) patients thought the Immersion PUVA schedule or travel was inconvenient and 28 (8.4%) did not want to receive AL. A total of 141 (24.2%) of those registered did not progress to randomisation; the main reasons were because they did not meet the eligibility criteria [N = 69 (48.9%)] or due to patient choice [N = 29 (20.6%)]. A total of 441 participants were randomised.

FIGURE 3.

Consolidated standards of reporting trials flow diagram.

The screened population and randomised populations were similar in respect of age, gender and ethnicity (Table 1).

Of those patients randomised, 220 (49.9%) were allocated to receive AL and 221 (50.1%) to receive Immersion PUVA. Of those allocated to AL, 201 (91.4%) were confirmed to have started treatment within 7 days post randomisation compared with 165 (74.7%) of those allocated to Immersion PUVA. A total of 132 (29.9%) participants were withdrawn or withdrew from the follow-up schedule or were reported as lost to follow-up, with 54 (24.5%) in the AL group and 78 (35.3%) in the Immersion PUVA group. A total of 441 (100.0%) participants were included in the ITT population.

Baseline characteristics

Patient characteristics were balanced across randomised treatment groups and are detailed in Tables 2 and 3. In summary, the study population had a median age of 46 years (range 18–81); 61.9% (n = 273) were female and 88.9% (n = 392) were of white ethnicity. The majority of participants had been suffering with CHE for more than 2 years [n = 311 (70.5%)], and 64.9% (n = 286) participants had predominantly hyperkeratotic CHE. Just over half of the patients were atopic, as verified by the presence of allergen-specific IgE [n = 227 (51.5%)]. The baseline DLQI score was dichotomised based on the NICE TA177 threshold for prescribing AL (≥ 15), and our results showed that the majority of participants recruited with severe CHE actually had a DLQI < 15 (n = 250, 56.7%), with a median [interquartile range (IQR)] of 13 (9–18).

Outcomes

Secondary objective – hand eczema severity index over 52 weeks

The results of fitting a linear mixed model to log(HECSI + 1) collected over the full 52 weeks show that there is no evidence of a difference between AL and Immersion PUVA at 24 weeks, with the estimate of the fold change (95% CI) equal to 0.92 (0.798 to 1.08) (Figure 4). At 52 weeks, there continues to be no evidence of a difference between AL and Immersion PUVA, with the estimated fold change (95% CI) equal to 1.27 (0.97 to 1.67) (see Figure 4). The estimated parameters of the full model are included in Table 40. The adjusted mean log(HECSI + 1) scores and corresponding differences by treatment arm are included in Table 41.

FIGURE 4.

Estimated fold change in HECSI + 1 scores (AL vs. Immersion PUVA), with ongoing treatment decision up to 24 weeks used as auxiliary variable.

Secondary objective – PeDeSI over 52 weeks

At baseline, just 16.1% (n = 71) were assessed as having sufficient knowledge, ability and confidence to manage their condition on their own according to the PeDeSI, with a similar distribution of education needs across the treatment groups (Figures 5 and 6). At 12 weeks, the proportion of participants with available data who had sufficient knowledge and education was equal to 26.2% (117/324), with 28.0% (49/175) in the AL group and 24.2% (36/149) in the Immersion PUVA group (Appendix 2, Table 45). The results observed at 24 weeks should be treated with caution due to the smaller number of participants who provided data at 24 weeks instead of at 52 weeks due to the protocol amendment.

FIGURE 5.

Distribution of PeDeSI categories over time, AL.

FIGURE 6.

Distribution of PeDeSI categories over time, Immersion PUVA.

After fitting an ordinal logistic mixed model to the PeDeSI scores at 12 weeks, the results show that there was no evidence of a difference in terms of educational need between treatment groups at 12 weeks, with an estimated odds ratio (AL vs. Immersion PUVA) for lower educational need (i.e. higher PeDeSI score) of 0.65 (CI 0.39 to 1.08) (Appendix 2 – Clinical results supplementary tables and figures, Table 46).

Exploratory objective – comparison of outcome measures

Figures 7, 8 and 9 show the distribution of the HECSI, mTLSS and DLQI scores, respectively within each PGA category according to the blinded assessor, with further detail provided in Appendix 2 – Clinical results supplementary tables and figures (Table 48). These assessments were taken across all patients and all time points with no adjustment for repeated assessments within participants in the analysis. In general, the PGA categories are reflected well by the HECSI and mTLSS scores, with little overlap of the HECSI and mTLSS IQRs for each PGA. Where the CHE was assessed as severe, the median (IQR) HECSI score was equal to 67 (44–324), and the median (IQR) mtLSS was equal to 13 (10–15). In contrast, the IQRs for the DLQI within each category of the PGA do overlap. Of note, the median (IQR) DLQI when the CHE is assessed as severe is equal to 11 (6–16).

FIGURE 7.

Box plot of HECSI by PGA categories over all time points.

FIGURE 8.

Box plot of mTLSS by PGA categories over all time points.

FIGURE 9.

Box plot of DLQI by PGA categories over all time points.

There was high positive correlation of the raw HECSI scores with the raw mTLSS scores over all time points, with an estimated correlation (95% CI) of 0.84 (0.83 to 0.86). Moderate positive correlation was observed for the raw DLQI scores compared with both the raw HECSI scores [estimated correlation (95% CI) = 0.56 (0.53 to 0.58)] and the raw mTLSS scores [estimated correlation (95% CI) = 0.65 (0.62 to 0.68)]. Scatterplots of the scores are in Appendix 2 – Clinical results supplementary tables and figures, Figures 16–18.

Exploratory objective – subgroup analyses

Figure 10 shows the estimated fold change in the HECSI + 1 scores between AL and Immersion PUVA. Estimates and CIs that lie fully to the left of the reference line of 1 indicate a benefit of AL compared with Immersion PUVA, while estimates and CIs that lie fully to the right of the reference line indicate a benefit of Immersion PUVA compared with AL. These analyses are exploratory and should be treated with caution because they are underpowered and conducted on the available case population. For each variable, there is overlap of each subgroup CI, suggesting that there is no evidence of a differential treatment effect for any of the subgroups.

FIGURE 10.

Forest plot of effect sizes within pre-planned subgroups (AL vs. Immersion PUVA).

Exploratory objective – remission and corticosteroid use

The end of remission is defined as no longer having a clear/almost clear PGA, and of the 132 participants who were observed to be clear/almost clear by the end of the treatment phase, 81.8% (n = 108) were confirmed to have a more severe PGA at a subsequent assessment, corresponding to 90.7% (n = 68) of those in the AL group and 70.2% (n = 40) in the Immersion PUVA group (Table 20).

TABLE 20 - Number of responders who remained clear/almost clear

	AL (%)	Immersion PUVA (%)	Total (%)
Participants remained in remission
Yes	7 (9.3)	17 (29.8)	24 (18.2)
No	68 (90.7)	40 (70.2)	108 (81.8)
Total	75 (100)	57 (100)	132 (100)

Figure 11 shows Kaplan–Meier curves for the time to the end of remission. Inspection of the curves leads to the conclusion that there is no evidence of a difference between treatment groups.

FIGURE 11.

Kaplan–Meier plot for the time to the end of remission, where end of remission is defined as no longer having clear/almost clear PGA.

In general, there seemed to be a higher proportion of participants allocated to AL who reported no use of corticosteroids compared with participants allocated to Immersion PUVA; however, these data are subject to high levels of missing data (Appendix 2, Figures 27 and 28, Table 49).

Treatment compliance

Aims and objectives

Within-trial analysis

The within-trial analysis was aimed at estimating the cost-effectiveness of the intervention(s) from the NHS perspective and the societal perspective. The analysis compared costs and consequences for over 12 and 52 weeks post start of treatment. The outcome measures were QALYs estimated from utility values collected during the trial, while costs were based on estimates from resources used by the trial participants (more details below). The analysis was performed using R studio^®.

Outcome measures

Quality-adjusted life-years were used as an outcome measure; these are a generic measure of health that take account of both quality and length of life, such that one QALY is equal to 1 year of life lived in a state of full health. 68 For the primary analysis, health-related QoL was estimated using responses from the EQ-5D-3L. 69 A further analysis mapped the responses of the DLQI to the EQ-5D-3L instrument to estimate QALYs (more details below).

Measurement and valuation of outcomes

Responses to both the EQ-5D-3L and the DLQI instrument were collected at baseline and weeks 12, 24, 36 and 52 post planned start of treatment. To estimate QALYs from the EQ-5D-3L instrument, we used standard UK tariff values (based on UK general population preferences) to estimate patients’ utility at each collected time point. QALYs were then calculated using the ‘area under the curve’ approach. 69 These methods use two adjacent time points to estimate QALYs for a specific time period. We used the following formula to estimate QALYs at week 12 (1) and week 52 (2).

Where EQ5D_Baseline, EQ5D₁₂, EQ5D₂₄, EQ5D₃₆ and EQ5D₅₂ are the EQ-5D scores at baseline, week 12, week 24, week 36 and week 52, respectively, while ti is the time in weeks between two given measurements over the 12- or 52-week period (t₁ = 12/52; t₂ = 12/52; t₃ = 12/52; t₄ = 16/52).

The DLQI responses were mapped to EQ-5D-3L using the method described by Ali et al. 70 This method was used by Blank et al. 71 The mapped responses were then used to estimate the DLQI QALYs using the area under the curve approach described above.

Cost

All healthcare resources used to estimate the costs of both interventions were collected via a health resource use questionnaire. This was aimed to collect information on resources used or costs incurred by the patients during their participation in the trial and follow-up period. The questionnaire was administered at weeks 12, 24, 36 and 52.

The questionnaire captures information to estimate the costs incurred by NHS in the provision of the treatment and other healthcare resource utilisation, including medications over the trial follow-up. It also included a section to capture patients’ out-of-pocket expenditures, including travel expenses, creams, medical aids such as gloves, and the potential cost of a day out of work to estimate the potential overall burden of both treatments on patients’ income.

Valuation of resource use data

Wherever possible, unit costs for resources were obtained from national sources: NHS reference costs were employed to value hospital resource use (e.g. outpatient attendances). Costs of medical personnel, such as general practitioner (GP) and nurse consultations, were obtained from the PSS Research Unit Costs of Health and Social Care. 72 Medication costs were taken from the British National Formulary (BNF) and the Prescription Cost Analysis (PCA) for England. 73,74 Table 26 contains the costs of the resources used during the trial. These unit costs were then used to estimate the total healthcare costs per patient during the intervention and follow-up period.

Treatment costs were then added to the estimated healthcare costs to estimate the total NHS and PSS costs. According to protocol, patients were to receive their allocated treatment up to week 12 depending on their HE status. However, patients could withdraw from their allocated treatment at any time point during the intervention and follow-up phase due to side effects, lack of effectiveness or other. A proportion of these patients were offered an alternative treatment option, some including AL or Immersion PUVA. Our estimates of treatment costs were then based on patients’ responses on the duration of their original randomised treatment and the offered alternative if applicable. Table 26 contains the resources used by participants with their unit costs. Resources used related to their allocated intervention are described in Table 27. Patients’ responses along with the details described in Tables 26 and 27 were used to estimate the healthcare costs of both interventions at weeks 12 and 52 (excluding out-of-pocket expenditure). These healthcare costs estimates comprise those incurred during the intervention phase, follow-up phase and pharmacological treatment alternatives offered to patients who stopped their initial allocated intervention and were offered an alternative. This information was captured monthly. All costs were updated to 2022 British pound sterling (GBP) prices using the healthcare index. 75

Personal and social or societal costs were collected at the same time as the resource use collection schedule (weeks 12, 24, 36 and 52). Information collected included items such as gloves and creams (including steroids and emollients), travel costs and loss of workdays because of HE.

Total costs for the base case analysis were those incurred by the NHS and PSS. These were calculated by adding all resources used by participants as reported in the questionnaires plus the treatment cost based on their initial treatment allocation and treatment alternatives if treatment was changed or stopped during the treatment and follow-up phase.

Secondary analysis was aimed at exploring the impact of both interventions including the patient perspective. To achieve the latter, we added to the total NHS and PSS costs the total out-of-pocket expenditure as reported by the participants. These costs included those due to loss in productivity, measured by loss of workdays because of HE. To estimate the average cost of a workday lost, we use the average wage or income per day for people of working age depending on their work status (i.e. full-time, part-time, self-employed or government-supported training). Loss of income was accounted for those patients who reported being at work during the follow-up period and indicated that they required time off work due to their skin condition. This information was obtained from the Office of National Statistics UK and the UKRI. 80,81

Population, temporality, discounting rates and cost-effectiveness threshold

The analysis adopted the ITT perspective. Alternative population analysis such as complete data was also performed as part of the sensitivity analysis (details below).

The within-trial analysis was carried out at the end of weeks 12 and 52. At these time points, we estimated the mean QALYs and costs associated with each treatment option and compared them to evaluate which strategy was considered cost-effective. Costs and QALYs were not discounted, as they incurred within a year from randomisation, as recommended by NICE. Cost-effectiveness was determined by comparing the estimated incremental cost-effectiveness ratio (ICER) against the NICE threshold value (λ). The ICER was estimated as follows (3):

The ICER represents the additional cost per QALY gained for each intervention compared with the next best alternative. 82 The more expensive intervention will be considered cost-effective as long as its ICER is below λ = £20,000 or within the £20,000–30,000 per QALY range.

Data cleaning and analysis

Data were analysed using R Studios^®. Face validity tests were conducted on data (e.g. to identify mis-spelled text) and checked against the source documents. Mis-spelled text or different connotations for the same item were found in the responses, particularly on the out-of-pocket expenditure data sets. Any corrections or standardised use of terms was done on a copy of the data set to avoid altering the original responses from the questionnaires.

Long-term cost-effectiveness analysis

Decision analytic modelling was undertaken to extrapolate costs and outcomes beyond the duration of the trial, irrespective of statistical significance in trial results. The model had a 10-year duration, as previous work had demonstrated that the ICER was likely to change marginally on a higher time horizon; an analytic 10-year horizon, therefore, appeared long enough to capture all the main costs and benefits of both treatments. 41

To achieve the latter, we developed a Markov decision analytic model to compare both interventions (AL vs. Immersion PUVA) in the management of severe chronic HE. The starting point was at the end of week 24, corresponding to the end of the intervention phase. This allows some data collected during the last 28 weeks of follow-up to be used to populate the model (more details below).

The model structure is shown in Figure 12. The main elements of the model were aimed to capture patients’ response to treatment at week 24 and every 3 months until 10 years. The model included three mutually exclusive health states depending on the status of the patients’ HE: clear/almost clear; mild/moderate; and severe. Death from natural causes at any stage of the above-mentioned states was also considered.

FIGURE 12.

Markov model structure.

Health states and treatment effects beyond the end of the intervention phase

The long-term impact of both interventions was sourced from the trial data. The proportion of patients considered to have a HE that was clear/almost clear, moderate and severe at week 24 was used to determine the initial proportion of patients in the Markov chain. Transitions from weeks 24 to 36 and then to week 52 were those observed in the trial. Transitions from week 52 onwards were assumed to be the same as those observed between weeks 36 and 52. According to the trial protocol, after week 24, all patients will have stopped their initially allocated treatment at randomisation. After this week, patients who required additional treatment followed what was considered standard practice. This includes ciclosporin A, methotrexate, acitretin and azathioprine but also intensified topical treatment with immunomodulatory agents such as glucocorticoids, calcineurin inhibitors or Jak (Janus kinase) inhibitors. The standard practice may also include treatment with Immersion PUVA or AL irrespective of the patient’s initial treatment allocation. Treatment decisions were reviewed every 4 weeks and continued or modified accordingly. After consulting with the clinical team of the trial, it was agreed that the clinical transitions between HE status observed between weeks 36 and 52 could be assumed to be carried on for transitions from week 52 onwards. This assumption implies that self-awareness, recognition of symptoms, treatment experience and standard of care management, if required, would allow patients to maintain or fluctuate between clear/almost clear and moderate, while patients who had severe HE after 52 weeks would either remain there or move towards moderate or clear. Other studies have assumed that the time to relapse for patients under AL was 24 weeks. Based on the data collected from the trial, most patients will likely fluctuate between clear/almost clear and moderate from 24 weeks onwards, while not many patients return to the severe state. The proportion of patients in the clear state for AL was 32%, while this proportion slightly increased to 33% at week 36 and 38% at week 52. Similarly, for Immersion PUVA, the proportion of patients with a clear/almost clear status at week 24 was 41%; at week 36, this proportion reduced to 33%, increasing again to 45% at week 52. The transition probabilities used for the model are shown in Appendix 3, Tables 51–54. The length of the model cycle was 3 months. To fit this with the actual data, and in consultation with the clinical team of the trial, we assumed that patient status at week 52 was equivalent to that observed at week 48.

Death for natural causes was considered from transitions beyond week 52. The proportion of patients dying from natural causes was obtained from the Office for National Statistics (ONS) life tables for all causes of mortality and was considered depending on the age of the patient.

Adverse events of treatment

Two of the main AEs reported by using AL are the potential increase of cholesterol and triglyceride levels and persistent headaches. Both AEs were considered during the trial design and were closely monitored by the treating physician. If the patient showed any of these events in a severe form, treatment dose was reduced, or treatment was suspended. In the case of discontinuation of AL as randomised treatment, alternative treatment according to standard of care was given. Given that these events were mainly observed, and action was taken during the intervention phase and follow-up period (up to 52 weeks), there were no concerns in terms of potential long-term implications or future dropouts after 52 weeks for these reasons.

Quality of life

Utility values collected at baseline, week 12 and week 24 were used to estimate QALYs at week 24. We follow the under-the-curve approach described above for this estimation. As the Markov model’s initial population was divided by HE status, we estimated QALYs at week 24 by this variable (clear/almost clear; moderate; and severe). These estimates were adjusted by baseline unbalances to avoid potential initial bias at the randomisation stage. The utility reported from the trial at week 36 was used for the initial cycle, while the utility reported at week 52 was assumed for the second and future cycles of the model (Appendix 4, Tables 56 and 57).

Costs

Costs up to week 24 were estimated based on those reported in the health economics questionnaires up to that week. To these costs, we added the treatment costs based on the treatment offered at randomisation but using the treatment continuation reported by patients to accurately cost the time under AL or Immersion PUVA.

Table 27 shows the items included in each treatment arm. As described above, if a patient had their treatment discontinued, they were offered alternative treatment according to standard care. These costs were estimated from the resource use questionnaires and may include AL or Immersion PUVA as well as other alternative treatments offered, according to the patient’s treatment plan. A treatment plan for these patients was reviewed and updated every 4 weeks up to week 36 and then at week 52. We adjusted the estimated costs between weeks 36 and 52 to fit with the model cycle duration (every 3 months). After discussion with the clinical team of the trial, the average costs during the last 4 months of the intervention phase (between weeks 36 and 48) were considered representative of those related to a long-term HE standard of care. Up to this point in time (and confirmed by the treatment continuation responses), it was unlikely that patients would be still receiving AL or Immersion PUVA, even as part of their standard care package. This was confirmed by the data analysis, as no patient remained under such treatments at week 52. As such, the adjusted costs between weeks 36 and 52 (adjusted to 48) were assumed as standard care costs and were used from this point onwards in the model.

Discounting

Discounting of costs and outcomes was used after week 52 according to NICE guidelines. 67 The discounting rate was 3.5%.

Model uncertainty

Parameter uncertainty was addressed through probabilistic sensitivity analysis (PSA) using Monte Carlo simulation. Transitions between HE states followed a Dirichlet distribution, while 3-monthly costs and QALYs followed a gamma and beta distribution, respectively (Appendix 3 – Long-term model transition probabilities, cost, QALYs parameters; Tables 51–55). From these iterations, we estimated the mean ICER comparing AL versus Immersion PUVA. We also developed a scatterplot and cost-effectiveness acceptability (CEAC) curves and calculated the INMB of both interventions. Cost-effectiveness was determined by comparing the ICER against the NICE recommended threshold of £20,000 per QALY gained. This was supported by the NMB, in which case the intervention with a positive or above-zero INMB was cost-effective. We performed one-way and scenario deterministic sensitivity analysis. These were: both interventions had equal costs after week 52; both interventions had equal costs and outcomes after week 52; and lastly, both interventions had equal costs, outcomes and transition probabilities between health states after week 52.

Expected value of perfect information

The long-term decision analytic cost-effectiveness model was used to estimate the value of further research on the use of AL or Immersion PUVA [expected value of perfect information (EVPI)]. We estimated a 5-year EVPI, assuming a 1-year HE prevalence of severe HE of 1.2% and a yearly incidence of 0.955 per 1000 people, based on Quaade et al. , and a £20,000 willingness to pay. 1 The 5-year population was discounted at 3.5% yearly rate.

Results

The overall within-trial analysis results indicate that AL is the cost-effective strategy. These results are robust, as the probability of cost-effectiveness is over 96% in favour of AL in all scenarios. The long-term (10 years) analysis indicates a slight advantage for Immersion PUVA; however, the small difference in costs and QALYs indicates that the results are uncertain, as only 50% of the iterations favour Immersion PUVA.

The within-trial estimated costs indicate that Immersion PUVA is more costly than AL. Treatment costs are the main drivers of Immersion PUVA costs (these include outpatient visits, the cost of therapy, medication and follow-up). Out-of-pocket expenditures are also higher for patients originally assigned to Immersion PUVA. Alternative treatments offered were higher for AL at week 52 (Table 29). The resource use breakdown can be found in Appendix 5, Tables 58–61.

The adjusted QALYS for both interventions at weeks 12 and 52 are shown in Table 29. Immersion PUVA’s QALYs are higher at both time points. The difference at week 12 is negligible (< 1 full day of health). Similarly, at week 52, Immersion PUVA participants reported an unadjusted increase in QALYs of 0.036, which is equivalent to around 13 days of full health.

Within-trial analysis

Long-term cost-effectiveness model results

The estimated costs at 10 years were £5432 for AL and £5361 for Immersion PUVA, while the estimated QALYs were 6.530 and 6.536, respectively, indicating that Immersion PUVA is dominant over AL and therefore is the cost-effective strategy (Table 38). However, the results are uncertain, as the probability of cost-effectiveness is equal between the two interventions (50%). The latter is due to the negligible difference in QALYs (< 0.01) and a small difference in costs (< £75) (see Table 38).

The scatterplot shown in Appendix 6 (see Figures 34 and 35) shows the uncertainty of the results, as the cloud of iterations is equally distributed between the two interventions, while the CEAC at £20,000 indicates that both interventions have an equal probability of being cost-effective.

The change in the results from the long-term sensitivity analysis in all the scenarios analysed (equal costs; equal costs and outcomes; and equal costs, outcomes and transition probabilities from week 52) showed a relatively small variation in costs and QALYs with no change in the cost-effectiveness of Immersion PUVA. Similarly, results remain uncertain, as only 51% of the iterations indicate that Immersion PUVA is the preferred option. The latter is given by the relatively bigger proportion of patients in the clear/almost clear and moderate states in the Immersion PUVA arm at the end of week 52 (see transition probabilities). Results of these analyses can be found in Appendix 7, Tables 62–64.

Expected value of perfect information

Considering all adults in the UK between the age of 15 and 75 years (49.3 million people), we estimated that 610,000 people may suffer from severe HE in one given year, with a yearly incidence of 46,950. 87 Using those estimates, the individual EVPI was £5145, while the total EVPI indicates that research in this area may be up to £4 billion in a 5-year time period.

Discussion

The within-trial and long-term cost-effectiveness analysis comparing AL versus Immersion PUVA indicate that AL is the cost-effective strategy at weeks 12 and 52, while the long-term analysis indicates that Immersion PUVA is cost-effective, but with a high degree of uncertainty.

The within-trial analysis results are, however, robust, as the probability of cost-effectiveness indicates that in 96% (or above) of the 10,000 iterations, AL is the cost-effective strategy, while the long-term analysis is uncertain, as the PSA analysis indicates that both interventions have equal probability of being cost-effective.

In both analyses, the cost difference is the driver of the cost-effectiveness results. The within-trial analysis estimates that AL is over £1300 cheaper during the intervention phase (up to week 12) and over £1000 cheaper at week 52 (intervention and follow-up phase). The latter reflects the relevant cost incurred by repeated outpatient visits to receive Immersion PUVA during the intervention phase. The follow-up phase between weeks 12 and 52 indicates that costs incurred by patients initially under Immersion PUVA are lower than for those under AL. As a result, the long-term costs of Immersion PUVA (at 10 years after the intervention) are almost equal to those of AL, as the cost difference between the two at 10 years is almost equal (£71 difference).

Within-trial analysis estimating QALYs based on patients’ DLQI responses offers similar results to those of the base case analysis, where AL is the cost-effective strategy. In contrast, however, the mean estimated QALY difference favours AL instead of Immersion PUVA. These estimates indicate that AL is the dominant strategy (less costly and more effective than Immersion PUVA).

The out-of-pocket analysis performed indicates that patients undertaking Immersion PUVA spend on average £287 more than patients under AL at 12 weeks and £539 over 52 weeks. The latter is likely to be related to the increased travel expenditure and/or potential loss of workdays to attend treatment. The increase in out-of-pocket expenditure reaffirmed the cost-effectiveness of AL during the treatment and follow-up phases.

As a limitation of this study, the results obtained from the 10-year analysis assume that patients who were relatively successful in managing their HE during the treatment and follow-up phase will be more likely to effectively treat their HE if they experience a recurrence. This means that patients are likely to remain or fluctuate between almost clear and moderate HE for the 10 years following the intervention, as they will be able to identify and manage their condition better along with their treatment physician. People with severe HE at the end of the follow-up period will likely remain there or move to moderate eczema. We made this assumption based on clinical expertise due to the lack of long-term data on the effectiveness of AL or Immersion PUVA. Other studies evaluating AL have assumed that patients relapse after 24 weeks and assumed retreatment of eligible patients with AL. 71 However, these assumptions contradict our observations in the trial period, which indicate that patients on average remain in either the clear/almost clear and/or the moderate state after 24 weeks post treatment. Furthermore, phototherapy guidelines generally do not recommend more than one Immersion PUVA course per year, and there is an upper limit of (lifetime) exposure sessions for hand and foot Immersion PUVA of around 200. We evaluated alternative scenarios to evaluate the potential impact of this assumption; on those scenarios, however, the results remain unchanged.

Although expected value of parameter information was not performed, it is clear that further research in the area investigating the long-term impact of these two interventions would help reduce the uncertainty of the long-term results, as indicated by the EVPI, who estimated that further research in the area up to £4 billion is considered good use of public funds. Future research including expected value of perfect partial information (EVPPI) would help determine which other parameters future research should focus on.

Sample size and primary outcome measure

Choosing the best-suited outcome measure is of high importance for a clinical trial. Within ALPHA, we have used several outcome measures. For the assessment of HE severity, the HECSI, mTLSS and PGA were used. The reason for this was that none of the outcome measures were the gold standard (see discussion below) and that we wanted to be in a position to compare our data with other HE trial results when different outcome measures have been used.

The HECSI was chosen as the primary outcome measure because it incorporated more details about the condition compared with PGA, which had been used in previous trials. 52,91–94 Furthermore, the HECSI is continuous rather than discrete or binary, which meant that a smaller sample size could be achieved for a desired level of power. There were few published data on the target treatment effect for the HECSI, but due to the lognormal distribution of the outcome, it was reasonable to base the sample size on a relative difference, or fold change. This was discussed at length with the clinical co-applicants, including demonstrating what different fold change values equated to in terms of absolute differences in the HECSI score. This led to an agreed target fold change of 1.3 (or 0.77). At the trial design stage, there was great uncertainty in the assumptions for the sample size calculation, which was based on the HECSI. The trial therefore included a pre-planned sample size re-estimation. This was conducted earlier than planned due to slower than anticipated recruitment, but the final analysis showed that the estimates of the CV obtained at the sample size re-estimation were consistent with the final trial results. Furthermore, the trial was originally powered at 80%, and the pre-planned sample size re-estimation provided confidence in the underlying sample size assumptions to ensure that the revised target would provide sufficient power.

Participants responding to treatment

Previous trials have reported on the efficacy of AL. Depending on the patient selection criteria, around half of the patients treated with AL 30 mg/day achieved a PGA of clear/almost clear after 12 or 24 weeks, respectively. In an open-label study by Dirschka et al. , the responder rate was 46.6%. 93 This study enrolled HE patients with PGA severe, and 91% of that cohort (n = 249) showed a hyperkeratotic HE morphology. Thaci et al. found in a cohort (n = 631) with moderate to severe HE at baseline (62% PGA severe; 34.1% PGA moderate), of which 61.2% had a hyperkeratotic morphology, that 31.7% and 29.8% of all patients receiving at least one dose of AL had a PGA status of clear/almost clear at week 12 and 24, respectively. 95 The initial AL study by Ruzicka et al. recruited 1032 participants with severe CHE, of whom more than 80% had hyperkeratotic eczema morphology: at the end of the treatment phase (maximum 24 weeks), 47.7% of patients in the 30 mg dose group and 27.5% in the 10 mg dose group achieved clear/almost clear compared with 16.6% in the placebo group. 52 The ALPHA trial was balanced regarding predominant eczema morphology (hyperkeratotic rhagadiform vs. vesicular) as well as atopy status. Participants for the ALPHA trial were eligible if they had a severe PGA according to the treating clinician at the time of randomisation. By the end of the treatment phase in ALPHA, 34.1% of AL participants were confirmed to have achieved a clear or almost clear response, but this lower rate compared with previous trials may be due to the differences in the patient population recruited.

Regarding Immersion PUVA, Brass et al. reported a clear/almost clear response rate to Immersion PUVA of 43% after 12 weeks in 23 participants. 91 We are not aware of other larger trials on Immersion PUVA in severe HE that used the PGA as outcome measure. In the ALPHA trial, the proportion of participants with available data who had achieved clear or almost clear according to the blinded assessor was equal to 23.6% (35/148) for the Immersion PUVA group. This difference in proportion may be due to the small sample size reported in Brass et al. 91

Ethnicity/skin of colour

Information on HE in skin of colour is sparse. Indeed, we failed to identify any publication or trial that specifically mentions skin of colour. However, it has been highlighted in several reports that eczema and other inflammatory skin diseases may be underestimated in severity due to difficulties in scoring erythema intensity in skin of colour. 96,97 The ALPHA trial included 11.6% of participants with different ethnic backgrounds, and the proportion observed in those randomised was similar to the proportion of the participants who were screened for participation. At the start of the trial, the issue of assessment challenges for skin of colour was raised by the PPI representative on the TSC. Following discussion with the clinical members of both the TMG and the TSC, a decision was made to incorporate skin colour as a minimisation factor, and to take photographs at baseline and week 12 for all consenting participants with skin of colour. There was already a pre-planned substudy to take photographs of 30% of consenting participants, and this remained for participants with white skin. Each photograph was assessed by three blinded assessors, and there was no evidence in our trial of a disadvantaged assessment regarding to people with darker skin colour. However, there does not exist a ‘gold standard’ regarding assessment of erythema, which is a morphological feature in most composite scores for inflammatory skin diseases, in this patient population. Furthermore, the photographic guide for HE was used to guide the central review; however, it was developed on photographs of white patients only. 98 A later study was conducted to determine whether the photographic guide could be used for HE self-assessment; however, the study only included participants with white skin due to the limitations of the photographic guide, stating that further research would be required for patients with darkly pigmented skin. 99 Within ALPHA, consent was also received for any photographs taken to be used in further research, which may include extending a photographic guide for patients of ethnic minority background.

Subtypes of hand eczema and therapy response

As mentioned above, a multitude of HE subtypes have been described, some based on morphology (dishydrotic, hyperkeratotic, rhagadiform, tylotic, nummular, psoriasiform) or assumed aetiology (irritant, atopic, contact allergic). In general, there is currently no strong evidence to support a relationship between morphology of HE and response to therapy. However, some important information is well known and accepted.

Firstly, the only viable solution to treat contact dermatitis to a relevant contact allergen is allergen avoidance. None of the treatments available will overcome allergen-triggered contact allergic inflammatory activity in the long term. Of note, allergic contact dermatitis is a common feature of occupational HE. In the ALPHA trial, non-avoidance of relevant contact allergens was an exclusion criterion. However, there have been challenges across UK centres regarding testing for allergic contact dermatitis (patch test), and this will be discussed below.

Secondly, AL has been described as showing higher response rates in hyperkeratotic rhagadiform as compared with dishydrotic or vesicular/pompholyx phenotypes. 52 However, that does not imply that the latter group fails to respond to AL. Studies are under way to investigate specifically the response to AL as compared with ciclosporin A in vesicular types of HE. 88

In ALPHA, the proportions of participants with hyperkeratotic versus vesicular HE types were balanced across both treatment groups. In the exploratory analysis, there was no evidence of a differential treatment effect by morphology on disease severity [see Forest plot (see Figure 10)]. However, an analysis exploring the relationship between morphology and outcome was not conducted because it was outside the trial aims and objectives. There is also the expert opinion that AL may work better on lesions in the palmar than on the dorsal side of the hands. Although dorsal/palmar lesion data were collected, the relationship between hand side and treatment group on treatment response was not a pre-specified analysis and therefore is the subject of future research.

Diagnosis/differential diagnosis

In cases where eczema is restricted to the palmoplantar area and when presenting with hyperkeratotic morphology, it can be difficult to distinguish between eczema and palmoplantar psoriasis. 100,101 Experience in clinical practice and throughout the ALPHA trial reflects that absence of other signs of psoriasis or eczema results in diagnostic uncertainties. In some patients, diagnostic biopsies were necessary to confirm the diagnosis. Other differential diagnoses that came up included palmoplantar keratodermas and Bazex syndrome. Palmoplantar psoriasis, however, causes the largest diagnostic challenge.

Allergic contact dermatitis

It is a generally accepted expert view that every patient with chronic hand and/or foot eczema should have a patch test assessment to exclude contribution of a relevant contact allergen to the clinical symptoms. Thyssen et al. reported full consensus regarding recommendation for a patch test in all patients with HE of more than 3 months duration and who are non-responsive to adequate treatment or if there is clinical suspicion of contact allergy. 14 However, the current situation in UK dermatology is not ideal, and the severe understaffing and waiting lists fail to allow appropriate patch testing for patients in need within a reasonable time frame. There is, therefore, underdiagnosis of contact allergies. It is mentioned above that avoidance therapy is an absolute necessity in the face of relevant contact sensitisations. However, the ALPHA trial aimed to mimic clinical practice and thereby followed a pragmatic approach, resulting in the following pathway: in the case of severe CHE without a strong patient history indicating a contact allergic trigger, treatment was started even without available patch test results. Although far from ideal, it was considered unethical to leave patients untreated who suffer from severe HE, impairing functioning in daily life and causing significant pain and itch. This is in line with routine clinical practice in the UK. This approach implies that a minority of the participants included in the trial may have suffered from a non-identified, relevant contact allergy. In addition, it is important to highlight that even with patch test clinics in place, these may not be supervised by experienced allergologists in many centres. In-depth knowledge in the area of contact allergy is necessary to be aware of potential allergens to screen for and to include relevant chemicals regarding co- and cross-sensitisations. As such, even in patients in whom patch test had been performed, appropriate batteries/panels may not have been tested due to lack of specialists and training in contact allergy in some centres in the UK. Furthermore, in-depth knowledge about cross-reacting chemicals and where relevant, contact allergens that may be expected to be encountered is necessary for optimal medical avoidance strategy advice.

Factors influencing hand eczema severity and response to treatment: lifestyle factors, atopy status and genetic background

Immersion 8-methoxypsoralen combined with ultraviolet A as comparator treatment arm

8-Methoxypsoralen combined with ultraviolet A is a well-recognised treatment for HE. Our survey performed among UK dermatologists supported PUVA being among the first-line therapies for severe chronic HE in the UK. 42 Debate is ongoing about the best modality to deliver PUVA, with the following options being available: systemic PUVA requiring oral administration of a photosensitising agent, Immersion (also referred to as bath or soak) PUVA requiring soaking of hands in 8-methoxypsoralen prior to Ultraviolet A (UVA) exposure, and cream and gel PUVA, where the photosensitising agent is applied prior to UVA exposure. The most widely used approach in UK centres and Europe is topical PUVA, mostly using PUVA soaks. There are no published trials comparing the different PUVA modalities for HE, but only smaller case series. 104 For palmoplantar psoriasis, a case series showed that both bath and systemic PUVA showed a potential benefit at the 4-week time point, with systemic PUVA showing rapid and possible increased efficacy at earlier time points but also more systemic side effects. 105 It is well documented that systemic administration of oral methoxypsoralen (8-MOP), a deoxyribonucleic acid (DNA) intercalating substance, comes with potential side effects and contraindications. A proportion of patients wish to avoid oral medication and are worried about side effects, and it is thus important to also offer patients the choice of Immersion PUVA. Numerous studies have shown the efficacy of topical PUVA in HE conditions. It has been suggested that atopic HE may respond better than hyperkeratotic HE variants. 106,107

Although the decision to use Immersion PUVA as the comparator was supported by the survey results, it was recognised that there may be challenges with implementation in the trial. Firstly, Immersion PUVA delivery schedule is limited in many UK dermatology centres due to staff shortage. While it was common practice in the past to offer slots to patients on Saturdays and evening times, treatment slots outside general working hours are no longer offered by most phototherapy units. This situation had significant consequences for recruitment, as patients with strict work-hour commitments (e.g. teachers) were unable to participate. Indeed, a large proportion of patients approached mentioned inability to attend Immersion PUVA treatment due to work, travel difficulties or other problems as the main reason for not participating in the trial. The Immersion PUVA treatment arm was also more affected by the COVID pandemic than the AL treatment arm.

The difficulty of accessing Immersion PUVA treatment is of concern. The trial shows that Immersion PUVA and AL led to a reduction in disease severity, and in the longer term there was no evidence of a difference in clinical effectiveness compared with AL as a first-line therapy. This means that Immersion PUVA could be considered as a potential first-line treatment choice for longer-term outcomes, particularly in patients where AL is not considered an appropriate treatment. However, compliance with Immersion PUVA was low, and so there is a need to improve accessibility to treatment, such as through home delivery solutions, which have been explored for ultraviolet B (UVB) treatment in a recent NIHR vitiligo trial. 108

Treatment feasibility – pregnancy prevention programme for alitretinoin

Alitretinoin is a retinoid and thus belongs to a group of drugs that are teratogenic. Pregnancy prevention programme (PPP) rules must be followed. The need for two independent contraceptive measures is of concern for some patients. The need for one monthly face-to-face appointment with pregnancy testing can be a challenge in some understaffed dermatology units. The COVID pandemic adopted some remote solutions, where pregnancy tests were sent to patients, who had to report back the results.

Pregnancies and ALPHA

In the context of the ALPHA trial, which prohibited any pregnancy while receiving trial treatments as intervention as well as during follow-up, four pregnancies were reported. Only two pregnancies occurred under retinoid therapy, one under AL and another one under Isotretinoin, which was given as standard-of-care medication during the follow-up period for acneiform lesions unrelated to the HE. Interestingly, none of the pregnancies occurred during the COVID pandemic/lockdown, during which it was difficult to perform in-attendance pregnancy tests.

With regard to PPP, the ALPHA trial followed standard of care. Lessons learned from the occurrences are the following: (1) patients who are assumed to be knowledgeable regarding pregnancy prevention (e.g. healthcare professionals) should be given the same attention and degree of information as all other patients; (2) exceptions to the PPP measures (e.g. BAD leaflet for Isotretinoin with exceptions) should be handled with utmost care. We have sufficient evidence that not being in a current relationship does not preclude ‘pregnancies’.

Assessment tools for hand eczema

Currently, there is no gold standard for HE assessment tools. This has prompted some researchers to use the initial photographic assessment guide, which has been the basis for the development of assessment tools, including PGA. 88,98 However, the photographic guide has significant shortcomings, including the absence of a ‘mild’ category due to lack of expert consensus on this item. HECSI has been used more recently in research, as it most accurately captures extent and morphology items. However, it requires some training and dermatology lesion knowledge. The items oedema and infiltration are difficult to distinguish. HECSI also does not allow separate assessment of each hand; the areas assessed include, for example, palm/dorsum or fingers of both hands. 51 The PGA is much faster to perform but very unclear in its criteria, and it only focuses on the most severely affected side of the most severely affected hand. mTLSS was also included in ALPHA to allow comparison with previous trials, including the Ruzicka et al. study. 52

We identified shortcomings in the QoL assessment, in particular the DLQI. This widely used score is derived as the sum of scores provided for 10 questions. Current NICE guidelines (TA177) require a score of 15 for eligibility for AL treatment. This cut-off is considered to be an unusually high threshold, given that any DLQI above 11 is considered as ‘very large effect on patient’s life’. 57 However, in ALPHA, only 43.3% of our participants with severe CHE had a DLQI score of 15 or above. Moreover, clinicians’ experience throughout the trial was that elderly patients, patients with pre-existing disabilities that restricted mobility, patients not working and patients who were perceived as ‘coping’ very well with their disease were ‘disadvantaged’. Marron et al. reported a median DLQI of 7 in a cohort of HE patients who showed increased anxiety and depression parameters. 9 Agner et al. reported in a cohort of 416 patients a significant correlation between HE severity as assessed by HECSI and DLQI. 109 A median score of 8 was considered to be in agreement with a reported 7.8 in severe occupational HE. 6 This further underlines that a DLQI of 15 or higher seems an unrealistic high threshold for a guideline not supported by clinical reality.

None of the current assessment tools considers fingernail changes. There is no specific tool [such as nail psoriasis severity index (NAPSI) for psoriasis] to score nail involvement in HE. However, HE affecting the nail fold is well recognised to result in visible nail changes, which can impair function and/or be distressing for patients upon social exposure. We here included nail assessment in a single participating centre as a ‘pilot’ assessment: results of this will be published elsewhere.

Photography

The agreement of the photographic assessments with the blinded PGA was consistent between treatment groups and ethnicity groups, but there was generally a low level of agreement (43.7% overall). Further research may be beneficial to further understand the reliability of the photographic guide and how it could be used in trial research. 98,99

Moreover, the extent of misclassification between the photoguide and blinded PGA further supports the use of HECSI, as it measures extent and morphology.

Time in remission and time to relapse

Results on time in remission/time to relapse should be treated with caution and considered only as exploratory. The analysis was based on a small number of participants who had achieved remission, and therefore, the estimated treatment effects are subject to bias.

Lessons learned: recruitment

From our work with patient representatives, we know that many people with severe HE self-manage at home. Based on the experience on the ALPHA trial, reaching patients who were not under the care of a GP or dermatologist posed many challenges on how best to ‘advertise’ the study outside health-care settings. Numerous avenues were explored (online/social media, newspaper, radio, TV, poster campaigns, magazines), and professions associated with HE were approached (hairdressing, building, garden/landscaping, health care/occupational health). We identified that communications/advertising and media required specialist expertise and a significant amount of time to manage, and this represents a specialty/role in its own right distinct from trial management. In future studies with a similar patient population, we would consider employing a member of staff dedicated to this role within the trial.

We know that many patients remain under their GP without referral, and recruitment/patient identification from primary care was incorporated; however, challenges existed as (1) there is no commonly used ‘read code’ on GP systems specifically for HE, using a search term of eczema would produce extensive lists with many not eligible for the study (eczema not on the hands); (2) opportunistic screening during standard care appointments relied upon every GP at a single practice being trained on the study; (3) we could only approach GP practices in areas associated with a secondary care recruiting centre. Many GP practices supported the study by displaying posters with information on the study self-referral website; however, this may have been for a limited duration only rather than the full duration of the recruitment period. Engagement with primary care networks and a structured approach to primary care recruitment, including payment for screening activities, is key, and we note that this has become much more established since ALPHA. Early liaison with clinical research networks and allocation of dedicated resource are required, and this understanding will be taken forward into future studies.

Access to treatment slots for Immersion PUVA treatment was a limiting factor at most participating centres, with a restriction on the number of Immersion PUVA ‘slots’ for trial participants alongside standard care treatment appointments. Where only one or two slots were available for trial participants, once these were in use, the participating centre was unable to randomise another participant until the course of treatment had been completed for the existing participant. This meant that our estimate of one to two per centre per month was not achievable at the majority of centres. Through ALPHA we have developed an excellent understanding of the pathway for patients with severe HE, which can be taken into future studies, and understanding the key questions to ask of participating centres when developing estimated recruitment rates.

Equipoise between the two treatment options proved difficult, as administration of the two treatments required an extremely different level of commitment from participants, and as discussed previously, this prevented a number of people from taking part, as they were unable to commit to the Immersion PUVA schedule. However, we must also acknowledge that a number of participants did not want to take AL (or any other oral medications) due to potential side effects or the requirement to complete a pregnancy prevention programme, and it is important to provide options suitable for all patients. Consideration of alternative ways to deliver Immersion PUVA would be of value in any future trials.

Patient and public involvement

Patients were involved in the design of the trial during development of the grant application, and throughout the trial via membership on the TMG and the TSC.

Feedback from our TMG patient representative and PPI group held in September 2019 detailed that patients often self-manage at home with over-the-counter creams for many months prior to seeking a GP appointment, and this is ‘the last resort’ as they have ‘tried everything else but nothing is working’. There is frustration at not being taken seriously, and despite meeting NICE criteria for referral to a secondary care dermatology centre, many receive repeated prescriptions for steroid creams prior to a referral being made, resulting in delays of months to years from first GP visit to seeing a dermatologist. At this point, patients describe feeling ‘desperate’, with their severe HE having had a significant impact on their daily lives, QoL and well-being for an extended period of time. Importantly, delay in medical attention to HE is recognised to be associated with a poorer prognosis. 102

When discussing the findings of ALPHA, our patient representative felt strongly that ‘getting early relief is really important’ as it had been such a long time managing at home, and the study results showing benefit of AL at 12 weeks would provide this ‘quick fix’, which would be so important.

Easy self-management is a key consideration, and generally a preference, particularly for those who have work or child-care commitments; however, it must be acknowledged that some patients are unable to, or choose not to, take AL, and having alternative treatment options available is important.

Equality, diversity and inclusion

We have previously discussed the considerations and steps taken in ALPHA regarding assessment of different skin types and the photography guide. In addition, ALPHA was able to maximise research engagement of underserved populations (i.e. low socioeconomic, minority ethnic and rural/urban) by opening a large number of centres across the UK representing the devolved nations, large urban city centres and rural communities, including large acute Trusts/Teaching Hospitals and smaller District General Hospitals (DGHs), and incorporating recruitment pathways/strategies that allowed the trial to be accessible to all people, including those not currently under the care of a healthcare professional.

We worked with our patient representatives to develop patient information materials in different accessible formats such as video, website, posters, summary sheets and content for social media platforms (Twitter, Instagram). ALPHA was the first Leeds CTRU trial to use Instagram as a social media platform after it was flagged that this was important to ensure trial information reached people with wider-ranging demographics (such as age). A significant amount of time was dedicated to developing the self-referral pathway via a dedicated website and e-mail account to ensure that it was simple and user friendly, and this has been used as an exemplar moving forward for other large multicentre studies.

Outlook

ALPHA delivers important evidence to feed into standard-of-care situations. It was a pragmatic real-life study with direct relevance to UK dermatology care. The main recommendations for which we provide evidence at this point include the following:

• Patients with severe chronic HE have a considerable benefit from treatment in specialised secondary care. Referrals should be encouraged.

• AL works faster and is more effective at week 12 when comparing Immersion PUVA with AL.

• AL is more cost-effective than Immersion PUVA.

• Immersion PUVA leads to a reduction in symptoms for chronic severe HE and has the potential to be cost-effective after 10 years post treatment. However, access to Immersion PUVA treatment is limited in the UK. There is a strong clinical need to improve access to this treatment for a larger group of patients.

• Current guideline for treatment based on a DLQI of 15 or more is too high and may prevent treatment access for patients with severe CHE.

• Assessment of HE severity requires assessor training.

• HECSI may be a useful primary outcome measure for HE trials, and the results from the ALPHA trial will be useful to inform future CHE trial designs.

Acknowledgements

Contributions of authors

Miriam Wittmann (https://orcid.org/0000-0003-2328-4926) (Associate Professor in Inflammatory Skin Diseases, Chief Investigator) contributed to the design, protocol development, protocol implementation and interpretation of data; was involved in the implementation of protocol and coordination of local approvals at Leeds Teaching Hospitals NHS Trust and Bradford Teaching Hospitals NHS Foundation Trust; contributed to drafting or revising for intellectual content and approval of monograph final version.

Isabelle L Smith (https://orcid.org/0000-0002-8326-1075) (Principal Statistician) contributed to the design, protocol development, protocol implementation, statistical design, statistical analysis of the clinical results and interpretation of data; contributed to drafting or revising for intellectual content and approval of the monograph final version.

Sarah Tess Brown (https://orcid.org/0000-0002-1840-3786) (Principal Statistician) was a co-applicant and contributed to the design, protocol development, protocol implementation, statistical design, and oversight and interpretation of data; contributed to drafting or revising for intellectual content and approval of the monograph final version.

Anna Berekméri (https://orcid.org/0000-0001-5894-8812) (ALPHA Clinical Coordinator) led design of the tape stripping substudy, including protocol development, data analysis and interpretation; developed and delivered the skin assessment scoring tool training package; blinded expert photography review panel; contributed to drafting and revising for intellectual content and approval of the monograph final version.

Armando Vargas-Palacios (https://orcid.org/0000-0002-6503-0134) (Research Fellow in Health Economics) contributed to protocol implementation, health economic analysis plan, data analysis and data interpretation; contributed to drafting or revising for intellectual content and approval of the monograph final version.

Lesley Sunderland (https://orcid.org/0000-0002-5807-530X) (General Practitioner) was a co-applicant and contributed to the design, protocol development, protocol implementation and interpretation of data; contributed to revising for intellectual content and approval of the monograph final version.

Amy Barker (https://orcid.org/0000-0003-1626-0934) (Patient Representative) was a co-applicant and contributed to the design, protocol development, protocol implementation and interpretation of data; contributed to drafting and approval of the monograph final version.

Fiona Cowdell (https://orcid.org/0000-0002-9355-8059) (Professor of Nursing and Health Research) was a co-applicant and contributed to the design, protocol development, protocol implementation and interpretation of data; contributed to revising for intellectual content and approval of the monograph final version.

Steven Ersser (https://orcid.org/0000-0001-6995-6121) (Head and Professor of Nursing and Dermatology Care) was a co-applicant and contributed to the design, protocol development, protocol implementation and interpretation of data; contributed to revising for intellectual content and approval of the monograph final version.

Rachael Gilberts (https://orcid.org/0000-0003-3326-7900) (Head of Trial Management) contributed to protocol implementation and interpretation of data; contributed to revising for intellectual content and approval of the monograph final version.

Cathy Green (https://orcid.org/0000-0003-0920-036X) (Consultant Dermatologist) was a co-applicant and contributed to the design, protocol development, protocol implementation and interpretation of data; carried out protocol implementation and coordination of local approvals and data acquisition for Ninewells Hospital, Dundee; contributed to revising for intellectual content and approval of the monograph final version.

Philip Hampton (https://orcid.org/0000-0001-6797-2697) (Consultant Dermatologist) was a co-applicant and contributed to the design, protocol development, protocol implementation and interpretation of data; carried out protocol implementation and coordination of local approvals and data acquisition for The Newcastle upon Tyne Hospitals NHS Foundation Trust; contributed to revising for intellectual content and approval of the monograph final version.

Catherine Smith (https://orcid.org/0000-0001-9918-1144) (Professor of Dermatology and Therapeutics) was a co-applicant and contributed to the design, protocol development, protocol implementation and interpretation of data; carried out protocol implementation and coordination of local approvals and data acquisition for Guy’s and St Thomas’ NHS Foundation Trust; contributed to revising for intellectual content and approval of the monograph final version.

Jane Nixon (https://orcid.org/0000-0003-1705-7698) (Professor of Tissue Viability and Clinical Trials Research) contributed to the design, protocol development, protocol implementation and interpretation of data; contributed to drafting or revising for intellectual content and approval of the monograph final version.

Disclosure of interests

Full disclosure of interests: Completed ICMJE forms for all authors, including all related interests, are available in the toolkit on the NIHR Journals Library report publication page at https://doi.org/10.3310/TWQC0141.

Primary conflicts of interest: Miriam Wittmann reports grants from Pfizer, UCB, British Skin Foundation, Psoriasis UK Foundation and Sanofi. Miriam Wittmann has received payment for lectures from Sanofi, Leo, Astra Zeneca, UCB, Euforea Masterclass, Novartis and Janssen. Miriam Wittmann reports membership of advisory boards for UCB, Leo, Biogen and Astra Zeneca. Miriam Wittmann reports board or committee membership for Journal of Investigative Dermatology, British Skin Foundation, and British Society of Investigative Dermatology. Sarah Brown reports grants from NIHR (NIHR151863, NIHR131475, NIHR201612). Anna Berekméri reports grants from Leeds Teaching Hospitals NHS Trust. Anna Berekméri reports travel support from Health Professional Scholarship, Health Care Provider Scholarship, and ESDR Travel Grant. Lesley Sunderland reports payment paid to employers for attending TMG meetings. Rachael Gilberts reports grants from NIHR (NIHR151863). Catherine Smith reports grants from the European Commission. Catherine Smith reports payments for speaking at various conferences. Jane Nixon reports membership of the HTA Commissioning Committee 2018–24.

No competing interests are declared by Isabelle L Smith, Sarah Tess Brown, Armando Vargas-Palacios, Amy Barker, Fiona Cowdell, Steven Ersser, Cathy Green and Philip Hampton.

Patient data statement

This work uses data provided by patients and collected by the NHS as part of their care and support. Using patient data is vital to improve health and care for everyone. There is huge potential to make better use of information from people’s patient records, to understand more about disease, develop new treatments, monitor safety, and plan NHS services. Patient data should be kept safe and secure, to protect everyone’s privacy, and it’s important that there are safeguards to make sure that they are stored and used responsibly. Everyone should be able to find out about how patient data are used. #datasaveslives You can find out more about the background to this citation here: https://understandingpatientdata.org.uk/data-citation.

Data-sharing statement

All data requests should be submitted to the corresponding author for consideration. Access to available anonymised data may be granted following review.

Ethics statement

The trial was conducted in accordance with the World Medical Association Declaration of Helsinki (www.wma.net/policies-post/wma-declaration-of-helsinki-ethical-principles-for-medical-research-involving-human-subjects/) and was reviewed and approved by the Yorkshire and Humber Research Ethics Committee (REC reference: 14/YH/1259 on 9 January 2015), the Health Research Authority (HRA) and Medicine and Healthcare products Regulatory Agency (MHRA).

Information governance statement

All personal information was handled in line with the Data Protection Act (2018) and General Data Protection Regulation (EU GDPR) 2016/679. Under Data Protection legislation, the University of Leeds is the Data Processor and Data Controller, and personal data are processed in accordance with their instructions. You can find out more about how personal data are handled here: https://ctru.leeds.ac.uk/ctru-comprehensive-privacy-guide/ and the contact details for the University of Leeds Data Protection Officer below:

• e-mail: DPO@leeds.ac.uk

• Telephone number: +44 (0)113 343 7641

• Postal address for data protection issues: University of Leeds, Room 11.72 EC Stoner Building, Leeds, LS2 9JT

Publications

The following associated publications to this report are listed below:

Smith IL, Brown S, Nixon J, Cowdell FC, Ersser S, Fernandez C, et al. Treatment of severe, chronic hand eczema: results from a UK-wide survey. Clin Exp Dermatol 2016. https://doi.org/10.1111/ced.13015

Berekméri A, Latzko A, Alase A, Macleod T, Ainscough JS, Laws P, et al. Detection of IL-36γ through noninvasive tape stripping reliably discriminates psoriasis from atopic eczema. J Allergy Clin Immunol 142(3):988–991.e4.

Smith IL, Gilberts R, Brown S, Fernandez C, Nixon J, Reynolds C, et al. Comparison of ALitretinoin with PUVA as the first-line treatment in patients with severe chronic HAnd eczema (ALPHA): study protocol for a randomised controlled trial. BMJ Open 2022;0:e060029. https://doi.org/10.1136/bmjopen-2021-060029

Disclaimers

This article presents independent research funded by the National Institute for Health and Care Research (NIHR). The views and opinions expressed by authors in this publication are those of the authors and do not necessarily reflect those of the NHS, the NIHR, the HTA programme or the Department of Health and Social Care. If there are verbatim quotations included in this publication the views and opinions expressed by the interviewees are those of the interviewees and do not necessarily reflect those of the authors, those of the NHS, the NIHR, the HTA programme or the Department of Health and Social Care.

Selection of sampled area

Lesional tape samples were taken from the most representative lesion of either the palmar or the dorsal hand, depending on the lesion localisation. In addition, samples were preferentially taken from non-treated sites, or at sites with no topical treatment for a minimum of 2 days. Lesions with oozing or broken skin (erosions, vesicles, bleeding/oozing fissures) were avoided. Sampling was stopped if any bleeding or oozing occurred during the procedure. Non-lesional samples were collected from the forearm of patients.

Sample processing and protein extraction

Samples for protein quantification by ELISA-based techniques were transferred tape by tape into an Eppendorf tube with 1.5 ml lysis buffer containing 20 mM Tris pH 7.4, 150 mM NaCl, 1% Triton X100, 5 mM EDTA and 1X protease inhibitor cocktail (Roche, Welwyn Garden City, UK). Following transfer, samples were labelled and incubated on ice for 30 minutes. Subsequently, samples were sonicated three times for 20 seconds, with 20-second intervals of cooling on ice between sonication steps in order to avoid overheating of sample and protein denaturation. Finally, samples were centrifuged at 15,000g (maximum rpm) for 10 minutes, and the supernatant was carefully collected. Samples obtained were stored at −80 °C.

The total soluble protein content of each tape stripping sample was determined using bicinchoninic acid assay (BCA – microplate procedure, Life Technologies Ltd, Paisley, UK).

Protein quantification