Surgical procedures and non-surgical devices for the management of non-apnoeic snoring: a systematic review of clinical effects and associated treatment costs

Diagnosis

The aim of investigating patients presenting with signs consistent with non-apnoeic snoring is to rule out more severe SDB disorders (i.e. UARS and OSAS) and other conditions. On initial screening, care is therefore taken to elicit evidence of any symptoms indicative of OSAS, for example EDS. This can be assessed using the Epworth Sleepiness Scale (ESS), which is a validated 24-point questionnaire that aims to assess the likelihood of falling asleep in a variety of situations. The score can be used to clinically subdivide patients into the normal range (ESS < 11), mild subjective daytime sleepiness (ESS 11–14), moderate subjective daytime sleepiness (ESS 15–18) or severe daytime sleepiness (ESS > 18). Depending on the history, in which care will be taken to assess patients presenting with ESS who snore for other underlying causes for the presenting symptoms, and examination findings, patients will be referred for one of several further procedures, including PSG or modified PSG, before formal treatment options are discussed. This is consistent with current guidelines from the Scottish Intercollegiate Guidelines Network on the management of obstructive sleep apnoea/hypopnoea syndrome in adults, which state that OSAS must be excluded before treatment is considered. 2

Sleep studies

The main purposes of sleep studies are to confirm or refute the diagnosis of OSAS or sleep-related problems and to assess their severity to guide management options. Sleep studies may therefore measure some aspects of ventilation either directly or indirectly to assess any possible compromise by upper airway obstruction and any consequences for sleep quality. Studies vary in their complexity from full PSG with recordings of 12–30 channels of various electrophysiological, breathing and movement signals to one-channel information studies such as oximetry. Recordings can be undertaken during part or all of a night’s sleep and may be performed in a sleep centre or at home, depending on the study complexity.

Assessment of the site of airway narrowing

Several techniques have been used to attempt to localise the level of the obstruction implicated in the generation of the snoring sound. Identification of the primary and secondary sites of obstruction may be used to determine the most appropriate treatment options and potentially provide some prediction of treatment outcome. The most commonly used techniques are:

• visual inspection

• fibreoptic endoscopy with or without Muller’s manoeuvre

• sedation (sleep) nasendoscopy

• upper airway pressure recordings during sleep

• imaging and cephalometric radiography

• acoustic analysis of the snoring sound.

Interventions for non-apnoeic snoring

Deciding which of the various treatment options is most appropriate for the management of non-apnoeic snoring depends on the level of airway obstruction and therefore the source of the snoring sound generation; the intensity of the snoring sound and the amount of social disruption caused; and the characteristics of the individual patient.

There are a number of conservative treatment options by which a reduction in snoring may be achieved. These include weight reduction,14 the use of sleep positioning devices53 and a reduction in alcohol intake. 54 These are generally indicated either as a first-line treatment option or as an adjuvant intervention depending upon the patient’s lifestyle risk factors for snoring and the pathophysiological mechanism of snoring sound production. Additionally, non-prescription (over-the-counter) medications such as nasal dilators are also currently available on the market. The effects of non-prescription treatments for non-apnoeic snoring and OSAS have been reviewed by the American Academy of Sleep Medicine Clinical Practice Review Committee. 55

The findings of this review highlighted the paucity and low quality of the evidence currently available to support the use of external nasal dilator strips (ENDS), internal nasal dilators (INDS), lubricant nasal and oral products, oral dietary supplements and magnetic pillows and mattresses in the treatment of both non-apnoeic snoring and OSAS. 55 The Clinical Practice Review Committee concluded that there was insufficient information on these products to develop standards of practice recommendations.

As the current assessment is limited to a review of the clinical effects of surgical procedures and non-surgical devices for the management of primary snoring, a brief outline of the available devices and surgical techniques used as treatment options is given in the next section.

Identification of studies

MEDLINE, EMBASE, CINAHL, the Cochrane Controlled Trials Register (CCTR) and NHS EED were searched for titles and abstracts of potentially relevant studies of clinical effects and treatment costs published between 1980 and 2007. No study design filters or language restrictions were applied to the search strategy.

The literature searches retrieved 1903 unique references after deduplication. All references were managed using Reference Manager software version 11. Full details of the search strategies are presented in Appendix 1.

Relevant studies were identified in two stages. Two reviewers screened titles and abstracts returned by the database searches independently. Full texts of any references that were considered relevant by either reviewer were obtained when available. The relevance of each paper was assessed according to the criteria set out in Chapter 2. Any discrepancies between the reviewers were resolved by recourse to the papers and, if necessary, a third reviewer was consulted. All duplicate papers were double-checked and excluded. The bibliographic details of excluded studies, along with the reason for their exclusion, are detailed in Appendix 2.

Review methods

Treatment costs

The costs of the surgical procedures and non-surgical devices included initial treatment costs as well as the ongoing costs of care associated with the interventions. The costs of diagnostic tests (i.e. PSG or modified PSG to exclude a diagnosis of sleep apnoea) were not included in the treatment costs. These will depend on the staff levels in the centre where the sleep study is undertaken, the number and complexity of the studies performed and the cost of the equipment and premises. However, it should be noted that use of a MAS may not always be associated with patients requiring a sleep study (i.e. PSG or modified PSG) before use of the device, which is necessary before either surgery or use of CPAP is initiated, and this is likely to impact considerably on the overall treatment costs.

No studies were identified by the searches that had assessed any of the costs associated with providing the interventions of interest. Therefore, to fulfil the scope of the review, all costs were estimated based on data from the NHS reference costs (2006),84 data from device manufacturers and clinical opinion. It was not possible to estimate a cost associated with RFA of either the soft palate or the tongue base as the costs of both the somnoplasty generator and electrodes were considered to be ‘commercial in confidence’ by the manufacturers (Gyrus ENT) and were not available in the public domain through either NHS purchasing suppliers or the NHS Supply Chain.

To provide a limited analysis of the rough cost estimates for the other surgical procedures it was assumed that each procedure would entail an initial consultation with an ENT surgeon, the specific inpatient or day-case surgical procedure time and device costs (where relevant), and a follow-up consultation. For UP3 it was assumed that the procedure would entail an inpatient stay of 1 day, as this is the most common length of stay, although patients who experience significant haemorrhage or pain may have a more protracted stay, increasing the costs associated with the procedure. For LAUP, as this is potentially a multistage procedure that can be conducted in one, two or three stages, it was assumed that each additional procedure was associated with one additional follow-up visit. For CPAP, as there are numerous machines and face masks currently available on the market, the mean costs of the machines and masks available were used to provide the acquisition cost and the cost of an annual replacement of the face mask. In accordance with advice from one of the manufacturers (ResMed) it was assumed that the CPAP machine has a device life of 7 years and that the face mask will be replaced on an annual basis. In terms of the use of MAS there are currently no published NHS costs of dental devices for the management of snoring and therefore the cost was estimated based on clinical opinion. It was assumed that the dentist provided a Thornton Adjustable Positioner^® (TAP), a device commonly used for the management of OSAS and snoring in the UK. Based on information provided in the Scottish Intercollegiate Guideline Network guideline on the management of obstructive sleep apnoea/hypopnoea syndrome in adults2 it was assumed that the device lasts for 2 years and that the patient would have one annual check-up. Under the new NHS dental contract a course of treatment is classified into a treatment band. Provision of a MAS was classified as band 3, as such treatment requires laboratory work (www.ic.nhs.uk), and 12 units of dental activity (UDA) were applied as appropriate to the band. 84 The national average reimbursement rate for a UDA is not known as this varies according to contract values negotiated locally by PCTs. Therefore, the value of a UDA was obtained from material published by Bath and North East Somerset Primary Care Trust (www.banes-pct.nhs.uk/documents/Board_Papers/2007/May/Agenda%20Item%2010%20Annex%201.pdf). This estimated the average reimbursement per UDA to be £20.91. Therefore, on the basis of provision of a MAS taking 12 UDA it was estimated that the cost of a MAS is approximately £250. The costs of both the CPAP machine and MAS were expressed as equivalent annual costs using the public sector discount rate of 3.5%.

Quantity and quality of research available

A total of 1903 titles and abstracts were screened for inclusion in the review of clinical effects and treatment costs. Of the titles and abstracts screened, 233 were ordered as full papers and assessed in detail. Six papers were unobtainable at the time of the assessment. The process of study selection is shown in Figure 2.

FIGURE 2.

Process of study selection.

A total of 27 studies reported in 30 publications met the inclusion criteria. The identified studies assessed a broad range of interventions but the evidence base was predominantly weighted (89%) towards studies that had evaluated surgical procedures,49,86–109 with only three studies on either MAS or CPAP. 110–112 Of the studies on surgical procedures, these could broadly be grouped into assessments of UP3 versus LAUP (n = 2),86–88 UP3 alone (n = 7),49,89–94 LAUP alone (n = 3),93,95,96 RFA of the soft palate or tongue base (n = 6 and n = 1 respectively)103–109 and palatal stiffening techniques (Pillar implants and injection snoreplasty) (n = 4 and n = 1 respectively). 97–102 No studies were identified that had assessed surgery to improve coincidental nasal obstruction alone, uvulectomy alone, DAUP or tongue-based suspension procedures.

In terms of studies assessing non-surgical devices, only three studies were identified that met the inclusion criteria. Two of these had examined the use of a MAS,110,111,113 while one had assessed the effects of CPAP versus a no-intervention control. 112

Although studies could broadly be grouped into intervention categories, for both UP3 and LAUP, closer inspection of the interventions evaluated within the studies designated to both of these categories revealed that there was often variability between the interventions assessed. This was partly because the procedures were sometimes performed in a modified manner91,94,95 or in conjunction with a second procedure. 92,114 More frequently, however, it was because of a lack of clear reporting on the procedures conducted, which did not allow comparisons to easily be drawn,89,92 or the methods of study design and analysis. 89,90,94 Three studies evaluating UP3 alone89,90,94 assessed two or more different procedures but pooled the post-treatment results across the intervention groups and reported the results as a one-group pre–post study. In all of these studies there were potential differences (the degree to which varied across the studies) between the other interventional procedures assessed [LAUP, cautery-assisted uvulopalatoplasty (CAUP) and uvulopalatal elevation palatoplasty (UEP)] and the broad category grouping to which the studies were assigned. The interventions within the UP3 alone category were therefore more heterogeneous and the results liable to be more confounded than is readily apparent from the somewhat simplistic nomenclature adopted for study grouping. For all of the other intervention groups, although there were differences between the study interventions within some categories (e.g. RFA versus bipolar RFA), these were more clearly reported, and the results were reported by treatment group or as between-treatment group comparisons as appropriate. Table 1 provides a summary of the identified studies by intervention group and specific intervention evaluated, and Table 2 displays a summary of the interventions identified by the number of studies and design.

The identified evidence base, as well as covering a relatively broad number of interventions, also differed in terms of study design, sample sizes and duration of follow-up. As could be expected given the interventions examined, the studies were predominantly surgical case series that reported both pre- and post-treatment data. Across the 27 studies, three RCTs were identified (11%), two CCTs (7.5%) and 22 (81.5%) pre–post studies. In total, 23 (85%) of the studies included only patients with a diagnosis of non-apnoeic snoring, whereas the remaining four included patients with OSAS as well as those with primary snoring but reported results separately for both patient groups. 92,96,109–111 Sample sizes across the studies were small, ranging from 9 to 231, with a median of 32. Similarly, the duration of follow-up was highly variable and ranged from 10 days to 5 years. Across the studies there was a tendency for objective outcome measures, such as PSG parameter results or objective measures of snoring sound intensity, to be assessed for a shorter follow-up duration and subjective measures to be assessed at longer follow-up times.

Where sufficient baseline prognostic factors were adequately reported, the patient population included in the studies was relatively homogeneous. All patients, in accordance with the inclusion criteria, had undergone either full PSG, modified PSG, oximetry or a sleep study (unspecified) to rule out a diagnosis of OSAS at baseline. The mean age of patients in the majority of studies was between 40 and 50 years, although across the studies the range was from 19 to 83 years. All of the studies included a higher percentage of male than female patients, with the percentage of male participants ranging from 56% to 100%. Furthermore, nearly all studies recruited patients with a body mass index (BMI) classified in the range of ‘overweight’ but not ‘obese’, although there was some variation in BMI across the studies, with a range from 18.3 kg/m² to 47.2 kg/m² observed.

The included studies had mainly been conducted in an otolaryngology department (or equivalent) setting in a number of different countries. Of the identified studies, five had been conducted in North America (Canada or the USA),92,96,102,104,112 seven in Scandinavia (Norway, Sweden or Finland),90,91,97,98,107,108,110,111 five in the UK,49,86–88,94,113 and four in Europe (Germany, Netherlands and a European multicentre study). 99,100,103,106,109

Of the remaining studies two were conducted in Thailand2,114 and one each in Turkey,89 Israel,95 Hong Kong101 and Australia. 105 The year of study publication ranged from 1994 to 2006.

Study quality

The included studies were predominantly small and of low quality. An indication of study quality according to quality indicators, grouped by study design, is given in Figures 3–5. A further assessment of study quality is provided in each specific section of the report according to intervention group.

FIGURE 3.

Summary of study quality indicators for the three randomised controlled trials (RCTs).

FIGURE 4.

Summary of study quality indicators for the two controlled clinical trials (CCTs).

FIGURE 5.

Summary of study quality indicators for the 22 pre–post studies.

Quantity and quality of the studies

Two parallel group studies comparing UP3 with LAUP reported in three publications86–88 were included. The study by Osman and colleagues86,87 was a RCT. Participants were studied at both short-term and long-term follow-up times with results from the two time points reported separately in two papers published in 200086 and 200387 respectively. The other study by Prasad and colleagues88 was a non-randomised CCT. Across the two studies there were a total of 107 non-apnoeic snoring patients, of whom 45 underwent UP3 and 62 underwent LAUP. Table 3 summarises the characteristics of the two studies.

The baseline test to rule out OSAS in the Osman study86,87 was an overnight sleep study using a DENSA Compact Sleep Apnoea Screening System; in the Prasad study88 it appeared to be oximetry. Participant baseline demographics in both of the studies were not reported separately for the UP3 and LAUP groups except for age in the Prasad study. The mean age of the participants ranged from 49.40 to 51.96 years. The proportion of men was 87% and 76% in the two studies respectively. In the Osman study nearly half of the patients were smokers (48.9%), whereas the Prasad study did not report smoking status. Neither of the studies reported baseline ESS scores or alcohol consumption levels.

In the LAUP procedure in the Osman study86,87 two linear scars were created in the oral surface of the soft palate from the junction of the hard and soft palate to the uvula, followed by laser uvulectomy, whereas in the Prasad study88 the procedure consisted of approximating the soft palate against the velopharyngeal isthmus and excising any redundant portion along the uvula. The number of LAUP procedures per patient was not reported in either study. In the UP3 procedure in the Osman study86,87 1 cm of the soft palate was removed with the uvula, as well as the tonsils if present; the posterior pillar was then sutured to the anterior pillar followed by mucosa-to-mucosa closure of the palate using sutures. In the Prasad study88 the UP3 procedure consisted of excising the uvula and a rim of soft palate and removing both palatine tonsils and suturing of the mucosa. Follow-up duration in the two studies ranged from 2 months to 56 months (or > 4 years).

Results

Neither of these two studies measured PSG parameter outcomes.

Only the study by Osman and colleagues86,87 reported objective outcomes. The results are presented in Tables 5 and 6. The number of patients with a change in the SI was analysed based on a subgroup of 38 patients at short-term follow-up (mean 3.4 months, range 2–11 months after the surgery),86 and the results are reported separately for the UP3 and LAUP groups. SI was reported in the long-term follow-up study87 based on a subgroup of 23 patients pooled from both UP3 and LAUP groups at short-term and long-term follow-up (median 45 months, range 29–56 months after the surgery) respectively. Snoring was defined as a spike in breathing sound intensity starting from baseline and returning back to baseline whose amplitude was greater than 50 dB SPL.

In the short-term follow-up study,86 after the surgical procedure there was no significant difference found between the two groups in terms of the proportion of patients who had either a higher postoperative SI or a lower postoperative SI. The authors stated that the reduction in SI in both the UP3 and LAUP groups was significant. Also, the pooled mean SI for both groups combined reduced from 154 snores/hour at baseline to 75.84 snores/hour postoperatively (p < 0.0001). In the long-term follow-up study,87 SI measured in 23 patients dropped from 135.6 snores/hour at baseline to 38.2 snores/hour at short-term follow-up and 21.7 snores/hour at long-term follow-up. The authors stated that from baseline to long-term follow-up the reduction in SI in both the individual UP3 group (measured in 11 patients) and the individual LAUP group (measured in 12 patients) was statistically significant, but no significant differences were found between the UP3 and LAUP groups.

The authors of this study also indicated that, at short-term follow-up, reduction in pre-and postoperative median snore loudness was significant at the 0.1% level in UP3 patients and at the 10% level but not the 5% level in LAUP patients; UP3 was better than LAUP at reducing the median snore loudness (p < 0.05). At long-term follow-up the reduction in median snore loudness in pooled data from 38 patients was significant at the 1% level; however, the mean reduction was modest, being 191 ± 108 mV (or 3.8 dB, SD 2.2 dB) for 95% confidence limits. At long-term follow-up there was no significant difference between the UP3 and LAUP groups in the reduction in snoring loudness.

In the Osman study86,87 the number of patients reporting an improvement in their snoring at short-term follow-up was the only subjective result reported. It was assessed by a questionnaire with the following response options: absent, a lot better, slightly better, unchanged or worse. At short-term follow-up after the surgery 89% of the patients in the UP3 group and 83% in the LAUP group reported an improvement; at long-term follow-up the improvement declined to 83% and 76% respectively, as two patients in the UP3 group and one in the LAUP group were rereferred by their GP for late recurrence of snoring. The significance of the improvement and the differences between the two groups were not reported.

Adverse events were only reported by Osman and colleagues. 86,87 The results are shown in Table 7. Postoperative pain scores were assessed using a linear analogue scale with 0 indicating no pain and 4 indicating worst pain. In the first 24 hours following the LAUP treatment patients had significantly less pain (mean pain score 1.3) than following UP3 (mean pain score 1.7). One patient who underwent UP3 had severe postoperative infection. One patient in the UP3 group and two in the LAUP group at long-term follow-up needed an unscheduled repeat procedure. All other adverse events were minor and the differences between the two treatment groups were not statistically tested.

In the Prasad study88 subjective snoring results were all partner assessed (see Table 8). None of these subjectively assessed outcomes was measured in the same way as in the Osman study.

Improvement in snoring was measured as the percentage of patients with snoring ‘abolished’, ‘a lot better’, ‘slightly better’, ‘unchanged’ or ‘worse’. The authors stated that residual snoring was significantly more frequent following LAUP than following UP3, with snoring being abolished in 19% and 15% of the UP3 group and LAUP group respectively. Partners considered snoring as being a lot better in 58% of the UP3 group and 28% of the LAUP group, as being slightly better in 11% of the UP3 group and 24% of the LAUP group, and as unchanged in 4% after UP3 and 21% after LAUP.

Difficulties falling asleep because of partner’s snoring was assessed with the following response options: ‘no’, ‘occasionally’, ‘sometimes’, ‘most often’ and ‘always’. The authors stated that after UP3 partners were significantly less likely to have trouble falling asleep, with 33% of the partners in the UP3 group and 24% in the LAUP group never having difficulties falling asleep. Occasionally having difficulties falling asleep was reported by 40% of the partners in the UP3 group and 22% in the LAUP group, ‘always’ was reported by 4% of the UP3 group and 16% of the LAUP group, and ‘most often’ was reported by none of the UP3 group and 19% of the LAUP group.

Sleep disturbance because of partner’s snoring was assessed with response options of ‘no longer disturbed’, ‘reduced frequency’, ‘remains unchanged’ or ‘increased’. Disturbance of sleep and the need to wake up were reduced very significantly after both UP3 and LAUP, with 59.3% of partners in the UP3 group and 39.4% of partners in the LAUP reporting no disturbed sleep, and 37.0% in the UP3 group and 33.4% in the LAUP group reporting reduced sleep disturbance. The differences between the two groups were not statistically significant; however, there was a trend for LAUP to be less efficacious than UP3 (p = 0.064).

The need to sleep in a separate room because of partner’s snoring was assessed with response options of ‘no longer need to sleep in a separate room’, ‘frequency reduced’, ‘remains unchanged’ or ‘increased’. After surgery, 81.5% of the partners in the UP3 group and 65.6% of the partners in the LAUP group no longer needed to sleep in a separate room; 11.1% of the partners in the UP3 group and 15.6% of the partners in the LAUP group had the need reduced. The statistical significance for either the difference between the UP3 and LAUP groups or the change pre- and postoperatively was not reported.

Quality of marital life was also assessed but the response options were not reported. In total, 40.7% of the patients in the UP3 group and 30.3% in the LAUP group recorded an improvement in quality of marital life, whereas in 33.3% of the UP3 group and 57.6% of the LAUP group quality of marital life remained unchanged; 26.0% of the UP3 group and 12.1% of the LAUP group were undecided. Both groups also showed an improvement in quality of life but the results were not reported. The authors indicated that significantly more partners in the UP3 group reported an improvement in both quality of marital life and quality of life.

Conclusions

Based on very limited evidence from two small poor-quality studies UP3 appears to be more efficacious than LAUP in improving some subjectively measured snoring outcomes and snoring-related problems. However, this effect was not consistently observed across all subjective outcome measures. Both UP3 and LAUP seemed to reduce the SI and snoring loudness; however, the effects did not differ between the two operations.

Quantity and quality of the studies

Seven pre–post studies49,89–94 that evaluated the effects of UP3 alone for non-apnoeic snoring met our inclusion criteria. One91 included both non-apnoeic snorers and patients with a diagnosis of OSAS but some outcomes were reported separately for non-apnoeic snorers. Study characteristics are summarised in Table 10.

Five of the studies assessed UP3 alone. 49,89,91–93 In the other two studies, Jones and colleagues94 randomised patients into two parallel groups, one which underwent traditional palatoplasty with uvulectomy (TP) and the other which underwent UEP, and Janson and colleagues90 randomised patients to either UP3 or conservative treatment. However, in both of these studies pre- and post-treatment data were pooled across the two intervention groups and reported as the difference in pre- and post-treatment means. No between-group comparisons were conducted. Both studies were therefore classified as pre–post studies. The parallel group study by Janson and colleagues90 included a total of 231 patients, of whom 155 underwent UP3. The study by Ozmen and colleagues89 included 162 patients. All of the remaining studies were quite small, having less than 60 patients each. Across the studies a total of 538 patients underwent UP3. Patient baseline characteristics in the seven studies were not reported in detail. At baseline in five of the studies49,89,92–94 patients had a PSG and in the remaining two studies90,91 patients underwent oximetry to rule out OSAS. In the four studies that reported baseline examinations to assess the site of airway narrowing the tests included clinical examination alone,90 examination by an otolaryngologist including examination of the nasopharynx, oropharynx and laryngopharynx,92 thorough physical and otolaryngological examinations with the use of flexible nasopharyngoscopy with Muller’s manoeuvre and a lateral cephalogram,93 and otolaryngological examinations with additional paranasal sinus CT if needed. 89 The other three studies did not report baseline tests of upper airway narrowing. In the study by Miljeteig and colleagues92 that contained non-apnoeic snorers and OSAS patients, the baseline characteristics in terms of age, BMI, ESS, gender composition and smoking status were not reported by subgroup. In the remaining six studies that assessed non-apnoeic snorers only, of the patients who underwent UP3 treatment, mean age at baseline ranged from 35.2 to 48.3 years, except in the study by Friberg and colleagues91 in which it was reported as a median of 26.1 years (range 18.3–35.4 years). The majority of patients in these six studies were male, with the percentage ranging from 56% to 100%. BMI in these six studies ranged from 25.2 kg/m² to 35.4 kg/m². Only two studies reported baseline ESS scores,94,123 and smoking status was reported in only one study. Follow-up duration of the participants in the studies varied markedly from 1 to 83 months. The mean follow-up duration in the study by Miljeteig and colleagues92 was 45 months (SD 20 months), but only sleep data were recorded separately for the non-apnoeic snorers, these outcomes being measured at 13 months after surgery.

Results

Friberg and colleagues91 undertook UP3 on 56 non-apnoeic snorers and reported both subjective results and objective sleep data measured at a median time of 63 months’ (range 37–83 months) follow-up. Postoperative sleep recording was performed in 53 patients. A significant worsening from baseline to follow-up was observed for the median oxygen desaturation index [ODI; increased from 0 (range 0–24) to 1 (range 0–24)]; the median duration of obstructive respiratory pattern [increased from 8% (range 0–43%) to 17% (range 0–74%)]; and the median nadir SaO₂ level [decreased from 92% (97–84%) to 88% (95–73%). Consistent with these changes in oxygen saturation levels and obstructive respiratory pattern, the authors indicated that, according to the postoperative recording results, six patients developed mild to moderate OSAS after the surgery.

In the study median BMI also increased significantly after surgery (p = 0.0005). Changes in both ODI and obstructive respiratory pattern were significantly correlated with changes in BMI (p = 0.0003). Five of the six patients in whom OSAS developed postoperatively had gained weight.

Of 51 patients who preoperatively reported EDS, the number (%) of patients reporting no EDS, improvement in daytime sleepiness or no change in EDS after the operation was 29 (57%), 8 (16%) and 14 (27%) respectively. All of the 56 patients in the study preoperatively complained of having loud snoring. At follow-up 15 (27%) reported no snoring at all, 34 (60%) reported less snoring and seven (13%) reported persistent snoring. Overall, 87% of the patients with snoring were improved or cured. In none of the patients did snoring become worse. Of those who still snored, the majority reported that the snoring was much less loud but more high pitched. Changes in BMI correlated significantly with the degree of postoperative snoring (p = 0.04) but not with the degree of EDS (p = 0.13) and snoring.

No serious postoperative complications were observed. Of the 56 patients, seven reported minor throat disturbances such as dryness or increased mucous secretion, and one reported a change in voice quality.

Janson and colleagues90 assessed 231 patients: 155 underwent one of three UP3 procedures and 76 underwent conservative treatment. Subjective outcomes were reported in the 155 patients who underwent a UP3 procedure. All of the 155 patients classified their symptoms on a 5-point scale, multiple choice questionnaire. The questionnaire consisted of 16 symptom-related questions. Postoperative outcomes were assessed in 105 patients at 3 months and 50 patients at 12 months. Only a mixture of results for the two time points was reported. Subjective outcomes measured were the percentage of patients who reported either moderate or occasional problems, or severe or frequent problems, with snoring, involuntarily falling asleep, falling asleep when relaxing, difficulty staying awake when driving, daytime tendency to take a nap, difficulty maintaining sleep or inducing sleep, awakening because of trouble breathing, daytime tiredness and morning headache. All of the measured outcomes showed a statistically significant improvement after UP3. A summary of the percentage of patients reporting each problem pre-and postoperatively is displayed in Table 12 below.

Limited results at 1 year of follow-up after UP3 versus conservative treatment were reported by the authors in graphs and minimally in the text. It was unclear whether the p-values stated in the text were for the moderate/occasional problems, severe or frequent problems, or the total of both categories. The specific figures from the graphs were therefore not extracted. The authors indicated that after treatment the UP3 patients had significantly lower symptom scores than the conservatively treated patients for tendency to take a nap in the daytime (p < 0.05), EDS, morning headache (p < 0.01), difficulty in maintaining sleep, snoring, and waking because of trouble breathing (p < 0.001); significantly more patients in the UP3 group also felt rested on awakening in the morning (p < 0.01).

The side effects of UP3 observed in this study were minor, with a total of 42 patients experiencing events. In total, 35% of the 42 patients had undergone the classical (standard) UP3 treatment and 23% had undergone either the CO₂ laser treatment or UP3 without tonsillectomy. The most frequent events were irritation in the throat (19 patients) followed by hypersecretion in the throat (four patients), nasal speech (three patients), nasal regurgitation (three patients) and tendency to vomit (one patient).

In the study by Jones and colleagues94 54 patients were randomised into two parallel groups to undergo either TP or UEP. Pre- and post-treatment data on 35 patients (16 in the TP group and 19 in the UEP group) were pooled across the two groups and reported as the difference between pre- and postoperative scores. At the first follow-up visit (mean 2.5 months) there was a significant decrease from baseline in the median (90th centile) measures of snore periodicity, from 87.5% (80–94%) to 84.0% (73–88%); however, this was not maintained at the second follow-up visit (mean 9.7 months), with a slight increase to 84.5% (73.5–91.5%). Acoustic parameters of the median (90th centile) snoring sound energy ratios for the frequency bands 0–200 Hz, 0–250 Hz and 0–400 Hz all demonstrated significant improvement from baseline at the first follow-up visit, being 74 (37–87), 83 (42–93), and 95 (56–97) at baseline, respectively, but reduced to 58.5 (18–83), 73 (18–86.5) and 85 (49.5–95.5) at the first visit respectively. Again, however, there was no significant difference from baseline values at follow-up visit two, with increases to 67 (14–89) in the 0- to 200-Hz band, 74 (27–93) in the 0- to 250-Hz band and 89 (50–97.5) in the 0- to 400-Hz band; of these, the snoring energy ratio of the 0- to 250-Hz band reached marginal significance. The authors indicated that a paired comparison of snore loudness between the early and late postoperative recordings showed a significant increase in snore loudness between recordings (p < 0.001).

The authors also reported that age, peak nasal inspiratory flow rate (measured as the best of three maximum rates), ESS and reported alcohol intake (units/week) were unrelated to either the early improvement or subsequent decline in the acoustic parameters after surgery.

At a median time of 21 months (range 11–26) after the surgery, based on 34 patients, 11.8% reported no improvement in snoring symptoms, 55.9% reported an improvement that was sustained for 1 year or less, and 32.3% reported an improvement that was sustained until the assessment time. The authors indicated that the objective acoustic outcomes were not correlated with the subjective outcomes on an individual patient level. Adverse events were not reported in this study.

In the study by Miljeteig and colleagues,92 which included both non-apnoeic snorers and patients with OSAS, only the objective outcomes of SI and sound intensity for 38 non-apnoeic snorers measured at a mean of 13 (SD 15) months’ follow-up were reported and compared with baseline values. The mean SI increased from 252 (SD 280) events/hour before surgery to 296 (SD 318) events/hour after surgery, but the change was not statistically significant. Changes in mean sound intensity from 64.8 dB (SD 7.2) at baseline to 65.4 dB (SD 7.0) post-treatment, and mean maximal sound intensity from 85.3 dB (SD 12.3) before surgery to 80.5 dB (SD 14.5) after surgery, were also not statistically significantly different. Adverse events were not reported in this study.

The study by Neruntarat93 assessed the effects of a reversible uvulopalatal flap. Data from 56 patients on subjectively assessed outcomes, cephalogram and PSG measured both at baseline and at a mean of 14.2 months (range 12–20) after the operation were analysed. Snoring sound was measured with a microphone placed at the trachea. Severity of snoring was graded by the patient’s bed partner or observer using a 10-cm VAS, with ‘no snoring’ and ‘severe snoring’ used as anchors. A 10-point VAS for postoperative pain, difficulty in swallowing and changes in speech was completed by the patients daily for 10 days postoperatively. Comparing the pre- and postoperative values, there were changes in the following variables: mean AHI score decreased from 3.2 (SD 1.2) to 3.0 (SD 1.8); sleep efficiency decreased from 88.6% (SD 8.6%) to 87.5% (SD 10.4%); mean lowest oxygen saturation increased from 86.4% (SD 9.6%) to 88.2% (SD 5.5%); and mean oxygen saturation increased from 95.2% (SD 1.8%) to 96.2% (SD 2.1%); however, none of these changes was significant. Significant changes were observed in the following variables: mean SI, from 245.8 (SD 40.8) events/hour to 42.5 (SD 20.7) events/hour; mean time spent in loud snoring, from 10.2% (SD 1.5%) to 3.8% (SD 2.8%); partner-assessed mean severity of snoring 10-point VAS score, from 8.2 (SD 3.4) to 2.6 (SD 1.4); and mean ESS, from 8.1 (SD 3.5) to 5.2 (SD 3.2). Snoring was considered to be cured if the bed partner-assessed VAS score was less than half of that at baseline. Based on this criterion the problem was eliminated in 88% (49/56) of patients. A positive and significant correlation was found between the changes in percentage of time spent in loud snoring and the changes in subjective severity of snoring VAS score (r = 0.42, p < 0.05), and between the changes in SI and the changes in subjective severity of snoring VAS score (r = 0.38, p < 0.05). Cephalometric data showed a significant decrease in two parameters: the length from posterior nasal spine to uvula (PNS-P) and the palatal width (PW). No significant changes were observed in the sella-nasion-point A angle (SNA), the sella-nasion-point B angle (SNB), the posterior airway space (PAS) or the length between mandibular plane and hyoid bone (MP-H). A summary of the pre- and post-treatment cephalometric data is displayed in Table 13.

There was minimal morbidity associated with treatment. After the procedure the mean VAS scores for pain, swallowing problems and speech problems decreased from 6.9, 4.4 and 2.2, respectively, on day 1 to 1.1, 0.5 and 0.2, respectively, on day 10. Most patients (52/56) had mild to moderate pain (VAS score ≤ 7) for 5–7 days after the procedure. Two (4%) patients reported transient nasal regurgitation and one (2%) patient reported foreign body sensation. Bleeding, dysphagia, infection and nasopharyngeal stenosis were not observed.

In their study Ozmen and colleagues89 aimed to assess the association between snoring and excess risk for cardiac events by specifically assessing the effects of UP3 on QT dispersion. A total of 162 patients underwent one of three surgical procedures: general UP3, LAUP or CAUP. The number of patients in each of the procedure groups was not reported.

Surgical procedure success was assessed by the patient’s partner using a six-level snoring assessment table (from 0 indicating no snoring to 5 indicating snoring with hypopnoea or apnoea). At follow-up nine (6%) patients had a score of 2 and were considered to have mild or moderate snoring. They were therefore judged to have no improvement. Surgery was therefore classified as successful in 153 patients (94%). ECG data of corrected QT interval parameters QTc min, QTc max and QTc dispersion were analysed based on 153 patients who postoperatively had a 6-level snoring assessment score of < 2. Significant changes were found in all of the three variables. A summary of the mean pre- and post-treatment scores is displayed in Table 14.

In this study no complications were observed.

Prichard and colleagues49 reported changes in objective measurements during sleep from baseline to 6 months post UP3 for 32 non-apnoeic snorers. No significant change in AHI was found after UP3. Snoring sound level was measured. The objective indices of snoring used were the proportion of sleep time during which the level exceeded 50 dB (P50) and the sound level (dB) above which 1% and 5% of the sound occurred (L1 and L5). A significant reduction in the snoring sound level was observed in the supine posture for all of the three indices, with the percentage of sleep time that exceeded 50 dB (median value) being 1.12 at baseline and 0.36 post treatment, the level above which 1% of the sound occurred (median value) being 50.6 dB preoperatively and 46.3 dB postoperatively, and the level above which 5% of the sound occurred (median value) being 43.6 dB preoperatively and 38.8 dB postoperatively.

Subjective snoring grade was assessed by the patient’s partner using the responses ‘nil’, ‘mild’, ‘moderate’ or ‘severe’, scoring 0, 1, 2 or 3 respectively. Decreases in snoring grades were observed. At baseline the percentage of partners reporting snoring being moderate and severe was 25% and 75% respectively; after surgery these values reduced to 18.8% and 12.5% respectively. Also, after surgery 15.6% reported no snoring and 53.1% reported mild snoring. The significance of the differences was not assessed statistically. No adverse events were reported.

A summary of the study results for PSG parameter outcomes, objectively assessed snoring sound outcomes, subjectively assessed snoring levels and adverse events are displayed in Tables 15–18 respectively. A further overall summary of the key study results is displayed in Table 19.

Conclusions

Limited evidence from poor-quality studies showed that, for the treatment of non-apnoeic snorers, UP3 seemed to improve subjectively assessed snoring and reduce snoring-related problems, EDS and sleeping difficulties. On objectively assessed outcomes UP3 seemed to reduce snoring volume in the supine sleeping position and the percentage of sleeping time spent in loud snoring, and reduce snoring sound energy ratios and snore periodicity at short-term follow-up. UP3 also seemed to reduce the palatal width and the distance between the posterior nasal spine and soft palate. Evidence on SI and oxygen saturation levels is equivocal.

Based on the limited results, adverse events associated with UP3 appeared to be minor, with postoperative throat disturbances and pain being the most common.

Quantify and quality of the studies

Three pre–post studies were identified that evaluated the effects of LAUP alone for the treatment of non-apnoeic snoring. Of the three studies, one by Lauretano and colleagues96 completed LAUP on a sample of 52 patients, but the study only included 12 non-apnoeic snorers. The other two studies, by Berger and colleagues95 and Neuruntarat,114 only included 14 and 32 patients respectively. The total number of non-apnoeic snorers in the three studies was therefore 58. In the study by Lauretano and colleagues96 the objective outcomes analysed for non-apnoeic snorers were based on only three patients who underwent a postoperative PSG test. In this study baseline data for non-apnoeic snorers were based on a total of 32 patients who were recruited, not just the 12 non-apnoeic snorers who completed treatment. The mean age of patients in the three studies ranged from 35.2 to 51.2 years. The proportion of male patients ranged from 71% to 91%. Mean follow-up duration ranged from 2 months to 10.1 months. Patients in all three studies had a PSG test at baseline and mean AHI was reported. Mean BMI in these studies ranged from 26.2 kg/m to 28.08 kg/m². None of the studies reported baseline ESS scores, smoking status or the level of alcohol consumption of patients. A summary of the study characteristics are displayed in Table 20.

Results

Berger and colleagues95 followed up patients at 4 weeks postoperatively and at a mean of 10.1 months (SD 7.9) after completion of LAUP treatment. At the short-term follow-up the snoring status in 79% of the patients was improved, in 14% it remained unchanged and in 7% it became worse. At the long-term follow-up the improvement declined, with 57% reporting improvement, 21% remaining unchanged and 21% becoming worse.

Assessment of sleep-related symptoms at long-term follow-up showed that 57% of the patients reported an improvement, whereas 36% reported no change and 7% felt worse. The changes were not tested for statistical significance. LAUP appeared to have no adverse effects on either AHI scores or the lowest SaO₂ level, with no significant difference in changes in either parameter observed between baseline and follow-up. The specific point estimates for both outcomes and the associated variance are reported in Table 22.

Postoperative pain was reported as lasting for a mean of 9.7 days (SD 3.5 days), with severe pain that precluded patients working lasting for 4.5 days (SD 3.1 days). Additionally, 36% of patients experienced persistent throat dryness or itching, and 17% had difficulty with nasal breathing.

In the study by Lauretano and colleagues,96 at 2-months’ follow-up the snoring status of the patients was improved in 75%, unchanged in 17% and worse in 8%. Patient satisfaction with LAUP was assessed on a 7-point scale, with 1 indicating very dissatisfied, 4 neutral and 7 very satisfied. Assessment indicated that 83% of the patients were satisfied with the results of the procedure and 17% were unsatisfied. On the basis of results from three patients, LAUP did not significantly affect either the mean or the lowest oxygen desaturation levels, with no significant differences in levels between baseline and follow-up. Adverse events were not reported in this study.

In the study by Neruntarat,114 at a mean follow-up duration of 5.2 months (range 4–6) partner-assessed snoring scores showed a significant decrease from 8.8 (SD 2.2) to 2.8 (SD 1.7) after surgery, and 91% of patients considered that their snoring was cured (i.e. the VAS score was less than half of that at baseline). Nasal obstruction as measured by a VAS decreased significantly from a mean of 6.2 (SD 3.2) Pa/ml/cm² at baseline to 1.6 (SD 0.8) Pa/ml/cm² after the surgery. The nasal symptoms of rhinorrhea, nasal pruritus and sneezing also showed similar decreases postoperatively, with changes from a mean of 5.8 (SD 1.4) to 2.1 (SD 1.8), 6.1 (SD 2.2) to 2.3 (SD 0.9) and 7.1 (SD 2.3) to 2.5 (SD 1.7) respectively. However, the significance of the changes was not tested. Postoperative pain assessed using a 10-point VAS was also mild to moderate, with a score of ≤ 4.

A summary of the PSG parameter and objective outcome results is displayed in Table 22. Additionally, Table 23 gives a summary of the subjective outcome results and Table 24 presents the adverse events associated with LAUP.

Conclusions

Based on poor quality pre–post studies with very small sample sizes, LAUP significantly reduced nasal resistance and improved subjective snoring scores. LAUP when used in conjunction with septoplasty and laser turbinectomy tended to improve nasal obstruction related symptoms. No major complications were reported with LAUP, except postoperative pain.

Quantity and quality of the studies

A total of seven studies assessed the effects of RFA techniques. 103–109 In six of the studies surgery was to the soft palate103–108 and in the remaining study it was to the tongue base. 109 The available evidence consisted of one RCT by Stuck and colleagues,103 which compared soft palate RFA with placebo; one CCT conducted by Cartwright and colleagues,104 in which soft palate RFA was compared with the short-term use of a MAS; and five pre–post studies, four of which assessed RFA to the soft palate105–108 and one that evaluated bipolar RFA of the tongue base. 109

Study sample sizes were small ranging from 9 to 45 participants, with a median of 20. Overall there was a total of 151 participants across the studies. Duration of follow-up ranged from 10 weeks to 1 year.

In six of the studies patients had undergone a PSG at baseline to rule out OSAS103,105,106,108,109,124 and in the remaining study patients had undergone a sleep study (type unspecified). 107 Investigations at baseline to assess the site of airway narrowing included clinical examination alone;104 clinical examination in conjunction with endoscopy plus or minus cephalometric radiography;103,107,108 lateral cephalometry alone;105 and endoscopy alone. 109 In one study baseline tests were not reported. 106 The age range of participants included in the studies was from 26 to 67 years. When reported, the percentage of male participants ranged from 66% to 95%. 105–109 In the five studies reporting baseline BMI the mean BMI was 25.7–27.1 kg/m², with a range from 22.0 to 34.6 kg/m², indicating that patients were mainly overweight but not obese. Baseline ESS scores ranged from 4.5 to 11.4 across the studies. None of the studies reported the number of participants who smoked or the levels of alcohol consumption. Study characteristics are summarised in Table 26.

The studies differed slightly in their aims and objectives. Stuck and colleagues103 aimed to assess the efficacy of soft palate RFA compared with placebo sham surgery in their small RCT. In the CCT conducted by Cartwright and colleagues104 the objective was to evaluate the effectiveness of soft palate RFA and compare this with the effects of using a MAS (Snore X oral appliance; Snorex, Fremont, CA) for a 2-week treatment period. Consequently, four out of the five outcomes assessed were reported as pre-and postintervention scores for the RFA group alone, with only one outcome reported for both groups with a between-group comparison being made. The pre–post studies conducted by Hukins and colleagues,105 Boudewyns and van de Heyning106 and Back and colleagues108 all aimed to assess the clinical effectiveness and adverse events associated with RFA palatoplasty for the treatment of non-apnoeic snoring. The aim of the study conducted by Haraldsson and colleagues107 was to evaluate whether RFA tissue reduction is associated with voice changes as a result of velopharyngeal insufficiency in patients with heavy snoring, whereas den Herder and colleagues109 assessed the adverse events and complications associated with bipolar RFA of the tongue base.

Results

Conclusions

From the small number of studies available the evidence suggests that at 1 year of follow-up after RFA of the soft palate there is a significant decrease in partner-reported levels of snoring severity. However, any treatment benefit with RFA may only be moderate as an unweighted mean decrease of only 3.24 points was observed for partner-reported snoring levels. Additionally, relapse rates may be quite high at longer follow-up times. There is no strong evidence to suggest that subjectively assessed partner-reported snoring levels are correlated with objective measures of snoring sound intensity. In all studies treatment with RFA had a positive effect on levels of ESS but in only two studies did these reach statistical significance. However, this is likely to be a reflection of the fact that patients’ baseline levels of ESS were within the range considered normal, and therefore a floor effect would be in operation. The majority of the available evidence suggests that RFA has no detrimental or positive effects upon any PSG parameters, such as AHI or sleep oxygen saturation levels. Additionally, there was no evidence either from cephalometric or from MRI-assessed measurements of the palate that RFA has any overall significant effect on palatal width or length. No serious adverse events were reported and overall complication rates across the studies were low.

Quantity and quality of the studies

Five studies reported in six publications assessed the effects of palatal stiffening techniques for palatal flutter snoring. 97–102 Four studies assessed palatal implants and one assessed injection snoreplasty. The available evidence base consisted of one small RCT that evaluated the effects of increased rigidity in palatal implants97 and four pre–post studies, three of which assessed palatal implants98–101 and one that assessed injection snoreplasty. 102

Overall across the studies a total of 134 participants were included. Study sample size was small ranging from 12 to 40 with a median of 27. In the four studies that assessed palatal stiffening techniques97–101 the duration of follow-up ranged from 3 to 12 months. In the study by Brietzke and Mair102 that assessed injection snoreplasty follow-up was at 10 days post each treatment.

In all five studies patients had undergone either PSG or oximetry at baseline to rule out OSAS. 97–102 Additionally, clinical examination to assess the site of airway narrowing had been performed in all five studies. 97–102 Other tests used in conjunction with examination were awake endoscopy,99,100,102 fibreoptic nasopharyngolaryngoscopy,97 fibreoptic laryngoscopy in conjunction with indirect layngoscopy98 and measurement of the soft palate size and uvula width. 101 The mean age of the participants across the studies was 38–41.1 years, with a range of 24–72 years. In the four studies reporting the gender composition of their study sample the percentage of male participants ranged from 60% to 100%. 97,98,101,102 Mean baseline BMI in four studies ranged from 25.1 kg/m² to 26.1 kg/m² with a range of 20.2–29.7 kg/m², indicating that the majority of patients were overweight but not obese. 97–101 Mean ESS scores were within general population norms ranging from 6.1 to 9.3 in the three studies reporting baseline sleepiness levels. 98–101 None of the studies reported the number of smokers or levels of alcohol consumption. A summary of the study characteristics of the five studies on palatal stiffening techniques is presented in Table 36.

Across the four studies assessing the effects of palatal implants the aims, objectives and treatment protocols were similar. 97–101 In the small RCT by Skjostad and colleagues97 the objective was to assess the safety and efficacy of the use of more rigid implants (rigidity 1.8) compared with standard implants (rigidity 1.0). The two effectiveness outcomes assessed were analysed as the difference between pre- and postoperative means for the two treatment groups separately, with comparisons between groups only being conducted for the number of extrusions. The aim of all three of the pre–post studies conducted by Nordgard and colleagues,98 Maurer and colleagues99 and Ho and colleagues101 was to assess the safety and effectiveness of Pillar implants for the treatment of primary snoring. The treatment protocol in all three studies consisted of three implants, with the first inserted in the midline at the trailing edge of the hard palate and the other two inserted approximately 5 mm laterally to the midline on each side. In all studies treatment was undertaken as one procedure under local anaesthesia. When reported (in two studies97,98) the mean operative time was approximately 8 minutes. Only one of the studies99,100 reported the need for a second procedure because of implant extrusion.

Brietzke and colleagues102 aimed to assess the effects of injection snoreplasty in their pre–post study. The treatment protocol consisted of a first injection to the soft palate midline of 2.0 ml of 1% sodium tetradecyl sulphate (10 mg/ml). For repeat procedures 3% Sotradecol™ (Bioniche Pharma; 30 mg/ml) was generally used and the site of injection modified to include the lateral areas of the soft palate as well as those not adequately stiffened by the previous procedure. Treatment sessions were performed 6–8 weeks apart, with a mean of 1.8 sessions per patient (range 1–4). Patients were followed up for 10 days post each procedure.

Results

Conclusions

From the limited evidence available it appears that, in the year after treatment with Pillar implants, there is a significant decrease in partner- and patient-reported levels of snoring severity. However, this treatment benefit appears to be relatively modest, with an unweighted mean decrease of 2.68 points on a 10-point VAS scale. None of the studies attained a mean post-treatment score of less than 4 and, in the two studies in which it was reported, only 23.5% and 22% of patients achieved a VAS score of < 3. There is no evidence to suggest that partner-reported levels of snoring severity are corroborated by objective measures of snoring sound intensity, as no significant differences in any measure of objective snoring sound intensity were reported. In all studies Pillar implants had a significant positive benefit on ESS scores, but these were within the range considered normal at baseline. The limited available evidence also suggests that the use of Pillar implants does not have any significant positive or negative effects on any PSG-measured indices and, where any changes were reported, these may well be related to the ‘noise’ associated with the measurement of these parameters. Postoperative pain and swallowing and speech difficulties were all rated as mild or mild to moderate, and the use of the technology does not appear to be associated with any moderate or serious adverse events.

Given the limited very poor-quality evidence available on the effects of injection snoreplasty it is not possible to draw any conclusions regarding the safety and efficacy of this treatment without consideration of substantive further evidence on the use of this technology.

Quantity and quality of the studies

Only one study, conducted by Series and Marc,112 assessed the effects of CPAP in the treatment of non-apnoeic snoring. The aim of the study was to evaluate the effect of discontinuing nasal continuous positive airway pressure (NCPAP) treatment on the snoring characteristics of patients. The study included 18 non-apnoeic snorers who were allocated to either NCPAP or a no-treatment control group. The control group were studied twice, at baseline and at 30 days’ follow-up. The NCPAP group were studied at baseline and then underwent a second sleep study to determine the level of NCPAP required to abolish snoring. Patients remained on treatment at a level that banished their snoring for a 30-day period. A third sleep study was obtained on the first night without CPAP after the 30 days of treatment. Six patients from the intervention group were additionally followed up in a fourth sleep study 8–10 days after the end of the treatment period.

Although there was a no-treatment control group in the study, no between-group comparisons were undertaken and therefore the study was classified as a two-group parallel pre–post study. The total length of follow-up was 38–40 days.

All patients had undergone an overnight PSG with snoring measurement at baseline to rule out a possible diagnosis of OSAS. The patient group included in the study ranged in age from 25 to 64 years and was predominantly male (88.9%). The mean BMI was 29.1 kg/m² and 22.2% were smokers. None of the patients had undergone previous treatment for snoring or were taking medication at the time of the study. Study characteristics are summarised in Table 44.

Results

In the NCPAP group the baseline snoring index score of 387 per hour (SD 150) was effectively abolished to a snoring index score of 15.1 per hour (SD 2.5), with a mean NCPAP level of 5.3 cmH₂O (SD 0.5) (range 3–8 cmH₂O). At post-treatment follow-up visit 1 (the first night without NCPAP) there was no significant difference between the baseline snoring index value [387 per hour (SD 150)] and the post-treatment value [320 per hour (SD 171)]. At post-treatment follow-up visit 2, no significant differences were observed in snoring index values from baseline, with snoring levels having returned to their pretreatment levels [387 per hour (SD 150) versus 374 per hour (SD 181.3)]. The percentage of TST spent with snoring levels > 60 dB SPL had significantly decreased from baseline at follow-up visit 1 [10.3% (SD 1.8) versus 7.4% (SD 1.5)]. This improvement was observed for sound pressure levels up to 69 dB SPL. However, by follow-up visit 2, post-treatment values were no longer significantly different from baseline values [10.3% (SD 1.8) versus 9.8% (SD 2.1)]. Likewise, at follow-up visit 1 the mean snoring level had significantly decreased from the baseline value of 66.4 dB SPL (SD 0.9) to 65.0 dB SPL (SD 0.9), but at follow-up visit 2 it was no longer significantly different from the baseline value, with a mean of 65.7 dB SPL (SD 0.98).

Both AHI and AI scores decreased significantly whilst patients were on therapy, from baseline values of 2.9 (SD 3.9) and 9.7 (SD 3.9), respectively, to 0.2 (SD 0.6) and 3.7 (SD 2.4) respectively. AHI scores, although increasing slightly from levels attained whilst on therapy, remained significantly lower than the baseline value at both follow-up visits 1 and 2, with values of 1.0 (SD1.5) and 1.1 (SD 1.5) respectively. AI values, in accordance with AHI scores, also increased from the levels attained whilst on therapy at both follow-up visits 1 and 2, with values of 8.2 (SD 4.2) and 6.9 (SD 4.8), respectively, no longer being significantly different from the baseline value. There were no significant differences in TST or sleep architecture parameters in the NCPAP treatment group compared with baseline values, across the treatment period and at follow-up visits 1 and 2. The mean values for TST and all sleep architecture parameters at baseline, on treatment and at follow-up visits 1 and 2 are presented in Table 46.

Patients in the no-treatment control group showed similar levels of baseline snoring characteristics to those in the NCPAP group. In the control group the snoring index remained unchanged from baseline at follow-up, with values of 380 per hour (SD 108) and 388 per hour (SD 201) respectively. The percentage of TST > 60 dB SPL was similar at baseline and at follow-up: 11.1% (SD 6.0) versus 10.5% (SD 2.3). Likewise, the mean snoring level did not change significantly, from 65.6 dB (SD 0.9) at baseline to 65.1 dB (SD 1.5) at follow-up. As expected there were no significant differences between baseline and follow-up in the control group in AHI scores, TST or sleep architecture parameters. The mean baseline and follow-up values for all reported outcomes are displayed in Table 46.

Three patients in the NCPAP group discontinued treatment after a few nights because of the discomfort of the device. These patients were replaced at this point in the trial by a further three patients with similar baseline snoring indices. No adverse events were reported in either group.

Conclusions

There is very little evidence available on the effects of CPAP for the management of non-apnoeic snoring. What evidence is available suggests that the use of CPAP may be effective in reducing levels of snoring, without having any adverse effects on PSG-measured parameters. However, given the poor-quality evidence available it is not possible to draw any definitive conclusions regarding the efficacy and safety of the use of CPAP for the management of non-apnoeic snoring.

Quantity and quality of the studies

Two pre–post studies reported in three publications assessed the effects of MAS. 110,111,113 One further study104 (reported in the section on RFA) compared RFA to the soft palate with the short-term use of a MAS. The results of this study have already been presented in the section on RFA and therefore are not reported again. However, the results are included in the summary section as part of the overall evidence on the effects of MAS.

The aim of the studies by Smith and Battagel113 and Fransson and colleagues110,111 was to assess the effects of using a MAS with a minimum mandibular advancement of 5 mm. The study by Smith and Battagel113 used an adjustable Herbst splint, whereas Fransson and colleagues110,111 used a monobloc MAS. Smith and Battagel113 included 35 non-apnoeic snorers who were followed up before and 1 month after fitting the MAS. In total, 29 patients were followed up at 1 month and therefore included in the assessment of subjective questionnaire-assessed outcomes, with 11 completing a baseline and follow-up unsupervised overnight domiciliary sleep recording. Fransson and colleagues110,111 included both patients with a diagnosis of OSAS and those with non-apnoeic snoring. A total of 77 patients were included in the study: 50 with OSAS and 27 non-apnoeic snorers. Only limited results were reported separately for the non-apnoeic snoring subgroup, with 13 non-apnoeic snoring patients included in the 6-month follow-up of objectively measured snoring scores and 27 assessed for ODI scores at 2 years’ follow-up.

The study by Smith and Battagel113 included oximetry at baseline to rule out a diagnosis of OSAS, whereas the study by Fransson and colleagues110,111 included limited PSG assessment to categorise patients by diagnosis. No tests to assess the site of airway narrowing were reported by Smith and Battagel,113 whereas fibreoptic pharyngoscopic examination with Muller’s manoeuvre, in addition to subjective assessment of the degree of hypopharyngeal collapse and the presence or absence of tongue base hyperplasia, were employed by Fransson and colleagues. 110,111 Across the 62 non-apnoeic snorers included in the two studies, the mean age ranged from 44 to 52 years (range 28 to 70 years) and the percentage of male participants ranged from 57% to 66.7%. The mean BMI ranged from 26.8 kg/m² to 28.4 kg/m², indicating a patient population that was overweight but not obese. In the study by Smith and Battagel113 the mean baseline ESS score was 9.4 (SD 5.0). Neither of the studies reported the smoking status of patients nor levels of alcohol consumption. A summary of study characteristics is presented in Table 48.