Of the 75 patients recruited, 5 dropped out and 70 (44 males, 26 females) completed the trial. The average age ranged from 24 to 79 (mean 46.45) and the BMI from 20.4 to 34 (mean 26.7). Of the 5 that dropped out, one did so because of strong gag reflex and therefore was unable to tolerate the device, one lady became pregnant, one participant had very poor dental hygiene and the remaining two withdrew for personal reasons. Of the 70 patients, 32 were simple snorers, and 38 had mild OSA with AHI between 5.05 and 14.8/h. The baseline demographic data is illustrated in Table 1. Snoring data was analysed in all patients, and the respiratory parameters more specifically were studied in the mild OSA sub-group.
Objective measurement of snoring is a complex and poorly defined entity. There are no agreed or published guidelines on which parameters of snoring that should be measured or what degree of change can be considered clinically relevant. Comparative analysis is difficult as within the literature, different trials report on different indices. WatchPat sleep study reports on the % time the individual snores at different thresholds with a fixed position of the recording mic at the sternal notch.
The objective change in snoring time was noted at snoring intensity threshold level of 40 dB (all snoring), 45dB (moderate snoring) and 50dB (epic snoring) and demonstrated an improvement in 66 (94.8%), 62 (88.5%) and 61 (87.1%) of the 70 patients respectively. Table 2 illustrates relation to the reduction of snoring time at three different intensities and demonstrate statistical significant improvement at each level. The average reduction in snoring at each of those intensity levels and the actual reduction in snoring time at 40, 45 and 50 dB levels were 40.8%, 46.6% and 40.9% respectively. As any improvement in snoring could be considered as a liberal entity, we therefore chose to further calculate an improvement of greater than 25% in the snoring to be a more stringent perspective and noted that 74.3% of the patients in our study had demonstrated an improvement of this calibre at 40dB.
Multivariate analysis was performed to identify any predictors of success for snoring improvement at 40dB threshold. Multiple logistic regressions were the method used to determine the predictors for improvement in snoring. AHI was included in the multivariate analysis, and there was no association found. Pre-treatment AHI was used as a continuous variable in the logistic regression analysis of sleep study parameters. However, to assess the effect of three different categories of AHI (<5, 5–10, >10), it was therefore also included as a categorical variable and remained non-significant.
There was a positive association with ESS (i.e. the higher chance of response, the higher the pre-therapy ESS with odds ratio of 1.194) and negative correlation with BMI (odds ratio of 0.0805). There was no association with demographic indices (age, sex, average alcohol intake, smoking), with clinical indices (neck collar size, Nasal pathology, tonsil size, Friedman tongue position) or endoscopic evaluation (endoscopic characteristics, Muller Manoeuvre-related collapse, simulated snoring or Esmarch manoeuvre). There was no obvious significant association between intensity levels of stimulation and reduction in snoring. However, this was not an objective for this particular study.
The AHI in this cohort of 70 patients ranged from 0.2–14.8/h, and the mean AHI value was found to be 5.94/h which dropped to 5.37/h following completion of eXciteOSA® therapy. Likewise, the mean ODI reduced from 4.92 pre-treatment to 4.73 post-treatment. However, in patients with mild OSA (n=38), the mean AHI dropped from 9.8 to 4.7/h and the ODI from 7.8 to 4.3/h (Fig. 3). These changes were statistically significant (p<0.001). It was noted that the greatest reduction of AHI was in the patients with the initial value of AHI greater than 10/h. We evaluated the impact of supine sleep position on our outcomes and noted that the amount of time spent in supine position was 46% and 42% pre- and post-treatment respectively, and this was not statistically significant.
There was a corresponding significant reduction in ESS from 9.0 to 5.1 (p<0.001) for this subgroup of 38 patients (Fig. 3).
When evaluating stimulus intensity (level range 1–15) utilised by participants, we found that the average intensity starting point was at level 6 (equivalent of 18mA), and by week 6, the average tolerable intensity had increased to 9.3 (Fig. 4). Device compliance was measured remotely by app utilisation, and this demonstrated a range of 59.5–95.2% with the average value of device utilisation being 83.3%.
Of the 70 patients, 61 completed the snoring VAS forms adequately. The pre-treatment mean snoring VAS dropped significantly (p<0.001) from 5.88 to 3.98 at week 5/6 of the treatment period, and this benefit was sustained at week 7/8 despite stopping the treatment (Fig. 5).
All 70 patients had completed the ESS form pre- and post-treatment, and the mean respective values were 8.4 and 6.14 and which was statistically significant (p < 0.001). Total PSQI (patient) and total bed partner PSQI decreased significantly (Table 3). With regard the PSQI, the most significant decrease was seen with components 1 (subjective sleep quality) and 5 (Sleep disturbance).
In terms of adverse effects secondary to the use of eXciteOSA® device, 11 participants had commented on experiencing mild problems at some stage during the 6-week therapy period. This included excess salivation by 10 (14.2%) patients, tongue tingling/discomfort by 7 (10%) patients, filling sensitivity by 3 (4.2%) patients, metallic taste and gagging sensation by 3 (4.29%) patients each and tightness in the jaw by 1 (1.43%) patient. In all cases, symptoms were very mild and only lasted during or part of the 20-min therapy time. There were no long-term problems reported.