Comparing the Side-lying Sleep Positioning to Back-lying in Infants With Cleft Palate

The primary aim of this project is to determine the clinical effectiveness in infants with cleft palate (CP) of side-lying as compared to back-lying sleep positioning in reducing oxygen desaturation resulting from obstructive sleep apnoea (OSA). This will be achieved:

• comparing oxygen saturation during sleep in the side- and back-lying positions at 1 month of age.
• comparing sleep quality between the side-lying and back-lying groups using a parental questionnaire.
• co-developing national recommendations with parents regarding sleep position for infants with CP.

A randomised controlled trial (RCT) of side-lying compared with back-lying sleep positioning in reducing oxygen desaturation resulting from OSA in infants with CP. The design and conduct of the study will benefit from lessons learned from the feasibility and other previous studies. All parents in the feasibility study indicated strong interest to participate in further studies evaluating the effects of sleep position. Infants meeting the eligibility criteria will be randomised to side-lying or back-lying in a ratio 1:1 using a minimisation routine incorporating a random element to reduce predictability. Minimisation factors will be clinical site and syndrome suspected or indicated (yes / no). Allocations will be delivered via a password-protected web-based system. The allocated position will only be used on the day(s) when the infant is monitored for the study purposes. Thereafter, parents will be free to revert back to the standard sleep position as advised by their cleft centre, should it be different than that used for the monitoring period. All centres represented at the preparatory meeting with the United Kingdom (UK) Lead Clinical Nurse Specialist group, confirmed that the side-lying position was recommended in some infants at their centre, irrespective of whether it was the standard advice used. As such, all centres had experience of the side-lying position. It was decided not to change the specific advice that centres give to parents regarding how to position the infant in a side-lying position, but any standard written or verbal information would be collected by the study documents.

An internal pilot will investigate participants' opinions about their participation in the study. Using bespoke questionnaires participants' experience will be assessed and it will be decided how it could be improved. After 6 months questionnaire results will be analysed and if necessary study amendments will be put in place to reflect participants' opinions.

Data from the feasibility study and published studies have reported estimates of the standard deviation (SD) of the primary outcome ODI-4 in the side-lying infants at four weeks to range from 8 to 11 units, with a higher SD observed in the back-lying group. The observed difference in mean ODI-4 between the side-lying and back-lying infant cohorts was 15 units (a standardised effect size of 0.91). Investigators considered a smaller but more realistic difference in means of five units to be a clinically important difference (SD 10), a standardised effect size of 0.5. The sample size calculation comparing two means with unequal variances for the primary outcome was, therefore, based on a standardised effect size of 0.5. To account for potential unequal variances in each group a variance ratio of 2 was used in the calculations. To detect a difference of 0.5 standard deviation score (SDS) with 80% power and alpha equal to 0.05 would require 96 infants to be monitored in each arm of the trial (a total of 192 participants). Informed by the multicentre feasibility and oximetry studies the sample size will be inflated to 244 participants in the RCT, to allow for potential attrition of 21%.

This will follow a pre-specified and approved statistical analysis plan. The primary analysis of the RCT data will use intention-to-treat. Baseline data will be analysed to assess the comparability of the demographic and clinical characteristics of the participants. Data from the trial arms will be compared using generalised linear models and adjusted for minimisation covariates where appropriate. Estimates of treatment effect size will be reported as differences in means for continuous outcomes, and risk ratios for dichotomous outcomes and reported along with 95% confidence intervals. Subgroup analyses will explore the effects of sleep position on infants with and without suspected associated syndromes through subgroup treatment interactions, at a stricter alpha level 0.01.