Medications for Obstructive Sleep Apnea In Children With Down Syndrome: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Medications for Obstructive Sleep Apnea In Children With Down Syndrome (NCT04115878), a randomized phase 2 study. The primary outcome is Obstructive Apnea-hypopnea Index (oAHI).

2. Clinical Question

3. Trial Design

The registry describes a phase 2, randomized, quadruple-masked crossover study, sponsored by the University of Arizona. It is marked completed. The record lists a start date of 2020-10-21 and a primary completion date of 2022-12-30. Enrollment is listed as 22. The crossover design means participants were intended to receive both high- and low-dose ato-oxy in a randomized order, rather than remain in just one dose group.

Each dose period was four weeks. In the high-dose period, participants began with oxybutynin 5 mg and atomoxetine 0.5 mg/kg/day, up to 40 mg, for one week; the atomoxetine dose was then increased to 1.2 mg/kg/day, up to 80 mg. In the low-dose period, the listed regimen was oxybutynin 5 mg and atomoxetine 0.5 mg/kg/day, up to 40 mg. The registry states that dosing was intended to occur approximately 30 minutes before bedtime.

Eligibility screening included non-polysomnography criteria, followed by an in-lab polysomnography night for eligible candidates. The screening polysomnography served as the baseline measure for oAHI and other sleep-study outcomes for participants who enrolled. Participants were to return for inpatient polysomnography and a health-related quality-of-life assessment on the final dosing night of each period. The registry states that participants who withdrew would not be replaced.

The record does not report a washout interval, how period or sequence effects were handled in the analysis, or the statistical model used for the primary endpoint. It also does not report a primary between-period analysis among the posted analyses.

4. Participant Flow and Accounting

The registry reports how participants moved through each stage of the study. All counts below are exactly as posted; missing counts are not inferred from other stages.

Before assignment: Enrollment was only 15 randomized participants rather than planned 24 due to recruitment delays related to the COVID-19 pandemic. 22 participants were enrolled but 7 did not qualify for randomization. 4 did not meet the requirement for apnea-hypopnea index \>5 on their baseline polysomnogram, 1 had hypoxemia unrelated to OSA on their baseline polysomnogram, 1 could not tolerate polysomnogram and 1 participant became adherent to CPAP between consent and baseline polysomnogram.

First Intervention

Second Intervention

5. Baseline Characteristics

The table reproduces the baseline characteristics posted in the registry, which describe the participants analysed at baseline and help show whether the randomized groups were comparable.

Population: Only participants with complete data were included in analysis. One participant did not have usable data on their final sleep study and is not included in these results.

6. Primary Endpoint

The registry lists the primary outcome as Obstructive Apnea-hypopnea Index (oAHI), measured over four weeks.

The primary endpoint is change in obstructive apnea-hypopnea index from baseline, measured over four weeks. oAHI counts obstructive apneas and hypopneas per hour of polysomnography recording. It is a rate: it relates the number of breathing events to time observed, allowing sleep-study recordings of different lengths to be expressed on a common per-hour scale.

Change from baseline represents the difference between a participant’s baseline measurement and a later measurement. In general, using change can focus an analysis on how the measured outcome shifts from the starting value, rather than only comparing final measurements. In this study, the baseline was the screening polysomnography night, and endpoint measurements were planned for the final dosing night of each dose period.

The registry reports mean changes of -3.79 events per hour for the high-dose period and -3.51 events per hour for the low-dose period, each with a standard deviation. These are descriptive period-level summaries. The registry does not report a primary contrast between the periods, its uncertainty, or a statistical test. Therefore, the summaries alone do not establish the size or precision of a high-versus-low dose difference in oAHI change.

Because this is a crossover study, a central comparison would ordinarily use the relationship between measurements from the same participant as well as the observed values. The registry record does not report the paired participant-level data or the method used to account for that relationship.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

Statistical interpretation depends on the design as well as the reported numbers. This trial is described as randomized and crossover: participants were intended to receive both dose conditions, in random order. The dose contrast is therefore not the same as a conventional parallel-group comparison in which each person belongs to only one treatment group. In a crossover study, measurements from one participant under the different conditions are usually related. An analysis that uses that within-participant pairing can differ in precision from a calculation that treats period summaries as independent groups.

What does a mean change describe?

A mean is an average across the observations included in a summary. The primary outcome is change from baseline in oAHI, expressed as events per hour. The registry reports mean changes of -3.79 for high dose and -3.51 for low dose, along with standard deviations. A negative mean change describes a lower average endpoint value than baseline. It does not by itself quantify the contrast between dose periods, because the two values are separate summaries rather than a reported estimate of the high-versus-low difference.

A standard deviation describes variation among observations. It is not a confidence interval and does not show how precisely the difference between conditions has been estimated. Precision for a comparison depends on the size and variability of the comparison data, and, in a crossover design, on the association between each participant’s measurements across periods. The registry does not report that association or a primary model accounting for it.

Why is there no primary comparison?

The registry record says values were missing or not estimable for a two-arm comparison for the rate endpoint. Thus, the record contains arm-level primary summaries but no primary comparative estimate, confidence interval, or test. A comparison based only on the two means would not recover the missing crossover analysis. Nor can a valid paired comparison be reconstructed from the group means and standard deviations alone, because the participant-level pairing information is not reported.

For a rate outcome such as oAHI, the measurement is events per unit of observation time. In general, rate data can have variability that differs from the simple assumptions of common continuous-outcome methods; over-dispersion means the observed variability is greater than a chosen model expects. Suitable analysis depends on the actual data and the prespecified method. The registry record does not report the statistical model, any handling of over-dispersion, or the assumptions used for the primary analysis.

How should the secondary comparisons be read?

The OSA-18 and arousal-index comparisons are unadjusted recalculations from posted summary statistics. The OSA-18 mean difference for low dose versus high dose is -0.82, with a 95% confidence interval from -15.19 to 13.55 and p = 0.911. The arousal-index mean difference is 1.33, with a 95% confidence interval from -4.412 to 7.072 and p = 0.65. These recalculations do not incorporate the paired crossover structure because the summaries do not provide the required within-participant information. The registry lists no reported analyses for these secondary outcomes.

A confidence interval gives a range of values compatible with the estimate under the calculation’s assumptions. Wider intervals indicate less precision. When an interval includes zero for a difference, the data and method do not rule out no difference, but that does not prove equality. A p-value measures how unusual results at least as extreme as the observed result would be under a specified null model and its assumptions. It does not give the probability that the null hypothesis is true, and a large p-value is not evidence that two conditions are equivalent.

What do the planning illustrations mean?

The sample-size illustrations use the observed effect as a planning reference, with rows for planning for 100%, 75%, or 50% of that effect. They assume a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation. The resulting figures are participants per arm for 80% power or participants per arm for 90% power. They are planning illustrations, not the trial’s design justification and not observed power. The registry does not report the original sample-size justification or a prospective power calculation.

The illustrations are based on a two-arm framework and observed summary effects. They should not be treated as a replacement for planning a crossover study, which requires assumptions about within-participant variability and correlation and about period effects. Those assumptions are not reported in the registry record. The illustrations also depend on the observed effect, which can be imprecise in a small study; planning around an uncertain estimate can yield unstable target sizes.

What is not available?

The registry does not report the primary statistical analysis plan, the analysis population, how incomplete crossover sequences were handled, or how missing outcome measurements were treated. It does not report a period or sequence analysis, a carryover assessment, or multiplicity adjustment for secondary outcomes. These omissions limit assessment of whether the posted descriptive summaries correspond to the intended randomized comparison and how much confidence to place in the recalculated secondary results.

10. Multiplicity, Missing Data and Analysis Populations

The registry record does not report a multiplicity plan. Multiplicity arises when several outcomes or comparisons are examined: interpreting each result as if it were the only test can make chance findings more likely across the collection of analyses. The record lists two secondary outcomes, but does not state whether they were prespecified for formal testing, whether any hierarchy was used, or whether p-values were adjusted. The unadjusted secondary recalculations should therefore not be read as confirmatory evidence.

The record also does not report how missing outcome data were handled. The study description says participants who withdrew would not be replaced, but does not state how incomplete measurements or incomplete treatment sequences were included in the analysis. It does not report an analysis population, such as whether results were based on all randomized participants or only those with measurements in both periods. The posted period summaries list 11 observations for each dose condition, while the safety summaries list 14 at risk in each group; the registry does not explain this difference in reporting populations.

For a crossover trial, missingness can be consequential because a participant may have a measurement in one period but not the other. The record does not describe how such partial data were treated or whether any imputation or sensitivity analysis was performed. It also does not report methods for period effects, sequence effects, or carryover. These are design-related considerations for crossover comparisons, not evidence that any particular bias occurred.

Without the analysis plan and a clear account of outcome availability, the posted summaries cannot establish which participants contributed to each analysis or how robust results are to missing data and multiple comparisons. These gaps are especially material for the primary endpoint, for which the record does not report a between-period comparison.

11. Safety Results

The registry reports serious adverse events as affected participants divided by participants at risk.

Deaths

Other (non-serious) adverse events

Participants with at least one non-serious event at or above the registry's 0% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: respiratory failure and death (7.1%).

The safety record lists 14 participants at risk in each dose group. Serious adverse events affected one participant in the high-dose group and none in the low-dose group. Other adverse events affected eight participants in each group. Deaths are listed for one participant in the high-dose group and none in the low-dose group. The record identifies “Respiratory failure and death” as a top serious adverse event term at 7.1%.

These are counts from the registry record, not estimates of comparative risk adjusted for the crossover design or exposure time. They should not be interpreted as proof that either dose caused or prevented an event. The record does not report event-level timing, attribution, severity criteria, exposure-adjusted rates, or a statistical comparison of safety outcomes. It also does not explain how the safety reporting groups relate to the outcome-summary counts. Given the limited counts, the figures are descriptive and cannot precisely characterize uncommon harms.

12. What the Result Does, and Does Not, Mean

The registry record shows mean oAHI changes from baseline for the high- and low-dose periods, as well as secondary outcome summaries and safety counts. The posted primary means are negative in both periods. However, the record does not report the primary high-versus-low contrast or its uncertainty. The two descriptive means are not enough to determine how large or precise a difference between dose conditions may be, particularly in a crossover study where within-participant measurements matter.

The secondary comparisons recalculated from the posted summaries are unadjusted, and their confidence intervals are broad. They do not establish equivalence, and they do not amount to a sponsor-reported crossover analysis. They also do not resolve the missing primary comparison. The safety counts describe events recorded in each group but do not establish causation or comparative safety.

More generally, a trial result is interpreted through both its estimate and its precision. A point estimate alone does not show how much uncertainty surrounds it, and a non-significant p-value does not prove that two conditions have the same effect. Conversely, the absence of a reported comparison is not evidence that no difference exists. The registry record does not provide enough inferential detail to characterize the primary dose contrast.

The results therefore describe what was posted, not a definitive conclusion about the relative effects of the two dose conditions. The record does not report a primary analysis, an analysis plan, or enough detail to evaluate missing data and crossover-specific methods. Any interpretation should keep those limits distinct from the descriptive outcome values and avoid treating planning illustrations or unadjusted secondary calculations as observed trial conclusions.

13. Why This Trial Matters Statistically

This study is statistically notable because it uses a randomized crossover design to compare two dose conditions of the same medication combination. Crossover designs can make efficient use of within-participant comparisons, since each participant can contribute measurements under both conditions. That potential advantage depends on having an analysis that uses the paired structure and addresses features such as period and sequence effects. The registry describes the design but does not report how those features were handled.

The record also illustrates why descriptive summaries are not a substitute for a direct estimate of the comparison of interest. The primary endpoint has period-specific means and standard deviations, yet the registry does not report a between-period estimate, confidence interval, or test. Without that contrast, readers cannot assess the primary comparison’s magnitude or precision from the registry results.

The distinction between a reported analysis and a recalculation is important as well. The secondary comparisons can be reproduced from posted summary statistics, but they are unadjusted and do not use the unavailable participant-level crossover information. Their numerical availability should not be mistaken for a prespecified or sponsor-reported analysis.

Finally, planning illustrations based on observed effects are not the same as the original design rationale or observed power. The registry does not report the sample-size justification. This makes the record a useful reminder that a trial’s design, analysis plan, and reporting details are needed to interpret its estimates—not just the outcome summaries.

14. Limitations

• The posted summary statistics are the only data these results rest on; patient-level data, full protocols and statistical analysis plans are not available.
• Recalculated results are simple, unadjusted calculations from posted summaries. They do not replace the sponsor's prespecified analysis.

Several trial-specific limitations affect interpretation. Most importantly, the registry does not report a primary between-period analysis for oAHI, despite posting mean changes for both dose conditions. The reason recorded is that values were missing or not estimable for a two-arm comparison of the rate endpoint. The primary effect estimate, confidence interval, and test are therefore unavailable in the registry.

The crossover design makes analysis details particularly important. The record does not state how within-participant pairing, period effects, treatment sequence, or possible carryover were handled. It also does not report whether the comparison used all randomized participants or a subset with complete measurements. The brief study description mentions screening and endpoint polysomnography, but does not provide outcome-level completeness or missing-data methods.

Secondary comparisons are recalculated from posted summary statistics and do not account for the crossover pairing. The registry lists no reported analyses for those outcomes, and it does not describe a multiplicity plan. These comparisons should consequently be interpreted as limited, unadjusted summaries rather than confirmatory tests.

The record’s outcome and safety summaries list different counts at risk, and it does not explain the relationship between those populations. Safety reporting is descriptive and lacks event attribution, timing, and an inferential comparison. The registry does not report a sample-size justification, an original statistical analysis plan, or a primary analysis model. These gaps constrain assessment of precision, robustness, and the intended analysis. They do not establish that a particular analysis was incorrect; they mean the registry record does not make those details available.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: Medications for Obstructive Sleep Apnea In Children With Down Syndrome — NCT04115878.
• Registry record last updated 2025-08-19; status last verified 2025-08.
• Results first posted 2025-08-19.
• Data retrieved from ClinicalTrials.gov on 2026-11-09.
• Group sizes, means, event counts and sponsor-reported p-values are reproduced as posted. Effect estimates, confidence intervals and p-values not attributed to the sponsor are independently calculated from the posted summary statistics.

18. Record Summary

NCT04115878 is a completed phase 2 randomized, quadruple-masked crossover study sponsored by the University of Arizona. It enrolled 22 children with Down syndrome and obstructive sleep apnea and compared high- versus low-dose atomoxetine and oxybutynin over four-week periods in randomized order.

The primary endpoint was change from baseline in oAHI. The registry posts a mean change of -3.79 events per hour for the high-dose period and -3.51 for the low-dose period, but no primary between-period analysis. Secondary outcomes were OSA-18 score and arousal index; unadjusted comparisons recalculated from posted summaries are imprecise and do not account for crossover pairing. The record also lists adverse-event and death counts, but does not report comparative safety analyses.