Mechanisms of Upper Airway Obstruction: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Mechanisms of Upper Airway Obstruction (NCT04322097), a non-randomized na study. The primary outcome is Pharyngeal Critical Pressure (Pcrit).

2. Clinical Question

3. Trial Design

The registry describes a completed, nonrandomized, parallel-model study of mechanisms of upper-airway obstruction. Its listed enrollment is 133. The study is associated with obstructive sleep apnea and was sponsored by the University of Pennsylvania. The recorded start date is 2020-06-03, and the primary completion date is 2024-05-10. The record lists no phase.

The study has two broad components. Study A, functional phenotyping during drug-induced sleep endoscopy, describes participants with diagnosed obstructive sleep apnea undergoing endoscopy to characterize determinants of upper-airway obstruction. CPAP and pharyngeal manometry are used during the described procedures, and jaw thrust and other positional maneuvers are performed as part of routine clinical standard of care.

Study B concerns participants with implanted Inspire hypoglossal nerve stimulation devices. The registry describes drug-induced sleep endoscopy and overnight polysomnography to assess upper-airway responses with and without stimulation. It also describes CPAP titration during endoscopy, upper-airway MRI, and tongue-force assessments. The record separates this component into responder and nonresponder groups, but the response definitions are incomplete in the registry record.

Allocation is nonrandomized and masking is listed as none. The parallel model label does not make Study A and Study B randomized or equivalent comparison groups. The record does not report a protocol rationale for these design choices, or enough detail to establish how participants entered each component. The listed intervention is hypoglossal nerve stimulation, but the study also includes physiological and procedural assessments rather than a simple intervention-versus-control structure.

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.

Overall Study

Analysis population for the primary outcome, as defined in the registry: Pcrit was unable to be measured in 20 patients (study A) and 2 patients (study B) due to technical issues and protocol deviations.

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.

6. Primary Endpoint

The registry lists the primary outcome as Pharyngeal Critical Pressure (Pcrit), measured over Collected during drug-induced sleep endoscopy (<1 day)..

The primary endpoint is pharyngeal critical pressure, abbreviated Pcrit, measured in cm H2O during drug-induced sleep endoscopy. It is recorded as a mean with standard deviation. Pcrit represents the pressure at which the upper airway closes during inspiration, so it provides a physiological measure of airway collapsibility rather than a direct measure of symptoms or a clinical event.

The registry describes the direction of interpretation: a higher Pcrit indicates greater collapsibility. A Pcrit below zero indicates that the airway remains open; a value at least zero indicates that the airway is closed. These thresholds describe how the measurement is interpreted, not results observed in this study.

Using Pcrit is relevant to the stated mechanistic aim because the study examines factors related to airway patency during sleep and the role of lingual muscles. In general, a continuous physiological measure can retain information across its range rather than reducing observations to categories. However, its interpretation depends on how it was measured and on the participants and conditions represented. The registry reports that Pcrit could not be measured in 20 Study A patients and 2 Study B patients because of technical issues and protocol deviations. Numerical group means, standard deviations, and other primary endpoint results are not reported in the registry record.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The registry describes Pcrit as a continuous measure summarized by a mean and standard deviation. These statistics answer different questions. A mean describes the central value of measurements in a group, while a standard deviation describes how dispersed individual measurements are around that mean. Neither statistic alone describes the full distribution, and neither is available as an observed Pcrit result in the registry record.

What does a mean Pcrit represent?

For a continuous endpoint, the mean combines participants’ measurements into a single summary on the endpoint’s original scale. Here, that scale is cm H2O. The endpoint description provides a physiological interpretation: higher Pcrit means greater airway collapsibility. A mean is useful for describing a group, but it can obscure variation between individuals. It also does not show whether measurements cluster around the mean or include a range of markedly different values. The registry does not report the observed means or standard deviations needed to evaluate those features.

What does the standard deviation add?

As a general statistical concept, the standard deviation describes spread in the observed measurements. It is not the same as the uncertainty in the estimated mean. The standard error, which reflects sampling uncertainty in a mean, depends on both the spread and the amount of information contributing to that estimate. No standard errors or confidence intervals for Pcrit are reported in the registry record. Readers therefore cannot use the record to judge how precisely any group mean or contrast was estimated.

Why does the study structure matter?

The record labels the allocation nonrandomized and says the arms are not a like-for-like control comparison. In general, random allocation can help balance participant characteristics between groups, while nonrandomized group membership can be related to other characteristics that also affect an outcome. This does not mean that every nonrandomized comparison is uninformative; it means that group differences require careful interpretation and that causal conclusions depend on design and analysis details. The registry does not report a detailed statistical plan or enough information to assess adjustment for potential differences between groups.

Study A and Study B also address different settings. Study A describes functional phenotyping during endoscopy in participants with diagnosed obstructive sleep apnea. Study B describes participants with implanted devices and measurements with and without stimulation, alongside other assessments. The record does not state that Study A serves as a control group for Study B. Nor does it report a formal estimate comparing stimulation conditions. The design description alone therefore should not be recast as a randomized comparison.

How should measurement thresholds be read?

The endpoint description says Pcrit below zero indicates an airway that remains open and Pcrit at least zero indicates closure. These are interpretive thresholds for the measure. They do not define a treatment response, and the available record does not state how many observed values were below, at, or above the threshold. Since Pcrit could not be measured in 20 Study A patients and 2 Study B patients because of technical issues and protocol deviations, the analyzed measurements may not represent every enrolled participant. The way missing measurements were handled is not reported in the registry record.

What do the planning assumptions mean?

The registry’s planning illustration is described as being computed from an observed effect, with a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation. It describes the effect to detect as a stated fraction of the observed effect. These are assumptions for planning illustrations, not evidence that the trial was designed that way and not observed power. No numerical planning counts or effect fractions are reported in the available registry information. In general, such calculations describe the sample size that would be associated with specified assumptions; they do not establish the probability that a completed study found a true effect.

The record lists seven outcome measures and two analyses, but it does not describe the statistical methods for those analyses in sufficient detail. It does not report whether models were adjusted, whether assumptions were checked, how repeated or related measurements were handled, or whether confidence intervals were calculated. These missing details limit independent appraisal of the analysis. No recalculated estimate is available from the posted primary endpoint statistics, because the needed group summaries are not reported.

10. Confidence Intervals for Each Arm

No two-arm comparison is shown for this trial, but each arm's mean can be given with a t-based 95% confidence interval computed from the posted mean, spread and number of participants.

R code and output

arm <- c("Study A - Functional Phenotyping During DISE", "Study B - HGNS Responders", "Study B - HGNS Nonresponders")
n   <- c(80, 12, 18)
m   <- c(1.26, -1.42, -1.06)
sp  <- c(3.29, 4.36, 2.62)
sd  <- sp      # posted standard deviations
for (i in seq_along(n)) {
  se <- sd[i] / sqrt(n[i]); h <- qt(0.975, n[i] - 1) * se
  cat(sprintf("CARM|%s|%d|%.3f|%.3f|%.3f|%.3f\n", arm[i], n[i], m[i], sd[i], m[i] - h, m[i] + h))
}

CARM  Study A - Functional Phenotyping During DISE  80  1.260  3.290  0.528  1.992
CARM  Study B - HGNS Responders  12  -1.420  4.360  -4.190  1.350
CARM  Study B - HGNS Nonresponders  18  -1.060  2.620  -2.363  0.243

11. Multiplicity, Missing Data and Analysis Populations

The registry lists seven outcome measures and two analyses. It does not report whether the analyses were prespecified in a detailed statistical plan, which outcome measures were included in each analysis, or whether any adjustment was made for multiple outcomes or comparisons. Multiplicity refers generally to the increased chance of at least one false-positive finding when many statistical hypotheses are tested. Whether an adjustment is appropriate depends on the questions and analysis plan; the record does not provide enough detail to assess that choice.

The primary endpoint is identified as Pcrit, but numerical results for it are not reported in the registry record. The record does not give the analysis populations, define which participants were included in each endpoint analysis, or describe exclusions beyond the stated inability to measure Pcrit in 20 Study A patients and 2 Study B patients. Those measurements were unavailable because of technical issues and protocol deviations. The method used to handle these missing values is not reported in the registry record.

In general, missing outcome data can affect both the amount of information available and the interpretation of a result, particularly if the reasons for missingness are related to the outcome or participant characteristics. The stated reasons identify technical issues and protocol deviations, but the record does not describe their nature or evaluate their consequences. It also does not report sensitivity analyses or a comparison of participants with and without measured Pcrit.

The record does not report an interim analysis plan, a multiplicity strategy, or a complete statistical analysis plan. Their absence from the registry record should not be taken as evidence that no such procedures existed; their details are simply not reported there.

12. 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 5% reporting threshold.

The registry’s adverse-event figures list Study A with 100 at risk for serious adverse events and 0 affected; the other adverse-event count is also 0, with 100 at risk. Study B lists 32 at risk for serious adverse events and 0 affected; the other adverse-event count is also 0, with 32 at risk. Deaths are listed as 0 in both groups, and no top serious-event terms are listed.

These are reported counts, not evidence that an event could not occur. The registry record does not specify the observation period for these figures or provide detail about how events were collected, defined, or adjudicated. The adverse-event figures should therefore be read within the limits of the registry reporting, rather than as a comprehensive estimate of safety or a comparison between equivalent groups. Because the study is nonrandomized and its components are not like-for-like control arms, these counts do not establish comparative safety between Study A and Study B.

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

The record establishes that the study examined mechanisms of upper-airway obstruction in obstructive sleep apnea and identified Pcrit as its primary endpoint. It describes physiological assessments during drug-induced sleep endoscopy, including CPAP and pharyngeal manometry in Study A, and device-related assessments with and without hypoglossal nerve stimulation in Study B. It also describes a hypothesis about the role of specific tongue muscles in maintaining airway patency during sleep.

Those design details do not establish what the study found. The registry record does not report numerical Pcrit means, standard deviations, a contrast between conditions or groups, or uncertainty measures for the primary endpoint. It therefore cannot show the magnitude or precision of any observed difference. The measurement thresholds give a way to interpret Pcrit values, but without the observed values they do not indicate how the study participants’ airways behaved on that scale.

The record also does not establish that stimulation caused any observed response. Study B includes assessments with and without stimulation, but allocation is nonrandomized, and the registry warns that its arms are not a like-for-like control comparison. Without reported numerical outcomes and a clear analysis plan, the descriptive study structure should not be interpreted as proof of a causal treatment effect.

The enrollment figure and adverse-event counts have distinct meanings. Enrollment is listed as 133, while the adverse-event record gives separate at-risk counts for Study A and Study B. The registry also reports that Pcrit could not be measured in 20 Study A patients and 2 Study B patients due to technical issues and protocol deviations. These figures should not be combined or treated as a complete accounting of the endpoint analysis population.

In short, the registry describes a mechanistic investigation and its measurement framework. It does not report enough primary endpoint results to determine the direction, size, statistical precision, or clinical significance of the observed data. A description of the hypothesis is not itself evidence that the hypothesis was supported.

14. Why This Trial Matters Statistically

This study is statistically notable for centering a mechanistic physiological measure rather than a direct clinical event. Pcrit is intended to quantify upper-airway collapsibility, and its interpretation is tied to whether the airway remains open or closes at the stated pressure thresholds. This kind of measure can help frame questions about how airway physiology varies under different assessment conditions, while still requiring careful attention to measurement quality and the population being assessed.

The record also illustrates why design labels and comparison structure matter. Although it uses a parallel-model label and describes more than one study component, it explicitly states that allocation is nonrandomized and the arms are not a like-for-like control comparison. A statistical interpretation should preserve that distinction rather than treating all groups as interchangeable or assigning causal meaning to differences without supporting design and analysis details.

Missing measurements are another important feature. Pcrit was unavailable for 20 Study A patients and 2 Study B patients because of technical issues and protocol deviations. Such information is relevant because the analyzed data may differ from the full set of participants, and the registry does not report how the missing measurements were handled. Statistical summaries cannot be fully assessed without knowing the analysis population and missing-data approach.

Finally, the record lists seven outcome measures and two analyses but does not describe a multiplicity strategy or detailed statistical plan. These omissions constrain interpretation of any potential findings. Most importantly, numerical results for the primary endpoint are not reported in the registry record. The study’s scientific question and design are described, but the posted information does not allow readers to assess the observed effect or its precision.

15. 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.

The main limitation for interpreting this record is that numerical results for the primary endpoint are not reported in the registry record. No Pcrit means, standard deviations, group contrasts, confidence intervals, or p-values are included in the available information. The direction and size of any observed pattern therefore cannot be assessed.

The study is nonrandomized, and the record cautions that the arms are not like-for-like control groups. Study A and Study B have different participant descriptions and assessment contexts. The registry does not report enough detail to determine comparability between groups or to evaluate adjustment for potential confounding. The responder and nonresponder descriptions are also incomplete: their definitions are truncated in the registry record, so the full classification criteria are not available.

Pcrit was not measurable in 20 Study A patients and 2 Study B patients because of technical issues and protocol deviations. The record does not describe whether missingness was addressed through exclusion, imputation, or another method, and it does not provide sensitivity analyses. It also does not provide the analysis populations needed to connect the enrollment figure to the endpoint results.

The registry lists seven outcome measures and two analyses, but does not report a detailed statistical analysis plan or multiplicity handling. The adverse-event figures have no stated observation period in the available information. These limitations concern what can be established from the registry record; they do not show that unreported analyses or procedures were absent.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Mechanisms of Upper Airway Obstruction — NCT04322097.
• Registry record last updated 2026-01-21; status last verified 2026-01.
• Results first posted 2026-01-21.
• 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.

19. Record Summary

The registry identifies a completed, nonrandomized study of mechanisms of upper-airway obstruction in obstructive sleep apnea. The listed enrollment is 133, and the sponsor is the University of Pennsylvania. The study includes functional phenotyping during drug-induced sleep endoscopy and assessments in participants with implanted hypoglossal nerve stimulation devices.

The primary endpoint is Pcrit, a continuous measure of upper-airway collapsibility collected during drug-induced sleep endoscopy and summarized as a mean with standard deviation. The record says measurements were unavailable for 20 Study A patients and 2 Study B patients because of technical issues and protocol deviations. Numerical primary endpoint results and details of missing-data handling are not reported in the registry record.

The registry lists seven outcome measures and two analyses. Its adverse-event figures report zero serious events, zero other events, and zero deaths in the listed Study A and Study B groups. The record does not describe a like-for-like control comparison or report sufficient outcome results to assess the study’s primary statistical findings.