Transnasal Cooling for Migraine: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Transnasal Cooling for Migraine (NCT04936061), a randomized na study. The primary outcome is Pain Relief.

2. Clinical Question

3. Trial Design

The registry describes a prospective, randomized, double-blind, sham-controlled, parallel-group study. Its stated purpose was to assess the safety, tolerability, and optimal dose of the CoolStat transnasal thermal regulating device for acute migraine treatment. The listed study status is terminated. The reason for termination is not reported in the registry record.

There were three listed groups: high-flow treatment, low-flow treatment, and sham flow with ambient air. The active groups used the CoolStat device, and the sham group used the CoolStat sham device. Each group description specifies a 15 minute transnasal air-flow session. The record does not provide further detail about how participants were assigned, how the flow settings were selected, or how the active and sham procedures were made indistinguishable.

Enrollment is listed as 24. The recorded start date is 2021-10-14, and the primary completion date is 2022-08-31. Phase is listed as NA. CoolTech LLC is the sponsor. These details describe the registry entry; they do not establish why the study stopped or whether every planned activity was completed.

The primary analysis is described as using a mixed logistic regression model, with pain relief without rescue medication at 2 hours as the outcome. No analysis results are posted, and the registry lists zero analyses. It does not report the model specification, the analysis population, the allocation ratio, or a statistical analysis plan. The participant counts attached to the posted primary outcome are 6 for high flow, 9 for low flow, and 9 for sham; the corresponding pain-relief counts are 3, 4, and 8.

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: A total of 87 subjects consented, 24 of whom were randomized to a treatment group. 63 subjects did not return to study sites for randomization during an active migraine attack during the 60-day enrollment period. 7 of those subjects reconsented for an additional 60- day enrollment period. Only two subjects who reconsented came to the clinic for randomization. 4 subjects were withdrawn post-consent because they became in eligible (3) or withdrew their consent (1).

Overall Study

Analysis population for the primary outcome, as defined in the registry: The primary analysis was based on a mixed logistic regression model using the variable Pain Relief without rescue medication at 2 hours post treatment as the outcome.

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 Pain Relief, measured over 2 hours.

The primary endpoint concerns pain relief at 2 hours after a CoolStat treatment session. The registry defines relief as a change from a moderate or severe migraine headache at baseline (Grade 2 or 3) to a mild headache or no headache (Grade 1 or 0). The primary analysis is specifically described as pain relief without rescue medication at that time point. The endpoint is expressed in the registry as a percentage, while the posted group results are participant counts.

A binary endpoint classifies each participant according to whether the defined outcome occurred. This can make the clinical meaning of an outcome straightforward to communicate: a participant either meets the stated pain-relief criterion or does not. Requiring no rescue medication in the primary analysis also distinguishes relief observed without that additional treatment from relief that follows its use. The registry does not give further rules about how rescue medication was handled in the analysis.

The reported counts are 3 participants in high flow, 4 in low flow, and 8 in sham. These are descriptive counts, not an estimate of the uncertainty or statistical significance of differences between groups. No confidence intervals, p-values, or primary-model estimates are reported in the registry record. The endpoint’s definition is available, but the posted results do not provide the inferential analysis described for it.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The registry describes a randomized, double-masked, parallel-group comparison of high-flow treatment, low-flow treatment, and sham ambient air. Randomization is intended to reduce systematic differences between groups, while masking is intended to limit the influence of treatment knowledge on participants’ experiences or outcome assessment. These are general purposes of such design features; the registry record does not report how randomization or masking was implemented or assessed.

What does a binary outcome represent?

The primary outcome is binary: a participant either meets the registry’s pain-relief definition or does not. The definition is a reduction from a moderate or severe headache at baseline to a mild headache or no headache at 2 hours, without rescue medication for the primary analysis. Binary outcomes are commonly analyzed with logistic regression, which models the probability of an event through the log odds. The registry identifies logistic regression as the planned approach, but does not report the model estimates or the exact contrast used.

The posted outcome values are counts of participants with the event. Counts are useful for seeing how many events were recorded in each group, but they do not quantify uncertainty. The group sizes are also important context: 6 in high flow, 9 in low flow, and 9 in sham. With small groups, a change in the outcome classification of a participant can have a noticeable influence on the group-level summary. This is a general statistical consideration, not a claim about what happened to any individual in this trial.

Why use a mixed logistic regression model?

A mixed logistic regression model combines a logistic model for a binary outcome with additional model components that can represent structure in the data. In general, mixed models may be useful when observations are related or grouped, depending on the design and the model specification. The registry names a mixed logistic regression model for the primary analysis, but does not describe its random or fixed effects, covariance structure, covariates, or handling of repeated observations. It is therefore not possible to determine from the record what features the mixed component was intended to address.

Model-based results ordinarily give an estimate for a specified comparison and an indication of uncertainty, such as a confidence interval. A confidence interval describes a range of values compatible with the data and model under the method’s assumptions; it is not a guarantee that the true effect lies within the interval. A p-value, when reported, addresses compatibility with a specified null hypothesis under the analysis assumptions; it does not measure clinical importance. Neither is posted for the primary endpoint in this registry record. The registry lists zero analyses.

What do the displayed counts establish?

The posted primary outcome counts are 3 for high flow, 4 for low flow, and 8 for sham. The counts can be reported directly, but they do not provide the adjusted analysis described by the registry. No independent recalculation of a comparative estimate is available from the record, and no sponsor analysis is posted. There is therefore no recalculated result to set alongside a sponsor result. A reader should not treat the counts as a substitute for an inferential model or as evidence of statistical significance.

The primary endpoint description uses a percentage as its measurement, but the posted group values are counts. No percentages are calculated in this interpretation. Reporting a percentage would require a computation from the counts and denominators, and the registry record does not post those percentages. Likewise, a difference between groups or a measure of relative effect would be a new calculation rather than a reported result.

How should planning assumptions be read?

The registry describes a sample-size planning illustration based on the observed effect, with a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation. It states that the effect to detect is a stated fraction of the observed effect. Such calculations are planning illustrations: they ask what sample size might be needed under chosen assumptions. They are not observed power, do not describe the trial’s original design, and cannot establish that the observed study was adequately powered. The record does not provide a sample-size justification for the trial’s design.

Planning calculations depend on the effect used as a target, the chosen significance level, allocation assumptions, and the statistical method. When based on an observed effect, the illustration also inherits uncertainty in that effect. The registry’s stated assumptions do not supply a reported inferential result for the primary endpoint, and no planning output values are included in this record summary. The sample-size framework should not be read as a substitute for an analysis of the trial data.

Overall, the statistical methods description is incomplete for reproducing or evaluating the primary analysis. The model family and endpoint are named, but the model details, analysis population, missing-data approach, covariate choices, and inferential results are not reported in the registry record.

10. Exact Estimates for Each Arm

No two-arm comparison is shown for this trial, but each arm's result can still be summarized with exact interval estimates computed from the posted counts. The Clopper-Pearson interval is exact and conservative; the Wilson interval is a common closer-to-nominal alternative; the Bayesian credible interval uses a uniform Beta(1, 1) prior. For small groups the three can differ noticeably.

R code and output

arm    <- c("High Flow Treatment", "Low Flow Treatment", "Sham Flow Ambient Air")
n      <- c(6, 9, 9)
events <- c(3, 4, 8)
for (i in seq_along(n)) {
  ci <- binom.test(events[i], n[i])$conf.int          # Clopper-Pearson exact interval
  cr <- qbeta(c(0.025, 0.975), 1 + events[i], 1 + n[i] - events[i])   # Bayesian, uniform Beta(1,1) prior
  wi <- (function(x, n) { p <- x / n; z <- qnorm(0.975); d <- 1 + z^2 / n
    c((p + z^2 / (2 * n) - z * sqrt(p * (1 - p) / n + z^2 / (4 * n^2))) / d, (p + z^2 / (2 * n) + z * sqrt(p * (1 - p) / n + z^2 / (4 * n^2))) / d) })(events[i], n[i])
  cat(sprintf("ARM|%s|%d|%d|%.1f|%.1f|%.1f|%.1f|%.1f|%.1f|%.1f\n", arm[i], n[i], events[i], 100 * events[i] / n[i],
              100 * ci[1], 100 * ci[2], 100 * wi[1], 100 * wi[2], 100 * cr[1], 100 * cr[2]))
}

ARM  High Flow Treatment  6  3  50.0  11.8  88.2  18.8  81.2  18.4  81.6
ARM  Low Flow Treatment  9  4  44.4  13.7  78.8  18.9  73.3  18.7  73.8
ARM  Sham Flow Ambient Air  9  8  88.9  51.8  99.7  56.5  98.0  55.5  97.5

11. Multiplicity, Missing Data and Analysis Populations

The registry lists 7 secondary outcomes and 10 outcome measures. The named secondary outcomes include pain relief, pain freedom, relief from migraine-associated symptoms, freedom from migraine-associated symptoms, and use of rescue medication. Pain relief and pain freedom each appear more than once. The record does not explain whether repeated entries refer to different time points, definitions, or analyses.

When a study evaluates multiple outcomes or multiple comparisons, the chance of obtaining at least one apparently notable result can increase. Statistical methods for addressing multiplicity may be specified in advance, but the registry record does not report whether any adjustment was planned or performed. No secondary analyses are posted, so there are no reported estimates or p-values to interpret across those outcomes.

The registry does not report how missing outcome data were handled, whether any participants were excluded from analyses, or how the analysis population was defined. It also does not report an intention-to-treat or per-protocol analysis population. The primary analysis description identifies the outcome and model family, but does not resolve these questions. Missingness can affect interpretation if participants with unavailable outcomes differ from those with recorded outcomes; this is a general concern, not evidence that such a pattern occurred in this trial.

Because no analyses are posted, the record does not permit assessment of multiplicity adjustment, missing-data assumptions, or sensitivity analyses. The group counts should consequently be treated as descriptive posted results rather than a complete account of the planned statistical evaluation.

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 0% reporting threshold.

Most frequent non-serious adverse event terms

The registry reports zero serious adverse events in high flow, low flow, and sham. It also reports zero deaths in each group. No serious adverse-event terms are listed. These entries describe the recorded counts; they do not establish that the intervention is free of risk or that uncommon harms could be ruled out.

For other adverse events, the registry records 3 participants affected in high flow, 3 in low flow, and 0 in sham. These are counts, not rates adjusted for group size or exposure. No further detail about event types, severity, timing, relatedness, or the method for collecting adverse events is reported in the registry record. The counts should not be used to infer that an event was caused by a device.

Serious and other adverse events are separate categories in the registry summary. A zero serious-event count means no serious events are recorded in the group-level result, while the other-adverse-event counts indicate recorded events in the specified category. Neither set of counts provides a comparative statistical safety analysis. No confidence intervals, p-values, or adjusted safety estimates are posted.

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

The posted results establish that the registry records pain-relief counts of 3 in high flow, 4 in low flow, and 8 in sham for the primary outcome, alongside group sizes of 6, 9, and 9. They also establish the recorded adverse-event counts: no serious adverse events or deaths in any group, and other adverse events affecting 3 participants in high flow, 3 in low flow, and 0 in sham. These are descriptive observations in the registry entry.

The results do not establish whether active treatment differs statistically from sham. The record does not post the mixed logistic regression results specified for the primary analysis, nor does it give confidence intervals or p-values. Without an inferential analysis, the precision of a treatment comparison cannot be assessed. The counts do not show whether a difference is compatible with chance variation, and they do not establish equivalence between groups.

The registry’s termination status does not explain why the study ended. The reason is not reported in the registry record, so it should not be attributed to a particular outcome or operational issue. Likewise, the stated purpose of assessing safety, tolerability, and optimal dose describes the study’s aim, not proof that those questions were resolved.

The primary endpoint is limited to a specified pain response at 2 hours and excludes rescue-medication use for the primary analysis. It does not, by itself, describe other aspects of migraine symptoms, longer-term outcomes, or outcomes beyond those listed in the record. The record lists secondary measures, but posts no analyses for them.

In short, the available entry supports a careful account of the trial design and recorded counts. It does not support a conclusion that one treatment is better, worse, or equivalent, and it does not provide the uncertainty estimates needed to judge the strength of a comparative result.

14. Why This Trial Matters Statistically

This trial illustrates why a registry’s outcome counts and its statistical analysis results answer different questions. Counts show how many participants met an outcome in each group. A model-based estimate and its uncertainty are needed to evaluate a comparison while accounting for the design and the analysis method. Here, the registry describes a mixed logistic regression model for the primary endpoint but posts no analyses.

The three-group design also makes it important to know which comparisons were planned and how multiple outcomes were handled. The registry lists high-flow, low-flow, and sham groups, as well as multiple secondary outcome entries. It does not report a multiplicity strategy or secondary analyses. Without this information, readers cannot determine how the statistical findings across outcomes were intended to be interpreted.

The small group sizes recorded for the primary outcome make uncertainty especially relevant. A count may appear clear while still providing limited information about the underlying event probability or a comparative effect. Confidence intervals and a complete account of the analysis would help characterize how precisely the study estimates its outcomes; neither is posted for the primary endpoint.

The entry also demonstrates that a trial’s stated aim and its posted evidence are not interchangeable. The record describes evaluation of safety, tolerability, and optimal dose, but the registry does not give an inferential analysis or an explanation for termination. A transparent statistical account would need to connect the planned model, analysis population, missing-data approach, and reported estimates. These details are not reported in the registry record.

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 the primary outcome is that the registry posts participant counts but no inferential analysis. Although a mixed logistic regression model is described, estimates, confidence intervals, p-values, and model specifications are not reported. The precision of a comparison therefore cannot be evaluated from the record.

The group sizes attached to the primary outcome are 6 for high flow, 9 for low flow, and 9 for sham. These limited counts constrain how much detail can be learned from the descriptive results, particularly without uncertainty estimates. No comparative effect should be inferred from the counts alone.

The registry does not report the reason for termination, the allocation procedure, the success of masking, the analysis population, or handling of missing outcomes. It also does not report the detailed device settings or explain whether the repeated secondary-outcome entries correspond to different assessment times or definitions. These gaps limit assessment of conduct, reproducibility, and the relationship between the stated design and the posted results.

Safety reporting is similarly limited to group-level counts. The registry gives no event descriptions for the other adverse events, and no inferential safety comparison is posted. The absence of recorded serious events in these groups should not be interpreted as ruling out risk. The record’s status and stated objectives provide context, but do not fill these reporting gaps.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Transnasal Cooling for Migraine — NCT04936061.
• Registry record last updated 2023-09-29; status last verified 2023-09.
• Results first posted 2023-09-29.
• 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

NCT04936061, titled “Transnasal Cooling for Migraine,” is a randomized, double-masked, sham-controlled, parallel-group study listed as terminated. The registry describes high-flow and low-flow active CoolStat sessions and a sham ambient-air session, each lasting 15 minutes. Enrollment is listed as 24; the phase is NA.

The primary endpoint is pain relief without rescue medication at 2 hours, defined by a decrease from moderate or severe headache at baseline to mild headache or no headache. The posted counts are 3 in high flow, 4 in low flow, and 8 in sham. The registry describes a mixed logistic regression model for the primary analysis but posts zero analyses and no inferential results.

Serious adverse events and deaths are recorded as zero in all groups. Other adverse events affected 3 participants in high flow, 3 in low flow, and 0 in sham. The reason for termination, analysis details, missing-data methods, and secondary-outcome analyses are not reported in the registry record.