Timolol Gel for Epistaxis in Hereditary Hemorrhagic Telangiectasia: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Timolol Gel for Epistaxis in Hereditary Hemorrhagic Telangiectasia (NCT04139018), a randomized phase 2 study. The primary outcome is Change in Assisted Epistaxis Severity Scale (aESS) Score From Baseline at 8 Week Follow-up.

2. Clinical Question

3. Trial Design

This was a randomized, parallel-group phase 2 clinical trial of timolol gel for epistaxis in hereditary hemorrhagic telangiectasia. The registry lists Washington University School of Medicine as sponsor, an enrollment of 27, a start date of 2019-10-20, a primary completion date of 2020-05-20, and a completed status. These registry entries describe the study record; they do not by themselves explain how participants were recruited or how treatment assignment was carried out.

The trial compared intranasal timolol gel 0.1% with placebo gel. The active-arm instructions describe 0.5 mL applied to each nostril twice daily by syringe, amounting to a 2 mg total daily dose. The placebo arm received gel with no active medication. The registry labels the allocation randomized and the model parallel. It describes the masking as quadruple, while the brief study summary calls the study double-blinded. Which individuals were masked, and how masking was maintained or assessed, are not reported in the registry record.

The primary outcome was change in assisted Epistaxis Severity Scale score from baseline to 8-week follow-up. The posted results give a median and full range for each arm, with 13 participants reported in each. The enrollment figure and outcome-analysis counts are not interchangeable: the primary outcome description says that a screen-fail participant in the timolol arm had no baseline data and was excluded. The registry record does not provide a participant-flow account that reconciles every stage from enrollment through outcome analysis.

The record lists one secondary outcome: the number of participants with an improved response on the Clinical Global Impression - Improvement scale. Its results are not available for a two-arm comparison in the posted data. The planned statistical methods, sample-size justification, allocation procedures, and any interim analyses are not reported in the registry record.

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: 272 participants total were assessed for eligibility; 245 met exclusion criteria

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Analysis was used with intention-to-treat principles, so all patients enrolled with baseline data were included. Note: the screen fail participant within the timolol gel arm had no baseline data say they were excluded from the analysis.

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 Change in Assisted Epistaxis Severity Scale (aESS) Score From Baseline at 8 Week Follow-up, measured over Baseline to 8-week follow-up.

The primary endpoint is the change in assisted Epistaxis Severity Scale score from baseline to 8-week follow-up. The aESS uses participants’ reports of epistaxis over the past 1 month. The registry describes the underlying ESS as covering frequency, duration, and intensity of nosebleeds, whether medical attention was sought, anemia, and blood transfusion. The overall scale ranges from 0 to 10, with higher severity categories corresponding to higher scores.

The endpoint is useful because it represents several features of nosebleed burden in a single measure and permits comparison of symptom scores over time. The record says the aESS has the same scoring and minimal important difference as the ESS, and gives an estimated minimal important difference of 0.71. That value is a stated benchmark for interpreting change, not a statistical test of the difference between randomized arms.

Change was calculated as the 8-week aESS score minus the baseline aESS score. Therefore, a positive change indicates a higher follow-up score than baseline, while a negative change indicates a lower follow-up score. The primary result is reported as a median change with the full range, rather than as a mean or a model-adjusted estimate. Because the instrument summarizes a participant-reported reference period, it should be understood as a measure of reported epistaxis burden over time, not a direct count of all nosebleeds during the entire trial.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The primary outcome is summarized separately for each randomized arm using the median change and full range. The registry record does not report the statistical test or model used to compare the arms. The following explanations describe the statistical concepts relevant to reading these summaries; they do not assert that unreported methods were used in this trial.

What does the median describe?

In general, the median is the middle value when observations are ordered. It describes a central participant-level result and is less influenced by unusually high or low observations than the mean. For a change outcome, the median change describes the center of the distribution of individual changes. It does not describe every participant, and it does not show the distribution’s shape. The reported medians are 2.32 units for timolol gel and 1.96 units for placebo gel.

The direction of the change matters. Here, change is defined as the score at 8-week follow-up minus the score at baseline. A positive value means a higher score at follow-up; a negative value means a lower score. Since higher aESS scores indicate more severe epistaxis, a positive change points toward a higher severity score relative to baseline. This interpretation concerns the scale’s direction, not a causal conclusion about treatment.

What does the full range show?

In general, the full range gives the smallest and largest observed values. The timolol arm’s reported range is 0.22 to 5.97, and the placebo arm’s range is -0.91 to 5.98. A range conveys the extent of observed variation, but it is sensitive to extreme observations and does not indicate how values are distributed between its endpoints. It is not a confidence interval and should not be read as a range of plausible treatment effects.

Because the arm ranges extend across many possible individual changes, they show that participant-level outcomes varied. They do not reveal the frequency of improvement, stability, or worsening. Nor does overlap between ranges provide a formal test of whether the randomized groups differ. Such a test would use a defined comparative method and would ordinarily report an effect estimate and uncertainty. Those comparative results are not reported in the registry record.

What would a confidence interval add?

In general, a confidence interval expresses uncertainty around an estimated quantity, such as an arm contrast. It can show which values are reasonably compatible with the observed data under the method’s assumptions. Its width is informative about precision, while its position relative to a meaningful benchmark can aid interpretation. No confidence interval for the between-arm primary outcome contrast is reported in the registry record. The full ranges posted for the arms are participant-level summaries and cannot take its place.

Why consider the outcome distribution?

In general, statistical methods should reflect the outcome’s scale and distribution. A median and range communicate a distribution without assuming that the mean and standard deviation fully capture it. If outcomes are skewed or contain influential extreme values, a method focused on means may not represent the data well. If a mean-based comparison is used, its assumptions and uncertainty need attention. The registry does not report the distribution of the individual aESS changes or the method used for a between-arm comparison, so it is not possible to assess those choices for this trial.

Over-dispersion is a general issue for count outcomes: observed variation can exceed what a simple count model expects. It is not the primary issue in interpreting this continuous scale-change summary. The registry does not describe an over-dispersion analysis, and none should be inferred from the posted medians and ranges.

How should the randomized comparison be understood?

Randomization is intended, in general, to create groups that can be compared without assigning treatment according to participants’ characteristics. A randomized design does not by itself guarantee balanced groups on every characteristic, especially when the study is limited in size. The registry identifies the trial as randomized and parallel, and the primary outcome results report 13 participants per arm. The allocation sequence method, concealment procedures, and baseline characteristics are not reported in the registry record. No adjusted estimate is available to account for baseline score or other characteristics.

The registry describes the trial as quadruple-masked, while its brief summary calls it double-blinded. Masking can reduce the influence of expectations on treatment delivery, reporting, or outcome assessment, depending on who is masked. The record does not identify the masked roles or provide details on whether masking was maintained. Thus, the masking labels can be reported, but their implementation cannot be evaluated from the registry information.

What do the planning illustrations mean?

In general, a sample-size calculation estimates how many participants per arm might be needed to detect a chosen effect under specified assumptions, such as a significance level, allocation, and statistical approximation. The registry’s planning description says its illustrations use the observed effect, a two-sided significance level of 0.05, equal allocation, and a normal approximation, with the effect to detect set to a stated fraction of the observed effect. Those are planning illustrations, not results from the randomized comparison, and they are not observed power. The trial’s own sample-size justification is not reported in the registry record. A planning exercise based on an observed effect also depends on that effect estimate and should not be read as independent confirmation of it.

The record does not report interim analyses, a prespecified statistical analysis plan, or a method for handling missing outcome data beyond the primary outcome population description. Any claims about those procedures would therefore go beyond the registry record.

10. Multiplicity, Missing Data and Analysis Populations

Multiplicity arises when a study assesses several outcomes, comparisons, or analyses. In general, testing many hypotheses can increase the chance of at least one apparently notable result by chance. Appropriate interpretation depends on which outcomes were prespecified, which analyses were performed, and whether any adjustment or hierarchy was planned. The registry lists the change in aESS as the primary outcome and one secondary outcome: the number of participants with an improved response on the Clinical Global Impression - Improvement scale. The method for addressing multiplicity is not reported in the registry record.

The posted secondary outcome does not have values available for a two-arm comparison. The registry record also lists no analyses. No adjustment strategy, outcome hierarchy beyond the primary and secondary labels, or additional exploratory analyses are reported. As a result, multiplicity procedures cannot be assessed from the available record; the absence of reported secondary results should not be interpreted as evidence of no effect on that outcome.

The primary outcome population description says the analysis followed intention-to-treat principles and included all enrolled patients with baseline data. It specifically notes that a screen-fail participant in the timolol gel arm had no baseline data and was excluded. This is a stated analysis population rule, but it is not a complete account of missing-data handling. The record does not report the number of missing follow-up scores, the reasons for missingness, whether any values were imputed, or whether sensitivity analyses were conducted.

In general, intention-to-treat analysis aims to retain participants in the groups to which they were randomized, preserving the basis of a randomized comparison. The description is qualified by the requirement for baseline data and the stated exclusion. It does not establish how every randomized participant was handled. The registry lists 27 enrolled participants and the primary outcome results show 13 participants in each arm; these figures should be read as distinct reporting details, not combined into a reconstructed participant flow. The record does not provide enough information to assess the impact of missing data on the result.

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 non-serious adverse event terms

The serious adverse event counts are reported by arm. The timolol gel arm lists 0 serious adverse events among 14 participants at risk; the placebo gel arm lists 0 among 13 participants at risk. Deaths are reported as 0 in each arm. These are the registry’s recorded counts for the stated safety groups and should not be interpreted as proof that no unrecorded or unobserved events occurred.

The record also lists 7 other adverse events in the timolol gel arm, with 14 participants at risk, and 3 in the placebo gel arm, with 13 participants at risk. These are event counts as posted, not percentages; no rates or comparative tests are reported in the registry record. The record does not provide event terms for the other adverse events in this summary, or explain whether each event was considered related to study treatment. The list of leading serious-event terms is empty.

Safety counts do not establish that a treatment caused an event, and an absence of serious events in these posted counts does not establish safety in broader populations or settings. The arm-level counts should be read alongside the limited size and duration of the trial, without converting them into a causal or definitive safety conclusion.

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

The posted results describe change in aESS score during the trial, separately for the timolol and placebo groups. The median change was 2.32 units in the timolol arm and 1.96 units in the placebo arm. Since the change is follow-up minus baseline and higher scores reflect more severe epistaxis, both medians indicate a higher score at follow-up than at baseline. The full ranges show variation among individual changes, including a negative value in the placebo arm.

These descriptive results do not answer, by themselves, whether timolol differs from placebo. A difference between arm medians is not a formal estimate of a randomized treatment effect, and no comparative confidence interval or p-value is reported. The data as posted therefore do not establish a statistically distinguishable benefit, a lack of benefit, or equivalence between the groups. They also do not show how many participants experienced a change considered important to them.

The registry describes a minimal important difference estimate of 0.71 for the aESS, matching the ESS scoring system. That threshold offers context for individual score changes, but the available arm summaries cannot show how many participants met it. Nor can the threshold replace an estimate of the between-arm contrast and its precision.

The study’s phase 2 status, enrollment of 27, and outcome-analysis counts are relevant context for interpreting the record, but they do not establish the certainty of a treatment effect. The primary outcome analysis included enrolled patients with baseline data and excluded the noted screen-fail participant without baseline data. Other missing-data procedures and sensitivity analyses are not reported. The secondary Clinical Global Impression - Improvement outcome has no available two-arm result.

The most defensible reading is therefore narrow: the registry reports a median increase in aESS score in each arm, with broad observed ranges, and does not report the inferential analysis needed to quantify the comparative effect. These results should not be generalized beyond the study population or treated as medical guidance.

13. Why This Trial Matters Statistically

This trial raises a useful statistical distinction between describing change within each arm and estimating a treatment contrast between randomized arms. A median change for timolol and a median change for placebo each summarize an arm’s outcomes. They do not, on their own, quantify the uncertainty in the difference attributable to assignment. A reader needs a comparative estimate and an account of its precision to assess the randomized question directly.

The endpoint also illustrates why the meaning of a score’s direction must be explicit. Change was calculated as follow-up minus baseline, and higher aESS scores indicate greater severity. Thus, positive median changes correspond to higher scores at follow-up. Without that definition, the numeric direction could easily be misread. The aESS’s one-month reference period also differs from the ESS description’s typical symptoms over the previous 3 months, an important feature when considering how the reported outcome represents symptoms over time.

The record highlights the importance of separating an instrument’s minimal important difference from a test of treatment effect. The stated estimate of 0.71 can inform discussion of the scale, but the posted medians and ranges do not identify how many participants crossed that benchmark or how precisely the arms differ. A clinically interpretable threshold and a statistically supported comparative estimate answer related but distinct questions.

Finally, the enrollment figure, outcome-analysis counts, and missing-baseline exclusion show why participant flow and analysis populations matter. The stated intention-to-treat approach is qualified by inclusion of those with baseline data, and other missing-data procedures are not described. Clear reporting of participant flow, prespecified analysis methods, and comparative uncertainty would help readers evaluate how robust the observed summaries are.

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.

The primary outcome results are descriptive: medians and full ranges are reported for each arm, but a between-arm effect estimate, confidence interval, and p-value are not reported in the registry record. This limits assessment of comparative effect and precision. The posted data also do not give individual participant outcomes, so the distribution of changes and the number of participants with improvement or worsening cannot be determined.

The analysis population description refers to intention-to-treat principles but specifies inclusion of enrolled participants with baseline data. It notes that a screen-fail participant in the timolol arm lacked baseline data and was excluded. The record does not report the handling of other missing follow-up data, reasons for missingness, or sensitivity analyses. The primary outcome results list 13 participants per arm, while the registry enrollment is 27; a detailed flow explaining all counts is not reported.

The endpoint is based on a participant-reported symptom scale. The description states that the aESS refers to epistaxis over the past 1 month, whereas the ESS description refers to typical symptoms over the previous 3 months. The record explains the aESS change calculation but does not provide further information about the implications of that reference-period distinction. Baseline arm scores and participant-level score changes are not reported.

The registry labels masking as quadruple, while the brief summary says double-blinded. The roles masked and procedures used to maintain masking are not reported. Allocation sequence generation and concealment are also not described. The secondary outcome lacks values for a two-arm comparison, and the statistical analysis plan, sample-size justification, and interim-analysis details are not reported in the registry record. These limitations constrain evaluation of both the analysis and the robustness of the reported findings.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: Timolol Gel for Epistaxis in Hereditary Hemorrhagic Telangiectasia — NCT04139018.
• Registry record last updated 2021-08-24; status last verified 2021-07.
• Results first posted 2021-08-24.
• 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

The registry record describes a completed randomized, parallel-group phase 2 trial of intranasal timolol gel versus placebo gel for epistaxis in adults with hereditary hemorrhagic telangiectasia. The active arm received timolol gel 0.1%, described as 0.5 mL in each nostril twice daily for a 2 mg total daily dose. The primary outcome was change in aESS score from baseline to 8-week follow-up.

The posted primary outcome results report median changes of 2.32 units in the timolol arm and 1.96 units in the placebo arm, with full ranges of 0.22 to 5.97 and -0.91 to 5.98, respectively. The record does not report a comparative effect estimate or inferential test. It lists one secondary outcome, but values are missing or not estimable for a two-arm comparison. Serious adverse event counts are 0 in both arms; other adverse event counts are also reported. Missing-data procedures beyond the stated baseline-data analysis population are not reported in the registry record.