OTO-313 in Subjects With Unilateral Subjective Tinnitus: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for OTO-313 in Subjects With Unilateral Subjective Tinnitus (NCT04829214), a randomized phase 2 study. The primary outcome is Percentage of Tinnitus Functional Index (TFI) Responders at Weeks 4 and at Week 8.

2. Clinical Question

The clinical question is therefore focused on a repeated, threshold-based measure of change rather than on the average change in a score at a single visit. Requiring the criterion at both specified weeks makes the primary outcome depend on meeting the threshold at both assessments. The registry record also lists daily tinnitus loudness, daily tinnitus annoyance, and Patient Global Impression of Change at Week 8 as secondary outcomes. Those measures address related aspects of the reported experience, but they do not replace the primary responder definition.

3. Trial Design

This was a randomized, triple-masked, parallel-group Phase 2 study of OTO-313 versus placebo in people with unilateral subjective tinnitus. The official title describes OTO-313 as a single intratympanic injection. The sponsor was Otonomy, Inc. The registry record reports enrollment of 153 and a completed study. The recorded study start was 2021-03-22, and primary completion was 2022-06-30.

The posted primary results use the OTO-313 Full Analysis Set and Placebo Full Analysis Set. These sets contain 77 and 76 participants, respectively. The primary endpoint is binary: a participant either met the specified responder criterion at both Week 4 and Week 8, or did not. The posted counts meeting the criterion are 20 for OTO-313 and 27 for placebo. The registry record also reports a sponsor analysis using the Mantel Haenszel method for a risk difference.

Randomization is intended, in general, to support a fair comparison by allocating participants to study groups according to a chance-based process. Masking is intended, in general, to limit the influence of knowledge about assignment on participants or study conduct. The term “triple-masked” is recorded, but which people or roles were masked is not reported in the registry record. Details of randomization procedures, allocation concealment, and any stratification factors are also not reported in the registry record.

The study includes safety analysis sets as well as full analysis sets for efficacy outcomes. The registry record does not explain how participants were assigned to these analysis populations, whether all randomized participants were included in each analysis, or how missing endpoint observations were handled. It also does not report the trial's original sample-size justification or statistical analysis plan. These gaps limit how fully the published comparisons can be reconstructed from 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: All subjects registered for this study signed an informed consent and entered a lead-in period. Following Lead-in, 153 subjects were randomized (=Full Analysis Set) and 151 subjects received study drug (=Safety Analysis Set). The most common reason for not being randomized was too low a score on the Tinnitus Functional Index (TFI)..

Randomization

Follow-up and Efficacy Analysis

Reasons for not completing this period:

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 Percentage of Tinnitus Functional Index (TFI) Responders at Weeks 4 and at Week 8, measured over Week 4 and Week 8 (both had to meet criterion for the subject to be considered a "responder").

The primary endpoint is the percentage of Tinnitus Functional Index (TFI) responders at Week 4 and Week 8, with both assessments required for a participant to count as a responder. The registry describes the TFI as a validated, 25-item questionnaire with an index score from 0 to 100; higher scores indicate a greater problem with tinnitus. A responder was defined as a participant with at least a 13-point improvement from baseline, and the improvement criterion had to be met at both time points.

This is a binary endpoint: each participant is classified as meeting or not meeting the responder definition. The registry's outcome title refers to a percentage, while the posted arm results give counts of participants meeting the criterion. The interpretation should therefore keep the two types of information distinct: the counts describe how many participants qualified in each analysis set, while the between-group analysis estimates a contrast in responder frequency.

A threshold-based responder endpoint can be useful because it expresses change in terms of a defined amount of improvement rather than only reporting a group average. Requiring the threshold at both specified weeks also means that a participant must satisfy the criterion at each assessment to be classified as a responder. In general, this kind of endpoint does not show the full distribution of individual score changes, and participants just below and just above the threshold are placed in different categories. The registry record does not report the rationale for choosing the 13-point threshold or explain how the threshold was selected for this study.

7. Results Posted in the Registry

• OTO-313 Full Analysis Set: 26.0% (20/77) versus 35.5% (27/76) on control; risk difference -9.6 percentage points (95% CI -24.1 to 5.0), risk ratio 0.73 (0.45–1.19); unadjusted two-proportion z-test p = 0.200.

8. Secondary Efficacy Endpoints

The registry also posts results for secondary outcomes. For each outcome below, the endpoint type, the test applied to the posted summary statistics, the result and its interpretation are given in turn. Secondary endpoints are usually not adjusted for multiple testing, so they support rather than confirm the primary finding.

Change From Baseline in Daily Tinnitus Loudness at Week 8

• Endpoint type: Continuous: a mean value per participant, compared as a difference in means.
• Test: Unadjusted two-sample z-test for the difference in means, using the posted standard deviations.
• Result:
  • OTO-313 Full Analysis Set: mean -0.67 versus -0.90 on control; difference 0.23 (95% CI -0.30 to 0.76), standardized effect d = 0.14; z-test p = 0.393.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between OTO-313 Full Analysis Set and the control. The interval gives the range of effects that remain compatible with the data.

Change From Baseline in Daily Tinnitus Annoyance at Week 8

• Endpoint type: Continuous: a mean value per participant, compared as a difference in means.
• Test: Unadjusted two-sample z-test for the difference in means, using the posted standard deviations.
• Result:
  • OTO-313 Full Analysis Set: mean -0.73 versus -1.11 on control; difference 0.38 (95% CI -0.21 to 0.97), standardized effect d = 0.20; z-test p = 0.206.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between OTO-313 Full Analysis Set and the control. The interval gives the range of effects that remain compatible with the data.

Patient Global Impression of Change at Week 8

• Endpoint type: Continuous: a mean value per participant, compared as a difference in means.
• Test: Unadjusted two-sample z-test for the difference in means, using the posted standard errors.
• Result:
  • OTO-313 Full Analysis Set: mean 0.32 versus 0.36 on control; difference -0.04 (95% CI -0.42 to 0.34), standardized effect d = -0.03; z-test p = 0.838.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between OTO-313 Full Analysis Set and the control. The interval gives the range of effects that remain compatible with the data.

The secondary outcomes address daily tinnitus loudness, daily tinnitus annoyance, and Patient Global Impression of Change at Week 8. Taken together, these measures provide additional perspectives on symptoms and perceived change, but they do not overturn or replace the primary result. Their overall pattern does not give a consistent basis for claiming a clear advantage for OTO-313 over placebo: the posted group summaries and unadjusted comparisons remain uncertain, with their intervals including no difference.

These outcomes relate to the primary endpoint because they concern change or perceived status over the study period, while the primary outcome classifies participants according to the TFI threshold at both specified weeks. Different measures can capture different aspects of an experience and need not yield identical estimates. A responder classification, an average change in a daily rating, and a global impression score are not interchangeable outcomes. The registry record does not report a prespecified hierarchy among the secondary measures.

Unadjusted secondary results carry less weight for confirmatory claims when multiple outcomes are examined. In general, testing several outcomes without accounting for multiplicity increases the chance that at least one result will appear notable by chance alone. The registry record does not report whether a multiplicity adjustment was used, whether secondary outcomes were intended as confirmatory, or how their interpretation was prioritized. These results are best treated as supportive or exploratory evidence, rather than as independent confirmation of the primary finding.

There is also limited information about the modeling of the secondary outcomes. The registry lists group summaries and recalculated unadjusted comparisons, but does not report sponsor analyses for these endpoints or provide full model specifications. For Patient Global Impression of Change, the posted group values are least-squares means with standard errors; the model and covariates underlying those estimates are not reported in the registry record. These limitations constrain comparisons across outcomes and make the secondary findings less definitive than a prespecified, fully documented analysis plan would allow.

Secondary outcomes registered for this trial, for reference (only those above were recalculated):

• Change From Baseline in Daily Tinnitus Loudness at Week 8
• Change From Baseline in Daily Tinnitus Annoyance at Week 8
• Patient Global Impression of Change at Week 8

9. Statistical Interpretation

10. Statistical Methods Explained

The primary endpoint is binary because it classifies each participant according to whether the TFI improvement criterion was met at both Week 4 and Week 8. Binary outcomes are commonly summarized by the number meeting the criterion in each group and compared using a measure of the difference in event frequency. In this trial, “event” means meeting the responder definition, not an adverse event. The posted counts are 20 for OTO-313 and 27 for placebo.

What does a risk difference represent?

In general, a risk difference compares the proportion experiencing an outcome between groups by subtracting one group's proportion from the other's. It is expressed on the probability scale, so a value of no difference corresponds to zero. The sponsor reports a risk difference of -10.2, with a 95% confidence interval from -26 to 5. The posted results do not describe the scale convention used for that reported value in more detail, so it is best to retain the registry's reported label and values rather than infer additional precision.

The recalculated comparison uses the posted summary counts and is unadjusted. Its p-value is 0.2. A p-value measures how unusual results at least as incompatible with a no-difference model would be under the assumptions of the test; it is not the probability that the null hypothesis is true, nor does it measure the size or clinical importance of an effect. The recalculated result is useful as a check based on the available summaries, but it does not replace the sponsor's analysis.

Why report a Mantel Haenszel analysis?

The registry identifies the sponsor's method as Mantel Haenszel and the parameter as a risk difference. In general, Mantel Haenszel procedures combine information across strata, often to account for a stratification variable while estimating an overall association. The registry record does not identify any strata, state whether the analysis incorporated stratification factors, or provide enough methodological detail to determine the exact implementation. Those details should not be inferred from the method name alone.

The sponsor's reported p-value is =0.385 and its confidence interval is reported at 95%. The simple recalculation and sponsor analysis are not identical procedures. A difference between their p-values can reflect differences in method or assumptions; the registry record does not provide a full explanation of the relationship between the analyses. Since the sponsor's result is the reported trial analysis, it should be read as the registry's primary inferential result, while the simple recalculation remains an unadjusted summary-based check.

How should the confidence interval be read?

In general, a confidence interval conveys uncertainty around an estimated effect. A wider interval indicates less precision than a narrower interval, and an interval that includes the value representing no difference does not rule out either direction of the effect. The sponsor's interval runs from -26 to 5, and it spans zero. This means that the posted result does not locate the group contrast precisely enough to exclude no difference under the reported interval procedure. The interval should not be read as a range containing a fixed proportion of individual participant outcomes.

The endpoint's threshold also matters statistically. In general, converting a continuous score change into a responder category discards information about how far each participant's change is above or below the threshold. The 13-point criterion makes the outcome clinically interpretable as a defined response, but participants with different score changes can receive the same classification, and small differences around the cutoff can alter classification. The registry record does not report a separate analysis of the full TFI score change as a primary outcome.

The secondary outcomes are continuous or score-based summaries. General methods for continuous outcomes compare group averages or model-based means, with uncertainty depending on variability and sample size. The posted secondary comparisons include unadjusted estimates and confidence intervals, but no sponsor-reported analysis is listed for them. One secondary result is reported as a least-squares mean with standard error, and the registry record does not describe the model that produced it. A standard error describes uncertainty in an estimated mean; it is not the same quantity as the standard deviation describing participant-to-participant spread.

These methodological distinctions help explain why the arm-level summaries, sponsor analysis, and recalculation should not be treated as interchangeable. The arm results describe observed summaries in the analysis sets; the sponsor analysis is the registry's stated inferential method; and the recalculation is a simple unadjusted comparison from posted summary statistics. The record does not include the full statistical analysis plan, model specifications, or detailed handling of missing data, so further reconstruction would require assumptions that are not documented.

11. Confidence Intervals and Precision

The sponsor's primary analysis estimates a risk difference of -10.2, with a 95% confidence interval from -26 to 5. This interval includes zero, the value corresponding to no difference on a risk-difference scale. It also extends to values on both sides of zero, so the direction of the underlying contrast is not established with precision by this result. The sponsor-reported p-value is =0.385.

In general, a confidence interval communicates uncertainty in an estimate under the method used to calculate it. It is not a guarantee that the true effect lies within the interval, and it does not describe the range of individual participant responses. The width of the interval indicates that a range of possible group contrasts remains compatible with the sponsor analysis. The registry record does not report a clinically important difference for this endpoint, so the interval cannot be judged against a stated clinical threshold.

The recalculated unadjusted interval for the risk difference runs from -0.241 to 0.05, with an unadjusted p-value of 0.2. This is a separate calculation from the sponsor's Mantel Haenszel analysis. The two intervals should not be combined or treated as if they were generated by one method. The registry record does not supply further information that would allow the sponsor's calculation to be reproduced in full.

Precision is also affected by the way the endpoint is defined. Requiring a response at both Week 4 and Week 8 creates a binary classification based on two assessments. The analysis-set sizes are 77 and 76, and the counts meeting the response rule are 20 and 27. These summaries do not reveal the distribution of TFI changes among participants who did not meet the rule, nor do they show how missing assessments affected classification. That information is not reported in the registry record.

12. Planning a Similar Trial: Sample Size

This is a planning illustration, not a re-evaluation of the original trial. It asks: if a new study wanted to detect a difference like the one observed, or a smaller one, how large would it need to be? It is deliberately not “observed power” of the finished trial, which is not a meaningful quantity.

The sample-size figures are planning illustrations based on the observed effect, not the trial's original design justification. For this binary primary endpoint, the listed figures are participants per arm. The stated assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. The effect to detect is the stated fraction of the observed effect. The trial's own original design assumptions are not reported in the registry record.

The 80% power column gives the illustrative number of participants per arm for a future study to have an 80% chance of detecting the stated effect under those assumptions, if that effect were the one that truly applied. The 90% power column uses a higher target chance of detection and therefore requires more participants under the same assumptions. For example, planning for 100% of the observed effect lists 366 participants per arm for 80% power and 489 for 90% power. Planning for 75% of the observed effect lists 664 and 889 participants per arm, respectively; planning for 50% lists 1,524 and 2,040.

Smaller effects are harder to distinguish from random variation, so the planning illustrations require more participants per arm as the assumed effect becomes smaller. These figures are not observed power and do not show the chance that the completed trial's result is correct. They are hypothetical planning calculations, and their relevance depends on the assumptions and effect size used. The registry record does not report that these assumptions were the basis for the trial's original enrollment target.

13. How the Sample Size Is Derived

Power is the probability that a study of a given size rejects the null hypothesis when a specific true effect exists. It depends on four things: the effect to detect, the variability of the outcome, the significance level (here a two-sided α of 0.05, so z_1−α/2 = 1.960) and the desired power (z_1−β = 0.842 for 80% and 1.282 for 90%).

Formula

n is the number of participants in each arm, p₀ and p₁ are the control and treatment proportions, and p̄ is their average.

Worked calculation for this trial

1. Control proportion p₀ = 0.355; treatment proportion p₁ = 0.260; average p̄ = 0.308.
2. First term: z_1−α/2 × √[2p̄(1−p̄)] = 1.96 × 0.6526 = 1.2791.
3. Second term: z_1−β × √[p₀(1−p₀) + p₁(1−p₁)] = 0.842 × 0.6491 = 0.5463.
4. n = (1.2791 + 0.5463)² / (-0.0955)² = 365.2, rounded up to 366 participants per arm for 80% power.

The smaller-effect rows in the table above repeat the same calculation with the effect multiplied by 0.75 and 0.5 (for ratios, applied to the logarithm of the ratio).

R code

alpha <- 0.05
z_a   <- qnorm(1 - alpha / 2)          # two-sided
p0 <- 0.355263157894737                          # control proportion
p1 <- 0.25974025974026                          # treatment proportion
n_per_arm <- function(power, frac = 1) {
  q1 <- p0 + (p1 - p0) * frac               # smaller effect = fraction of the observed difference
  pbar <- (p0 + q1) / 2
  z_b <- qnorm(power)
  ceiling(round((z_a * sqrt(2 * pbar * (1 - pbar)) + z_b * sqrt(p0 * (1 - p0) + q1 * (1 - q1)))^2 / (q1 - p0)^2, 6))
}
for (f in c(1, 0.75, 0.5))
  cat(sprintf("%3.0f%% of observed effect: n per arm = %s (80%% power), %s (90%% power)\n", 100 * f,
              format(n_per_arm(0.80, f), big.mark = ","), format(n_per_arm(0.90, f), big.mark = ",")))

# Check against R's built-in function (unrounded n per arm)
cat("power.prop.test:", round(power.prop.test(p1 = p0, p2 = p1, power = 0.80)$n, 1), "\n")

# Simulation: how often does the z-test reject at the 80%-power sample size?
set.seed(4829214)
n    <- n_per_arm(0.80)
reps <- 20000
x0 <- rbinom(reps, n, p0); x1 <- rbinom(reps, n, p1)
ph <- (x0 + x1) / (2 * n)
z  <- ((x1 - x0) / n) / sqrt(ph * (1 - ph) * 2 / n)
cat(sprintf("Simulated power at n = %s per arm: %.3f (%d replicates)\n", format(n, big.mark = ","), mean(abs(z) > z_a, na.rm = TRUE), reps))

Output from running the code

100% of observed effect: n per arm = 366 (80% power), 489 (90% power)
 75% of observed effect: n per arm = 664 (80% power), 889 (90% power)
 50% of observed effect: n per arm = 1,524 (80% power), 2,040 (90% power)
power.prop.test: 365.2 
Simulated power at n = 366 per arm: 0.802 (20000 replicates)

The simulation draws many pairs of event counts at the planned sample size and counts how often the two-proportion z-test rejects the null. It should land close to the target power; at very small required sample sizes it can fall below it, because the normal approximation is rough there.

14. Analyses Reported by the Sponsor

The registry reports the sponsor's primary comparison as a Mantel Haenszel risk difference of -10.2, with a 95% confidence interval from -26 to 5 and p =0.385. That is the sponsor's stated analysis and should be kept distinct from the simple unadjusted recalculation based on posted summary statistics, which gives p = 0.2. The recalculation is a check using the available arm-level information; it is not a reproduction of the sponsor's method.

The discrepancy flag is false. Even so, the analyses use different stated approaches, and the record does not give enough detail to reconstruct every step of the Mantel Haenszel calculation or explain its relationship to the unadjusted calculation. The sponsor's result should be used when describing the trial's reported inferential analysis. The recalculated result can help readers understand what a straightforward comparison of the posted summaries yields, but it should not be presented as the sponsor's test or as evidence that the two methods are interchangeable.

Both results should be interpreted cautiously. The sponsor's confidence interval includes zero, and the reported p-value does not meet a conventional threshold for statistical significance. Neither result establishes equivalence or proves that a difference is absent. The registry record does not report an equivalence margin, a non-inferiority framework, or the full analysis plan.

15. Multiplicity, Missing Data and Analysis Populations

Multiplicity arises when a study evaluates more than one outcome, time point, subgroup, or analysis. In general, conducting many statistical tests without an adjustment increases the chance of at least one apparently positive result arising by chance. This trial has a primary responder endpoint and three listed secondary outcomes, and the primary definition itself requires assessment at both Week 4 and Week 8. The registry record does not report a multiplicity strategy, a testing hierarchy, or whether the secondary outcomes were intended for confirmatory inference.

The posted secondary comparisons are unadjusted. They can describe patterns in the recorded summaries, but their p-values should not be interpreted as if each were the only test conducted. Without a reported multiplicity plan, the role of these results in formal inference is unclear. The record also does not state whether any additional analyses or outcomes were conducted but not posted.

Missing data are another important consideration for a responder endpoint that requires results at two follow-up assessments. The registry record does not report the number or pattern of missing TFI assessments, the rules used to classify participants with missing observations, or whether an imputation method was applied. It also does not describe sensitivity analyses examining alternative assumptions about missing outcomes. These omissions matter because the observed responder count can depend on how missing assessments are handled.

The full analysis sets are named, and the posted primary results report 77 participants in the OTO-313 set and 76 in the placebo set. The registry record does not explain the criteria for membership in these sets, whether they correspond to all randomized participants, or how exclusions were handled. It also does not provide a detailed accounting of participants who were randomized, received the injection, completed follow-up, or contributed data to each endpoint. The safety results use separately named safety analysis sets, but membership rules are not reported.

The original statistical analysis plan, including any prespecified covariate adjustment or interim analysis approach, is not reported in the registry record. Therefore, it is not possible to determine from the registry whether the posted analyses followed all planned procedures or whether other analyses were considered. These missing details limit reproducibility and the strength of conclusions that can be drawn from the registry results alone.

16. 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 2% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: arteriosclerosis (1.3%), myocardial ischaemia (1.3%).

The safety analysis sets include 77 participants assigned to OTO-313 and 76 assigned to placebo. The posted results report serious adverse events affecting 2 participants in the OTO-313 Safety Analysis Set and 0 in the Placebo Safety Analysis Set. No deaths are reported in either group. The listed serious adverse event terms are Arteriosclerosis and Myocardial ischaemia, each reported as 1.3%.

These counts describe what the registry record reports; they do not establish that an event was caused by the study intervention. The record does not provide event narratives, timing, severity details beyond the serious classification, or an assessment of relatedness. It also does not report whether any serious event led to discontinuation or how the events were managed.

The “other” adverse event counts are 23 in the OTO-313 Safety Analysis Set and 24 in the Placebo Safety Analysis Set. The registry record does not provide additional detail about the types or severity of those other events in the information recorded. Safety counts should be read as descriptive observations within the trial, not as a definitive comparison of risk. The analysis sets are limited in size, and the registry record does not provide a longer-term safety follow-up summary.

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

The primary result describes a comparison of TFI responder counts under a specific definition: participants needed at least a 13-point improvement from baseline at both Week 4 and Week 8. The posted counts were 20 in the OTO-313 Full Analysis Set and 27 in the Placebo Full Analysis Set. The sponsor's analysis estimated a risk difference of -10.2, with a 95% confidence interval from -26 to 5 and p =0.385. These results do not show a statistically clear difference between the groups under the reported analysis.

That finding does not establish that OTO-313 and placebo have identical effects. The confidence interval includes no difference and values in either direction, so the estimate remains uncertain. Nor does the result prove that a clinically meaningful difference is absent: the registry record does not report a threshold for clinical importance against which to assess the interval. A result that does not meet a conventional significance threshold should not be translated into a claim of equivalence.

The endpoint also has a specific scope. It records whether a participant crossed a TFI improvement threshold at both stated visits; it does not show every participant's score trajectory, the magnitude of change among responders, or the distribution of change among nonresponders. The secondary measures provide related but distinct information and are reported without a documented multiplicity strategy. They should not be treated as independent proof of benefit or lack of benefit.

The trial's design and registry results do not support medical advice. They describe one completed Phase 2 randomized, triple-masked comparison in people with unilateral subjective tinnitus. The registry does not provide all details needed to assess missing observations, analysis-set membership, or the complete statistical plan. The results are therefore best understood as a bounded account of the posted evidence, with uncertainty and reporting limitations retained rather than resolved by inference.

18. Why This Trial Matters Statistically

This trial is statistically informative because its primary outcome uses a clear, threshold-based responder definition that requires improvement at both Week 4 and Week 8. That design choice makes the endpoint easy to state, but it also means that the outcome depends on classification around a cutoff and on having information from both assessments. The posted counts and group comparison illustrate why endpoint definition and missing-data handling matter when interpreting binary outcomes.

The registry presents both the sponsor's Mantel Haenszel analysis and an unadjusted recalculation from summary statistics. Keeping those analyses distinct is an important statistical practice: they are based on different procedures and should not be combined into a single result. The sponsor's interval includes no difference, while the simple recalculation gives a different p-value. The registry does not fully document the sponsor method's implementation, so transparency about what is reported and what is not reported is essential.

The trial also illustrates how secondary outcomes can add context without automatically strengthening a primary conclusion. Multiple measures can be useful, but an unreported multiplicity plan makes it difficult to judge the evidential status of unadjusted secondary comparisons. Likewise, the sample-size illustrations can inform future planning, but they are based on an observed effect rather than the trial's original design assumptions and are not observed power.

Finally, the registry record separates efficacy and safety analysis sets and reports both kinds of outcomes, but does not explain all population and missing-data decisions. These are not minor technical details: they shape what the estimates represent and how reproducible the analysis is. The statistical value of the record is therefore both in the results it reports and in making visible where further methods information would be needed for a more complete assessment.

19. 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 limitations specific to this trial follow from the registry information. The primary endpoint requires a participant to meet the TFI threshold at both Week 4 and Week 8, but the record does not report missing assessment counts or how missing values were treated. Without those details, it is difficult to assess whether the responder counts depend on assumptions about incomplete follow-up.

The registry names full analysis sets for the primary endpoint and safety analysis sets for adverse events, but does not explain how participants qualified for each set or whether the efficacy sets include everyone randomized. It also does not provide a full participant accounting across randomization, treatment, and endpoint assessment. This limits understanding of the populations represented by the reported summaries.

The sponsor's primary method is identified as Mantel Haenszel, but the strata, implementation details, and analysis plan are not reported in the registry record. The sponsor's reported result and the unadjusted recalculation are therefore not fully reproducible from the posted information. The record does not explain why the analyses differ or specify whether additional adjustments were prespecified.

The original sample-size rationale and design assumptions are not reported. The available planning illustrations are calculated from observed effects and should not be taken as a record of the study's design. The registry also does not report a multiplicity plan for the secondary endpoints, a clinical-importance threshold for the primary risk difference, baseline group characteristics, or detailed eligibility criteria. These omissions constrain assessment of generalizability, confirmatory status, and the practical interpretation of the estimated effect.

For safety, the record gives serious adverse event counts and selected terms but does not provide narratives, relatedness assessments, or detailed follow-up. The completed status and dates are recorded, but they do not resolve these reporting gaps. Conclusions should remain limited to the posted analyses and their uncertainty.

20. Related Tutorials

21. Related Statistical Calculators

22. Sources

• ClinicalTrials.gov: OTO-313 in Subjects With Unilateral Subjective Tinnitus — NCT04829214.
• Registry record last updated 2023-01-06; status last verified 2022-12.
• Results first posted 2023-01-06.
• 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.

23. Record Summary

NCT04829214 is a completed randomized, triple-masked, parallel-group Phase 2 study sponsored by Otonomy, Inc. It compared a single intratympanic injection of OTO-313 with placebo in people with unilateral subjective tinnitus. The registry reports enrollment of 153, a study start of 2021-03-22, and primary completion on 2022-06-30.

The primary endpoint was TFI responder status at both Week 4 and Week 8, requiring at least a 13-point improvement from baseline at each assessment. Posted responder counts were 20 for OTO-313 and 27 for placebo. The sponsor's Mantel Haenszel risk difference was -10.2, with a 95% confidence interval from -26 to 5 and p =0.385. The interval includes no difference, and the registry record does not establish equivalence or absence of an effect.

Three secondary outcomes and safety results are also recorded. Important analysis details, including missing-data handling, full analysis-set rules, multiplicity procedures, and the original sample-size justification, are not reported in the registry record.