Study Evaluating the Equivalence of GDC-229 and Metronidazole Vaginal Gel 0.75% in the Treatment of Bacterial Vaginosis: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Study Evaluating the Equivalence of GDC-229 and Metronidazole Vaginal Gel 0.75% in the Treatment of Bacterial Vaginosis (NCT03091777), a randomized phase 3 study. The primary outcome is Clinical Cure.

2. Clinical Question

3. Trial Design

The registry describes a completed phase 3, randomized, parallel-group trial with quadruple masking. The official title also describes it as multicenter, double-blind, and vehicle controlled. These terms indicate that participants were allocated among concurrent study groups and that masking was intended across multiple groups involved in the study; further details about who was masked or how allocation was implemented are not reported in the registry record.

The three arms were GDC-229, reference metronidazole vaginal gel 0.75%, and GDC-229 vehicle placebo gel. All were described as vaginal gels applied as directed. The record lists enrollment of 871 participants. The primary endpoint results list 245 participants in the GDC-229 arm, 245 in the reference arm, and 243 in the vehicle placebo arm. Those endpoint denominators should be distinguished from total trial enrollment; the record does not explain all differences between enrollment and the endpoint population.

The study start date was 2017-03-10, primary completion was 2017-11-15, and status is completed. The record's summary says the study compared the test drug with the marketed reference drug and sought to establish that both were better than placebo. The posted primary analyses report superiority comparisons against placebo, while the title also names therapeutic equivalence as an objective. The original protocol's detailed statistical plan, including how equivalence was to be assessed, is not reported in the registry record.

The primary endpoint is a binary participant-level outcome: each participant either met the clinical-cure definition or did not. The results give counts of participants with clinical cure rather than a continuous measurement. No secondary outcomes are posted 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: All subjects who met the entry criteria were randomized and enrolled into the study.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Modified Intent-to-Treat (mITT) population: All randomized subjects who received study treatment and returned for at least one post-baseline visit, but excluding those who demonstrate a positive test result for other concomitant vaginal or cervical infections at baseline (e.g. C. trachomatis, N. gonorrhoeae) or who have a baseline Nugent score\<4.

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 Clinical Cure, measured over Day 21-30.

Clinical cure was defined as resolution of clinical signs and symptoms, assessed at Day 21-30. The outcome is reported as a count of participants, so it indicates how many participants in each endpoint group met the cure definition during that assessment timeframe. The registry record does not provide a more detailed operational definition of resolution or list the individual signs and symptoms used to determine cure.

General: A binary endpoint is useful when the clinical question is whether a defined state or event occurred for each participant. It can be summarized as counts and as a proportion of the assessed group. Comparisons of these proportions describe differences in the frequency of the endpoint between groups, while uncertainty intervals help show how much imprecision remains.

For this trial, the endpoint directly reflects the stated clinical question about resolution of signs and symptoms. Its interpretation depends on the specified assessment window and on which participants were included in the modified Intent-to-Treat population. The registry gives the population definition, but does not report further details about endpoint adjudication or the handling of participants whose clinical status could not be assessed.

7. Results Posted in the Registry

• Test Drug: 55.1% (135/245) versus 52.7% (129/245) on control; risk difference 2.4 percentage points (95% CI -6.4 to 11.3), risk ratio 1.05 (0.89–1.23); unadjusted two-proportion z-test p = 0.587.
• Vehicle Placebo Gel: 30.5% (74/243) versus 52.7% (129/245) on control; risk difference -22.2 percentage points (95% CI -30.7 to -13.7), risk ratio 0.58 (0.46–0.72); unadjusted two-proportion z-test p = <0.001.

8. Statistical Interpretation

9. Statistical Methods Explained

What does the endpoint count represent?

The primary outcome is binary and is reported as counts of participants meeting the clinical-cure definition. The recalculation uses the endpoint counts and denominators in each comparison to estimate a risk difference, meaning the difference between the two groups' observed proportions. In the GDC-229 versus reference comparison, the contrast is GDC-229 minus reference. In the vehicle placebo versus reference comparison, it is vehicle placebo minus reference. The sign therefore depends on the order of the groups, and the comparison direction should be kept in mind when interpreting a positive or negative estimate.

General: A risk difference is an absolute contrast between event proportions. It is distinct from a risk ratio, which compares one proportion with another as a relative measure. Both describe the same binary outcome from different perspectives, and neither alone establishes causation outside the design and analysis assumptions.

What was reported by the sponsor?

The registry lists corrected chi-squared tests for GDC-229 versus vehicle placebo and reference metronidazole gel versus vehicle placebo. Both are labeled superiority analyses, and both have p < 0.001. The registry does not give a sponsor-reported estimate or confidence interval for either comparison, and it does not list a direct sponsor-reported comparison of GDC-229 with reference drug. The study title's equivalence objective should not be confused with a posted equivalence test result.

General: A chi-squared test evaluates whether observed categorical outcome distributions differ between groups, subject to the method's assumptions. A corrected version modifies the test procedure. The registry identifies the corrected chi-squared method but does not describe the correction in more detail. Recalculated estimates based on posted group summaries can help describe the data, but they do not replace the sponsor's analysis when a discrepancy is flagged.

Why do confidence intervals matter?

A point estimate gives one summary of the observed contrast, but it does not show its precision. The recalculated GDC-229 versus reference interval includes zero and covers negative as well as positive differences. That means the summary data do not pin down the direction of the contrast with high precision. It does not demonstrate equivalence. To make an equivalence conclusion, the relevant interval must be interpreted against an appropriate prespecified equivalence margin, which is not reported in the registry record.

General: A confidence interval is generated by a method that, over repeated studies under its assumptions, would cover the underlying parameter at its stated confidence level at the corresponding frequency. For an individual study, it is best read as a range of values reasonably compatible with the data and method, not as a probability statement about the fixed true value.

What does the sample-size illustration describe?

The planning figures use the observed effect as a starting point and show scenarios for detecting the full observed effect or smaller fractions of it. They are illustrations, not the trial's documented original design calculations. The registry record does not report the original sample-size justification or assumptions. General: Power is the probability that a specified study design would detect an effect of a chosen size under the assumed model and significance threshold. It depends on the effect being targeted, the outcome variability, allocation, and analysis approach.

For a binary endpoint, over-dispersion can matter if outcomes vary more than the assumed model allows. General: When over-dispersion is present, ordinary model-based uncertainty can be understated, making confidence intervals too narrow and tests too optimistic unless the analysis accounts for it. The registry record does not report an over-dispersion assessment, and the posted binary counts do not establish that such an issue was present.

Which analysis details remain unknown?

The record identifies the modified Intent-to-Treat population and gives its inclusion and exclusion description, but the available definition is truncated. Handling of other missing outcomes, protocol deviations, and analysis exclusions is not reported in the registry record. Neither a multiplicity strategy nor interim analyses are reported. These omissions limit assessment of whether the posted tests account for all planned comparisons and analysis decisions.

10. Confidence Intervals and Precision

The recalculated GDC-229 versus reference risk difference is 0.025, with a 95% confidence interval from -0.064 to 0.113. Since zero lies within this interval, the data are compatible with a small negative contrast, no contrast, or a positive contrast within the interval's limits under the recalculation method. The interval is not evidence of equivalence on its own: equivalence depends on a justified margin, which is not reported in the registry record.

For vehicle placebo versus reference, the recalculated risk difference is -0.222, with a 95% confidence interval from -0.307 to -0.137. This interval lies below zero and indicates a lower observed cure proportion in the vehicle placebo group than in the reference group under this contrast definition. The sponsor's corrected chi-squared test for reference versus placebo reports p < 0.001. The registry does not report sponsor confidence intervals for these comparisons.

General: Confidence intervals communicate both the direction and precision of an estimate. A wider interval indicates greater uncertainty about the size of the contrast; an interval spanning zero for a risk difference does not establish that the groups are identical. Conversely, an interval that excludes zero is not a measure of clinical importance. Interpretation also depends on the prespecified analysis and the population represented by the data.

11. 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 planning figures are participants per arm for these binary-outcome comparisons, not total events. The planning assumptions stated in the registry-based calculation are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. The effect to detect is set to the stated fraction of the observed effect. The original trial's own design assumptions and sample-size justification are not reported in the registry record.

The 80% power column describes the illustrative number of participants per arm needed for a design to have an 80% chance of detecting the selected effect under those assumptions. The 90% power column describes the corresponding number for a 90% chance. These are planning targets under a model, not guarantees about a future study's result. The accompanying curve illustrates how the planning requirement changes as the target effect changes; smaller effects require more participants because they are harder to distinguish from random variation.

Values based on the full observed effect are particularly dependent on the observed estimate, which may itself be uncertain. Scenarios based on smaller fractions illustrate the consequences of planning for a more modest effect. No sample-size illustration can establish what effect a later study will observe, and these figures are not observed power. The registry record does not report whether the actual trial was designed using these assumptions or whether its original design used a different endpoint, margin, or method.

12. 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 = 0.527; treatment proportion p₁ = 0.551; average p̄ = 0.539.
2. First term: z_1−α/2 × √[2p̄(1−p̄)] = 1.96 × 0.7050 = 1.3817.
3. Second term: z_1−β × √[p₀(1−p₀) + p₁(1−p₁)] = 0.842 × 0.7048 = 0.5931.
4. n = (1.3817 + 0.5931)² / (0.0245)² = 6,502.9, rounded up to 6,503 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 <- 0.526530612244898                          # control proportion
p1 <- 0.551020408163265                          # 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(3091777)
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 = 6,503 (80% power), 8,706 (90% power)
 75% of observed effect: n per arm = 11,573 (80% power), 15,492 (90% power)
 50% of observed effect: n per arm = 26,062 (80% power), 34,888 (90% power)
power.prop.test: 6502.9 
Simulated power at n = 6,503 per arm: 0.797 (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.

13. Analyses Reported by the Sponsor

The registry reports the sponsor's corrected chi-squared superiority analyses for GDC-229 versus vehicle placebo and reference metronidazole gel versus vehicle placebo, each with p < 0.001. It does not report a sponsor estimate, confidence interval, or p-value for a direct GDC-229 versus reference comparison. In particular, the posted sponsor analyses do not constitute a reported test of therapeutic equivalence between the active-drug groups.

The recalculated comparisons use the posted counts and denominators to obtain simple unadjusted risk differences and associated uncertainty. The registry record flags a discrepancy between sponsor analyses and recalculations. Where a discrepancy applies, the sponsor's own analysis takes precedence; the recalculation is a separate descriptive check and should not be presented as the sponsor's result. The record does not provide enough detail to identify the specific source of every discrepancy or to reconcile different methods.

The sponsor's placebo comparisons and the unadjusted recalculations answer related but not necessarily identical analytical questions. A comparison showing a difference from placebo is not the same as a direct comparison between the two active treatments, and it does not by itself establish equivalence.

14. Multiplicity, Missing Data and Analysis Populations

The primary outcome record contains two sponsor-reported superiority comparisons, each against vehicle placebo. The registry record does not report whether these comparisons were part of a prespecified testing sequence, whether multiplicity was controlled, or what family of hypotheses was considered. Without that information, the reported p-values can be described as posted, but the overall error rate across planned tests cannot be assessed from the registry record.

The record identifies the primary analysis population as modified Intent-to-Treat. It includes randomized participants who received study treatment and returned for at least one post-baseline visit, excluding those with positive baseline tests for other concomitant vaginal or cervical infections. The examples in the definition are truncated, and further exclusion details are not reported in the registry record. The outcome denominators are 245 for GDC-229, 245 for reference drug, and 243 for vehicle placebo.

How missing primary outcome measurements were handled, whether any imputation was used, and whether additional populations such as per-protocol or safety populations were used for efficacy analyses are not reported in the registry record. The registry also does not describe sensitivity analyses for missing outcomes, protocol deviations, or participants excluded from the mITT population. Those details matter because different analysis populations and missing-data methods can produce different estimates, especially in a study evaluating equivalence.

No secondary outcome measures are posted. Interim analyses, stopping rules, and the original statistical analysis plan are not reported in the registry record. These gaps prevent a full assessment of the prespecified analysis strategy and of how the posted primary comparisons relate to all analyses planned by the sponsor.

15. 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 1% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: intussusception (0.3%), cholecystitis acute (0.3%), sickle cell anaemia (0.3%), supraventricular tachycardia (0.3%).

The serious adverse event counts are 2 among 287 participants at risk in the GDC-229 group, 0 among 285 in the reference group, and 2 among 292 in the vehicle placebo group. Deaths are listed as 0 in each group. The registry also lists 44 participants with other adverse events in the GDC-229 group, 37 in the reference group, and 35 in the vehicle placebo group; these are separate from the serious-event counts.

The record lists Intussusception, Cholecystitis Acute, Sickle Cell Anaemia, and Supraventricular Tachycardia among the top serious adverse event terms, each reported as 0.3%. The record does not specify which group is associated with each listed term in the provided adverse event summary, so group attribution should not be inferred from that list.

These counts describe events reported during the trial and are not, by themselves, evidence that a study drug caused an event or that the groups had comparable safety. The registry entry does not provide further details about event timing, relatedness assessment, severity beyond the serious classification, or statistical comparisons of adverse events. Small counts also leave substantial uncertainty about the underlying frequency of rare events.

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

The posted results show clinical-cure counts of 135 of 245 for GDC-229, 129 of 245 for reference metronidazole gel, and 74 of 243 for vehicle placebo at Day 21-30. The sponsor reports superiority tests against placebo for both active-drug arms, with p < 0.001 for each. Those are the analyses the registry reports. The simple recalculation for GDC-229 versus reference yields an interval that includes zero, but that alone neither proves equality nor establishes equivalence.

Equivalence is a specific statistical claim: the uncertainty around the direct between-treatment effect must be sufficiently limited relative to a prespecified margin that defines differences considered acceptably small. The registry record does not report that margin or a sponsor-reported direct equivalence analysis. Therefore the title's use of “equivalence” should be read as the study's stated objective, not as a conclusion demonstrated by the results posted in the registry.

General: Statistical significance and clinical importance are different concepts. A small p-value addresses compatibility with a null hypothesis under an analysis; it does not tell readers whether an effect is important to patients or whether a design met an equivalence criterion. A non-significant result for a direct contrast also does not establish that two treatments are the same.

The findings are limited to the endpoint, timeframe, and population described in the registry. The primary endpoint concerns resolution of clinical signs and symptoms at Day 21-30, and the analysis population excludes some participants with baseline concomitant infections. The registry does not provide complete detail about missing outcomes, the full protocol, or the full statistical analysis plan. No medical conclusion or treatment recommendation follows from the posted summary alone.

17. Why This Trial Matters Statistically

This trial illustrates why study objectives, analysis methods, and posted results need to be distinguished. The official title names therapeutic equivalence, while the sponsor-reported primary analyses listed in the registry are superiority comparisons of each active-drug group with vehicle placebo. These comparisons address whether each group differs from placebo under the reported tests; they do not directly answer whether GDC-229 and reference metronidazole gel are equivalent.

The example also shows the importance of reporting an equivalence margin and an uncertainty interval for the direct contrast. A non-significant recalculated comparison cannot substitute for a planned equivalence analysis. Without the margin and sponsor's direct analysis, readers cannot assess whether the evidence meets the study's stated equivalence objective.

The record includes both sponsor-reported tests and recalculations from group summaries, and it flags a discrepancy. Presenting the two distinctly avoids treating an unadjusted recalculation as if it were the sponsor's prespecified analysis. It also underscores the value of access to detailed methods, analysis populations, missing-data decisions, and multiplicity plans when interpreting confirmatory trials.

Finally, the sample-size illustrations emphasize that planning depends on the effect a study is intended to detect. Smaller target effects require larger samples under the stated planning assumptions. Such illustrations are not a substitute for the original design rationale, which is not reported in the registry record.

18. 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 registry entry does not report the original sample-size justification, equivalence margin, full protocol, or complete statistical analysis plan. Consequently, the intended equivalence criterion and its relationship to the posted analyses cannot be evaluated. The record lists sponsor-reported superiority tests against placebo but no sponsor-reported direct test of GDC-229 against reference metronidazole gel.

The mITT population definition is incomplete in the registry record: the description ends during an example of concomitant infections. The handling of missing endpoint measurements, protocol deviations, and exclusions beyond the stated definition is not reported. This limits assessment of who contributed to each analysis and whether missingness could have affected the estimated contrasts.

The sponsor's reported analyses provide p-values but no effect estimates or confidence intervals. Recalculations from the group counts are unadjusted, and the record flags a discrepancy. The exact source and scope of the discrepancy are not described. The sponsor's own analysis takes precedence where results disagree, but the available summary does not allow a full reconciliation.

The registry does not report multiplicity adjustment, interim analyses, or detailed adverse-event methods. It lists serious event counts and selected terms, but does not provide the group attribution for each listed term in the summary or fuller information on relatedness and timing. These limitations constrain interpretation of both efficacy and safety; they do not establish that a particular bias or error occurred.

19. Related Tutorials

20. Related Statistical Calculators

21. Sources

• ClinicalTrials.gov: Study Evaluating the Equivalence of GDC-229 and Metronidazole Vaginal Gel 0.75% in the Treatment of Bacterial Vaginosis — NCT03091777.
• Registry record last updated 2020-01-07; status last verified 2019-09.
• Results first posted 2019-10-14.
• 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.

22. Record Summary

NCT03091777 is a completed phase 3 randomized parallel-group trial of GDC-229, reference metronidazole vaginal gel 0.75%, and vehicle placebo for bacterial vaginosis. The registry lists enrollment of 871 and a primary endpoint of clinical cure, defined as resolution of clinical signs and symptoms at Day 21-30 in a modified Intent-to-Treat population.

Posted cure counts were 135 of 245 for GDC-229, 129 of 245 for reference drug, and 74 of 243 for vehicle placebo. Sponsor-reported corrected chi-squared superiority comparisons against placebo each have p < 0.001. A direct sponsor-reported equivalence analysis is not reported in the registry record. The record flags a discrepancy with recalculations, so the sponsor's analysis takes precedence where results disagree.