Recurrent Bacterial Vaginosis and Vaginal Acidifying Gel Trial: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Recurrent Bacterial Vaginosis and Vaginal Acidifying Gel Trial (NCT00545181), a randomized na study. The primary outcome is Recurrent Bacterial Vaginosis.

2. Clinical Question

The clinical question is whether adding vaginal acidifying gel to metronidazole is associated with a different recorded recurrence outcome than metronidazole alone. The registry describes this as a randomized investigator-blinded controlled trial. Its stated hypothesis was that adding gel would decrease recurrence within 3 months; the posted results should be read as an estimate of the observed difference, not as confirmation of that hypothesis.

The population details are limited. The record identifies bacterial vaginosis and the trial title refers to recurrent disease, but it does not report the full inclusion criteria, participant characteristics, or the procedures used to establish the condition before treatment. The result therefore applies most directly to the participants included in this trial, whose fuller clinical profile is not reported in the registry record.

The endpoint used two diagnostic approaches: recurrence by Amsel's criteria or by Nugent's criteria. This makes the question about recurrence classified using either of the stated criteria, rather than a time-to-event comparison. The record does not describe a separate question about symptom severity, duration of recurrence, or subsequent treatment.

3. Trial Design

This was a randomized, parallel-group trial with single masking. Its official title describes it as investigator-blinded. The registry lists Indiana University as the sponsor and records the trial as completed. The recorded start was 2007-09 and the primary completion was 2009-07. Phase is listed as NA.

The two groups received metronidazole plus acidifying vaginal gel, or metronidazole alone. The control description specifies oral metronidazole antibiotic therapy alone. The gel is identified as RepHresh. The design therefore compares adding the gel to the antibiotic regimen with the antibiotic regimen without gel; it does not compare metronidazole with no treatment.

Enrollment is recorded as 55. The posted outcome counts are 29 participants in the metronidazole-plus-gel arm and 25 in the metronidazole-alone arm. The registry does not explain how enrollment relates to the outcome denominators, or report the analysis population used for the endpoint. Reasons for any difference between enrollment and those arm totals are not reported in the registry record.

The primary outcome was recurrence of bacterial vaginosis within 3 months, assessed using either Amsel's or Nugent's criteria. The registry posts one outcome measure and one analysis. It does not give details of allocation concealment, the identities of people masked, follow-up procedures, adherence, or the handling of participants with incomplete outcome information. These design details should not be inferred from the randomized and single-masking labels alone.

The registry record does not report the original sample-size justification, the trial's planned statistical analysis methods, or whether interim analyses were planned or performed. The planning figures associated with the observed result are separate illustrations and should not be interpreted as the trial's own design assumptions.

4. Participant Flow and Accounting

The registry reports how participants moved through each stage of the study. All counts below are exactly as posted; missing counts are not inferred from other stages.

Overall Study

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 Recurrent Bacterial Vaginosis, measured over 3 months.

The primary endpoint is the number of participants with recurrent bacterial vaginosis at 3 months. A recurrence could be identified by either Amsel's or Nugent's criteria. In the posted results, 7 participants in the metronidazole-plus-gel group and 7 in the metronidazole-alone group are recorded as having recurrence.

Amsel's criteria define bacterial vaginosis through a set of clinical and laboratory findings. The registry describes four findings: homogeneous gray-white vaginal discharge, vaginal pH >4.7, clue cells comprising at least 20% of vaginal epithelial cells on wet-prep microscopy, and a positive amine odor test after addition of 10% KOH. The criteria require 3 of these 4 findings.

Nugent's criteria use microscopy and bacterial scoring, including lactobacillus, gardnerella, and curved gram-variable rods. A score of at least 7 is described as indicative of bacterial vaginosis. The registry does not report how many recurrences were identified by each diagnostic approach, whether both were applied to every participant, or how disagreements between them were handled.

Using a defined diagnostic endpoint makes recurrence measurable for a comparison between randomized groups. Combining the two stated criteria also means the result reflects the registry's either-criteria definition, rather than a result reported separately for each diagnostic method. The timeframe is 3 months; the record does not provide a longer-term recurrence endpoint.

7. Results Posted in the Registry

• Metronidazole Alone: 28.0% (7/25) versus 24.1% (7/29) on control; risk difference 3.9 percentage points (95% CI -19.6 to 27.4), risk ratio 1.16 (0.47–2.86); unadjusted two-proportion z-test p = 0.747.

8. Statistical Interpretation

9. Statistical Methods Explained

For a binary endpoint, each participant is classified as having or not having the event during the specified follow-up. Here, the event is recurrent bacterial vaginosis by either Amsel's or Nugent's criteria within 3 months. The arm counts and denominators support estimates of the event proportion in each group and comparisons between those proportions.

What does the risk difference express?

The risk difference subtracts one group's event proportion from the other's. In this comparison, the reported calculation is metronidazole alone minus metronidazole plus gel. Its estimate is 0.039. A value of zero would correspond to equal observed proportions under this definition; positive or negative values indicate which group has the larger proportion. The 95% confidence interval is -0.196 to 0.274. Because it extends on both sides of zero, the data do not pin down the direction of the difference.

The risk difference is expressed on the probability scale. It can help describe the absolute separation between groups, but it is sensitive to the baseline event probability and the population being studied. It should be read together with the arm-specific proportions and the interval, not as a stand-alone measure of certainty.

How do relative measures differ?

The risk ratio compares the event proportions by division, while the odds ratio compares the odds of the event. Both are relative measures: a value of 1 corresponds to no relative difference. The recalculated risk ratio is 1.16, with interval 0.471 to 2.855; the recalculated odds ratio is 1.222, with interval 0.361 to 4.135. Each interval includes 1 and covers a broad range, so neither measure gives a precise estimate of the relative association.

Risk ratios and odds ratios are not interchangeable with risk differences. They describe the comparison on different scales, and the odds ratio can be less intuitive as an absolute description. With a binary event, it is useful to consider the group proportions alongside these relative estimates. No adjusted estimates are posted, so these recalculations are unadjusted comparisons based on the reported summary counts.

What does the p-value say?

The recalculated unadjusted p-value is 0.747. The registry reports a corrected chi-squared analysis with p = 0.75. A p-value describes how unusual data at least as discrepant as those observed would be under a specified model in which the groups have no difference, subject to the test assumptions. It is not a measure of the size of an effect, and it does not give the probability that a treatment works or that there is no effect.

The registry identifies the sponsor's method as a corrected chi-squared test. The full implementation details are not reported. The recalculation is a simple unadjusted comparison from the posted summary statistics; it is not an adjusted model and does not incorporate participant-level covariates. Since the p-values are close and the record does not flag a discrepancy, the recalculated result can be read alongside the sponsor analysis, without treating the two methods as identical.

Why do intervals matter?

A confidence interval conveys uncertainty around an estimated effect. In repeated use of a specified method under its assumptions, intervals constructed at the stated confidence level are intended to cover the true parameter at that rate. For an individual trial, the interval gives a range of values compatible with the data and method; it is not a guarantee that the true effect lies within that particular interval.

The wide intervals for the risk difference and relative measures indicate limited precision. They include values associated with differences in either direction as well as the no-difference value. A non-significant p-value and an interval that includes the null value do not demonstrate equivalence. Equivalence or non-inferiority requires a prespecified margin and an analysis designed for that question; such information is not reported in the registry record.

These methods address the numerical comparison of the posted outcome counts. They do not resolve missing-data concerns, establish that the endpoint was measured consistently, or account for features of the trial that are not reported. The analysis population, missing outcome handling, covariate adjustment, and any planned multiplicity procedures are not reported in the registry record.

10. Confidence Intervals and Precision

The risk difference estimate is 0.039, and its 95% confidence interval runs from -0.196 to 0.274. This interval is broad relative to the estimate and crosses zero. In practical statistical terms, the posted counts do not provide a narrowly bounded estimate of the absolute difference in recurrence proportions. The interval permits a difference in either direction, including no difference.

The relative estimates are also imprecise. The risk ratio is 1.16, with a confidence interval from 0.471 to 2.855; the odds ratio is 1.222, with a confidence interval from 0.361 to 4.135. Each interval includes 1, the value corresponding to no relative difference, and extends over a broad range. These estimates should not be interpreted as establishing a stable relative effect.

A confidence interval is a statistical expression of uncertainty based on the observed data and the method used. It is not a range of outcomes for individual participants, and it does not assign equal probability to every value within the interval. Its interpretation also depends on assumptions, including how the data were collected and analyzed. The registry does not report the method used for the recalculated intervals.

The limited number of posted outcome observations contributes to the uncertainty, but the interval—not a single p-value—best conveys how much the estimated difference is constrained by these data. A p-value of 0.747 does not make the estimate precise, and it does not establish equivalence. The record reports no equivalence margin or non-inferiority design. The result is therefore best described as an uncertain comparison with a wide range of compatible differences.

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 sample-size figures are planning illustrations based on the observed effect, not the trial's own original design assumptions. They give participants per arm for this binary endpoint. They are not total-event counts. The assumptions listed for the illustrations are a two-sided significance level of 0.05, equal allocation between the two arms, a normal approximation, and an effect to detect set at the stated fraction of the observed effect.

The 80% power column gives the number of participants per arm indicated by those assumptions for an 80% chance of detecting the specified effect if that effect were the true one and the assumptions held. The 90% power column gives the corresponding planning number for a 90% chance. Power is a property of a design under an assumed effect and analysis; it does not describe the probability that the observed trial result is true.

The rows represent planning for 100%, 75%, and 50% of the observed effect. For the 100% row, the figures are 2,028 participants per arm for 80% power and 2,714 participants per arm for 90% power. For the 75% row, they are 3,562 and 4,767 participants per arm, respectively. For the 50% row, they are 7,914 and 10,594 participants per arm, respectively. Smaller effects require more participants under the same planning assumptions, which is why the figures increase as the effect fraction decreases.

These figures are not observed power. They are conditional planning illustrations, and the original trial's design assumptions and sample-size justification are not reported in the registry record. The curve should be interpreted in the same way: it reflects how the planning requirement changes with the assumed effect, not a finding about the trial's actual power or its treatment comparison.

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 = 0.241; treatment proportion p₁ = 0.280; average p̄ = 0.261.
2. First term: z_1−α/2 × √[2p̄(1−p̄)] = 1.96 × 0.6209 = 1.2169.
3. Second term: z_1−β × √[p₀(1−p₀) + p₁(1−p₁)] = 0.842 × 0.6203 = 0.5220.
4. n = (1.2169 + 0.5220)² / (0.0386)² = 2,027.2, rounded up to 2,028 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 <- 0.241379310344828                          # control proportion
p1 <- 0.28                          # 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(545181)
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 = 2,028 (80% power), 2,714 (90% power)
 75% of observed effect: n per arm = 3,562 (80% power), 4,767 (90% power)
 50% of observed effect: n per arm = 7,914 (80% power), 10,594 (90% power)
power.prop.test: 2027.2 
Simulated power at n = 2,028 per arm: 0.804 (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 a corrected chi-squared analysis with p = 0.75. A simple unadjusted recalculation from the posted group counts gives p = 0.747. These values are close, and the record does not indicate a discrepancy between the sponsor's result and the recalculation.

The sponsor's analysis is the registry's reported test. The recalculated estimates add an unadjusted risk difference of 0.039, with a 95% confidence interval from -0.196 to 0.274, as well as unadjusted relative measures. They make the magnitude and uncertainty easier to describe, but they are not participant-level adjusted results and do not replace the sponsor's analysis.

The registry does not report the full details of the corrected chi-squared procedure or a sponsor-reported effect estimate and confidence interval. Differences in methods and rounding can affect reported values; the record does not permit a more detailed reconciliation. Both p-values indicate that the posted analysis does not provide strong evidence against a no-difference model, but neither is proof that the groups are equivalent.

14. Multiplicity, Missing Data and Analysis Populations

The registry posts one primary outcome measure and one analysis, with recurrence of bacterial vaginosis at 3 months as the endpoint. It reports no secondary posted outcomes. This describes what appears in the registry record; it does not establish that no other outcomes were collected or examined.

The record does not report whether there were multiple planned comparisons, whether any multiplicity adjustment was specified, or whether the posted analysis was selected from a larger set of analyses. A single posted analysis does not by itself establish that multiplicity was not relevant. The original statistical analysis plan and any protocol amendments are not reported in the registry record.

Information about missing outcome data is also not reported. The record does not state whether all randomized participants were assessed at 3 months, how participants with unavailable endpoint data were handled, or whether any imputation or sensitivity analysis was used. It does not identify an intention-to-treat, per-protocol, or other analysis population. The arm totals for the outcome are reported, but the relationship between those totals and the recorded enrollment is not explained.

These omissions matter because the interpretation of a randomized comparison depends not only on the event counts but also on which participants were included and how missing outcomes were treated. No conclusion about the direction or extent of bias from missing data can be drawn without that information. Likewise, it is not possible to verify whether the reported analysis followed a prespecified plan.

Interim analyses, stopping rules, and any monitoring plan are not reported in the registry record. They should not be inferred from the randomized design or the completed status. The information available supports description of the posted endpoint and its numerical comparison, but not a full audit of the analysis strategy.

15. Safety Results

The registry reports serious adverse events as affected participants divided by participants at risk.

Other (non-serious) adverse events

Participants with at least one non-serious event at or above the registry's 5% reporting threshold.

The registry records zero serious adverse events among 29 participants at risk in the metronidazole-plus-gel group and zero among 25 participants at risk in the metronidazole-alone group. It also records zero other adverse events in the corresponding groups. No serious adverse event terms are listed.

These figures describe what was recorded in the registry; they do not establish that adverse events could not occur. The number of participants represented is limited, and the record does not provide detailed definitions, collection procedures, observation periods for safety, or information about events that may not have been reported. Death information is not reported in the registry record.

Zero recorded events should therefore be read as an absence of events in the posted counts, not as proof of safety or as a comparison demonstrating equal safety. The record does not provide enough detail to assess the completeness of adverse-event ascertainment or to characterize individual event types beyond the absence of listed serious terms.

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

The posted results show 7 recorded recurrences in each group: 7 of 29 participants receiving metronidazole plus gel and 7 of 25 receiving metronidazole alone. The recalculated proportions are 0.241 and 0.28, respectively. The estimated risk difference is 0.039 for metronidazole alone minus the combination arm, but its 95% confidence interval ranges from -0.196 to 0.274.

The confidence interval includes zero and spans differences in both directions. The relative-effect intervals also include the no-difference value. Thus, the posted data do not identify a precise difference between the groups. They also do not demonstrate equivalence, because no equivalence margin or equivalence analysis is reported. A non-significant result is not evidence that two regimens have identical effects.

The registry's stated hypothesis was that adding acidifying gel would decrease recurrence within 3 months. The observed counts and their uncertainty do not provide a clear, precise estimate in support of a particular direction. The result should be described as an uncertain comparison based on the posted outcome counts, not as proof that the gel does or does not change recurrence.

The finding is specific to the endpoint definition and timeframe recorded: recurrence by either Amsel's or Nugent's criteria within 3 months. It does not establish results for longer follow-up, other measures of bacterial vaginosis, particular participant subgroups, or outcomes not posted in the registry. The record also lacks detail on participant characteristics, analysis population, and missing data, limiting assessment of how broadly the result applies.

The trial's randomized design is relevant to the comparison, but randomization alone does not answer questions about adherence, endpoint ascertainment, incomplete follow-up, or the precision of the estimate. Those details are not reported. The most defensible statistical conclusion is that the posted comparison is inconclusive about the direction and size of a difference, with considerable uncertainty around the estimates.

17. Why This Trial Matters Statistically

This trial illustrates why event counts, effect estimates, and uncertainty need to be considered together. The groups each had 7 recorded recurrences, yet the denominators differ. The resulting proportions and the risk difference describe the comparison more directly than event counts alone, while the confidence interval shows how uncertain that estimate remains.

It also illustrates the distinction between failing to detect a difference and demonstrating no difference. The sponsor's reported p-value is 0.75, and the unadjusted recalculated p-value is 0.747. Neither value establishes equivalence. The broad intervals for the absolute and relative estimates show that the posted data leave substantial uncertainty about the size and direction of a possible difference.

Planning requirements depend on the assumed effect size. The planning illustrations show that smaller target effects require more participants under the stated assumptions. This is a useful statistical distinction: a study can be randomized and still have an imprecise estimate when the amount of information available is limited. The illustrations are not the original trial plan and do not measure the trial's observed power.

The example also emphasizes how registry reporting shapes interpretation. The endpoint definition is described, but the record does not report detailed handling of missing outcomes, analysis populations, or the original statistical plan. Those omissions do not invalidate the posted counts, but they restrict how fully readers can assess the analysis and its robustness.

For biostatisticians and clinical researchers, the central lesson is to avoid treating a p-value as a complete summary. The estimated absolute difference, relative measures, confidence intervals, endpoint definition, and reporting limitations each contribute to an appropriately cautious reading of this randomized comparison.

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 record provides limited detail about the participants. It names bacterial vaginosis and frames the study around recurrence, but full eligibility criteria and baseline participant characteristics are not reported. The result therefore cannot be examined across clinical subgroups or related to the characteristics of those enrolled.

The outcome is defined as recurrence by either Amsel's or Nugent's criteria, but the record does not report how many events came from each method, whether both methods were applied consistently, or how discordant assessments were resolved. The timing and completeness of endpoint assessment beyond the 3-month timeframe are also not described.

Although enrollment is listed as 55, the posted arm outcome totals are 29 and 25. The registry does not explain the relationship between enrollment and these outcome denominators. It does not report the analysis population, missing outcome counts, reasons for missingness, or methods for handling unavailable assessments. These details limit assessment of attrition and the population represented by the estimates.

The study is recorded as randomized and single-masked, and its title describes investigator blinding. Allocation concealment and the specific roles that were masked are not reported. The record also does not describe adherence, treatment exposure, or whether participants received additional interventions during follow-up.

The posted comparison has wide confidence intervals, and the planning illustration is based on the observed effect rather than the original study design. The trial's sample-size justification, original statistical plan, interim analyses, and any multiplicity procedures are not reported in the registry record. Finally, the safety entries are limited to event counts and do not provide detailed ascertainment methods. These limitations constrain precision and interpretation without changing what the registry reports.

19. Related Tutorials

20. Related Statistical Calculators

21. Sources

• ClinicalTrials.gov: Recurrent Bacterial Vaginosis and Vaginal Acidifying Gel Trial — NCT00545181.
• Registry record last updated 2010-01-12; status last verified 2010-01.
• Results first posted 2010-01-07.
• 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

NCT00545181 is a completed randomized, parallel-group, single-masked trial sponsored by Indiana University. It compared metronidazole plus vaginal acidifying gel with metronidazole alone for recurrent bacterial vaginosis. Enrollment is recorded as 55; the outcome arm totals are 29 and 25.

The primary endpoint was recurrence within 3 months by either Amsel's or Nugent's criteria. The registry records 7 events in each arm. The recalculated risk difference is 0.039, with a 95% confidence interval from -0.196 to 0.274; the registry reports a corrected chi-squared p-value of 0.75.

The estimates are imprecise and do not establish a clear difference or equivalence. Details of the analysis population, missing data, original sample-size justification, and full statistical plan are not reported in the registry record.