HPV Vaccine Therapy in Reducing High-Grade Cervical Lesions in Patients With HIV and HPV: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for HPV Vaccine Therapy in Reducing High-Grade Cervical Lesions in Patients With HIV and HPV (NCT03284866), a randomized phase 3 study. The primary outcome is Occurrence of Cervical High-grade Squamous Intraepithelial Lesions (HSIL) or Cervical Cancer.

2. Clinical Question

The question is about the recorded occurrence of cervical high-grade squamous intraepithelial lesions (HSIL) or cervical cancer during a defined follow-up period. It is not, by itself, a question about every possible effect of vaccination, nor does it address outcomes outside the stated endpoint and follow-up. The registry lists additional outcomes, including later cervical HSIL or invasive cancer, HPV clearance, vaccine-type HPV infections, cervical cytology, and vulvar HSIL or cancer. Those outcomes provide other aspects of the study, but their presence does not replace the primary endpoint or supply its missing comparison.

The study is described as a phase III randomized trial with a parallel design and double masking. Its enrollment was 535. The registry status is completed. Those design features describe how the trial was organized; they do not establish what the primary outcome showed. The posted record identifies a planned rate-based comparison, but does not include the numerical primary results needed to assess the direction, size, or precision of the difference between arms.

3. Trial Design

This was a randomized, double-masked, parallel-group phase III trial comparing recombinant human papillomavirus nonavalent vaccine with saline. The registry reports enrollment of 535. Participants were assigned to Arm I, the recombinant HPV 9-valent vaccine arm, or Arm II, the saline arm. The trial is listed as completed. The start date is 2019-07-31, and the primary completion date is 2024-11-26.

The vaccine arm description names Gardasil 9 and specifies intramuscular administration at baseline, 4, and 26 weeks, in the absence of disease progression or unacceptable toxicity. The saline arm description specifies intramuscular placebo vaccine at the same scheduled visits. Both arm descriptions include sample collection for laboratory biomarker analysis. The registry lists laboratory biomarker analysis as an intervention of type “other,” alongside the vaccine and saline interventions.

The official title frames the study as vaccination to reduce cervical HSIL among HIV-infected women participating in an HPV test-and-treat program. The brief summary describes an interest in reducing high-grade lesions and notes that lesions may develop or return after removal. That context explains the clinical subject of the trial; it is not evidence that either outcome occurred at a particular rate in this study.

The primary endpoint covers the period after the week 4 study visit to week 52 post-randomization. The registry says analyses were performed separately on the intention-to-treat and per-protocol sets. It does not report, in the record, the detailed definitions of those sets, how participants were assigned to them, the randomization procedure, concealment methods, or the procedures used to maintain masking. Detailed eligibility criteria and the sample-size justification are also not reported in the registry record.

The registry names seven secondary outcome areas and lists 14 outcome measures and 9 analyses. These include later cervical HSIL or invasive cancer, predictors of sustained absence of cervical HSIL and HPV clearance, incident vaccine-type HPV infections, abnormal cervical cytology, and vulvar outcomes. The number of listed outcomes and analyses does not show which outcomes were statistically significant, nor does it describe how multiplicity was handled. Specific numerical results for these secondary outcomes are not reported in the registry record.

4. Participant Flow and Accounting

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

Before assignment: 1256 patients underwent screening for determination of eligibility for the study. 535 patients were identified as eligible and enrolled

Overall Study

Analysis population for the primary outcome, as defined in the registry: Analyses were performed on intention-to-treat and per protocol set separately.

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.

Population: All randomized participants

6. Primary Endpoint

The registry lists the primary outcome as Occurrence of Cervical High-grade Squamous Intraepithelial Lesions (HSIL) or Cervical Cancer, measured over After week 4 study visit to week 52 post-randomization.

The primary endpoint is the occurrence of cervical high-grade squamous intraepithelial lesions (HSIL) or cervical cancer from after the week 4 study visit through week 52 post-randomization. The registry specifies the unit as events per 100 person-years. It describes estimating an event rate for each arm and comparing the rates between the vaccine and saline groups.

Person-time rates express events in relation to the time participants contribute under observation. In general, this can be useful when participants contribute different lengths of follow-up: the rate accounts for exposure time rather than treating every participant as having an identical observation period. A rate is not the same quantity as the proportion of participants who experienced an event. Interpreting the endpoint therefore requires attention to its stated unit and observation window.

The combined endpoint includes cervical HSIL or cervical cancer. It is an event-based measure, not a measure of lesion severity change, time lived without disease, or HPV clearance alone. The registry also describes a time-to-event analysis from week 4 to week 52, which uses event timing as well as whether an event was recorded. These analyses address related but distinct summaries of the endpoint.

The record says point estimates and 95% Poisson confidence intervals will be used for the arm-specific rates, with Poisson regression to compare event rates. It also describes Kaplan-Meier summaries and a log-rank comparison for time to event. These methods identify how the endpoint was to be analyzed, but the arm-specific estimates, intervals, and comparison results are not reported in the registry record. Without those values, the posted information cannot establish the observed event rate or how precisely the groups differed.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The registry describes several complementary ways to summarize the primary endpoint. These methods answer related questions, but they use different statistical summaries. The record specifies methods without supplying the resulting estimates or tests.

What does a person-time rate describe?

The endpoint is measured in events per 100 person-years. In general, a person-time rate relates the number of recorded events to the total time participants contribute under observation. This approach can accommodate differing follow-up durations, rather than requiring every participant to contribute an identical observation period. It does not directly state the probability that an individual participant will experience an event. A rate and a participant-level risk are different quantities and should not be treated as interchangeable.

The registry says the rate will be estimated separately for the vaccine and placebo arms, with a point estimate and 95% Poisson confidence interval for each. A point estimate is the single summary value calculated from the observed data. In general, a confidence interval communicates uncertainty around that estimate under the method used. It gives a range of values compatible with the data and model at its stated confidence level; it is not a guarantee that the true rate lies within the interval. The actual arm-specific rates and intervals are not reported in the registry record.

Why use Poisson regression?

The registry identifies Poisson regression as the method for comparing the arms with respect to event rate. In general, Poisson models are used for counts of events, often with an exposure-time term when the question concerns rates. The model can estimate a between-arm rate comparison while accounting for the amount of person-time represented in the data. The registry does not report the specific model specification, any covariates, diagnostics, or the resulting comparison. These details should not be inferred from the name of the method alone.

Count data may show more variability than a simple Poisson model expects, a feature often called over-dispersion. In general, over-dispersion can make model-based uncertainty estimates too small if it is not appropriately addressed. Whether it was assessed, whether it was present, and whether any adjustment was used are not reported in the registry record. This is a general methodological consideration, not a claim about the trial’s data.

How does time-to-event analysis differ?

The registry also says time to event from week 4 to week 52 will be described with the Kaplan-Meier method for each arm and compared using the log-rank test. In general, Kaplan-Meier estimates describe the event-free experience over time while allowing for participants whose follow-up ends before an event is observed. Such participants are commonly described as censored at their last relevant observation, but the specific censoring rules used in this trial are not reported in the registry record.

The log-rank test is a general method for comparing time-to-event patterns between groups. It evaluates the evidence for a difference across the follow-up period rather than reporting only an event rate per unit of person-time. It does not, by itself, quantify the size of a treatment effect. A full interpretation would require the underlying curves or other estimates and their uncertainty, none of which are reported in the registry record.

The rate and time-to-event approaches can complement one another because they summarize the event experience differently. They may not yield identical numerical summaries, and their interpretation depends on the data and model assumptions. The record does not provide enough results to compare the methods’ findings or determine whether they supported a consistent conclusion.

What do the analysis populations mean?

The registry states that analyses were performed separately in the intention-to-treat and per-protocol sets. In general, an intention-to-treat analysis seeks to compare participants according to their randomized assignment, while a per-protocol analysis focuses on participants meeting specified protocol conditions. These concepts are broad; the exact rules for defining either population in this trial are not reported in the registry record. The record also does not provide the population-specific outcome estimates or explain how deviations from the assigned regimen were handled.

Results from these populations can address different aspects of a trial. Differences between them may arise from the population definitions and the participants included, not necessarily from a distinct underlying effect. For that reason, their definitions and results should be read together. Here, the record confirms separate analyses but does not enable that comparison.

How should the planning illustrations be read?

The sample-size assumptions describe planning illustrations computed from the observed effect, not the trial’s own design. They specify a two-sided significance level of 0.05, equal allocation between arms, and a normal approximation, with the effect to detect set as a stated fraction of the observed effect. For time-to-event tables, the planning quantity is the total number of events using the Schoenfeld approach; other tables give participants per arm.

In general, such calculations explore how a hypothetical design performs under stated assumptions. They are not observed power, do not establish that the trial was adequately planned, and do not change the interpretation of the trial’s observed results. The record does not report the trial’s own sample-size justification or the assumptions used in its original design. The illustrations should therefore be treated as planning context, not as a retrospective verdict about the study.

10. Multiplicity, Missing Data and Analysis Populations

The registry lists 14 outcome measures and 9 analyses, with secondary outcomes covering several clinical and virologic areas. These include occurrence of cervical HSIL or invasive cancer from weeks 52–104, predictors of sustained absence of cervical HSIL and HPV clearance, incident cervical vaccine-type HPV infections, abnormal cervical cytology, and prevalent or incident vulvar HSIL or cancer. The list indicates that the study addressed more than the primary endpoint, but it does not describe which comparisons formed a statistical family or how results were prioritized.

Multiplicity arises in general when several outcomes, time points, groups, or analyses are tested. Conducting many tests can increase the chance of finding an apparently notable result by chance. Whether and how that concern was addressed depends on the prespecified analysis plan and any adjustment strategy. A multiplicity adjustment, hierarchical testing plan, or other strategy is not reported in the registry record. The listed number of outcomes and analyses is not sufficient to infer that an adjustment was or was not used.

Missing-data handling is also not reported in the registry record. The record does not state how missing endpoint assessments, incomplete follow-up, or loss to follow-up were handled, or whether sensitivity analyses were conducted. For time-to-event analysis, specific censoring rules are not reported. These details matter because they describe how incomplete observations contribute to an analysis; their absence prevents assessment of the robustness of the reported methods.

The registry says the primary analyses were performed separately in intention-to-treat and per-protocol sets. It does not provide detailed definitions of those sets, the analysis population sizes, or their results. The labels alone do not establish how protocol deviations, discontinuations, or missing outcomes were handled. Numerical results for the listed secondary outcomes are also not reported in the registry record, so their pattern and precision cannot be.

11. Safety Results

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

Deaths

Other (non-serious) adverse events

Participants with at least one non-serious event at or above the registry's 0% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: neoplasms benign, malignant and unspecified (incl cysts and polyps) - other (3.0%), pregnancy loss (0.8%), anemia (0.4%), blood and lymphatic system disorders - other (0.4%), gastrointestinal disorders - other (0.4%).

The adverse-event record lists Arm I, recombinant HPV 9-valent vaccine, with 265 participants at risk in the serious-event and other-event entries, and Arm II, saline, with 270 at risk in each entry. It records serious adverse events affecting 16 participants in the vaccine arm and 20 in the saline arm. Other adverse events are recorded for 20 participants in the vaccine arm and 22 in the saline arm. The listed deaths are 4 and 3, respectively.

These are reported counts, not evidence that an event was caused by the assigned vaccine or placebo. The registry figures alone do not establish attribution, timing, duration, severity beyond the table category, or whether an individual experienced more than one event. The table’s at-risk counts are not necessarily a description of follow-up time or exposure to each intervention.

The registry also lists serious-event terms and percentages, including “NEOPLASMS BENIGN, MALIGNANT AND UNSPECIFIED (INCL CYSTS AND POLYPS) - OTHER” at 3.0%, “PREGNANCY LOSS” at 0.8%, and “ANEMIA” at 0.4%. It additionally lists “BLOOD AND LYMPHATIC SYSTEM DISORDERS - OTHER,” “GASTROINTESTINAL DISORDERS - OTHER,” and “SEPSIS,” each at 0.4%. These terms and percentages describe entries in the safety record; they do not establish a treatment-related pattern or comparative risk. Interpretation requires the detailed event definitions, denominators, and attribution information, which are not reported in the registry record.

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

The registry establishes that this was a randomized, double-masked, parallel-group phase III trial comparing a nonavalent HPV vaccine with saline among women described as participating in an HPV test-and-treat program. It specifies a primary endpoint of cervical HSIL or cancer occurrence over a defined period and identifies rate-based and time-to-event methods. It also records participant counts for adverse-event categories and lists secondary outcome areas.

What the record does not establish is the result of the primary comparison. Arm-specific event rates, 95% confidence intervals, Poisson regression results, Kaplan-Meier estimates, and log-rank results are not reported in the registry record. The trial’s completed status does not supply those missing estimates. Without them, this page cannot describe the observed direction or magnitude of the primary comparison or its statistical precision.

In general, even a reported difference between arms would need to be interpreted in context. Statistical uncertainty, endpoint definition, follow-up, and the analysis population all affect what a result can support. A confidence interval can help show precision, while the endpoint’s event-rate unit must not be confused with a participant-level probability. The actual numerical results needed for those judgments are not included.

The listed safety counts likewise do not establish causation. Counts of serious events, other events, and deaths are descriptive entries in the registry record. They do not by themselves show whether an event was related to an intervention, and they do not summarize every aspect of safety experience.

Planning illustrations based on an observed effect, if displayed elsewhere on the page, should not be read as observed power or as proof that the original design was justified. They are calculations under stated assumptions. The original sample-size rationale is not reported in the registry record. Overall, the available record supports a description of the trial framework and analysis plan, not a conclusion about whether the intervention changed the primary outcome.

13. Why This Trial Matters Statistically

This trial is statistically instructive because its primary endpoint is a rate rather than a simple participant-level percentage. Rates relate event counts to person-time, and the registry specifies Poisson confidence intervals and Poisson regression for the arm comparison. That makes clear why endpoint units and exposure time matter when reading a result. A rate should not be interpreted as though it were automatically the proportion of participants with an event.

The registry also describes a time-to-event analysis using Kaplan-Meier summaries and a log-rank test. Considering both a rate comparison and an analysis of event timing can show how different statistical summaries address the same broad clinical endpoint. But methods alone are not findings: without the estimates, intervals, curves, or test results, readers cannot determine what those analyses showed.

The separation of intention-to-treat and per-protocol analyses is another useful point for statistical reading. In general, the populations address different analysis questions, and their definitions affect interpretation. The registry confirms that both were analyzed separately, but does not provide the definitions or results needed to compare them. This illustrates why naming an analysis population is not a substitute for explaining who was included.

The trial also lists a range of secondary outcomes. That breadth matters statistically because multiple outcomes and analyses raise questions about prioritization and multiplicity. The record gives counts and names, but not the plan for controlling or interpreting multiple comparisons. Readers should not infer a strategy from the outcome list alone.

Finally, the distinction between observed trial findings and planning illustrations is essential. Sample-size calculations based on an observed effect can explore hypothetical design requirements, but they are not observed power and do not recreate the trial’s original rationale. The registry record is a useful example of how design and analysis methods can be described more fully than the numerical findings needed to judge the primary result.

14. Limitations

• The posted summary statistics are the only data these results rest on; patient-level data, full protocols and statistical analysis plans are not available.
• Recalculated results are simple, unadjusted calculations from posted summaries. They do not replace the sponsor's prespecified analysis.

The central limitation for interpreting this trial from the registry record is that the primary numerical results are not reported. The endpoint definition and statistical methods are described, but the arm-specific rates, confidence intervals, and between-arm comparisons are absent. Consequently, no conclusion about the observed primary effect or its precision can be drawn from this record.

Important operational details are also missing. The record does not report detailed eligibility criteria, the randomization procedure, allocation concealment, or the practical procedures used to maintain masking. It does not give the exact definitions of the intention-to-treat and per-protocol populations, nor does it report their sizes or results. These omissions limit assessment of how the trial design was implemented and how the analysis populations were constructed.

The registry does not report the detailed handling of missing observations, incomplete follow-up, censoring, or sensitivity analyses. It also does not specify the Poisson model structure, whether covariates were used, or whether model assumptions were assessed. Such details would help readers evaluate the analyses, but they cannot be reconstructed from the method names alone.

There are further limits for interpreting the wider outcome set. The record lists secondary outcomes, 14 outcome measures, and 9 analyses, but does not report their numerical results or a multiplicity strategy. The safety entries list counts and selected event terms, but do not provide enough detail to establish causation, timing, or the degree of overlap among events. The original sample-size justification is not reported in the registry record. These limits are specific to the information available in this registry summary and should not be mistaken for evidence that a particular procedure was or was not used.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: HPV Vaccine Therapy in Reducing High-Grade Cervical Lesions in Patients With HIV and HPV — NCT03284866.
• Registry record last updated 2026-10-28; status last verified 2026-10.
• Results first posted 2026-10-28.
• Data retrieved from ClinicalTrials.gov on 2026-11-09.
• Group sizes, means, event counts and sponsor-reported p-values are reproduced as posted. Effect estimates, confidence intervals and p-values not attributed to the sponsor are independently calculated from the posted summary statistics.

18. Record Summary

The registry describes NCT03284866, a completed phase III randomized, double-masked, parallel-group trial with enrollment of 535. It compared recombinant HPV 9-valent vaccine with saline among women with HIV participating in an HPV test-and-treat program. The scheduled intramuscular vaccinations were at baseline, 4, and 26 weeks, and both arms included sample collection for laboratory biomarker analysis.

The primary endpoint was cervical HSIL or cervical cancer occurrence after the week 4 visit through week 52 post-randomization, expressed as events per 100 person-years. The registry describes Poisson rate estimation and comparison, alongside Kaplan-Meier and log-rank methods for time to event. It states that intention-to-treat and per-protocol analyses were performed separately.

Arm-specific primary estimates and comparison results are not reported in the registry record. The record lists 14 outcome measures and 9 analyses, and includes adverse-event counts and selected serious-event terms. Detailed missing-data methods, multiplicity handling, analysis-population definitions, and the original sample-size justification are not reported in the registry record.