A Sub-study of Cabotegravir (CAB) and Rilpivirine (RPV) in Human Immunodeficiency Viruses (HIV)-Infected Participants: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for A Sub-study of Cabotegravir (CAB) and Rilpivirine (RPV) in Human Immunodeficiency Viruses (HIV)-Infected Participants (NCT05896761), a randomized phase 3 study. The primary outcome is Concentration at End of Dosing Interval (Ctau) of CAB LA Following Administration of CAB LA + RPV LA Q8W.

2. Clinical Question

The study concerns a change in injection site among people already receiving long-acting treatment in the parent A2M study. Its primary endpoint is a drug concentration measure, not a direct measure of symptom change or a broad test of virologic efficacy. The posted estimates therefore address pharmacokinetics in the analyzed participants and time points. They do not, by themselves, answer every question about tolerability, virologic outcomes, or longer-term experience.

3. Trial Design

This is a Phase 3, randomized, open-label, parallel-design sub-study. The registry lists enrollment of 118 and describes the study as completed. The recorded start date is 2021-10-28, and primary completion is recorded as 2022-08-23. ViiV Healthcare is listed as the sponsor.

Participants were adults with HIV-1 already enrolled in the ATLAS2M study, also called A2M. The official title specifies that participants had received at least three years of gluteal injections. The sub-study assessed long-acting cabotegravir and rilpivirine injections administered into the vastus lateralis muscle in the thigh. Two regimens are described for the thigh injection phase: CAB LA 600 mg with RPV LA 900 mg Q8W, and CAB LA 400 mg with RPV LA 600 mg Q4W. The primary endpoint specifically concerns the Q8W regimen.

Randomized allocation and parallel design are recorded, while masking is listed as none. The registry describes thigh and gluteal injection phases, and reports safety results for both phases and dosing schedules. The primary pharmacokinetic population was based on receiving at least one thigh injection and having an evaluable PK parameter estimate; it was not defined simply by enrollment. The numbers analyzed vary by endpoint and time point. The registry record does not report the randomization procedure, allocation concealment, or a detailed statistical plan, so those features cannot be characterized further.

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: A total of 118 participants were enrolled in the study.

Thigh Injection Phase (Up to Week 16)

Reasons for not completing this period:

Return to Glu Inj Phase (Weeks 16 to 24)

Analysis population for the primary outcome, as defined in the registry: Sub-study pharmacokinetic (PK) parameter Population included all participants who received at least one CAB and/or RPV thigh injection and had at least one evaluable PK parameter estimate during the sub-study. Only those participants with data available at specified time points have been analyzed.

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 Concentration at End of Dosing Interval (Ctau) of CAB LA Following Administration of CAB LA + RPV LA Q8W, measured over Weeks -8,-7,-4 and Pre-dose on Day 1(Pre-Thigh Gluteal Injection);2 Hours post-dose on Day 1,Pre-dose on Week 8,Weeks 1 and 4 post-dose(First Thigh Injection);2 Hours post-dose on Week 8 and Weeks 9, 12 post-dose,Pre-dose on Week 16(Last Thigh Injection).

The primary endpoint is cabotegravir concentration at the end of the dosing interval, denoted Ctau, following CAB LA plus RPV LA Q8W. Ctau is a trough measure: it describes concentration near the end of a dosing interval, before the next dose. This is relevant to pharmacokinetics because it characterizes drug exposure at a consistent point in the dosing schedule. The endpoint is reported as a geometric mean in micrograms per milliliter, with geometric coefficient of variation as the measure of spread.

The registry gives Ctau summaries for the first thigh injection, the last thigh injection, and the pre-thigh gluteal injection. The analyzed population included participants who received at least one CAB and/or RPV thigh injection and had at least one evaluable PK estimate. Only participants with data at the specified time points were analyzed. Thus, the reported group sizes can differ across measurements.

Concentration is a pharmacokinetic outcome, not a direct measure of virologic response or tolerability. It helps characterize drug exposure in the regimen and injection-site context studied. The registry record does not report an exposure threshold or a clinical decision rule for interpreting a particular Ctau value. Nor does it specify a target range or a margin for judging similarity between thigh and gluteal administration. Those omissions limit how far the concentration results can be translated into other clinical conclusions.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The primary outcome is a continuous pharmacokinetic measure summarized using a geometric mean. Geometric means are commonly used for concentration data when values are positive and variability is more naturally described on a multiplicative scale. In general, a geometric mean summarizes the central level after accounting for that scale; it is not the same summary as an arithmetic mean. The registry also reports geometric coefficient of variation, a relative measure of dispersion. A larger coefficient indicates greater variability relative to the geometric mean. It does not describe the range of every participant’s measurements.

What does the ratio express?

The sponsor reports ratios of geometric least square means comparing thigh-phase measurements with the pre-thigh gluteal measurement. A ratio of 1 represents equal estimated geometric means on the ratio scale. A value below 1 indicates that the numerator estimate is lower than the reference estimate; a value above 1 indicates that it is higher. These are comparisons of estimated group summaries, not individual-level changes. The reported ratio estimates are 0.926 for the first thigh comparison and 0.929 for the last thigh comparison. Their 90% confidence intervals are 0.831 to 1.03 and 0.816 to 1.06, respectively.

How should the confidence intervals be read?

In general, a confidence interval expresses uncertainty in an estimate under the analysis assumptions. A narrower interval indicates greater precision than a wider interval on the same scale. The reported intervals include 1, so values corresponding to no ratio difference remain among the estimates compatible with the data under the method used. The registry record does not report the model details, the confidence-interval construction, or a decision margin. It would therefore be inappropriate to infer an equivalence conclusion from the interval alone. A confidence interval is not a prediction interval for an individual participant and does not mean that every value inside it is equally likely.

What is the role of the secondary recalculation?

The HIVTSQc comparison is a simple, unadjusted recalculation from posted summary statistics. It estimates a mean difference of 1.9 between the Q4W and Q8W groups. Its 95% confidence interval is -3.962 to 7.762, and its unadjusted p-value is 0.525. In general, the mean difference expresses the separation between the group averages, while the confidence interval indicates the uncertainty in that estimated separation. Since the interval includes zero, differences in either direction remain compatible with the recalculation. The p-value is not the probability that the null hypothesis is true, nor does it measure the size or importance of an effect. No sponsor-reported analysis for this outcome is listed in the record.

Why do planning illustrations differ from trial results?

The planning illustrations use the observed HIVTSQc effect as a reference and show hypothetical participants per arm for specified fractions of that effect, under a two-sided significance level of 0.05, equal allocation, and a normal approximation. For 100% of the observed effect, the illustration gives 1,074 participants per arm for 80% power and 1,438 for 90% power. For 75% of the observed effect, the corresponding figures are 1,909 and 2,555. For 50% of the observed effect, they are 4,294 and 5,749. These are planning illustrations, not observed power and not the trial’s own sample-size justification. The registry record does not report the study’s sample-size assumptions or justification.

In general, planning for a smaller effect requires a larger sample to achieve the same target power, all else held constant. That principle explains the direction of these illustrative values, but the values should not be read as a recommendation or as proof that the study was underpowered. They depend on the stated assumptions and use an observed effect that is itself uncertain. The record does not report the design effect or target effect used by the sponsor.

Why can variability matter?

General statistical explanation: over-dispersion means that observed variability is greater than a model would expect under its assumed variance structure. When data are more variable than assumed, standard errors and confidence intervals can be too small if the analysis does not account for that extra variation. The primary endpoint reports substantial geometric coefficients of variation, but the registry record does not describe an over-dispersion assessment or a model adjustment for it. The coefficients alone do not establish that the analysis mishandled variability.

The analysis populations also matter. The primary PK population required exposure to a thigh injection and an evaluable PK parameter, and only people with data at particular time points were included in those summaries. This is not necessarily the same population as the enrollment group or the populations used for safety and other secondary endpoints. The record’s endpoint-specific populations should guide interpretation of each estimate, rather than assuming that every result applies to every enrolled participant.

10. Analyses Reported by the Sponsor

The primary comparisons in the registry are sponsor-reported ratios of geometric least square means: 0.926 for the first thigh measurement versus the pre-thigh gluteal measurement, and 0.929 for the last thigh measurement versus the pre-thigh gluteal measurement. Their respective 90% confidence intervals are 0.831 to 1.03 and 0.816 to 1.06. These comparisons are not simple ratios recalculated from the displayed geometric means; they are reported as their own analysis estimates.

No simple two-group recalculation is reported for the primary endpoint. The registry record does not provide the method details needed to reproduce the sponsor’s model-based estimates from the summary statistics alone. The intervals include 1, but no equivalence margin or p-value is reported for these comparisons. Accordingly, they describe the sponsor’s estimated relative concentrations and their uncertainty, without establishing equivalence.

The HIVTSQc mean difference is a separate unadjusted recalculation from the posted group summaries, and no sponsor-reported analysis for that outcome is listed. The recalculation should not be substituted for an analysis specified by the sponsor. The record does not flag a discrepancy between a recalculation and a sponsor analysis for the primary outcome; it also does not offer a reproducible sponsor model specification that would allow an independent reconstruction of the reported ratios.

11. Multiplicity, Missing Data and Analysis Populations

The registry record lists numerous secondary outcomes and 34 outcome measures overall, but it does not report a multiplicity-adjustment plan. It is not reported in the registry record whether a hierarchy of hypotheses, family-wise error procedure, false-discovery procedure, or other strategy was prespecified. Without that information, nominal or unadjusted secondary findings should be interpreted cautiously: examining many outcomes can increase the chance of at least one apparently unusual result. This is a general statistical concern, not evidence that any particular finding is false.

The record also does not report whether interim analyses were conducted, whether stopping rules were used, or what sample-size justification informed the study design. These design and analysis details should not be inferred from the enrollment figure or from the planning illustrations. The illustrations are based on the observed HIVTSQc effect and stated assumptions; they are not the trial’s design calculation.

Missing-data handling is not described in sufficient detail to characterize imputation, censoring, or sensitivity analyses across the outcomes. For the primary PK endpoint, the stated population includes participants with an evaluable estimate, and only those with data at the specified time points were analyzed. This makes the time-point-specific group sizes relevant to interpretation. The record does not report how missing PK measurements were handled or whether missingness was assessed as informative. For virologic outcomes, the FDA Snapshot algorithm is named, but the record does not provide enough detail to reconstruct all endpoint-level handling of missing or unavailable measurements.

Analysis populations differ by purpose. The PK population requires thigh injection exposure and an evaluable PK parameter; safety summaries use the groups and phases described for adverse events; and the secondary endpoints may have their own definitions. The record does not report a complete population flow or the reasons participants may be absent from an endpoint analysis. As a result, the enrollment figure should not be treated as the denominator for every estimate, and findings from one analysis population should not automatically be generalized to another.

12. 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 5% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: covid-19 (2.0%).

The registry reports serious adverse event figures separately for thigh and gluteal injection phases and for the two dosing schedules. In the Q8W thigh phase, 54 participants were at risk and 0 were recorded with a serious adverse event; in the Q4W thigh phase, the corresponding figures were 64 and 0. In the Q8W gluteal phase, 51 were at risk and 1 was recorded with a serious adverse event. In the Q4W gluteal phase, 61 were at risk and 0 were recorded with a serious adverse event. Deaths are recorded as 0 in each of these four groups.

The record identifies COVID-19 as a top serious adverse event term, at 2.0%. It does not provide enough detail in the registry record used for this analysis to reconcile that percentage with the phase-specific serious-event counts or to determine the term’s exact analysis denominator. The figures are descriptive and do not establish that injection site or dosing schedule caused a difference in risk. They also do not establish the absence of other safety concerns: the record reports other adverse events, and the number of participants affected by them differs across groups. The serious-event figures should be interpreted within the stated phases, risk groups, and registry definitions, not as comparative evidence from an adjusted safety model.

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

The primary result describes cabotegravir trough concentrations in a particular setting: adults with HIV-1 already enrolled in A2M, receiving long-acting CAB and RPV, with Q8W thigh administration compared with a pre-thigh gluteal measurement. The sponsor-reported geometric least square mean ratios are close to 1, and both 90% confidence intervals include 1. This indicates that the estimates are compatible with equal relative geometric means as well as values on either side of 1. It does not establish equivalence, because the registry record does not report an equivalence margin or the associated decision framework.

The endpoint is pharmacokinetic. It measures concentration at a dosing-interval trough, not viral suppression, resistance, injection-site tolerability, or overall treatment satisfaction. Those topics appear among the secondary outcomes, but a concentration comparison cannot settle them. Similarly, absence of a serious adverse event in a listed group does not mean that all adverse events were absent or that there is no safety uncertainty.

The analyses also have limits of scope. PK summaries include participants with evaluable measurements at the specified time points, and the corresponding group sizes differ. The participants were already receiving treatment in the parent study and had a history of gluteal injections, so the findings should not automatically be extended to populations or treatment circumstances not described in the registry record. The record does not report all details needed to assess missing measurements, multiplicity, or model assumptions.

The appropriate reading is therefore specific: the posted estimates characterize selected concentrations and outcomes in this sub-study, with the stated uncertainty. They do not alone prove that thigh and gluteal administration are interchangeable, establish an efficacy conclusion, or predict an individual participant’s experience.

14. Why This Trial Matters Statistically

Statistically, this study is notable for separating a pharmacokinetic question from the broader clinical outcomes that may surround it. Its primary endpoint is a trough concentration, summarized geometrically, while the secondary outcomes span safety, virologic status, resistance, discontinuation, and treatment satisfaction. Keeping those estimands distinct helps prevent a concentration result from being treated as a proxy for every clinical outcome.

The comparison also illustrates why estimates and uncertainty should be read together. The sponsor-reported ratios are near 1, but their 90% confidence intervals extend across 1. The point estimates alone would give an incomplete account: the intervals indicate that the relative concentration is not estimated with enough precision to pin down a narrow range around the point estimates. At the same time, including 1 does not establish equivalence without a specified margin and analysis framework.

The study’s existing A2M population and its pre-thigh gluteal measurements provide a particular context for interpreting the results. This is not a comparison against no treatment, and the primary question is not a general test of whether CAB and RPV are effective. Its design and endpoint focus attention on exposure following a change in injection site, while the registry’s secondary measures address other aspects of the experience.

Finally, the record is a reminder that registry summaries may report outcome estimates without all the detail needed to reproduce the analysis or assess multiplicity and missing-data decisions. Those omissions do not invalidate the posted numbers, but they constrain claims that can be made from them. A careful statistical reading preserves the distinction between measured results, model-based comparisons, and questions that remain unanswered in the record.

15. Limitations

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

The primary endpoint is limited to the Q8W regimen and to cabotegravir trough concentration. Although the record describes both Q8W and Q4W thigh regimens, it does not report a primary Q4W Ctau comparison in the registry record used for this analysis. The endpoint therefore should not be treated as a direct comparison of the two thigh dosing schedules.

The PK analysis population is defined by receiving at least one CAB and/or RPV thigh injection and having an evaluable PK parameter. Only participants with data at the specified times were analyzed. This selection and the variation in time-point-specific group sizes can limit generalizability and precision. The registry record does not report the reasons for unavailable measurements, a missing-data sensitivity analysis, or a complete participant flow for each endpoint.

The study is open-label, and masking is listed as none. The primary endpoint is a measured concentration, but the record does not describe laboratory procedures or other safeguards that would permit further assessment of measurement bias. It also does not report the statistical model details behind the sponsor’s geometric least square mean ratios, an equivalence margin, or a multiplicity plan. These omissions restrict independent assessment of the comparison and of secondary findings.

The study population had prior experience with gluteal injections in the parent A2M study, and the official title specifies at least three years of such injections. Results may not apply to people without that treatment history or to other clinical settings. The primary completion date is recorded, but the registry entries used for this analysis do not give a detailed follow-up description beyond the stated endpoint timeframes. They also do not report the sample-size justification or interim-analysis plan. These limitations concern interpretation of the registry record; they do not establish that the study’s conduct or analyses were inadequate.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: A Sub-study of Cabotegravir (CAB) and Rilpivirine (RPV) in Human Immunodeficiency Viruses (HIV)-Infected Participants — NCT05896761.
• Registry record last updated 2023-09-18; status last verified 2023-08.
• Results first posted 2023-09-18.
• 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.

19. Record Summary

NCT05896761 is a completed Phase 3 randomized, open-label, parallel-design sub-study sponsored by ViiV Healthcare. It enrolled 118 adults with HIV-1 who were participating in A2M; the official title describes participants as having received at least three years of gluteal injections. The study assessed long-acting cabotegravir and rilpivirine administered into the thigh using Q8W and Q4W regimens.

The primary endpoint was Q8W cabotegravir Ctau, summarized as a geometric mean with geometric coefficient of variation. Sponsor-reported ratios compared first and last thigh measurements with the pre-thigh gluteal measurement, with 90% confidence intervals that include 1. The record also lists secondary outcomes spanning safety, virologic measures, resistance, and treatment satisfaction. Interpretation is limited by the endpoint-specific populations and by statistical details, including multiplicity and missing-data methods, that are not reported in the registry record.