PREPARE (A5361s) Ancillary Study of REPRIEVE (A5332): Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for PREPARE (A5361s) Ancillary Study of REPRIEVE (A5332) (NCT03070223), a randomized phase 3 study. The primary outcome is Physical Function: Rate of Change in Chair Rise Rate.

2. Clinical Question

The registry describes the study as investigating the impact of pitavastatin on physical and muscle function. Its background notes that aging with HIV is associated with earlier frailty or disability, including loss of physical and muscle strength and walking speed. The record also describes uncertainty about whether statins might affect these impairments and notes that statins can increase the risk of muscle aches and pains. Those statements provide the study rationale; they are not findings from this primary endpoint analysis. The statistical question addressed by the posted primary result is specifically about longitudinal chair rise rate, not every aspect of physical function, frailty, or muscle symptoms.

3. Trial Design

PREPARE (A5361s) is an ancillary study of REPRIEVE (A5332), with the official title “Pitavastatin to REduce Physical Function Impairment and FRailty in HIV (PREPARE).” The registry identifies it as a phase 3 study with randomized allocation, quadruple masking, and a parallel design. The interventions are pitavastatin and placebo. The arm descriptions clarify that participants were randomized to these assignments in the main REPRIEVE study. Thus, the ancillary comparison uses the treatment assignments of the parent study to examine physical-function outcomes.

The recorded enrollment is 602. The study start date is 2017-03-14, the primary completion date is 2023-08-21, and the status is completed. The primary endpoint analysis population is specified as all enrolled participants with any chair rise rate data available. The posted arm counts for this endpoint are 316 for pitavastatin and 285 for placebo. These are endpoint-specific counts, not a statement that every enrolled participant contributed data at every scheduled assessment.

The primary outcome is the rate of change in chair rise rate, measured at entry and months 12, 24, 36, 48, and 60. The registry says that linear mixed effects models for repeated measures were used to estimate annualized slopes and the between-group difference in slopes. This is a longitudinal analysis: it uses the pattern of measurements over follow-up to estimate change over time, rather than treating each measurement occasion as an unrelated comparison.

The record lists 18 outcome measures, including the primary measure and secondary measures covering gait speed, grip strength, balance, modified SPPB score, DASI-related impairment, frailty phenotype, and physical activity. It lists 12 analyses. These counts alone do not describe which analyses were prespecified, how they were prioritized, or whether inferential procedures were adjusted across outcomes. Those details 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.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: All participants enrolled with any chair rise rate data available.

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 participants enrolled.

6. Primary Endpoint

The registry lists the primary outcome as Physical Function: Rate of Change in Chair Rise Rate, measured over Entry and months 12, 24, 36, 48, 60.

The primary endpoint is the annualized rate of change in chair rise rate, expressed in rises/minute per year. Chair rise rate is calculated as the number of chair stands performed divided by the time taken to complete 10 chair stands. The measure therefore reflects both the number of stands completed and the time required to perform the test. The registry states that participants unable to attempt the test were assigned the worst time, defined as 9 seconds/stand for every stand not done.

Measurements were scheduled at entry and months 12, 24, 36, 48, and 60. A linear mixed effects model for repeated measures estimated an annualized slope for each treatment group and the difference between the groups’ slopes. The registry defines a negative annualized rate of change as decline over time and a positive rate as improvement. The endpoint is therefore intended to describe a trajectory of physical performance across repeated assessments, not just performance at one visit.

Chair rise performance is relevant to the study’s broad focus on physical function, muscle function, and frailty. The registry background discusses loss of physical and muscle strength as part of the concerns motivating the ancillary study. Chair rise rate offers a specific, quantified measure to analyze over time. It does not by itself represent every component of physical function or frailty, and the result should be interpreted as evidence about this endpoint in the analyzed population, not as a complete assessment of all the outcomes named in the study rationale.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

What is being estimated?

The primary endpoint is continuous: chair rise rate is recorded as a quantity, and the analysis estimates mean annualized change. The registry reports the treatment contrast as a net mean difference in slope. A slope represents the direction and rate of change across time. Here, the sign convention is explicit: a negative annualized rate of change reflects decline, and a positive rate reflects improvement. The between-group contrast asks whether those estimated trajectories differ.

Interpreting a slope contrast requires distinguishing it from a comparison of raw measurements. The reported -0.10 is a difference in annualized rates of change, not a statement that one group completed a particular number of stands at a particular visit. It summarizes the modeled difference in trajectories during the scheduled follow-up. The registry does not provide arm-specific visit means or a separate treatment contrast at each visit in the primary result.

Why use a repeated-measures model?

A linear mixed effects model for repeated measures is the stated method. In general, repeated-measures methods are designed for data in which the same participants can contribute observations at multiple times. Measurements from one person are usually related, rather than independent. A mixed model can represent the pattern of observations across time and estimate an average trend while accounting for that within-person structure under the model assumptions. The registry describes using the model to estimate annualized rates of change and the interaction between slope and treatment group.

The treatment-by-slope interaction is the feature that expresses whether the rate of change differs by assigned group. In general, an interaction means that the association between one factor and the outcome varies according to another factor. In this setting, the relevant question is whether the time trend differs for pitavastatin and placebo. The posted net mean difference is the resulting group contrast in slope. The record does not give the model formula, covariance structure, estimation procedure, or details of any covariates, so those aspects are not reported in the registry record.

How should the confidence interval and p-value be read?

The reported 95% confidence interval for the net mean difference extends from -0.30 to 0.10. In general, a confidence interval communicates uncertainty around an estimate. Its width reflects how precisely the data and model estimate the contrast, while its endpoints indicate values compatible with the analysis under its assumptions. The interval includes zero, so the posted result is compatible with no difference in slopes as well as with differences in either direction. The interval is not a range containing a specified proportion of individual participants’ outcomes.

The reported p-value is 0.31. In general, a p-value is calculated under a null hypothesis, commonly that there is no treatment-group difference, and measures how unusual the observed result or a more extreme one would be under that hypothesis and the test assumptions. It is not the probability that the null hypothesis is true, and it does not indicate the clinical importance of the estimate. The registry identifies the analysis as a superiority analysis. The record does not report a prespecified significance threshold or further testing details, so no additional inferential specifications should be assumed.

What does the endpoint calculation imply?

The registry defines chair rise rate as the number of chair stands performed divided by the time to complete 10 chair stands. It further states that inability to attempt the test is handled by assigning the worst time, 9 seconds/stand for every stand not done. This rule is relevant to how the endpoint is constructed because it assigns a value for stands that could not be attempted. The registry does not describe how other kinds of missing assessments were handled, or whether additional transformations or checks were used.

The repeated assessments are scheduled at entry and months 12, 24, 36, 48, and 60. Annualization expresses the estimated trajectory as a rate per year, which gives the treatment contrast a common time scale. In general, a longitudinal slope can draw information from observations collected at different visits; it is not necessarily calculated by subtracting only the first and last measurements. The exact model specification and the handling of participants with incomplete visit sequences are not reported in the registry record.

What can planning illustrations tell readers?

General sample-size planning asks how many participants might be needed to detect a specified effect under stated assumptions. Such calculations are planning illustrations, not observed power and not evidence that a study had a particular chance of detecting its actual effect. The registry describes an illustration based on the observed effect, a two-sided significance level of 0.05, equal allocation, and a normal approximation; the effect to detect is a stated fraction of the observed effect. The registry also distinguishes participant counts per arm for non-time-to-event outcomes from total event counts for time-to-event outcomes. The actual illustration results are not reported in the registry record summarized for this endpoint, so no sample-size values can be interpreted from that description alone.

The design enrollment is recorded as 602, whereas the endpoint-specific groups have counts of 316 and 285. These quantities have different meanings: enrollment concerns the study overall, while the posted endpoint counts describe participants contributing any chair rise rate data. They do not establish how many observations each participant posted or how many were available at each assessment. In general, the amount and pattern of missing data can affect the precision and interpretation of longitudinal estimates. Missing-data assumptions and sensitivity analyses for this endpoint are not reported in the registry record.

10. Analyses Reported by the Sponsor

The registry reports the primary comparison as a Mixed Models Analysis, with a net mean difference of -0.10 rises/minute per year, a 95% confidence interval from -0.30 to 0.10, and p = 0.31. It also posts mean annualized rates of change separately for the pitavastatin and placebo groups. The group-level estimates and the net contrast address related but distinct quantities: the former describe estimated slopes within each arm, and the latter is the modeled difference between arms.

The posted record does not identify a separate recalculated estimate for this comparison. The reported mixed-model result is therefore the inferential estimate to use when describing the primary analysis; a calculation from arm summaries would not reproduce the repeated-measures model without the underlying data and model specification. The registry does not give sufficient detail to reconstruct the model, including its covariance structure and any covariate adjustment. No independent recalculation should be treated as a replacement for the reported analysis.

The confidence interval includes zero and the reported p-value is 0.31. In statistical terms, the posted analysis does not establish a clear difference in annualized chair rise rate between the assigned groups. This reading is limited to the primary endpoint and the stated model result. It does not determine the findings for secondary outcomes, nor does it establish that the treatment groups had identical trajectories. The interval retains uncertainty about the size and direction of the difference.

11. Multiplicity, Missing Data and Analysis Populations

The registry lists 18 outcome measures and 12 analyses, with the chair rise rate trajectory identified as the primary outcome and a set of physical-function and frailty measures listed as secondary outcomes. When several outcomes or analyses are examined, multiplicity can matter: in general, multiple statistical tests can increase the chance of obtaining at least one apparently notable result by chance. The registry record does not report whether an adjustment for multiple testing was planned or applied, how the secondary outcomes were prioritized, or whether any hierarchical testing strategy was used. The primary p-value should therefore be interpreted as the reported result for that endpoint, without assuming a broader multiplicity procedure.

The endpoint population is defined as all enrolled participants with any chair rise rate data available. The group counts are 316 for pitavastatin and 285 for placebo. This definition does not mean that all participants had observations at every scheduled time point. The record does not report visit-specific completion, the number or timing of missing observations, reasons for missingness, or the distribution of missing data by group.

The stated model is a linear mixed effects model for repeated measures. Such models can use repeated observations when participants have incomplete follow-up, under assumptions about the data and missingness. The registry does not specify the assumptions used, whether imputation was performed, or whether analyses were conducted to assess sensitivity to missing-data assumptions. Those methods are not reported in the registry record and should not be inferred from the model name alone.

The record does not report interim analyses, stopping rules, a detailed statistical analysis plan, or a sample-size justification for the primary endpoint. It gives general assumptions for a sample-size planning illustration, but no endpoint-specific planning values are included in the registry record for this result. Consequently, neither the rationale for the enrolled sample size nor any adjustment for interim monitoring can be established from the registry record.

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

The posted result estimates the difference between pitavastatin and placebo in the annualized rate of change in chair rise rate. The estimated net mean difference is -0.10 rises/minute per year, and the 95% confidence interval extends from -0.30 to 0.10. The interval includes zero, and the reported p-value is 0.31. These statistics indicate that the primary analysis did not establish a statistically clear difference between groups in this modeled endpoint.

That interpretation is not the same as proving that the groups were identical. A confidence interval that includes zero allows no difference as one possibility, but it also allows differences in either direction. The estimate is uncertain, and the interval gives readers a sense of that uncertainty under the model. The posted result does not justify turning a non-significant comparison into a claim that there was no effect, nor does it establish a treatment benefit.

The result concerns a particular measurement, population, and analysis: chair rise rate, among enrolled participants with any such data available, using a repeated-measures mixed model. It is not a direct estimate of every aspect of physical function or frailty. The registry lists other outcomes, including gait speed, grip strength, balance, modified SPPB score, DASI-related impairment, frailty phenotype, and physical activity, but the primary result does not answer those separate questions.

Nor does this slope comparison describe every participant’s experience. It is a group-level estimate of average change over the scheduled follow-up. The record does not provide individual-level trajectories, the spread of individual responses, or visit-specific treatment contrasts. Its stated rule for participants unable to attempt the chair rise test is part of the outcome definition, but it does not resolve the reporting gaps about other missing assessments.

The registry’s background discusses possible pathways and concerns related to statins, including muscle aches and pains. The primary endpoint result does not test all of those mechanisms or report the complete safety profile. The adverse-event information in the record is limited, and no serious or other non-serious event counts at risk or affected are listed for the groups. The posted chair rise comparison should therefore be kept distinct from broad conclusions about safety, mechanism, or clinical effects.

13. Why This Trial Matters Statistically

This study is statistically informative because its primary question concerns change over time rather than a single endpoint measurement. Repeated-measures analysis is suited to a longitudinal question: the target is whether the average chair rise trajectory differs between randomized groups. Reporting an annualized slope and a treatment-group contrast makes the estimand explicit and helps distinguish a change-rate comparison from a cross-sectional difference.

The endpoint also illustrates why outcome definitions matter. Chair rise rate combines the number of completed stands with the time needed to perform the test, and the registry specifies a rule for participants unable to attempt it. This operational definition helps readers understand what the modeled outcome represents. It also highlights that the analysis is specific to the chosen measure and cannot stand in for every domain of physical function or frailty.

The primary estimate is accompanied by a confidence interval and p-value, allowing interpretation to include uncertainty rather than focus on a point estimate alone. The interval includes zero, and the reported p-value is 0.31. Those features make it important to distinguish lack of a statistically clear difference from proof of equivalence or identical outcomes. The registry does not report an equivalence or non-inferiority framework for this endpoint.

Finally, the study record highlights practical reporting questions that matter in longitudinal trials: the relationship between overall enrollment and endpoint-specific counts, the handling of incomplete repeated observations, and the treatment of multiple outcomes. The registry specifies the analysis population and model class, but several details needed to judge missingness and multiplicity are not reported in the registry record. These distinctions are important when interpreting the scope and precision of 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 primary endpoint counts are 316 in the pitavastatin group and 285 in the placebo group, while overall enrollment is 602. The analysis population includes participants with any chair rise rate data available. The registry does not report the number of participants contributing data at each scheduled assessment, the number with complete follow-up, or reasons for absent measurements. This limits assessment of how representative the analyzed data are of all enrolled participants and how incomplete follow-up may affect the estimated slopes.

The statistical method is described as a linear mixed effects model for repeated measures, with annualized slope and treatment-group difference estimated. The model’s detailed specification is not reported: the record does not identify its covariance structure, covariates, estimation method, or other modeling choices. These details can matter to interpretation of a longitudinal estimate, so the model cannot be independently reproduced from the registry record alone.

The reported 95% confidence interval for the net mean difference, -0.30 to 0.10, includes zero. The estimate therefore remains uncertain in both direction and magnitude. The registry does not report a prespecified clinically important difference for chair rise rate or an equivalence margin. The result cannot be interpreted as evidence of equivalence or as ruling out effects of a particular practical importance.

The registry lists multiple outcome measures and analyses but does not report a multiplicity strategy. It also does not report endpoint-specific sample-size justification, interim analyses, or sensitivity analyses for missing data. The posted primary result should be interpreted with these reporting limits in mind, and it should not be generalized to secondary outcomes that are not described by this estimate. The reported assignment and masking characteristics are available, but the record does not specify who was masked.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: PREPARE (A5361s) Ancillary Study of REPRIEVE (A5332) — NCT03070223.
• Registry record last updated 2024-10-28; status last verified 2024-09.
• Results first posted 2024-10-28.
• 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

PREPARE (A5361s) is a completed, phase 3 ancillary study of REPRIEVE (A5332) involving people with HIV-1 infection. The registry records randomized allocation, quadruple masking, a parallel design, and enrollment of 602. Participants were assigned in the main study to pitavastatin or placebo.

The primary endpoint was annualized change in chair rise rate, assessed at entry and months 12, 24, 36, 48, and 60. The endpoint analysis included participants with any chair rise rate data available, with group counts of 316 for pitavastatin and 285 for placebo. A repeated-measures linear mixed effects model was used to estimate annualized slopes and the between-group difference in slopes.

The registry reports a net mean difference of -0.10 rises/minute per year, with a 95% confidence interval from -0.30 to 0.10 and p = 0.31. The interval includes zero, so the primary analysis does not establish a statistically clear difference in this endpoint. Details about missing-data methods, multiplicity procedures, and the model specification are not reported in the registry record.