GRACE: A Study to Compare the Effectiveness, Safety and Tolerability of PREZISTA (Darunavir)/Ritonavir by Gender and Race When Administered With Other Antiretroviral Medications in Human Immunodeficiency Virus (HIV) Positive Women and Men.: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for GRACE: A Study to Compare the Effectiveness, Safety and Tolerability of PREZISTA (Darunavir)/Ritonavir by Gender and Race When Administered With Other Antiretroviral Medications in Human Immunodeficiency Virus (HIV) Positive Women and Men. (NCT00381303), a non-randomized phase 3 study. The primary outcome is Number of Viral Load (VL) < 50 HIV-1 RNA Copies/mL (Time to Loss of Virologic Response[TLOVR]) Subjects by Sex.

2. Clinical Question

3. Trial Design

GRACE is identified as a phase 3, open-label, multicenter trial. The registry records a nonrandomized allocation and a single-group model, with no masking. Its official title describes a 48-week treatment period, and the enrollment is listed as 429. The record lists the study as completed. The recorded study start is 2006-11 and primary completion is 2008-11.

Participants received darunavir and ritonavir in combination with an individually optimized background regimen. The intervention description specifies darunavir 600 mg twice daily and ritonavir 100 mg twice daily for 48 weeks. Additional antiretroviral agents were selected by investigators using resistance testing and prior treatment history. As the background regimen could be individually selected, treatment was not simply a fixed combination identical for every participant.

The primary outcome was reported by sex, with female and male groups. Those groups do not constitute randomized comparison arms: participants were not assigned by chance to sex, and the study did not include a randomized control arm. Comparisons across sex or race can describe outcomes observed in the enrolled groups, but they cannot by themselves separate the influence of treatment from differences in participant characteristics, background regimens, or other factors.

The study’s stated purpose included evaluating effectiveness, safety, and tolerability by gender and race. The primary endpoint focused on virologic response, while secondary outcomes included viral-load response by race, results for an etravirine subgroup, and changes in CD4+ cell count using observed values and last observation carried forward. The record lists 10 outcome measures and 1 analysis. A protocol-level description of allocation procedures, eligibility details, or a sample-size justification is 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: Intention to Treat (ITT)

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 Number of Viral Load (VL) < 50 HIV-1 RNA Copies/mL (Time to Loss of Virologic Response[TLOVR]) Subjects by Sex, measured over Week 48.

The primary endpoint counted participants with HIV-1 RNA below 50 copies/mL at week 48, reported by sex. It used the time-to-loss-of-virologic-response, or TLOVR, response algorithm and the intention-to-treat population. A count of responders is a binary-outcome summary: each participant is classified as a responder or non-responder under the specified rules. The endpoint is intended to capture whether viral suppression was maintained according to those rules, rather than only whether a measurement at one visit crossed the threshold.

The registry description states that a participant was considered a responder at a time point and subsequent time points when response criteria were met. Treatment discontinuation, rebound, or certain treatment changes could lead to non-response classification. Intermittent missing values were treated as response only when the immediately preceding and following visits were responses. A new, unplanned antiretroviral agent was also part of the non-response criteria, except when used for tolerability.

General: a prespecified response threshold makes the endpoint interpretable as a count of participants meeting a defined virologic criterion. The TLOVR approach also incorporates later treatment and response history, so it is not equivalent to counting only observed viral-load measurements at week 48. The primary endpoint counts by sex are not reported in the registry record. Consequently, the endpoint definition can be described, but the observed response frequency and any difference between groups cannot be quantified from the record.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The registry’s primary endpoint is a binary response count. Each participant is classified under the TLOVR definition as a responder or non-responder at the relevant time point, and the count is reported by sex. The threshold is viral load below 50 HIV-1 RNA copies/mL at week 48. The record names the intention-to-treat population, but the primary counts and any model-based or unadjusted comparison are not reported in the registry record.

What does a binary endpoint represent?

General: a binary endpoint has two classification states, such as meeting or not meeting a response criterion. A count reports how many people fall into one state; an event proportion would relate that count to the number included in the analysis. A count alone is not a complete comparison when group sizes differ, and a group contrast requires an explicitly defined measure and its uncertainty. No response proportions or contrast estimates should be inferred when the primary counts are absent.

The TLOVR definition extends beyond a single measurement. The registry says that response at a time point carries forward to subsequent time points when the response conditions remain satisfied. Discontinuation at a time point, a rebound, or specified treatment changes can classify a participant as a non-responder. Intermittent missing measurements count as response only when both the immediately preceding and following visits were responses. The description also identifies rebound at an earlier time point as a non-response circumstance. This makes the endpoint dependent on the response history and specified rules, not just the week 48 laboratory value.

Why does the nonrandomized design matter?

General: random assignment is used to make treatment groups comparable, on average, with respect to both measured and unmeasured characteristics. In a nonrandomized study, subgroup differences may instead reflect group composition or other differences in care. GRACE is recorded as nonrandomized, single-group, and unmasked. Thus the female and male outcome groups are not randomized treatment groups, and their comparison cannot be interpreted as a randomized estimate of a treatment effect. The individually optimized background regimen adds another source of treatment variation because its components were selected according to resistance testing and prior treatment history.

The registry record does not report whether the primary sex comparison was adjusted, which covariates would have been included, or what statistical model was used. It also does not report a confidence interval or p-value for that comparison. These omissions mean readers cannot evaluate the modeling assumptions, adjustment strategy, or precision from the registry record.

How should planning calculations be understood?

General: a sample-size calculation is a planning exercise that depends on assumptions about the outcome, target effect, significance level, power, and allocation. It is not an observed trial result and is not evidence that the study achieved a particular level of power. The registry record does not report the trial’s sample-size justification. Any planning illustration based on an observed effect should be distinguished from the trial’s original design and should not be presented as observed power. The planning assumptions described for illustrations use a two-sided significance level of 0.05, equal allocation, and a normal approximation; they are not reported as the trial’s own design assumptions.

General: over-dispersion is relevant when observed variability exceeds what a chosen statistical model expects. It can make uncertainty estimates too small if the model does not allow for that extra variability. The primary endpoint is a binary count, and the registry record does not report a dispersion estimate or an over-dispersion adjustment. It would therefore be inappropriate to claim that such an adjustment was made or was needed.

What do the secondary outcomes add?

The registry lists secondary outcomes for the number of participants with viral load below 50 copies/mL by race, an etravirine subgroup, descriptive viral-load summaries, and change from baseline in CD4+ cell count. CD4+ changes are described using observed values and last observation carried forward. General: observed-value analyses use measurements actually available at the relevant time, while last observation carried forward substitutes a participant’s last recorded value for a later missing value. These approaches can produce different summaries when follow-up measurements are missing. The record does not report numerical results for these outcomes or enough detail to assess the assumptions behind each analysis.

10. Exact Estimates for Each Arm

No two-arm comparison is shown for this trial, but each arm's result can still be summarized with exact interval estimates computed from the posted counts. The Clopper-Pearson interval is exact and conservative; the Wilson interval is a common closer-to-nominal alternative; the Bayesian credible interval uses a uniform Beta(1, 1) prior. For small groups the three can differ noticeably.

R code and output

arm    <- c("Female", "Male")
n      <- c(287, 142)
events <- c(146, 83)
for (i in seq_along(n)) {
  ci <- binom.test(events[i], n[i])$conf.int          # Clopper-Pearson exact interval
  cr <- qbeta(c(0.025, 0.975), 1 + events[i], 1 + n[i] - events[i])   # Bayesian, uniform Beta(1,1) prior
  wi <- (function(x, n) { p <- x / n; z <- qnorm(0.975); d <- 1 + z^2 / n
    c((p + z^2 / (2 * n) - z * sqrt(p * (1 - p) / n + z^2 / (4 * n^2))) / d, (p + z^2 / (2 * n) + z * sqrt(p * (1 - p) / n + z^2 / (4 * n^2))) / d) })(events[i], n[i])
  cat(sprintf("ARM|%s|%d|%d|%.1f|%.1f|%.1f|%.1f|%.1f|%.1f|%.1f\n", arm[i], n[i], events[i], 100 * events[i] / n[i],
              100 * ci[1], 100 * ci[2], 100 * wi[1], 100 * wi[2], 100 * cr[1], 100 * cr[2]))
}

ARM  Female  287  146  50.9  44.9  56.8  45.1  56.6  45.1  56.6
ARM  Male  142  83  58.5  49.9  66.7  50.2  66.2  50.2  66.2

11. Multiplicity, Missing Data and Analysis Populations

The registry lists several outcomes: the primary viral-load response by sex, secondary viral-load outcomes by race and in an etravirine subgroup, and CD4+ cell-count changes using observed values and last observation carried forward. It lists 10 outcome measures and 1 analysis. A multiplicity adjustment, a hierarchy for interpreting outcomes, or a prespecified distinction between confirmatory and exploratory outcomes is not reported in the registry record. Without that information, the relationship among the multiple outcome questions and the handling of repeated statistical testing cannot be evaluated.

The primary population is described as intention to treat. The record does not report further details about which participants were included in each analysis, how participants with incomplete follow-up were handled beyond the stated TLOVR rules, or whether any analysis population differed from the enrolled population. The endpoint description does specify how certain missing viral-load visits were classified: intermittent missing values were considered response only if the immediately preceding and following visits were responses. It also identifies discontinuation, rebound, and some treatment changes as non-response circumstances.

For CD4+ outcomes, the registry names both observed-value and last-observation-carried-forward approaches. General: an observed-value summary can be based on a subset with measurements available, whereas last observation carried forward fills a missing later value with an earlier measurement. The latter assumes that the last observed value is a reasonable stand-in for the unobserved later value; that assumption may not hold in every setting. The record does not report missingness counts, reasons for missing data, sensitivity analyses, or a broader missing-data plan. Interim analyses and stopping rules are not reported in the registry record.

12. 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.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: pneumonia (3.5%), chest pain (2.8%), pneumocystis jiroveci pneumonia (2.8%), angina pectoris (1.4%), myocardial infarction (1.4%).

The registry reports safety counts separately for female and male participants. For female participants, 287 were at risk in the serious-event group and 47 were affected; the separate other-event figures list 287 at risk and 73 affected. For male participants, the corresponding serious-event figures are 142 at risk and 33 affected, while the separate other-event figures are 142 at risk and 33 affected. These are registry counts, not a randomized comparison. The record does not report deaths.

The listed serious adverse-event terms and percentages are Pneumonia, 3.5%; Chest Pain, 2.8%; Pneumocystis Jiroveci Pneumonia, 2.8%; Angina Pectoris, 1.4%; Myocardial Infarction, 1.4%; and Non-Cardiac Chest Pain, 1.4%. These percentages are reported as registry entries. The denominator and whether participants could contribute to more than one listed term are not reported in the registry record, so the percentages should not be added or treated as mutually exclusive categories.

General: safety counts describe events recorded under the study’s reporting definitions; they do not establish that treatment caused an event. Comparisons by sex are also affected by the nonrandomized design. The registry does not report enough detail to interpret event timing, duration, severity beyond the serious-event classification, or relatedness to study treatment.

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

GRACE was designed to evaluate effectiveness, safety, and tolerability of darunavir with ritonavir and an optimized background regimen, including outcomes by sex and race. Its primary endpoint was a week 48 viral-load response count under the TLOVR algorithm. The trial record supplies the endpoint definition, study design, participant enrollment, and some safety summaries, but the primary responder counts and comparative estimates are not reported in the registry record.

Accordingly, the record does not support a numerical statement about the size or direction of the primary difference between female and male participants. It also does not support a conclusion about statistical precision for that endpoint, because confidence intervals and p-values are not reported in the registry record. The presence of a defined endpoint and an analysis population does not substitute for reporting the endpoint results.

The design further limits the meaning of subgroup contrasts. The study was nonrandomized, single-group, and open-label. Outcomes by sex or race are not randomized treatment comparisons, and the investigator-selected background regimen could vary according to treatment history and resistance testing. Any observed subgroup contrast would therefore require careful interpretation and could not, on its own, establish that sex or race caused an outcome difference.

The listed serious adverse-event counts and terms describe what the registry reports, but they do not establish causal attribution. Details needed for fuller safety interpretation, including deaths, are not reported in the registry record. The secondary outcomes are named but their numerical results are not reported in the record.

General: a study result is most informative when the endpoint definition, observed estimate, uncertainty, analysis population, and design are considered together. Here, the endpoint and design can be characterized, but the primary effect estimate and its uncertainty cannot. The findings should therefore be understood as a record of the study question and available registry summaries, rather than as a basis for asserting a treatment benefit, lack of benefit, or difference between groups.

14. Why This Trial Matters Statistically

This trial illustrates why a clearly defined outcome and a reported outcome estimate are both necessary for statistical interpretation. GRACE specifies a binary virologic response threshold, a week 48 assessment, and the TLOVR rules that determine responder status over time. Those details clarify what the primary endpoint was intended to measure. Yet without the responder counts and comparison estimate, the size and precision of the primary result cannot be assessed from the registry record.

It also illustrates the distinction between subgroup reporting and randomized comparison. The study’s stated interest in sex and race is reflected in its outcome definitions, but its nonrandomized, single-group design means that subgroup contrasts are not protected by random assignment. The optimized background regimen was selected according to individual treatment history and resistance testing, so treatment context could differ among participants. This is important when describing what an observed contrast could mean.

The secondary outcome list also highlights analytic choices. CD4+ cell-count changes are described with both observed-value and last-observation-carried-forward approaches. General: different missing-data approaches can yield different summaries because they make different use of incomplete follow-up. Naming an approach is useful, but assessment also requires details on missingness and the assumptions behind the analysis, which are not reported in the registry record.

Finally, safety summaries need context. Counts and event-term percentages are informative descriptions, but they do not on their own establish causation or allow a complete comparison between groups. For clinical research readers, this record is a reminder to distinguish endpoint definition from endpoint results, subgroup patterns from randomized effects, and safety listings from causal conclusions.

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 central limitation for interpreting effectiveness is that the primary response counts by sex, effect estimate, confidence interval, and p-value are not reported in the registry record. Without these results, the magnitude, direction, and precision of the primary comparison cannot be assessed. Numerical results for the listed secondary outcomes are also not reported in the registry record.

The study design limits causal interpretation. Allocation was nonrandomized, the model was single-group, and masking was absent. The female and male groups are observational subgroups rather than randomized arms. The individually optimized background regimen was chosen by investigators based on resistance testing and prior treatment history, so treatment context was not necessarily uniform. An adjusted comparison and the details of any adjustment are not reported in the registry record.

The primary endpoint applies a specific response algorithm that incorporates treatment discontinuation, rebound, intermittent missing values, and certain treatment changes. That algorithm makes the classification rule explicit, but the registry record does not report the number of participants affected by each rule or provide enough information to examine alternative classifications. For CD4+ outcomes, observed-value and last-observation-carried-forward approaches are listed; missingness patterns and sensitivity analyses are not reported in the registry record.

Safety interpretation is also limited. The record lists serious-event counts by sex and percentages for selected serious-event terms, but does not report deaths. The denominator for the term percentages, event timing, attribution, and participant overlap across terms are not reported in the registry record. A sample-size justification, interim analyses, and a detailed statistical plan are not reported in the registry record. These gaps constrain assessment of precision, multiplicity, and the robustness of the analyses.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: GRACE: A Study to Compare the Effectiveness, Safety and Tolerability of PREZISTA (Darunavir)/Ritonavir by Gender and Race When Administered With Other Antiretroviral Medications in Human Immunodeficiency Virus (HIV) Positive Women and Men. — NCT00381303.
• Registry record last updated 2014-04-21; status last verified 2014-04.
• Results first posted 2010-10-19.
• 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

GRACE (NCT00381303) is recorded as a completed phase 3, open-label, multicenter, nonrandomized, single-group study. It enrolled 429 participants with HIV and evaluated darunavir with ritonavir alongside an investigator-selected optimized background regimen over 48 weeks. The primary endpoint was the number of participants with viral load below 50 HIV-1 RNA copies/mL at week 48 under the TLOVR algorithm, analyzed by sex in the intention-to-treat population.

The record lists secondary viral-load and CD4+ cell-count outcomes, including analyses by race and an etravirine subgroup. Primary responder counts, comparative estimates, and numerical secondary results are not reported in the registry record. Safety entries include serious-event counts by sex and selected serious-event terms with percentages; deaths are not reported. The nonrandomized subgroup design and missing numerical efficacy results limit the conclusions that can be drawn from the registry record.