Studies of the Ocular Complications of AIDS (SOCA)--Monoclonal Antibody CMV Retinitis Trial (MACRT): Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Studies of the Ocular Complications of AIDS (SOCA)--Monoclonal Antibody CMV Retinitis Trial (MACRT) (NCT00000135), a randomized phase 2, phase 3 study. The primary outcome is Mortality Rate.

2. Clinical Question

3. Trial Design

The Monoclonal Antibody CMV Retinitis Trial was a randomized, double-masked, parallel-group study. Its registry phase is listed as phase 2 and phase 3, and its status is completed. The record lists enrollment of 209 and identifies Johns Hopkins Bloomberg School of Public Health as the sponsor.

One group received MSL-109 and the other received placebo. Both interventions are described as intravenous infusions every 2 weeks at 60 mg. The brief summary characterizes MSL-109 as adjunct therapy intended for evaluation in controlling CMV retinitis. The record does not give additional treatment details, such as treatment duration or criteria for modifying or stopping treatment.

The registry lists a start date of 1995-09 and a primary completion date of 1996-08. For the primary outcome, all patients enrolled were followed for a 17 month period or until a common study closing date. The record does not explain how follow-up ended for individual participants or how time at risk was handled when follow-up ended.

The posted arm counts are 104.0 for MSL-109 and 105.0 for placebo. The comparison is reported as MSL-109 relative to placebo. Further information about the randomization sequence, allocation concealment, masking implementation, protocol deviations, and analysis populations is not reported in the registry record. The record also does not describe the original trial design assumptions or its planned statistical analysis.

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: Two hundred and nine patients with acquired immunodeficiency syndrome and active CMV retinitis were enrolled in a multicenter, phase 2/3, randomized, placebo controlled clinical trial. Patients received adjuvant treatment with MSL-109, 60mg intravenously every 2 weeks, or placebo,.

Overall Study

Reasons for not completing this period:

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 Mortality Rate, measured over All patients enrolled were followed for a 17 month period or until a common study closing date.

The primary endpoint is mortality rate, expressed as deaths per person-year. A rate relates events to accumulated time at risk, rather than simply reporting the number of participants with an event divided by the number assigned to a group. This form of outcome can be useful when people contribute different amounts of observation time, because the time component is part of the quantity being compared.

The registry specifies follow-up for all enrolled patients for a 17 month period or until a common study closing date. The posted rate is 0.68 deaths per person-year for MSL-109 and 0.31 deaths per person-year for placebo. These are the reported group rates; the registry does not provide arm-specific confidence intervals or a detailed account of how each participant's observation time was determined.

Mortality was selected as the single primary outcome measure listed in the record. The stated trial objective concerned the efficacy and safety of MSL-109 as adjunct therapy for CMV retinitis, while the registered primary measure captures mortality rather than a measure of retinitis progression. Other clinical outcomes and secondary results are not posted in the registry record. The endpoint therefore answers a specific rate-comparison question and does not, by itself, describe every potential benefit or harm relevant to the treatment.

7. Results Posted in the Registry

• MSL-109: 0.68 versus 0.31 (deaths per person-year) on control; rate ratio 2.19 (95% CI 1.55–3.11, approximate, Poisson model); p ≈ <0.001.

8. Statistical Interpretation

9. Statistical Methods Explained

Why use an event rate?

An event rate uses both the number of events and the amount of time participants were observed. Its unit is deaths per person-year. In general, expressing an outcome as a rate is appropriate when follow-up time contributes to the opportunity to experience an event. It distinguishes a rate comparison from a simple comparison of event proportions, which does not include time at risk in its denominator.

The registered timeframe states that all enrolled patients were followed for a 17 month period or until a common study closing date. The registry does not give participant-level follow-up times, event dates, or details about the construction of exposure time. Consequently, the posted summary supports interpretation of the reported rates but not a reconstruction of those rates from individual records.

What does the rate ratio express?

The comparison is reported as a rate ratio, with MSL-109 as the group of interest and placebo as the comparator. A rate ratio compares the event rate in one group with the rate in the other. The posted estimate is 2.194. It is a relative measure, whereas 0.68 and 0.31 deaths per person-year are rates in the endpoint's original units. Both forms are useful: the rates show the scale of the outcome, and the ratio summarizes their relative relationship.

A ratio is not a count of additional deaths and does not directly state how many participants experienced the outcome. Nor should it be interpreted without its uncertainty interval. The registry identifies a Poisson analysis, which is a standard modeling approach for event rates. The record does not provide enough detail to assess adjustments, diagnostics, or whether another model was considered.

What do the confidence interval and p-value contribute?

The reported 95% confidence interval for the rate ratio extends from 1.548 to 3.109. A confidence interval conveys uncertainty in the estimated comparison under the statistical model. It is not a guarantee that the true value lies in the interval, nor does it incorporate uncertainty from possible bias, incomplete reporting, or model misspecification.

The unadjusted p-value is <0.001. In general, a p-value assesses how unusual data at least as discrepant as those observed would be under a specified null model and its assumptions. It is not a probability that the null model is true. The record does not state an adjusted p-value, a multiplicity procedure, or the complete testing strategy, so the reported value should be described as unadjusted rather than as a result adjusted for other analyses.

Why can over-dispersion matter?

In general, a basic Poisson model relates the variability of event counts to their expected counts. Over-dispersion means the observed variability is greater than that model allows. If over-dispersion is present and not accounted for, standard errors and confidence intervals may be too narrow, and p-values may appear more compelling than warranted. This is a general consideration for Poisson rate analyses, not a finding about this trial: dispersion information and model diagnostics are not reported in the registry record.

The record marks the comparison as Poisson, but does not describe whether the model included participant characteristics, stratification, or other adjustment. It also does not report the analysis code or data needed to test alternative assumptions. Interpretation therefore rests on the posted estimate and interval while recognizing that the registry record gives limited detail about model evaluation.

What is and is not reported about the analysis?

The record lists one primary outcome measure and no posted secondary outcomes. It lists no analysis records, but it does post a mortality rate comparison, its confidence interval, and an unadjusted p-value. These entries provide a summary of the primary result without a full statistical analysis plan or report. The protocol's prespecified model, handling of missing follow-up, analysis population, covariate plan, and any interim analyses are not reported in the registry record.

These distinctions matter because an estimate, its uncertainty, and the process used to select or conduct analyses are related but separate pieces of evidence. The posted comparison can be read as the registry's reported primary rate result; the unavailable methodological details limit assessment of how fully it follows a prespecified plan.

10. Confidence Intervals and Precision

The rate ratio estimate is 2.194, and its 95% confidence interval runs from 1.548 to 3.109. The interval communicates the uncertainty around the relative rate estimate under the analysis model. It is narrower than an interval that would encompass a much broader range of relative comparisons, but it still allows for meaningful variation in the magnitude of the estimated relationship.

The interval lies on the side of higher rates in the MSL-109 group relative to placebo; it does not include a value indicating equal rates. This is consistent with the reported unadjusted p-value of <0.001, although the two statistics summarize evidence in different ways. The p-value is not a measure of precision or effect magnitude. The interval is not a prediction of what would occur for an individual participant, and it does not account for potential biases or limitations in the registry record.

The arm-specific rates are 0.68 and 0.31 deaths per person-year, but arm-specific confidence intervals are not reported. The record also does not provide dispersion estimates or detail about the model's standard error calculation. Thus, the reported confidence interval is the available measure of precision for the relative comparison, and its interpretation depends on the analysis assumptions. It should not be treated as resolving uncertainty about unreported outcomes, missing data, or the generalizability of the result.

11. Planning a Similar Trial: Sample Size

This is a planning illustration, not a re-evaluation of the original trial. It asks: if a new study wanted to detect a difference like the one observed, or a smaller one, how large would it need to be? It is deliberately not “observed power” of the finished trial, which is not a meaningful quantity.

The planning figures are participants per arm for this event-rate outcome, not total events. The registry's general assumption for these illustrations is a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. The planning effect is defined as the stated fraction of the observed effect. The trial's original design assumptions are not reported in the registry record.

For planning for 100% of the observed effect, the figures are 43 participants per arm for 80% power and 57 participants per arm for 90% power. For planning for 75% of the observed effect, they are 81 participants per arm for 80% power and 108 participants per arm for 90% power. For planning for 50% of the observed effect, they are 195 participants per arm for 80% power and 260 participants per arm for 90% power.

In plain terms, 80% power is the assumed chance of detecting the specified effect under the planning model if that effect is present; 90% power uses a higher target chance under the same type of assumptions. Smaller effects require more participants to detect, as reflected in the increasing planning figures as the assumed fraction of the observed effect becomes smaller. These figures are planning illustrations, not observed power.

12. How the Sample Size Is Derived

Power is the probability that a study of a given size rejects the null hypothesis when a specific true effect exists. It depends on four things: the effect to detect, the variability of the outcome, the significance level (here a two-sided α of 0.05, so z_1−α/2 = 1.960) and the desired power (z_1−β = 0.842 for 80% and 1.282 for 90%).

Formula

n is the number of participants in each arm, λ₀ and λ₁ are event rates per patient-year, and T is follow-up per participant in years. This assumes Poisson counts and equal follow-up.

Worked calculation for this trial

1. Control rate λ₀ = 0.3100; treatment rate λ₁ = 0.6800 events per patient-year; follow-up T = 1.42 years.
2. n = (1.96 + 0.842)² × (1/0.3100 + 1/0.6800) / (1.42 × [ln(2.194)]²) = 42.2, rounded up to 43 participants per arm for 80% power.

The smaller-effect rows in the table above repeat the same calculation with the effect multiplied by 0.75 and 0.5 (for ratios, applied to the logarithm of the ratio).

R code

alpha <- 0.05
z_a   <- qnorm(1 - alpha / 2)          # two-sided
lambda0 <- 0.31                      # control rate, events per patient-year
lambda1 <- 0.68                      # treatment rate
tt      <- 1.41666666666667                       # follow-up per participant, years
n_per_arm <- function(power, frac = 1) {
  l1 <- lambda0 * (lambda1 / lambda0)^frac   # smaller effect = fraction of the observed log rate ratio
  ceiling(round((z_a + qnorm(power))^2 * (1 / lambda0 + 1 / l1) / (tt * log(l1 / lambda0)^2), 6))
}
for (f in c(1, 0.75, 0.5))
  cat(sprintf("%3.0f%% of observed effect: n per arm = %s (80%% power), %s (90%% power)\n", 100 * f,
              format(n_per_arm(0.80, f), big.mark = ","), format(n_per_arm(0.90, f), big.mark = ",")))

# Simulation 1: Poisson counts. Simulation 2: over-dispersed counts (variance = 1.5 x mean), analysed with the matching variance.
set.seed(135)
n    <- n_per_arm(0.80)
reps <- 20000
m0 <- lambda0 * n * tt; m1 <- lambda1 * n * tt
x0 <- rpois(reps, m0); x1 <- rpois(reps, m1)
z  <- log(x1 / x0) / sqrt(1 / x0 + 1 / x1)
cat(sprintf("Simulated power, Poisson counts, n = %s per arm: %.3f\n", format(n, big.mark = ","), mean(abs(z) > z_a, na.rm = TRUE)))
phi <- 1.5
y0 <- rnbinom(reps, mu = m0, size = m0 / (phi - 1)); y1 <- rnbinom(reps, mu = m1, size = m1 / (phi - 1))
z2 <- log(y1 / y0) / sqrt(phi * (1 / y0 + 1 / y1))
cat(sprintf("Simulated power, over-dispersed counts (phi = 1.5), same n: %.3f\n", mean(abs(z2) > z_a, na.rm = TRUE)))

Output from running the code

100% of observed effect: n per arm = 43 (80% power), 57 (90% power)
 75% of observed effect: n per arm = 81 (80% power), 108 (90% power)
 50% of observed effect: n per arm = 195 (80% power), 260 (90% power)
Simulated power, Poisson counts, n = 43 per arm: 0.834
Simulated power, over-dispersed counts (phi = 1.5), same n: 0.657

The first simulation draws Poisson counts and should land close to the target power. The second allows extra variability (variance 1.5 times the mean, a purely illustrative value), which lowers the power at the same sample size, so rate trials usually need more participants than the Poisson formula suggests.

13. Multiplicity, Missing Data and Analysis Populations

The registry lists one primary outcome measure, mortality rate, and no posted secondary outcomes. It does not report whether the study protocol included additional outcomes that were not posted, or whether multiple statistical tests were planned or conducted. A multiplicity strategy, such as a prespecified approach for controlling the chance of false-positive findings across several tests, is not reported in the registry record.

The reported p-value is unadjusted. That description matters because the record does not show whether other comparisons or analyses were considered alongside the primary result. Without a stated testing hierarchy or adjustment procedure, the p-value should not be described as adjusted for multiplicity. The record also gives no interim analysis information or stopping rules, so these aspects of the analysis process cannot be assessed from the registry entry.

Details about missing outcome information and the analysis population are not reported in the registry record. The stated follow-up endpoint is a 17 month period or a common study closing date, but the record does not explain how incomplete follow-up, withdrawal, or other missing data were handled. It does not specify whether the analysis followed an intention-to-treat principle, included participants with available outcome data, or used another population definition.

These omissions do not negate the posted summary, but they limit evaluation of how participants and time at risk entered the analysis. They also prevent assessment of whether missingness could have affected the estimated rate. A complete account would require the prespecified analysis plan and details about outcome ascertainment and follow-up, none of which are reported in the registry record.

14. 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 0% reporting threshold.

The serious adverse event listing gives 104 participants at risk in the MSL-109 group and 105 in the placebo group, with 0 serious adverse events affected in each group. For other adverse events, the listing gives 104 at risk and 0 affected for MSL-109, and 105 at risk and 0 affected for placebo. No serious adverse event terms are listed.

These are registry entries for the adverse-event categories shown; they should not be read as a complete description of all safety outcomes or as evidence that no adverse effects occurred. The record does not provide event terms beyond the empty serious-event listing, event definitions, collection methods, or an explanation of the reporting scope. Deaths in the adverse-event listing are recorded as unavailable, so those entries do not provide a death count. Mortality is reported separately as the primary rate outcome.

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

The result means that the registry reports a higher mortality rate in the MSL-109 group than in the placebo group for the stated follow-up, with a rate ratio of 2.194 and a 95% confidence interval from 1.548 to 3.109. The unadjusted p-value is <0.001. These statistics characterize the reported comparison under its analysis assumptions; they do not establish why the rates differed.

The result does not show that MSL-109 caused the difference. Randomization is an important design feature, but the registry record does not provide sufficient information about implementation, follow-up completeness, the analysis population, or model checking to evaluate every potential source of bias. The posted result is also specific to mortality. It does not describe the full course of CMV retinitis, other clinical outcomes, or all aspects of safety.

The rates are population-level summaries and do not predict an individual participant's outcome. The confidence interval describes uncertainty in the relative rate estimate under the model, not uncertainty about all possible limitations of the evidence. Likewise, the small unadjusted p-value is not the probability that a treatment effect is absent or present, and it does not determine clinical importance.

The registry record lists no secondary outcomes and provides limited methodological detail. Accordingly, interpretation should remain focused on the primary mortality-rate comparison and the precision reported for it. Broader conclusions about outcomes not posted, mechanisms, or use in populations beyond the trial are not established by these registry results.

16. Why This Trial Matters Statistically

This trial illustrates why the choice of outcome scale matters. Mortality is presented as an event rate per person-year, which incorporates time at risk rather than treating the outcome solely as a proportion of participants. A rate ratio then summarizes the relative comparison while the group rates retain the endpoint's original units. Reading both measures together helps distinguish the scale of the outcome from its relative relationship.

It also demonstrates the complementary roles of an estimate, a confidence interval, and a p-value. The rate ratio of 2.194 gives the reported relative comparison; the 95% confidence interval from 1.548 to 3.109 conveys its model-based uncertainty; and the unadjusted p-value of <0.001 summarizes evidence against a specified no-difference model under the analysis assumptions. None of these statistics alone describes bias, clinical importance, or the reliability of unreported analyses.

The record is also a useful example of the limits of registry summaries. It identifies the design and primary endpoint and posts a comparison, yet details such as the analysis population, missing-data handling, and model diagnostics are not reported in the registry record. Statistical interpretation must account for both what the numerical result says and what methodological context is unavailable.

Finally, the planning illustrations distinguish power calculations from evidence observed in a completed study. They show how a specified effect size and planning assumptions relate to participant requirements, not whether the trial had a particular probability of success after results were observed. This separation helps prevent planning calculations from being mistaken for a retrospective measure of evidential strength.

17. 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.
• Rate comparisons use a Poisson approximation unless the sponsor posted confidence limits; real event counts are often over-dispersed.

The registry record gives a concise result but limited supporting detail. It does not report the full eligibility criteria, participant characteristics, outcome ascertainment procedures, or participant-level time at risk. Those omissions make it difficult to judge how broadly the enrolled population is described or to reconstruct the mortality rates independently.

The primary outcome is reported with group rates and a relative comparison, but arm-specific confidence intervals and dispersion information are not reported in the registry record. Although the comparison is identified as Poisson, model diagnostics, covariate adjustment, and any sensitivity analyses are not described. The unadjusted p-value is available, but a complete statistical analysis plan and multiplicity approach are not reported.

The record does not explain how incomplete follow-up or missing outcome information was handled, nor does it define the analysis population. The stated follow-up period ends at 17 months or a common study closing date, but the process for determining exposure time and handling participants who did not complete follow-up is not described. Interim analyses and stopping rules are also not reported.

Only one outcome measure is listed, and no secondary results are posted. The adverse-event entries show zero affected participants in the listed serious and other categories, but do not include serious event terms or explain the scope of safety reporting. These gaps mean the registered results support a focused interpretation of the mortality comparison, rather than a comprehensive appraisal of all clinical outcomes or safety evidence.

18. Related Tutorials

19. Related Statistical Calculators

20. Sources

• ClinicalTrials.gov: Studies of the Ocular Complications of AIDS (SOCA)--Monoclonal Antibody CMV Retinitis Trial (MACRT) — NCT00000135.
• Registry record last updated 2015-11-17; status last verified 2015-07.
• Results first posted 2015-11-17.
• 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.

21. Record Summary

The completed randomized, double-masked, parallel-group trial evaluated intravenous MSL-109 against placebo as adjunct therapy for CMV retinitis in people with HIV infections. Both interventions are described as infusions every 2 weeks at 60 mg. The primary outcome was mortality rate in deaths per person-year, with follow-up for a 17 month period or until a common study closing date.

The posted rate was 0.68 for MSL-109 and 0.31 for placebo. The reported rate ratio was 2.194, with a 95% confidence interval from 1.548 to 3.109 and an unadjusted p-value of <0.001. The registry record lists no secondary outcomes and provides limited detail on the statistical plan, missing data, and analysis population. The result should be interpreted as the posted primary rate comparison, with those reporting limitations in view.