Safety and Efficacy of Ranibizumab in Diabetic Macular Edema: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Safety and Efficacy of Ranibizumab in Diabetic Macular Edema (NCT01131585), a randomized phase 3 study. The primary outcome is Change in Best-Corrected Visual Acuity (BCVA) From Baseline to Month 12.

2. Clinical Question

3. Trial Design

This was a randomized, double-masked, parallel-group Phase IIIb study. The official title describes a 12-month comparison of laser photocoagulation used alongside ranibizumab intravitreal injections against laser photocoagulation monotherapy, followed by a 12-month follow-up period. The registry record lists enrollment of 128 and identifies Novartis Pharmaceuticals as the sponsor. The study status is terminated.

The two groups received active laser photocoagulation. One group also received ranibizumab 0.5 mg, while the other received sham injections. The arm descriptions specify treatment at baseline and scheduled injections at 30, 60 and 90 days, with additional treatment allowed when needed under minimum intervals. The record does not explain the practical details of masking or how the sham procedure maintained masking; those details are not reported in the registry record.

The listed start date is 2010-06, and the listed primary completion date is 2011-07. The study was designed around change in best-corrected visual acuity from baseline to month 12. However, the registry record states that no patient completed the 12-month observational period because the study was terminated early. This makes the distinction between the planned endpoint and the information available at termination especially important.

The posted arm results include 85 participants in the active laser and ranibizumab group and 43 in the active laser and sham-injection group. These are the group sizes attached to the endpoint results, not evidence that every participant completed the intended follow-up. The registry record does not report the original sample-size justification, the reasons for early termination, or the details of how randomization was generated and concealed. Those design details should not be inferred from the randomized label alone.

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: Full Analysis Set consisted of all participants who received at least one application of study treatment and had at least one post-baseline assessment for BCVA. No patient completed the 12 months observational period due to study early termination; therefore, the last observation carried forward (LOCF) method was used with data from 11.1 months.

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 Change in Best-Corrected Visual Acuity (BCVA) From Baseline to Month 12, measured over 12 months.

The primary endpoint was mean change in best-corrected visual acuity (BCVA) from baseline to month 12, measured in letters. BCVA accounts for a participant’s refraction when visual acuity is assessed. Testing used Early Treatment Diabetic Retinopathy Study-like visual acuity charts, with the participant seated at a distance of 4 meters. A higher score at month 12 relative to baseline indicates improvement in the measured visual acuity.

A mean change is a continuous outcome: it summarizes the average change in the measured score within a group. Comparing group means gives an estimate of the difference in average change between the treatment strategies. It is useful for describing the overall direction and size of change while retaining more information than a simple classification of participants as improved or not improved. It does not, on its own, show how individual participants’ outcomes varied or how many experienced a particular degree of change.

The endpoint was planned at month 12, but the record says no patient completed the 12-month observation period because of early termination. The posted population description refers to last observation carried forward, though the description is incomplete in the registry record. Therefore, the reported endpoint should not be read as a complete set of measurements collected at the planned month-12 visit. The full details of the handling of missing month-12 measurements are not reported in the registry record.

7. Results Posted in the Registry

• Active Laser Photocoagulation and Ranibizumab: mean 6.50 versus 1.40 on control; difference 5.10 (95% CI 2.25 to 7.95), standardized effect d = 0.62; z-test p = <0.001.

8. Statistical Interpretation

9. Statistical Methods Explained

What does a mean change describe?

BCVA is a continuous measurement expressed in letters. For each participant with measurements available, change is assessed relative to baseline; the group mean summarizes those changes. A mean is useful for comparing average results, but it does not show the full distribution. Participants can have changes above and below the mean, and the same mean can arise from different patterns of individual outcomes. The standard deviation gives a summary of spread around the mean: the posted values are 8.6 letters in the ranibizumab group and 7.3 letters in the sham-injection group.

The endpoint’s direction is defined in the record: a higher BCVA number at month 12 in reference to baseline indicates improved BCVA. Consequently, the positive reported mean difference of 5.1 letters indicates a higher average change in the active laser and ranibizumab arm than in the active laser and sham-injection arm. This is a comparison of group averages, not a statement about each participant or a measure of the proportion who improved.

How is the mean difference read?

The reported mean difference compares the arm mean changes: 6.5 letters in the combination arm and 1.4 letters in the comparator arm. The posted estimate is 5.1 letters. Its direction depends on the order of the groups: the combination arm is described first, followed by the comparator. The estimate is unadjusted and is based on the posted summary statistics. It does not account for covariates or any additional structure unless that is explicitly included in an analysis, and such an adjusted analysis is not reported in the registry record.

A general explanation of an unadjusted comparison is that it contrasts the observed group summaries without modifying them for other measured characteristics. In a randomized trial, random assignment is intended to support a fair comparison, but randomization does not remove the need to describe the actual analysis population, incomplete data, or the uncertainty in the estimate. The registry record does not report enough detail to evaluate adjustments, covariate handling, or the full analysis plan.

What does the confidence interval contribute?

A confidence interval is a range of values generated by a statistical procedure to express uncertainty around an estimate. For the mean difference, the posted 95% confidence interval runs from 2.253 to 7.947 letters. In general, a narrower interval indicates greater precision under the calculation’s assumptions, whereas a wider interval indicates more uncertainty. The interval is not a range containing every individual participant’s result; it concerns the estimated difference in group means.

The interval is above zero, which is consistent with a positive difference under the unadjusted calculation. A confidence interval does not correct for bias from missing outcome data, early termination, or selection into the analysis population. Its interpretation also depends on the assumptions behind the calculation. It should therefore be considered together with the incomplete month-12 follow-up described in the registry record.

What does the p-value mean?

The reported unadjusted p-value is p < 0.001. In general, a p-value describes how unusual a test statistic at least as extreme as the observed one would be under a specified null model. It is not the probability that the null hypothesis is true, and it does not measure the size or clinical importance of a difference. Nor does it tell readers whether missing data or early termination have biased the estimate.

The registry record lists zero analyses, so the p-value should be understood as belonging to the unadjusted calculation rather than a documented sponsor analysis. The record does not report the sponsor’s full inferential method, any prespecified testing hierarchy, or a multiplicity procedure. The p-value alone cannot answer whether those considerations were addressed in the trial’s original analysis.

Why does over-dispersion matter?

General statistical context: for a continuous endpoint, the spread of individual observations matters because greater variability makes a group mean less precise when other features are held constant. The posted standard deviations describe variation in BCVA change, but the available record does not provide individual-level observations or distributions. That limits assessment of departures from the assumptions used in a simple mean comparison. The term over-dispersion is more often used for count data; it refers generally to variability greater than a model expects. It is not an identified finding for this continuous endpoint.

What is important about the analysis population?

The endpoint description defines the Full Analysis Set as participants who received at least one application of study treatment and had at least one post-baseline BCVA assessment. This definition indicates that the endpoint summary was not necessarily based on every randomized participant. The record does not provide a separate count of participants with a completed month-12 measurement, nor a full account of exclusions from the analysis population.

The endpoint description states that no patient completed the 12-month observational period because of early termination, then begins to describe last observation carried forward. This approach generally uses an earlier measurement in place of a missing later measurement, but its use can rely on assumptions about what would have happened after the last observed assessment. Here, the description is truncated, and the full implementation, timing of carried-forward values, and sensitivity analyses are not reported in the registry record. These uncertainties matter because the planned outcome is explicitly a month-12 change.

The record also does not report an interim-analysis plan, a stopping rule, or a reason for termination. Those details should not be inferred from the status alone. In general, early termination and missing endpoint observations can affect precision and interpretation even when an unadjusted comparison produces a small p-value. The posted analysis therefore needs to be read as a limited summary of the data available under the recorded analysis population and incomplete follow-up.

10. Confidence Intervals and Precision

The reported mean difference in BCVA change is 5.1 letters. Its 95% confidence interval extends from 2.253 to 7.947 letters. A confidence interval expresses uncertainty around the estimated group difference under the statistical procedure used; it does not describe the range of individual participants’ changes. In general, the farther the interval’s endpoints are from one another, the less precisely the mean difference is estimated.

Both endpoints of this interval are above zero. Under the unadjusted comparison, the interval is therefore consistent with a positive average difference. The reported p-value is p < 0.001, also from an unadjusted comparison. These statistical summaries address sampling uncertainty under their calculation assumptions. They do not account for unreported features of the original analysis, and they do not resolve uncertainty introduced by missing month-12 measurements.

The registry record says that no patient completed the 12-month observational period because of early termination. It identifies a Full Analysis Set based on treatment received and at least one post-baseline assessment, and its endpoint description begins to refer to last observation carried forward. The complete missing-data description is not reported in the registry record. Thus, the confidence interval should not be read as capturing all uncertainty about how the observed data relate to the intended 12-month endpoint.

The posted standard deviations, 8.6 and 7.3 letters, describe variability within the groups. The confidence interval concerns the estimated difference between group means, not the spread of individual outcomes. No adjusted confidence interval or sponsor analysis is reported in the registry record. Precision in the unadjusted calculation is only one part of judging the result; incomplete follow-up and analysis choices remain relevant limitations.

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, not total events. They are illustrations based on the observed mean difference, not the trial’s original design assumptions. The stated planning assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. The effect to detect is set to the stated fraction of the observed effect. The trial’s own original sample-size assumptions are not reported in the registry record.

For planning to detect 100% of the observed effect, the display gives 41 participants per arm for 80% power and 55 participants per arm for 90% power. For planning to detect 75% of the observed effect, it gives 72 participants per arm for 80% power and 97 participants per arm for 90% power. For planning to detect 50% of the observed effect, it gives 162 participants per arm for 80% power and 217 participants per arm for 90% power.

In general, power is the chance that a specified statistical test will detect an effect of a chosen size if that effect is present, under the assumptions used in the calculation. The 80% and 90% columns represent different target probabilities for detecting the specified effect. A higher target power generally requires more participants, and smaller effects need more participants to detect than larger effects, other assumptions being equal. The curve should be read as a visual representation of the same planning relationship, not as an additional estimate of what happened in this trial.

These figures are planning illustrations, not observed power. They do not establish the trial’s actual power, enrollment target rationale, or the adequacy of the original design. Their interpretation is also limited by reliance on the observed effect as the reference for the planning scenarios.

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, Δ is the difference in means to detect and σ is the common standard deviation.

Worked calculation for this trial

1. Observed difference Δ = 5.100; pooled standard deviation σ = 8.190.
2. n = 2 × (1.96 + 0.842)² × 8.190² / 5.100² = 40.5, rounded up to 41 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
delta <- 5.1                      # observed difference in means
sigma <- 8.18962758616044                     # pooled standard deviation
n_per_arm <- function(power, frac = 1) {
  ceiling(round(2 * (z_a + qnorm(power))^2 * sigma^2 / (delta * frac)^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 = ",")))

# Check against R's built-in function (t-based, so very slightly larger)
cat("power.t.test:", round(power.t.test(delta = delta, sd = sigma, power = 0.80)$n, 1), "\n")

# Simulation: exact two-sample t-test using simulated means and variances
set.seed(1131585)
n    <- n_per_arm(0.80)
reps <- 20000
m_diff <- rnorm(reps, delta, sigma * sqrt(2 / n))             # difference in sample means
s2     <- sigma^2 * rchisq(reps, 2 * (n - 1)) / (2 * (n - 1))   # pooled sample variance
tstat  <- m_diff / sqrt(s2 * 2 / n)
cat(sprintf("Simulated power at n = %s per arm: %.3f (%d replicates)\n", format(n, big.mark = ","), mean(abs(tstat) > qt(0.975, 2 * (n - 1))), reps))

Output from running the code

100% of observed effect: n per arm = 41 (80% power), 55 (90% power)
 75% of observed effect: n per arm = 72 (80% power), 97 (90% power)
 50% of observed effect: n per arm = 162 (80% power), 217 (90% power)
power.t.test: 41.5 
Simulated power at n = 41 per arm: 0.794 (20000 replicates)

The simulation draws many sets of group means and variances at the planned sample size and counts how often the two-sample t-test rejects the null. It should land close to the target power; at very small required sample sizes it can fall below it, because the normal approximation is rough there.

13. Multiplicity, Missing Data and Analysis Populations

The registry record lists one outcome measure and no posted analyses. The primary outcome is change in BCVA from baseline to month 12. No secondary outcomes are posted. A multiplicity strategy for handling several endpoints, comparisons, or analysis time points is not reported in the registry record. The absence of multiple posted outcomes does not establish what was included in the original protocol or statistical analysis plan.

The record does not report an interim-analysis plan, a stopping rule, or a statistical rationale for termination. The study status is terminated, and the endpoint description says that no patient completed the 12-month observation period because of early termination. The reason for termination and its timing relative to review of accumulating results are not reported in the registry record. No assumptions about interim monitoring or stopping decisions should be made from the status label alone.

The stated Full Analysis Set included participants who received at least one application of study treatment and had at least one post-baseline BCVA assessment. This is the only analysis-population definition described for the primary endpoint. A separate intention-to-treat analysis population, per-protocol population, or safety analysis population definition is not reported in the registry record. The posted arm counts for the endpoint are 85 and 43, but the record does not give a full participant flow or explain how those endpoint counts relate to all randomized participants.

The endpoint description begins to identify last observation carried forward after stating that no participant completed the 12-month observation period. The full description is truncated in the registry record. The exact timing of the last observations, the missing-data rules, and any sensitivity analyses are not reported. It is also not reported how many endpoint measurements were missing at each time point or how withdrawals and other unavailable assessments were handled.

Missing-data assumptions are important because a month-12 endpoint may require information that was not collected at month 12. In general, carrying an earlier observation forward treats that observation as a substitute for the later outcome; this can affect estimates if outcomes would otherwise change. The posted confidence interval and p-value do not by themselves assess the consequences of that assumption.

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 5% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: retinal detachment (fellow eye) (4.7%), hypoglycaemia (2.4%), coronary artery disease (2.3%), coronary artery stenosis (2.3%), vitreous haemorrhage (fellow eye) (2.3%).

The registry record lists serious adverse events in 14 of 85 participants in the active laser and ranibizumab group, and in 5 of 43 participants in the active laser and sham-injection group. Other adverse events are listed for 32 of 85 participants in the ranibizumab group and 10 of 43 participants in the sham-injection group. These are counts of participants affected among the respective groups at risk, as recorded; they are not a comparison adjusted for exposure time or other factors.

Among the listed serious adverse-event terms, the record includes retinal detachment (fellow eye) at 4.7%, hypoglycaemia at 2.4%, coronary artery disease at 2.3%, coronary artery stenosis at 2.3%, vitreous haemorrhage (fellow eye) at 2.3%, and concomitant disease progression at 2.3%. The record does not provide enough detail in these entries to infer the timing, duration, severity beyond the serious classification, or relationship of each event to study treatment.

Deaths are not reported in the registry record. The record does not state a definition of serious adverse events in the posted summary, provide exposure-adjusted rates, or give a complete account of event timing. Group counts alone should not be interpreted as establishing a difference in safety between the arms.

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

The posted primary-outcome summaries show a higher mean change in BCVA in the active laser and ranibizumab arm than in the active laser and sham-injection arm. The mean changes are 6.5 and 1.4 letters, respectively, and the unadjusted mean difference is 5.1 letters. Its 95% confidence interval is 2.253 to 7.947, with p < 0.001. These are statistical descriptions of the posted results for the endpoint, not a complete account of every planned analysis.

The result does not establish what would have been observed if all participants had completed the planned 12-month observation period. The registry record says no patient completed that period because the study ended early. It begins to describe last observation carried forward, but the missing-data details are incomplete. Consequently, uncertainty from incomplete follow-up and the method used to represent unavailable month-12 measurements is not fully characterized by the confidence interval.

The average difference also does not mean that every participant in one group changed by the same amount, or that every participant in that group had a higher score than a participant in the other group. Group means and their standard deviations summarize different features of the outcome. The record does not post individual-level results, responder counts, a subgroup analysis for proliferative diabetic retinopathy, or secondary outcome results.

The unadjusted p-value describes the comparison under its statistical assumptions; it is not the probability that a treatment effect is absent, and it does not measure clinical importance. The registry record does not report an adjusted sponsor analysis, multiplicity plan, or full statistical analysis plan. The observed estimates should therefore be interpreted in the context of the available summary data and the trial’s limitations, rather than as definitive evidence about all patients or all follow-up circumstances.

In general, randomized comparisons aim to assess differences between assigned treatment strategies. Here, early termination and incomplete endpoint follow-up constrain how closely the posted results can answer the planned month-12 question. The findings describe the available registry results; they do not by themselves establish a broader treatment conclusion.

16. Why This Trial Matters Statistically

This trial illustrates why a reported treatment-group difference cannot be interpreted separately from how the endpoint was collected. The primary measure was planned at month 12, but the registry record states that no patient completed the 12-month observational period. That makes the distinction between a planned time point and the information actually available especially important for interpreting the posted mean changes.

It also illustrates the difference between a result’s statistical precision and its completeness. The unadjusted mean difference has a confidence interval from 2.253 to 7.947 and a reported p-value of p < 0.001. Those statistics summarize uncertainty under the calculation’s assumptions. They do not explain the effects of early termination, establish the missing-data procedure, or replace a full prespecified analysis.

The record’s analysis population is defined by treatment received and having at least one post-baseline BCVA assessment. That definition, together with the incomplete account of last observation carried forward, raises practical questions about which measurements contributed to the month-12 result and how unavailable assessments were handled. These are not minor technical details: the endpoint is a change over time, and the reported result depends on what information represents that change.

Finally, the sample-size illustrations show how planned participant numbers depend on the effect size selected and the desired power, under stated assumptions. They are not evidence of the original design rationale or of the trial’s achieved power. The registry record does not report the original sample-size justification or the full analysis plan. Together, these features make the study a useful example of reading effect estimates, uncertainty, missing-data information, and design context as connected parts of a clinical-trial result.

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.

Early termination is a central limitation: the registry record says no patient completed the planned 12-month observational period. The reason for termination and the circumstances around the decision are not reported in the registry record. The posted month-12 outcome therefore should not be assumed to represent complete measurements at the planned visit.

The endpoint description identifies last observation carried forward, but the text is incomplete. Details of implementation, the timing of available observations, and any sensitivity analyses are not reported. Without those details, the influence of missing data on the estimated mean changes and confidence interval cannot be assessed from the registry summary alone.

The registry record lists the Full Analysis Set definition but does not provide participant flow, counts of missing BCVA measurements by visit, or a detailed account of withdrawals. The posted endpoint arm sizes are 85 and 43; the record does not explain all movement from randomization to the endpoint analysis. A complete protocol, statistical analysis plan, or sponsor analysis is not reported.

The primary comparison presented is an unadjusted recalculation from posted summary statistics. No additional adjusted analyses are listed. The record does not report the original sample-size justification, a multiplicity plan, interim analyses, stopping rules, or an analysis of the proliferative diabetic retinopathy subgroup. These omissions limit assessment of how the reported comparison relates to the planned analysis and any additional questions the trial intended to address.

The safety summary provides counts and selected serious-event terms, but deaths are not reported in the registry record, and event timing, exposure-adjusted rates, and treatment attribution are not described in the posted summary. These limitations mean that the adverse-event counts should be read descriptively, not as a complete comparative safety analysis.

18. Related Tutorials

19. Related Statistical Calculators

20. Sources

• ClinicalTrials.gov: Safety and Efficacy of Ranibizumab in Diabetic Macular Edema — NCT01131585.
• Registry record last updated 2012-08-30; status last verified 2012-08.
• Results first posted 2012-08-17.
• 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 registry record describes a randomized, double-masked, parallel-group Phase IIIb study of active laser photocoagulation with ranibizumab 0.5 mg versus active laser photocoagulation with sham injections in people with visual impairment due to diabetic macular edema. It lists enrollment of 128 and identifies Novartis Pharmaceuticals as sponsor. The study status is terminated.

The primary endpoint was mean change in BCVA from baseline to month 12, measured in letters. The posted mean changes were 6.5 in the ranibizumab group and 1.4 in the sham-injection group. The unadjusted mean difference was 5.1 letters, with a 95% confidence interval from 2.253 to 7.947 and p < 0.001.

The record states that no patient completed the 12-month observation period because of early termination. It identifies a Full Analysis Set and begins to describe last observation carried forward, but the missing-data description is incomplete. The original sample-size justification, full analysis plan, and reason for termination are not reported in the registry record.