A Maintenance Study With Niraparib Versus Placebo in Patients With Platinum Sensitive Ovarian Cancer: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for A Maintenance Study With Niraparib Versus Placebo in Patients With Platinum Sensitive Ovarian Cancer (NCT01847274), a randomized phase 3 study. The primary outcome is Progression-Free Survival (PFS) in Cohort With Germline BReast CAncer Gene (BRCA) Mutation (gBRCA).

2. Clinical Question

3. Trial Design

This was a phase 3, randomized, double-blind, placebo-controlled, parallel-group study. The registry describes the study as multicenter and identifies quadruple masking. The planned allocation ratio was 2:1 for niraparib and placebo. Overall enrollment was 596, and the study status is completed. The registry lists a start date of 2013-06-21 and a primary completion date of 2016-04-22.

Eligible participants had platinum-sensitive ovarian cancer and had responded to their most recent chemotherapy containing a platinum agent. The study description identifies two qualifying tumor or genetic groups: participants with a germline BRCA mutation and participants with high-grade serous histology. The primary endpoint result is specifically for the germline BRCA mutation cohort, rather than the entire enrolled population. In that cohort, the registry reports 138 participants in the niraparib arm and 65 in the placebo arm for the primary outcome.

Niraparib and placebo were administered once daily continuously during a 28-day cycle. The study description says health-related quality of life was assessed using the Functional Assessment of Cancer Therapy - Ovarian Symptom Index, the European Quality of Life scale, 5-Dimensions, and a neuropathy questionnaire. Safety and tolerability assessments included clinical review of adverse events, physical examinations, electrocardiograms and safety laboratory values.

The registry identifies progression-free survival as the primary efficacy objective. It lists additional outcomes related to subsequent therapy, chemotherapy-free interval, progression-free survival 2 and overall survival, with outcomes reported by germline BRCA status for several measures. The original statistical analysis plan, detailed allocation procedures and the trial’s own sample-size justification 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.

Before assignment: A total of 596 participants were enrolled in the study. The results presented are based on the data cut-off date of 31 March 2021 (which aligns with the time of the study unblinding) and the post-unblinding safety data until the end of study (01-April-2021 to 26-December-2021).

Only the first 6 of 11 groups are shown.

Main Study (Upto 7years 7months 6 Days)

Reasons for not completing this period:

FE Sub-study, Period1 (Day1)

Reasons for not completing this period:

FE Sub-study, Washout 1 (Up to Day 7)

Reasons for not completing this period:

FE Sub-study, Period 2 (Day 1)

Analysis population for the primary outcome, as defined in the registry: Intent-to-treat population was defined as all randomized participants with participants analyzed according to the study drug assigned via randomization

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: The baseline analysis population was the intent-to-treat population, defined as all randomized participants with participants analyzed according to the study drug assigned via randomization

Only the first 6 of 8 groups are shown.

6. Primary Endpoint

The registry lists the primary outcome as Progression-Free Survival (PFS) in Cohort With Germline BReast CAncer Gene (BRCA) Mutation (gBRCA), measured over From date of randomization to the earliest date of disease progression or death from any cause, up to 7 years 7 months and 4 days.

The primary endpoint was progression-free survival in the cohort with a germline BRCA mutation. It measures the time from randomization until the earliest of disease progression or death from any cause. A participant’s time therefore ends at whichever of those events occurs first. The registry specifies that the endpoint was assessed from randomization and gives a follow-up timeframe of up to 7 years 7 months and 4 days.

Progression was assessed using computed tomography or magnetic resonance imaging. The registry describes imaging at baseline, every 8 weeks through cycle 14, and then every 12 weeks until treatment discontinuation. Objective progression was determined through blinded central radiologic and clinical review using Response Evaluation Criteria in Solid Tumors, version 1.1. The record defines progressive disease as at least a 20% increase in the sum of target-lesion diameters, using the smallest sum on study as the reference.

General statistical context: a time-to-event endpoint uses both whether an event occurred and the time until it occurred. It can account for participants whose event has not been recorded by the end of their available follow-up, rather than treating all such participants as if they experienced an event at the same time. The median is the time by which half the population’s estimated event-time distribution has reached the event. Progression-free survival is used to describe the period without recorded progression or death; it is distinct from overall survival, which concerns death alone.

7. Results Posted in the Registry

• gBRCA Niraparib vs gBRCA Placebo: hazard ratio 0.27 (95% CI 0.17–0.41), reported p < 0.001 (as reported in the registry).

8. Secondary Efficacy Endpoints

The registry also posts results for secondary outcomes. For each outcome below, the endpoint type, the test applied to the posted summary statistics, the result and its interpretation are given in turn. Secondary endpoints are usually not adjusted for multiple testing, so they support rather than confirm the primary finding.

Time to First Subsequent Therapy in Cohort With Germline BRCA Mutation (gBRCA)

• Endpoint type: Time to event: compared as a hazard ratio.
• Test: Hazard ratio as reported by the sponsor (Log Rank); it was not recalculated.
• Result:
  • gBRCA Niraparib vs gBRCA Placebo: hazard ratio 0.57 (95% CI 0.41–0.78), reported p < 0.001 (as reported in the registry).
• Interpretation:
  • The 95% confidence interval excludes no difference (a ratio of 1), so the hazard of the event is statistically lower for gBRCA Niraparib vs gBRCA Placebo than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

Time to First Subsequent Therapy in Cohort With No Germline BRCA Mutation

• Endpoint type: Time to event: compared as a hazard ratio.
• Test: Hazard ratio as reported by the sponsor (Log Rank); it was not recalculated.
• Result:
  • Non-gBRCA Niraparib vs Non-gBRCA Placebo: hazard ratio 0.58 (95% CI 0.45–0.74), reported p < 0.001 (as reported in the registry).
• Interpretation:
  • The 95% confidence interval excludes no difference (a ratio of 1), so the hazard of the event is statistically lower for Non-gBRCA Niraparib vs Non-gBRCA Placebo than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

Chemotherapy-Free Interval in Cohort With Germline BRCA Mutation (gBRCA)

• Endpoint type: Time to event: compared as a hazard ratio.
• Test: Hazard ratio as reported by the sponsor (Log Rank); it was not recalculated.
• Result:
  • gBRCA Niraparib vs gBRCA Placebo: hazard ratio 0.39 (95% CI 0.27–0.56), reported p < 0.001 (as reported in the registry).
• Interpretation:
  • The 95% confidence interval excludes no difference (a ratio of 1), so the hazard of the event is statistically lower for gBRCA Niraparib vs gBRCA Placebo than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

The registry lists secondary outcomes covering time to first subsequent therapy, chemotherapy-free interval, progression-free survival 2 and overall survival, with several outcomes separated by germline BRCA status. It reports 48 secondary outcomes. Numerical comparisons are available in the posted results for selected measures, including time to first subsequent therapy in the germline BRCA and no-germline-BRCA cohorts, and chemotherapy-free interval in the germline BRCA cohort. Across those reported comparisons, the median estimates are longer in the niraparib groups and the reported hazard ratios are below 1.

This broad direction is consistent with the primary endpoint’s reported comparison, but it does not establish that every secondary outcome had the same pattern. The registry record does not provide numerical estimates for every listed secondary measure in the information summarized for this analysis. In particular, the outcome list includes progression-free survival 2 and overall survival by germline BRCA status, but their results should not be inferred from the selected measures with numerical estimates.

Secondary outcomes can help describe other aspects of follow-up and provide context for the primary finding. They are not interchangeable with progression-free survival: time to subsequent therapy, chemotherapy-free interval, later progression or death, and overall survival have different event definitions and clinical interpretations. Their follow-up, populations and analysis details may also differ.

General statistical context: when many outcomes are tested, the chance of at least one apparently notable result arising by chance can increase. Adjustment procedures or a prespecified hierarchy can address this concern, but multiplicity handling is not reported in the registry record. Unless the analysis plan establishes otherwise, secondary results should therefore be treated as less definitive than a prespecified primary comparison. Their p-values and estimates are informative descriptions, but they should not be read as independent confirmation of the primary finding.

Secondary outcomes registered for this trial, for reference (only those above were recalculated):

• Time to First Subsequent Therapy in Cohort With Germline BRCA Mutation (gBRCA)
• Time to First Subsequent Therapy in Cohort With No Germline BRCA Mutation
• Chemotherapy-Free Interval in Cohort With Germline BRCA Mutation (gBRCA)
• Chemotherapy-Free Interval in Cohort With No Germline BRCA Mutation
• Progression-Free Survival 2 in Cohort With Germline BRCA Mutation (gBRCA)
• Progression-Free Survival 2 in Cohort With No Germline BRCA Mutation
• Overall Survival in Cohort With Germline BRCA Mutation (gBRCA)
• Overall Survival in Cohort With No Germline BRCA Mutation

9. Statistical Interpretation

10. Statistical Methods Explained

What does a hazard ratio describe?

General statistical context: a hazard is the rate at which an event occurs among participants who have not yet experienced it, at a particular point in follow-up. A hazard ratio compares that rate between the groups across the analysis. The registry reports a hazard ratio of 0.27 for niraparib versus placebo for primary progression-free survival in the germline BRCA mutation cohort. The value is below 1, so the reported comparison is in the direction of a lower progression-or-death event rate in the niraparib group over the analyzed follow-up.

A hazard ratio is not an absolute risk, a difference in median time, or a probability that a participant will avoid progression. Its interpretation as a single summary assumes that it is reasonable to express the group contrast with one relative rate over time. The registry record reports the estimate and Log Rank method, but further details of the model and its assumptions are not reported in the registry record.

What does the Log Rank method do?

General statistical context: the Log Rank test compares the timing of events across groups while using follow-up information from participants at risk at different times. It is commonly used for time-to-event comparisons, including outcomes where participants have different lengths of observed follow-up. The registry identifies Log Rank as the method for the reported hazard-ratio analysis. Detailed implementation choices, including any stratification or covariate adjustment, are not reported in the registry record.

The reported p-value, p < 0.001, is evidence against the test’s null hypothesis under the test’s assumptions. It does not quantify the size of the treatment-group contrast and does not establish that every participant has the same event-time experience. Its meaning is best considered alongside the hazard ratio, confidence interval and median estimates.

How should medians and confidence intervals be read?

General statistical context: the median event time is the point at which the estimated event-time distribution reaches its midpoint. The primary median estimates are 21.0 months for niraparib and 5.5 months for placebo. They summarize the distributions and are not predictions for any individual participant. A 95% confidence interval describes uncertainty in an estimate under the statistical procedure: in repeated use of the same procedure under its assumptions, intervals constructed in this way would cover the underlying quantity at the stated frequency.

The registry reports a lower confidence limit of 12.9 months for the niraparib median, but an upper limit is not reported in the registry record. For placebo, the reported interval is 3.8 to 7.2 months. The reported hazard-ratio interval is 0.173 to 0.410. These intervals refer to distinct estimates and should not be treated as interchangeable. The hazard-ratio interval describes uncertainty in the relative event-rate contrast; the median intervals describe uncertainty in each group’s median time.

Why does censoring matter?

General statistical context: in time-to-event analyses, some participants may not have had the event when their observed follow-up ends. Their exact event time is therefore not known; such follow-up is generally treated as censored at the last applicable observation. Time-to-event methods can use their observed time without assuming they had an event at the end of follow-up. Valid interpretation depends on the censoring and analysis assumptions. The registry describes the endpoint and follow-up timeframe, but detailed censoring rules and missing-data procedures are not reported in the registry record.

The distinction is important because a median or hazard ratio is estimated from the timing information available, not simply from counting participants with an event at one fixed time. The registry’s results should therefore be understood as estimates over the recorded follow-up and analysis, rather than complete event histories for every randomized participant.

What do the planning illustrations represent?

General statistical context: the sample-size illustrations use the observed effect as a reference and show planning scenarios for detecting stated fractions of that effect. For the time-to-event outcomes, the displayed requirement is a total event count. It is not a count of participants in each arm. The calculations use a two-sided significance level of 0.05, equal allocation between the two arms and a normal approximation. The trial’s own original design assumptions and sample-size justification are not reported in the registry record.

The 80% power column describes a planning scenario in which a study with the listed event requirement would have an 80% chance of detecting the specified effect if that effect and the calculation assumptions were correct. The 90% column represents a higher chance of detection under the same conditions. These are conditional planning probabilities, not guarantees. Smaller effects are harder to distinguish from random variation and therefore require more events for the same target power. Because the calculations are based on the observed effect, they should not be mistaken for the assumptions used to design the trial.

11. Confidence Intervals and Precision

The primary hazard-ratio estimate is 0.27, with a 95% confidence interval from 0.173 to 0.410. The interval lies below 1 and expresses uncertainty in the relative event-rate comparison under the reported analysis. Its span makes clear that the point estimate is not exact: values across the interval are compatible with the data under the confidence-interval procedure and its assumptions. The registry reports p < 0.001 for the Log Rank analysis.

The group median estimates provide another view of the result. Niraparib’s median progression-free survival is reported as 21.0 months, with a lower confidence limit of 12.9 months; an upper confidence limit is not reported in the registry record. Placebo’s median is 5.5 months, with a 95% confidence interval from 3.8 to 7.2 months. The missing upper limit for the niraparib median means its reported interval cannot be described fully from the registry record. It should not be inferred from the hazard-ratio interval.

General statistical context: confidence intervals communicate precision for a particular estimate, not the range of outcomes for individual participants. Their width depends on the information available and the method used. The primary cohort sizes reported for this outcome are 138 in the niraparib group and 65 in the placebo group, but those counts alone do not determine the interval widths; event timing and censoring also matter in a time-to-event analysis.

The median estimates and hazard ratio answer related but different questions. The medians locate the midpoint of each group’s event-time distribution, while the hazard ratio summarizes the relative event rate over follow-up. Neither measure alone captures every feature of the distributions. The registry does not report further details needed to assess the stability of the proportional hazard interpretation or the precision of other unreported contrasts.

12. 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 illustrations are based on the observed effect and are not the trial’s original design calculations. Their stated assumptions are a two-sided significance level of 0.05, equal allocation between the two arms and a normal approximation. For time-to-event endpoints, the figures are the total number of events required, not participants per arm. The original sample-size justification and assumptions used to plan the trial are not reported in the registry record.

The 80% power column gives the event requirement for a planning scenario intended to have an 80% chance of detecting the specified effect, assuming the calculation conditions hold. The 90% power column gives the requirement for a scenario targeting a 90% chance. A higher target power generally requires more information. The rows use the observed effect and fractions of that effect; planning for a smaller effect requires more events to maintain the same power target.

These figures are planning illustrations, not observed power. They do not describe the probability that the completed trial’s result is correct, nor do they establish that the trial was designed using the observed effect. For time-to-event calculations, event accrual is the relevant quantity in the illustration; the number of participants needed to obtain those events depends on how often events occur and how long participants are followed. The registry does not report the trial’s original assumptions for translating event requirements into enrollment or follow-up targets.

The accompanying curve should be read as a visual representation of planning requirements under the stated scenarios, rather than as additional trial outcomes. Each point is conditional on the calculation assumptions and the effect size represented. The numbers are not a substitute for a protocol-based design calculation, particularly where the original design assumptions are unavailable.

13. 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

d is the total number of events needed (Schoenfeld's formula) with 1:1 allocation and proportional hazards.

Worked calculation for this trial

1. Hazard ratio HR = 0.270, so ln(HR) = -1.3093.
2. d = 4 × (1.96 + 0.842)² / [ln(0.270)]² = 18.3, rounded up to 19 events in total 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
hr <- 0.27                              # observed hazard ratio
events_needed <- function(power, frac = 1) ceiling(round(4 * (z_a + qnorm(power))^2 / log(hr^frac)^2, 6))   # Schoenfeld, 1:1 allocation
for (f in c(1, 0.75, 0.5))
  cat(sprintf("%3.0f%% of observed effect: events = %s (80%% power), %s (90%% power)\n", 100 * f,
              format(events_needed(0.80, f), big.mark = ","), format(events_needed(0.90, f), big.mark = ",")))

# Simulation: given d total events under proportional hazards, events in the treatment arm are Binomial(d, hr / (1 + hr)).
# The score (log-rank type) statistic is (D1 - d / 2) / sqrt(d / 4).
set.seed(1847274)
d    <- events_needed(0.80)
reps <- 20000
D1 <- rbinom(reps, d, hr / (1 + hr))
z  <- (D1 - d / 2) / sqrt(d / 4)
cat(sprintf("Simulated power with %s events: %.3f (%d replicates)\n", format(d, big.mark = ","), mean(abs(z) > z_a), reps))

Output from running the code

100% of observed effect: events = 19 (80% power), 25 (90% power)
 75% of observed effect: events = 33 (80% power), 44 (90% power)
 50% of observed effect: events = 74 (80% power), 99 (90% power)
Simulated power with 19 events: 0.795 (20000 replicates)

The simulation uses the fact that, with d events in total and proportional hazards, the number of events in the treatment arm is binomial. It is an approximation that conditions on the total number of events, and it should land close to the target power.

14. Group-Sequential Design Analysis

Many time-to-event trials, especially in oncology, stop or report early when interim data are overwhelming, so a single fixed-sample number understates the real design problem. The calculation below is an illustrative group-sequential design built from this trial's posted control median and hazard ratio. The registry record does not state the planned number of looks or the spending function, so the design is a planning illustration, not a reconstruction of the protocol. Assumed: one-sided α = 0.025, 90% power, three equally spaced looks on information, Lan-DeMets O'Brien-Fleming spending, 24 months of accrual, 60 months in total and a 1% monthly dropout hazard.

Under these assumptions the design needs about 30 participants in total, with the boundaries below.

Fixed design versus group-sequential alternatives

Group-sequential designs cost a small increase in maximum sample size and events compared with the fixed design, in exchange for the chance to stop early for overwhelming efficacy. O'Brien-Fleming spending keeps early boundaries very conservative and the final boundary close to the fixed-design value, whereas Pocock spending spends more alpha early.

Sensitivity to a smaller true effect

Planning for exactly the observed effect is optimistic, because observed effects in published trials tend to be somewhat larger than the true effect. The table repeats the main design (three looks, O'Brien-Fleming spending, 90% power) for smaller true hazard ratios.

For context, this trial enrolled 596 participants.

Cross-check, upper boundaries: gsDesign 3.71, 2.511, 1.993 | rpact 3.71, 2.511, 1.993. Agreement between the two packages confirms the boundary calculation.

R code and output

suppressWarnings(suppressMessages({ library(gsDesign); library(rpact) }))
med_control <- 5.5          # control-arm median, months (converted from the posted unit)
hr          <- 0.27          # hazard ratio used for planning
lambdaC     <- log(2) / med_control
eta         <- 0.01               # dropout hazard per month (assumption)
acc_months  <- 24                 # accrual duration (assumption)
study_months<- 60                 # total study duration (assumption)

design <- function(k, sfu, beta, h = hr) gsSurv(k = k, test.type = 1, alpha = 0.025, beta = beta, sfu = sfu,
  lambdaC = lambdaC, hr = h, hr0 = 1, eta = eta, gamma = 1, R = acc_months, T = study_months, minfup = study_months - acc_months)

# Main design: 3 looks, Lan-DeMets O'Brien-Fleming spending, one-sided alpha 0.025, 90% power
x <- design(3, sfLDOF, 0.10)
cat(sprintf("N_TOTAL|%s\n", format(ceiling(sum(x$gamma * x$R)), big.mark = ",")))
for (i in 1:3) cat(sprintf("GS|%d|%.1f|%.0f|%.3f|%.5f|%.5f\n", i, x$T[i], x$n.I[i], x$upper$bound[i], pnorm(-x$upper$bound[i]), cumsum(x$upper$spend)[i]))
cat("\n"); print(gsBoundSummary(x))

# Comparison of designs
cmp <- list(list("Fixed design, one analysis", 1, sfLDOF), list("Two looks, O'Brien-Fleming", 2, sfLDOF),
            list("Three looks, O'Brien-Fleming", 3, sfLDOF), list("Three looks, Pocock", 3, sfLDPocock))
cat("\n")
for (b in c(0.20, 0.10)) for (d in cmp) {
  y <- design(d[[2]], d[[3]], b)
  cat(sprintf("DES|%s|%d|%.0f|%s|%.1f\n", d[[1]], round(100 * (1 - b)), tail(y$n.I, 1), format(ceiling(sum(y$gamma * y$R)), big.mark = ","), tail(y$T, 1)))
}

# Sensitivity: what if the true effect is smaller than the observed one? (3 looks, O'Brien-Fleming, 90% power)
cat("\n")
for (fr in c(1, 0.75, 0.5)) {
  y <- design(3, sfLDOF, 0.10, exp(log(hr) * fr))
  cat(sprintf("SENS|%.2f|%.3f|%.0f|%s\n", fr, exp(log(hr) * fr), tail(y$n.I, 1), format(ceiling(sum(y$gamma * y$R)), big.mark = ",")))
}

# Independent cross-check of the boundaries with rpact
r <- getDesignGroupSequential(kMax = 3, alpha = 0.025, beta = 0.10, sided = 1, typeOfDesign = "asOF", informationRates = x$timing)
cat(sprintf("\nCross-check, upper boundaries: gsDesign %s | rpact %s\n", paste(round(x$upper$bound, 3), collapse = ", "), paste(round(r$criticalValues, 3), collapse = ", ")))

Method: LachinFoulkes 
   Analysis               Value Efficacy
  IA 1: 33%                   Z   3.7103
      N: 22         p (1-sided)   0.0001
  Events: 8        ~HR at bound   0.0715
  Month: 16    P(Cross) if HR=1   0.0001
            P(Cross) if HR=0.27   0.0338
  IA 2: 67%                   Z   2.5114
      N: 30         p (1-sided)   0.0060
 Events: 16        ~HR at bound   0.2830
  Month: 26    P(Cross) if HR=1   0.0060
            P(Cross) if HR=0.27   0.5603
      Final                   Z   1.9930
      N: 30         p (1-sided)   0.0231
 Events: 24        ~HR at bound   0.4413
  Month: 60    P(Cross) if HR=1   0.0250
            P(Cross) if HR=0.27   0.9000



Cross-check, upper boundaries: gsDesign 3.71, 2.511, 1.993 | rpact 3.71, 2.511, 1.993

15. Analyses Reported by the Sponsor

The registry reports the sponsor’s primary analysis as a Log Rank hazard ratio of 0.27, with a 95% confidence interval from 0.173 to 0.410 and p < 0.001. The posted primary result gives the same hazard-ratio estimate and interval, and no discrepancy is flagged between the posted analysis and the recalculation. Accordingly, the sponsor’s reported analysis is the central result to interpret.

A recalculation based on posted summary statistics is necessarily limited to the information in those summaries. It cannot reproduce analysis details that are not reported, such as any unlisted stratification, covariate adjustment or data-handling decisions. For time-to-event outcomes, summary medians and other aggregate figures do not contain the full event and censoring histories. The registry does not report enough detail to reconstruct every aspect of the sponsor’s analysis independently.

The planning illustrations are separate from the sponsor’s analysis. They use the observed effect under stated planning assumptions and do not show the study’s original design rationale. They should not be used to replace the reported inferential result or to characterize the trial’s achieved power.

16. Multiplicity, Missing Data and Analysis Populations

The primary endpoint is identified as progression-free survival in the germline BRCA mutation cohort, and the registry reports a Log Rank analysis. The primary population is described as intent-to-treat, with all randomized participants analyzed according to the study drug assigned through randomization. The record does not provide a full statistical analysis plan or enough detail to establish every analysis population and decision rule.

Multiplicity is relevant because the registry lists numerous secondary outcomes, including measures of subsequent therapy, chemotherapy-free interval, progression-free survival 2 and overall survival, some reported separately by germline BRCA status. The number of posted secondary outcomes is 48. Whether the sponsor used a formal testing hierarchy, adjusted significance levels, or another multiplicity procedure is not reported in the registry record. Without that information, the reported secondary p-values should not automatically be interpreted as confirmatory evidence at the same level as the primary analysis.

Missing-data and censoring procedures also affect interpretation. The primary endpoint is time-to-event, so participants without a recorded progression or death by the end of observed follow-up may contribute censored follow-up. The registry gives the endpoint definition, assessment schedule and follow-up timeframe, but detailed rules for censoring, missed assessments, withdrawal, missing covariates and sensitivity analyses are not reported in the registry record. It is not possible to determine from the registry summary how robust the estimates were to alternative assumptions about missing or incomplete observations.

The registry’s intent-to-treat description supports reading participants according to random assignment for the primary analysis. It does not, by itself, answer how participants with incomplete endpoint assessments were handled or whether additional analysis populations were used. Detailed protocol deviations, exclusions, interim analyses and the rationale for any decisions are not reported in the registry record. These omissions limit independent assessment of the full analysis strategy, without changing the numerical results that the registry does report.

17. Safety Results

The registry reports serious adverse events as affected participants divided by participants at risk.

Deaths

Other (non-serious) adverse events

Participants with at least one non-serious event at or above the registry's 5% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: thrombocytopenia (13.2%), anaemia (7.7%), urinary tract infection (6.2%), acute myeloid leukaemia (4.4%), small intestinal obstruction (3.8%).

The registry reports serious adverse events for several groups. In the germline BRCA mutation safety groups, 51 of 136 niraparib participants and 9 of 65 placebo participants were reported as affected by serious adverse events. The registry also lists serious adverse event counts for the non-germline-BRCA groups and separate study subgroups; those groups have different definitions and should not be combined casually. The number at risk is included for each group, and the serious-event count should be read against that group’s own at-risk count.

The most frequently listed serious-event terms include thrombocytopenia at 13.2%, anaemia at 7.7%, urinary tract infection at 6.2%, acute myeloid leukaemia at 4.4%, small intestinal obstruction at 3.8% and ascites at 3.8%. These are registry-reported terms and percentages; they do not provide a complete account of all adverse events or establish that a particular event was caused by treatment.

General statistical context: serious adverse event counts describe events classified as serious under the study’s reporting criteria. They are not equivalent to all adverse events, and comparisons can be affected by exposure time and group composition. The registry summary does not provide the detailed event definitions, exposure-adjusted rates or causal assessments needed for a full safety comparison. These figures should therefore be interpreted as reported counts and terms, not as a complete measure of comparative safety.

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

The registry’s primary result describes a randomized comparison of niraparib maintenance with placebo for progression-free survival in the germline BRCA mutation cohort. The reported hazard ratio is 0.27, with a 95% confidence interval from 0.173 to 0.410 and p < 0.001. The reported medians are 21.0 months for niraparib and 5.5 months for placebo. Together, these statistics show the direction and estimated magnitude of the observed group comparison over the recorded follow-up, with uncertainty represented by the confidence intervals.

They do not mean that every participant experienced the same duration of progression-free survival, or that the median is an individual prediction. A hazard ratio is a relative time-to-event summary, not an absolute probability of progression or death. The result also does not, by itself, quantify effects on quality of life, subsequent therapy, overall survival or other outcomes. Those require their own results and careful attention to their analysis populations and uncertainty.

The primary conclusion applies to the germline BRCA mutation cohort as defined in the registry, not automatically to all people with ovarian cancer or all participants with high-grade serous histology. The broader study enrolled 596 participants, but the primary comparison is reported for 138 niraparib and 65 placebo participants in the specified cohort. Generalization beyond that population requires evidence about the relevant groups and their results.

Randomization and masking are important design features, but they do not eliminate uncertainty, missing-data concerns or limitations in reporting. The registry does not report the original sample-size justification, detailed multiplicity plan or full missing-data procedures. The reported results are therefore best understood as the sponsor’s analysis as recorded, rather than a complete independent reconstruction of every analytic decision.

These findings are statistical descriptions, not medical advice. They do not determine what treatment an individual should receive. A treatment decision requires clinical context and evidence beyond a single registry comparison.

19. Why This Trial Matters Statistically

This trial illustrates how a randomized maintenance study can use a time-to-event primary endpoint to compare groups after response to platinum-containing chemotherapy. Progression-free survival combines progression and death into a defined endpoint and preserves information about when events occur. The registry also describes a blinded assessment process for progression using imaging and central radiologic and clinical review, which is relevant to how an endpoint is measured.

The statistical result is expressed through complementary summaries. The medians put the observed event-time distributions on a months scale; the hazard ratio summarizes the relative event rate during follow-up; and the confidence interval makes uncertainty explicit. Reading these measures together is more informative than focusing only on the p-value. Their different meanings also illustrate why a hazard ratio should not be translated into a median-time difference or individual prognosis.

The trial’s germline BRCA mutation subgroup provides a focused primary analysis within a broader study population that also included participants with high-grade serous histology. That distinction matters statistically: the overall enrollment is not the same as the primary analysis population, and results for one cohort should not be generalized automatically to another. The registry’s secondary outcomes extend the range of questions but also introduce multiplicity and interpretation concerns.

Finally, the record illustrates the difference between planning illustrations and a trial’s original design. Event-based calculations can show how information requirements change under different effect assumptions, but they cannot reconstruct the study’s protocol rationale when that rationale is absent. Clear reporting of analysis plans, missing-data rules and multiplicity procedures would help readers evaluate how the reported estimates were generated and how much confidence to place in secondary findings.

20. 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.

Several limitations arise from the registry information available for this trial. The primary endpoint is reported for the germline BRCA mutation cohort, rather than as a single result for all 596 enrolled participants. The primary group sizes are 138 for niraparib and 65 for placebo. The result therefore needs to be interpreted within this cohort and should not be assumed to describe every participant eligible for the broader study.

The registry reports a 95% confidence interval for the primary hazard ratio, but the upper confidence limit for the niraparib median is not reported in the registry record. That makes the reported precision for this median incomplete. The registry does not include the complete event and censoring data needed to reproduce the time-to-event analysis independently, nor does the available description specify all analysis details, such as stratification or covariate adjustment.

The original sample-size justification and the trial’s design assumptions are not reported in the registry record. The planning illustrations are based on the observed effect and stated assumptions, so they cannot be treated as the original design rationale. Detailed handling of multiplicity, missing data, interim analyses and protocol deviations is also not reported in the registry record. These gaps limit assessment of the analysis strategy, particularly for the many secondary outcomes.

The safety summary reports serious adverse event counts and selected event terms, but it does not provide all details needed for a comprehensive comparison, such as exposure-adjusted rates or full causal assessments. The record’s safety groups include subgroups beyond the primary efficacy cohort, so the group labels and denominators matter. The reported figures should not be combined across differently defined groups without a clear analysis plan.

21. Related Tutorials

22. Related Statistical Calculators

23. Sources

• ClinicalTrials.gov: A Maintenance Study With Niraparib Versus Placebo in Patients With Platinum Sensitive Ovarian Cancer — NCT01847274.
• Registry record last updated 2023-06-02; status last verified 2023-05.
• Results first posted 2019-05-01.
• 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.

24. Record Summary

The registry describes a completed, phase 3, randomized, double-blind, placebo-controlled, parallel-group study of niraparib maintenance in platinum-sensitive ovarian cancer. Overall enrollment was 596. The primary outcome was progression-free survival in the germline BRCA mutation cohort, defined as time from randomization to progression or death from any cause.

For the primary comparison, median progression-free survival was 21.0 months with niraparib and 5.5 months with placebo. The reported hazard ratio was 0.27, with a 95% confidence interval from 0.173 to 0.410 and p < 0.001 by Log Rank analysis. The registry also lists secondary outcomes and serious adverse event summaries.

The record does not report the trial’s original sample-size justification, full multiplicity procedures or detailed missing-data rules. Those omissions limit assessment of the broader analysis strategy, particularly for secondary outcomes.