Comparison of IV Topotecan/Docetaxel to Docetaxel Alone in Second-Line Stage IIIB/IV Non-Small Cell Lung Cancer: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Comparison of IV Topotecan/Docetaxel to Docetaxel Alone in Second-Line Stage IIIB/IV Non-Small Cell Lung Cancer (NCT00065182), a randomized phase 2 study. The primary outcome is Median Time of Overall Survival.

2. Clinical Question

The central question was whether the posted overall-survival results differed between the two randomized regimens in this population. The combination arm had a reported median of 30.7 weeks, with a 95% confidence interval from 26.1 to 35.9 weeks. The docetaxel arm had a reported median of 28.6 weeks, with a 95% confidence interval from 23.6 to 36.4 weeks. The registry record also gives a sponsor-reported hazard ratio of 1.007, with a 95% confidence interval from 0.813 to 1.248 and p = 0.946 for the log-rank analysis.

The statistical question is narrower than whether either regimen is appropriate for an individual. These results describe a comparison within the randomized trial population. They do not, by themselves, establish that the regimens are equivalent, or resolve how the results apply to people outside the recorded eligibility context. The registry record does not report a prespecified equivalence or non-inferiority question.

3. Trial Design

The registry record describes a randomized, parallel-group study with no masking. Participants were assigned to IV topotecan plus IV docetaxel or IV docetaxel alone. Enrollment was 399, with 202 participants in the combination arm and 197 in the docetaxel arm for the primary outcome. The registered population was people with advanced non-small cell lung cancer who had received one previous chemotherapy, excluding TAXOTERE or HYCAMTIN.

The brief study description states that the purpose was to compare efficacy and safety of a weekly combination regimen with one drug alone. The official title calls the study an open-label multicenter randomized phase III trial, while the phase entry in the registry record is PHASE2. The record does not explain this difference. No further details about the number or location of study centers are reported in the registry record.

The listed primary outcome is median time of overall survival, in weeks, with a stated timeframe of up to one year from Day -1 (randomization). Overall survival is defined as time from randomization to death. The endpoint description says that it occurs when all randomized participants had at least one year of follow-up past their date of randomization to treatment. The defined analysis population is intent-to-treat, including all randomized participants.

The registry record lists secondary outcomes covering one-year survival, time to progression, response, response duration, time to response, quality of life, and adverse events and hematologic toxicities. It reports 11 outcome measures and 10 secondary outcomes, but does not explain that difference. Detailed results for those secondary outcomes are not reported in the registry record. The study status is completed. Its recorded start date is 2003-08-14 and its primary completion date is 2007-08-30. Information about the original sample-size justification, interim analyses, or the full statistical analysis plan is not reported in the registry record.

4. Participant Flow and Accounting

The registry reports how participants moved through each stage of the study. All counts below are exactly as posted; missing counts are not inferred from other stages.

Before assignment: Based on the interim analysis results, the independent data monitoring committee (IDMC) recommended termination of the combination group due to increased toxicity as well as a low likelihood of demonstrating an overall survival benefit at the final analysis.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Intent-to-treat (ITT) population was defined as all participants randomized to one of the two treatment regimens.

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 Median Time of Overall Survival, measured over Up to one year from Day -1 (randomization).

Overall survival measures the time from randomization until death. The registry record reports its median in weeks. A median is the point in the observed time distribution at which half of the participants have experienced the event and half have not yet experienced it, under the relevant time-to-event analysis. This is a general statistical description; the registry definition specifically identifies death as the event and randomization as the starting point.

Time-to-event endpoints are useful when participants can have different lengths of follow-up and when the timing of an event matters. A median gives a time-based summary that is often easier to interpret than an event rate alone. It does not describe the full survival experience: participants' event times may vary, and the median does not show the shape of the survival distribution or the proportion alive at every time point.

The recorded timeframe is up to one year from Day -1 (randomization), and the outcome description says that all randomized participants had at least one year of follow-up past randomization. The reported medians were 30.7 weeks for IV topotecan plus IV docetaxel and 28.6 weeks for IV docetaxel. The record also reports a hazard ratio, which summarizes a relative time-to-event comparison using a different scale from the medians. The endpoint definition and the hazard-ratio result answer related, but not identical, statistical questions.

7. Results Posted in the Registry

• IV Topotecan + IV Docetaxel vs IV Docetaxel: hazard ratio 1.01 (95% CI 0.81–1.25), reported p = 0.946 (as reported in the registry).
• IV Topotecan + IV Docetaxel vs IV Docetaxel: hazard ratio 0.98 (95% CI 0.79–1.21) (as reported in the registry).

8. Statistical Interpretation

9. Statistical Methods Explained

Time-to-event analysis concerns both whether an event occurs and when it occurs. For this trial, the event was death, and the time origin was randomization. Participants may have different follow-up times, so an analysis of survival time uses information about event timing and follow-up rather than reducing every participant to a simple event/no-event outcome at one fixed time. This is a general explanation of the method class; the registry record does not describe all implementation details.

What does a median survival time describe?

A median is a location summary of the survival-time distribution. It identifies the time at which the estimated survival distribution reaches the halfway point. In this record, the group medians are 30.7 weeks for the combination and 28.6 weeks for docetaxel alone. A median is not the average survival time, and it does not tell the reader how survival differs at all other times. Two groups can have similar medians while having different patterns at other points in follow-up, or have different medians while their uncertainty remains substantial.

The record states that the primary outcome was measured up to one year from Day -1 (randomization). It also says that the overall-survival outcome occurs when all randomized participants have at least one year of follow-up past randomization. Those descriptions set the stated follow-up context, but they do not provide a full account of how censoring or incomplete follow-up was handled in the analysis. Those details are not reported in the registry record.

What does the hazard ratio add?

A hazard is the instantaneous event rate among participants who have not yet experienced the event, conditional on remaining under observation and event-free up to that time. A hazard ratio compares the hazards between groups. Under a proportional-hazards interpretation, the ratio is treated as a relative contrast that is constant over time. This is a general explanation; the registry record does not report whether proportional hazards were assessed or whether the assumption was considered appropriate.

The sponsor-reported log-rank analysis gives a hazard ratio of 1.007, a 95% confidence interval from 0.813 to 1.248, and p = 0.946. The log-rank test is a standard method for comparing time-to-event distributions between groups. It uses the ordering of events over follow-up, rather than comparing only the reported medians. The record identifies the log-rank method for this analysis. It does not provide further information about test details, stratification, adjustments, or diagnostics.

A separate reported hazard ratio is 0.977, with a 95% confidence interval from 0.788 to 1.210. The method and p-value for that estimate are not reported in the registry record. It should therefore not be assigned the log-rank method or a significance result by inference. The two confidence intervals both include 1, but they are separate reported estimates and the record does not explain their analytical relationship.

How should the p-value be understood?

A p-value measures how incompatible the observed data are with a specified null model, under the assumptions of the test. It is not the probability that the null hypothesis is true, nor does it measure clinical importance. The reported p = 0.946 belongs to the log-rank analysis. It does not show that the groups are exactly alike. Interpretation should consider the estimated effect and its confidence interval, not the p-value in isolation.

The registry record labels the analysis as a superiority analysis. A superiority test assesses evidence against a no-difference null hypothesis, but failure to reject that null does not establish equivalence. Equivalence and non-inferiority require prespecified margins and corresponding analyses. No such margin is reported in the registry record.

Why can over-dispersion matter in other outcomes?

Over-dispersion is a general issue in count-data analysis: observed variability can exceed what a simple model assumes. When that happens, standard errors based on the narrower variance assumption may be too small, which can make confidence intervals too narrow and tests too optimistic. The primary endpoint in this trial is time-to-event, not a count outcome. The registry record does not report an over-dispersion analysis for the primary endpoint, and no such analysis should be presumed.

The posted primary population is intent-to-treat, defined as all randomized participants. This generally preserves the randomized group comparison by analyzing participants according to assigned group, although the registry's description does not state how every practical complication was handled. The record does not report missing-data procedures, censoring rules, covariate adjustment, interim analyses, or the original statistical analysis plan. Those omissions mean that the posted methods can be described only to the level explicitly recorded: a median overall-survival endpoint, a log-rank comparison with a reported hazard ratio and p-value, and a second hazard-ratio estimate without a listed method or p-value.

10. Confidence Intervals and Precision

A confidence interval describes a range of values compatible with the data and the statistical procedure, under its assumptions. It conveys uncertainty around an estimate; it is not a range containing a fixed proportion of individual participants' survival times. The 95% intervals for the group medians are 26.1 to 35.9 weeks for IV topotecan plus IV docetaxel and 23.6 to 36.4 weeks for IV docetaxel. Each interval indicates uncertainty in its corresponding median estimate.

The log-rank hazard ratio is 1.007, with a 95% confidence interval from 0.813 to 1.248. The interval includes 1 and extends below and above it, so the estimate is compatible with relative hazards in either direction under the reported analysis. The second hazard ratio, 0.977, has a 95% confidence interval from 0.788 to 1.210, which also includes 1. These intervals do not prove that the true hazard ratio is exactly 1. Rather, they show that the data and procedure do not isolate a single precise value.

Precision and statistical significance are related but distinct. A narrow interval indicates more precision than a broad interval, while the position of an interval relative to a null value bears on a conventional significance interpretation. Here, the intervals leave room for effects in either direction. Their width is relevant when judging what magnitude of difference the trial can distinguish, and a non-significant result does not itself indicate that clinically meaningful differences have been ruled out. No clinical equivalence margin is reported, so the confidence intervals cannot be interpreted as an equivalence test.

Confidence intervals also depend on the analysis assumptions and the completeness of the reported methods. Details about censoring, missing data, proportional hazards, and the second hazard-ratio method are not reported in the registry record. That limits a fuller assessment of precision. The appropriate reading is therefore to retain the estimates and intervals as posted, while recognizing that they express statistical uncertainty rather than certainty about the underlying treatment contrast.

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 event counts, not participants per arm. Each value is the total number of events used in a time-to-event planning illustration. This differs from a design display for a non-time-to-event outcome, which may express participants per arm. The two sets of event-count figures correspond to the two reported hazard-ratio estimates: 1.007 and 0.977. Their planning values differ because the effect estimate used as the basis for planning differs.

The illustrations assume a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. The effect to detect is set at 100%, 75%, or 50% of the observed effect. The trial's own original design assumptions and sample-size justification are not reported in the registry record; these are planning illustrations based on the observed effect, not a reconstruction of the original protocol.

The 80% power column represents the event count that the planning calculation associates with an 80% chance of rejecting the null hypothesis if the specified effect and assumptions are correct. The 90% power column uses a higher target chance of detecting that effect under those same planning assumptions. These are design targets, not probabilities that the posted result is correct and not a measure of the power actually achieved by the completed trial.

For each reported hazard ratio, the planning display gives separate event totals for detecting 100%, 75%, and 50% of the observed effect, at both power targets. Smaller fractions of the observed effect mean smaller differences to distinguish from the null. Smaller effects generally require more events, and therefore may require more participants or longer follow-up, depending on event occurrence and study design. The plotted planning curve should be read as showing this relationship across the stated effect fractions; it is not an observed outcome curve. This is not observed power.

The totals can appear large because the calculations concern total events for time-to-event planning, not a per-arm headcount. They should not be compared directly with the enrollment of 399 as if they represented the number of people actually randomized. The planning display is an illustration of assumptions and sensitivity to effect size, not evidence that the trial had a particular achieved 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

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 = 1.007, so ln(HR) = 0.0070.
2. d = 4 × (1.96 + 0.842)² / [ln(1.007)]² = 645,212.6, rounded up to 645,213 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 <- 1.007                              # 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(65182)
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 = 645,213 (80% power), 863,757 (90% power)
 75% of observed effect: events = 1,147,045 (80% power), 1,535,568 (90% power)
 50% of observed effect: events = 2,580,851 (80% power), 3,455,027 (90% power)
Simulated power with 645,213 events: 0.800 (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.

13. Analyses Reported by the Sponsor

The registry record reports two sponsor analyses of the primary time-to-event comparison. One is identified as a log-rank analysis: its hazard ratio is 1.007, its 95% confidence interval is 0.813 to 1.248, and its p-value is 0.946. The other reported hazard ratio is 0.977, with a 95% confidence interval of 0.788 to 1.210; a method and p-value are not reported for that estimate.

The posted estimates are sponsor-reported analyses, rather than an independent re-estimation of the hazard ratio from participant-level data. The planning calculations are separate illustrations based on the observed effects; they do not replace the sponsor's primary analysis. The registry record does not explain why the two hazard-ratio estimates differ or provide enough information to reconstruct the second analysis. They should therefore remain separately identified and should not be blended or treated as interchangeable.

No discrepancy between a sponsor result and a recalculated primary estimate is flagged in the registry record. Accordingly, interpretation centers on the sponsor-reported results and their stated uncertainty. Where an analysis method or p-value is absent, that absence remains important: it cannot be filled in from the other analysis. The posted log-rank result does not establish equivalence, and the second estimate cannot be given a significance interpretation without a reported test or p-value.

14. Multiplicity, Missing Data and Analysis Populations

The registry record lists several secondary outcome areas, including one-year survival, time to progression, response rate and duration, time to response, quality of life, adverse events, serious adverse events, and hematologic toxicities. When a study assesses multiple outcomes or makes multiple statistical comparisons, the chance of at least one nominally significant result can increase. This is a general statistical concern known as multiplicity. Adjustments or a prespecified hierarchy can help control how multiple tests are interpreted, but the registry record does not report whether any multiplicity strategy was planned or applied.

The primary outcome is identified as median time of overall survival, and the primary analysis population is defined as intent-to-treat: all participants randomized to either regimen. The record identifies a log-rank analysis and reports a hazard ratio, confidence interval, and p-value. It does not report whether the primary analysis was adjusted for any covariates, whether there were stratification factors, or how secondary outcomes were analyzed. The statistical analysis plan is not reported in the registry record.

Missing-data handling and censoring rules are also not reported. For a time-to-event endpoint, the treatment of participants whose event time is not observed during follow-up is important to interpreting the analysis. The endpoint description states the follow-up timeframe and defines death as the event, but does not give operational censoring details. The record does not report the number of participants with incomplete endpoint information or the reasons for any such incompleteness.

The registry record does not describe interim analyses, stopping rules, or any adjustment for repeated looks at accumulating data. It also does not report analysis populations beyond the stated intent-to-treat population for the primary endpoint. These are limits on what can be inferred from the record, not evidence that a particular procedure was or was not used. Consequently, the reported p-value should be interpreted as the result of the identified log-rank analysis, without assuming unreported multiplicity control or additional analysis procedures.

15. 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: febrile neutropenia (3.6%), pneumonia (3.5%), anaemia (3.0%), dehydration (3.0%), dyspnoea (2.5%).

The safety results report serious adverse events among 200 participants at risk in the IV topotecan plus IV docetaxel group and 195 at risk in the IV docetaxel group. Serious adverse events affected 58 participants in the combination group and 49 in the docetaxel group. The record also reports other adverse events among 181 of 200 at risk in the combination group and 166 of 195 at risk in the docetaxel group. These are counts of participants, not event totals, and the serious-event counts should not be confused with the separate count of other adverse events.

The listed serious adverse-event terms and percentages include febrile neutropenia at 3.6%, pneumonia at 3.5%, anaemia at 3.0%, dehydration at 3.0%, dyspnoea at 2.5%, and pulmonary embolism at 2.1%. The record does not state in this summary whether these listed terms are exhaustive, how events were attributed, or the full event-specific denominators and timing. Counts and percentages describe what was recorded; they do not establish that a drug caused an event.

Deaths are also reported in the safety information: 170 in the combination group and 165 in the docetaxel group. These figures are not identified as treatment-related deaths in the record, so they should not be interpreted as such. The adverse-event figures are descriptive safety data and do not, by themselves, establish a comparative safety effect. The record does not provide a statistical comparison of these safety counts in the information.

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

The primary outcome results describe overall survival in a randomized comparison of IV topotecan plus IV docetaxel with IV docetaxel alone among participants with pretreated advanced non-small cell lung cancer. The reported medians were 30.7 weeks and 28.6 weeks, respectively. The sponsor-reported log-rank hazard ratio was 1.007, with a 95% confidence interval from 0.813 to 1.248 and p = 0.946. The second reported hazard ratio was 0.977, with a 95% confidence interval from 0.788 to 1.210, but no method or p-value was reported for it.

Statistically, the log-rank analysis did not identify a significant difference under its reported test. Its confidence interval includes values on either side of 1, so the data do not pinpoint the relative hazard to one direction. This is not proof that the regimens have identical effects. Nor does a non-significant result show that a difference of meaningful size has been excluded. The registry record does not report an equivalence margin or a non-inferiority design that would support either conclusion.

The result also does not establish how treatment effects vary among participant subgroups, because subgroup results are not reported in the registry record. The posted primary analysis applies to the intent-to-treat population as defined in the record. It should not be extended automatically to populations or treatment contexts that are not described in the eligibility summary. The record describes participants with advanced disease and one previous chemotherapy, excluding TAXOTERE or HYCAMTIN; the findings do not directly answer questions beyond that recorded context.

Safety counts show recorded adverse events and serious adverse events in each group, but they do not establish causation or a complete comparative safety profile. Secondary outcome areas are listed, but detailed results for those outcomes are not reported in the registry record. Overall, the appropriate statistical conclusion is limited to the posted estimates, their intervals, the stated test result, and the recorded safety counts. Clinical decisions require considerations beyond this statistical description.

17. Why This Trial Matters Statistically

This trial illustrates why a time-to-event result should be read through more than one statistic. The medians provide an interpretable time summary for each group, while the hazard ratio and log-rank test compare event timing across follow-up. These measures are related but not interchangeable. A reader who focuses only on the two median values may miss the uncertainty in the comparison, while a reader who focuses only on a p-value may overlook the range of effects compatible with the confidence interval.

The reported log-rank hazard ratio of 1.007 is close to the no-relative-hazard value of 1, but its 95% confidence interval, 0.813 to 1.248, includes values on both sides. That combination illustrates the distinction between an estimate and its precision: the point estimate alone is not the full result. The second reported hazard ratio, 0.977, has a similarly two-sided interval, 0.788 to 1.210, and lacks a reported method and p-value. Keeping those details visible prevents an unsupported interpretation of the second estimate.

The sample-size illustrations add a separate lesson. Event-based planning for time-to-event outcomes is not the same as a count of randomized participants, and assumptions about the effect to detect affect the required event total. The illustrations use fractions of the observed effect and stated power targets; they are not evidence of the trial's achieved power or its original design rationale.

Finally, a non-significant superiority test does not answer an equivalence question. A study can fail to find a difference while leaving uncertainty about the magnitude of effects. The absence of a reported equivalence margin, full analysis plan, and details on missing data means that statistical restraint is essential. The record supports interpretation of the posted results, but not claims beyond the analyses and outcomes it reports.

18. Limitations

• The posted summary statistics are the only data these results rest on; patient-level data, full protocols and statistical analysis plans are not available.
• Recalculated results are simple, unadjusted calculations from posted summaries. They do not replace the sponsor's prespecified analysis.

The registry record gives limited detail about the statistical analysis. It identifies the primary endpoint, the intent-to-treat population, and a log-rank analysis for one hazard-ratio result, but does not report the full analysis plan, original sample-size justification, interim analyses, censoring rules, or missing-data procedures. The method and p-value for the second hazard-ratio estimate are also not reported. This prevents a complete assessment of how the analysis was specified and implemented.

There is a difference between the phase entry and the official title: the phase is listed as PHASE2, while the official title describes a phase III trial. The registry record does not explain the difference. The record also reports 11 outcome measures and 10 secondary outcomes without explaining the difference. Those inconsistencies should be noted rather than resolved by assumption.

The trial is described as open-label, with no masking. This is a design feature of the recorded study. The primary outcome is time to death, but no details about blinded endpoint review or other protections against bias are reported in the registry record. The description identifies randomization and a parallel-group model, but does not provide further information about allocation procedures or stratification.

Only the primary overall-survival results and selected safety counts are detailed in the record. Results for the listed secondary outcomes are not reported in the registry record. The safety summary does not give a full account of event timing, attribution, severity for all events, or comparative statistical testing. The study population is described as having advanced non-small cell lung cancer after one previous chemotherapy excluding TAXOTERE or HYCAMTIN; applicability outside that recorded context is not established by these results.

19. Related Tutorials

20. Related Statistical Calculators

21. Sources

• ClinicalTrials.gov: Comparison of IV Topotecan/Docetaxel to Docetaxel Alone in Second-Line Stage IIIB/IV Non-Small Cell Lung Cancer — NCT00065182.
• Registry record last updated 2019-06-24; status last verified 2019-03.
• Results first posted 2019-06-24.
• 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.

22. Record Summary

The registry record describes a completed randomized, parallel-group, open-label study comparing IV topotecan plus IV docetaxel with IV docetaxel alone in pretreated advanced non-small cell lung cancer. Enrollment was 399. The recorded study start date is 2003-08-14, and the primary completion date is 2007-08-30.

The primary endpoint was median overall survival in weeks, defined from randomization to death, in an intent-to-treat population. The reported medians were 30.7 weeks and 28.6 weeks. The log-rank hazard ratio was 1.007, with a 95% confidence interval from 0.813 to 1.248 and p = 0.946. A second hazard ratio of 0.977, with a 95% confidence interval from 0.788 to 1.210, is also reported without a method or p-value.

The record includes selected serious and other adverse-event counts and lists secondary outcome areas. It does not report full statistical-plan details, missing-data procedures, the original sample-size justification, or detailed results for the listed secondary outcomes. The recorded phase entry is PHASE2, while the official title describes a phase III trial; this difference is unexplained.