Efficacy and Safety of IV Diclofenac (DIV075V)for Pain After Elective Orthopedic Surgery: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Efficacy and Safety of IV Diclofenac (DIV075V)for Pain After Elective Orthopedic Surgery (NCT00507026), a randomized phase 3 study. The primary outcome is Sum of the Pain Intensity Differences (SPID) Over 24 Hours.

2. Clinical Question

The trial addressed acute postoperative pain after elective orthopedic surgery. It compared repeated intermittent intravenous diclofenac (DIC075V) with intravenous ketorolac and placebo. The main statistical question was whether the mean 24-hour Sum of the Pain Intensity Differences (SPID) differed between each active-treatment group and placebo in the registry’s stated analysis population.

The design therefore allows two distinct comparisons with placebo; it does not, in the reported primary analyses, establish a direct statistical comparison between diclofenac and ketorolac. The endpoint is a continuous measure built from pain-intensity changes over time, rather than a simple assessment at one time point. The registry lists additional pain and rescue-medication outcomes, but the primary question centers on SPID over 24 hours. These results describe the posted trial data and their statistical uncertainty; they do not by themselves establish how an individual participant would respond.

3. Trial Design

This was a randomized, quadruple-masked, parallel-group phase 3 study. Its official title describes an active-and-placebo-controlled evaluation of repeated dosing of DIC075V against parenteral ketorolac and placebo in acute postoperative pain after elective orthopedic surgery. The registry lists postoperative pain as the condition and the study status as completed. Enrollment was 277. The recorded study start was 2007-07-25, and primary completion was 2008-10-14.

The three groups were diclofenac (DIC075V), ketorolac, and placebo. For the primary endpoint, the posted group summaries give a mean SPID of 577.0 with a standard deviation of 570.9 for diclofenac, 563.2 with a standard deviation of 586.2 for ketorolac, and 28.0 with a standard deviation of 428.4 for placebo. The corresponding group sizes were 145, 60, and 72. The reported endpoint population was the ITT population as defined in the record: randomized participants who received at least 1 dose of study medication.

Randomization is intended to support comparable treatment groups on average, while masking is intended to limit influences from knowledge of assigned treatment. The parallel model means participants were assigned to a group rather than crossing between treatments, according to the design description. The registry record does not report details of the randomization procedure, allocation concealment, masking roles, treatment schedule, site structure, or the original sample-size justification. Those details should not be inferred from the design labels alone.

The registry lists 15 outcome measures and 10 analyses. Secondary outcomes include pain-intensity differences over time, pain relief, rescue-medication use and timing, and participant global evaluation. The primary endpoint remains the 24-hour SPID result; the presence of secondary outcomes does not establish which were prespecified as confirmatory or how their results were adjusted for multiple testing.

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: Total 432 participants signed the inform consent form (ICF). Out of which 155 participants were not admitted into the study, 277 actually enrolled into the study and assigned to study treatment.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: ITT population included all participants who were randomized into the study and who received at least 1 dose of study medication.

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: Intent-To-Treat (ITT) population included all participants who were randomized into the study and who received at least 1 dose of study medication.

6. Primary Endpoint

The registry lists the primary outcome as Sum of the Pain Intensity Differences (SPID) Over 24 Hours, measured over Over 24 hours post first dose.

The primary endpoint was the Sum of the Pain Intensity Differences over 24 hours after the first dose, measured in mm*hours. Pain intensity was assessed on a visual analog scale from 0 mm, meaning no pain, to 100 mm, meaning the worst possible pain. At each post-dose time point, pain-intensity difference (PID) was defined as baseline pain intensity minus the post-baseline score. A positive PID therefore represents a decrease in reported pain intensity from baseline.

SPID combines those changes over time. Each post-dose PID is multiplied by the time since the preceding assessment, and the resulting values are summed over the specified period. The stated possible range is -2400 to 2400, and a higher SPID indicates better pain reduction under this measure. This construction captures both the direction and duration of changes rather than relying on a single pain score. It can therefore reflect whether improvement is sustained across the assessment period, while remaining a summary measure that may conceal different patterns over time.

The endpoint is continuous, and the registry reports its mean and standard deviation for each group. The ITT population for this outcome included randomized participants who received at least 1 dose of study medication. The registry record does not provide the detailed schedule of pain assessments or describe how missing measurements were handled in the SPID calculation, so those elements cannot be evaluated from the posted endpoint description.

7. Results Posted in the Registry

• Diclofenac (DIC075V): mean 577.00 versus 28.00 on control; difference 549.00 (95% CI 413.25 to 684.75), standardized effect d = 1.04; z-test p = <0.001.
• Ketorolac: mean 563.20 versus 28.00 on control; difference 535.20 (95% CI 356.89 to 713.51), standardized effect d = 1.06; z-test p = <0.001.

8. Statistical Interpretation

9. Statistical Methods Explained

The primary outcome is a continuous measure, so the registry reports a mean and standard deviation for each group. A mean describes the group’s average SPID. A standard deviation describes how dispersed individual values are around that average; it is not the uncertainty in the group mean. The posted standard deviations are substantial in all groups, indicating variation in participant-level SPID values. A mean difference is the contrast between two group averages, expressed in the endpoint’s units.

What does the registry’s ANCOVA analysis mean?

The sponsor’s reported method for the two primary comparisons was ANCOVA, or analysis of covariance. In general, ANCOVA compares group outcomes while accounting for specified covariates. Adjustment can improve precision or account for measured differences, depending on the model and variables included. The registry record does not report which covariates were included, how the model was specified, or the adjusted effect estimates and confidence intervals. The reported formal results are superiority analyses of diclofenac versus placebo and ketorolac versus placebo, each with p < 0.001. Those are the sponsor’s reported analyses, not the simple calculations from the posted means.

What do the recalculated comparisons represent?

The recalculated mean differences are simple, unadjusted comparisons of the posted group means. For diclofenac versus placebo the estimate is 549.0, and for ketorolac versus placebo it is 535.2. Their 95% confidence intervals are 413.3 to 684.7 and 356.9 to 713.5, respectively. The recalculated p-values are <0.001 for both comparisons. These calculations do not reproduce ANCOVA because the model details and adjusted estimates are not reported. They are useful as transparent summaries of the group-level data, but they should not be mistaken for the sponsor’s adjusted treatment effects.

The pooled standard deviations used for those recalculations are 528.1 for diclofenac versus placebo and 506.2 for ketorolac versus placebo. A standardized difference is also reported as 1.04 and 1.057, respectively. In general, standardization expresses a mean difference relative to a measure of variability, allowing a scale-free description of separation. It does not replace the original endpoint units or establish clinical importance. No clinical threshold for interpreting SPID differences is reported in the registry record.

How should p-values and confidence intervals be understood?

In general, a p-value assesses how compatible the observed data are with a specified null hypothesis under a statistical model. It does not measure the probability that the null is correct, and it does not convey the magnitude of an effect on its own. A confidence interval gives a range of effect values compatible with the data and method at the stated confidence level. Wider intervals represent less precision than narrower intervals, though interpretation also depends on the outcome scale and study design. The recalculated intervals are unadjusted and should be distinguished from intervals for the sponsor’s ANCOVA results, which are not reported.

For a two-group continuous-outcome comparison, the standard error reflects uncertainty in the estimated difference. In general, it depends on the variability of the outcome and the amount of information in each group. The posted standard errors for the group means are 47.41 and 50.49 for the diclofenac and placebo groups, and 75.68 and 50.49 for the ketorolac and placebo groups, respectively. The smaller ketorolac group has a larger standard error for its mean than the larger diclofenac group, consistent with the general relationship between group size and precision. These are group-mean standard errors, not the standard deviations of participant outcomes.

Why can variability matter?

SPID is built from repeated pain assessments, and participant responses can vary. In general, high outcome variability makes a mean difference harder to estimate precisely unless there is enough information to offset that variability. Differences in assessment patterns, rescue-medication use, or missing observations could also affect how an endpoint summarizes experience, but the registry record does not provide enough detail to evaluate those factors. The recalculation uses the posted summary statistics and does not make adjustments for any of them.

What comparisons are covered?

The sponsor reports separate superiority tests against placebo. The posted results do not report a primary inferential test directly comparing diclofenac with ketorolac. The fact that the two active groups have similar observed means is not itself evidence that they are equivalent or that their difference is statistically negligible. Establishing equivalence or noninferiority requires a specified framework and margin; such a result is not reported in the registry record. Likewise, the registry record does not report whether the two placebo comparisons or the wider set of outcomes were handled with a multiplicity adjustment. The two small p-values should therefore be interpreted as the reported comparisons, not as proof that the overall family of analyses had a particular error rate.

10. Confidence Intervals and Precision

The recalculated 95% confidence interval for diclofenac versus placebo is 413.3 to 684.7 mm*hours. For ketorolac versus placebo, it is 356.9 to 713.5 mm*hours. Both intervals lie above zero and describe substantial positive differences in mean SPID on the endpoint’s scale. Their widths also show that the exact magnitude is not known with the precision of a single point estimate. The interval for ketorolac is wider than the interval for diclofenac in these unadjusted recalculations.

In general, a confidence interval is a range of values compatible with the observed data and the calculation method, under the method’s assumptions. It is not a range containing every participant’s response, and it does not mean that there is a 95% probability that the true effect lies within this particular calculated interval. The intervals are unadjusted calculations from the posted summary statistics; they are not reported ANCOVA confidence intervals. The registry record does not report confidence intervals for the sponsor’s adjusted estimates.

Precision depends on both variability and information in the groups. The outcome standard deviations are 570.9 for diclofenac, 586.2 for ketorolac, and 428.4 for placebo. The group sizes are 145, 60, and 72, respectively. The larger standard errors associated with the ketorolac comparison are consistent with less precise estimation than the diclofenac comparison, while the intervals also reflect the outcome variability and the calculation method. These intervals quantify uncertainty about the mean contrast, not uncertainty about every possible clinical interpretation.

The endpoint aggregates pain differences across the full 24-hour period, and its mean difference does not identify when the groups diverged or how consistently individual participants improved. A clinical threshold for a meaningful SPID difference is not reported in the registry record. Precision around the statistical contrast therefore should not be treated as a judgment about practical importance.

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 illustrations based on the observed effect, not the trial’s original design justification. The original sample-size assumptions are not reported in the registry record. For each active-treatment comparison, the figures describe participants per arm for 80% power and participants per arm for 90% power under the stated planning assumptions. They are not the number enrolled or analyzed in this trial.

The planning illustration assumes a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. It considers effects equal to 100%, 75%, and 50% of the observed effect. For diclofenac versus placebo, the listed participants per arm for 80% power are 15, 26, and 59 across those effect assumptions; for 90% power they are 20, 35, and 78. For ketorolac versus placebo, the corresponding figures are 15, 25, and 57 for 80% power, and 19, 34, and 76 for 90% power.

In plain terms, 80% power means that a study designed under those assumptions would have an 80% chance of detecting an effect of the specified size at the stated significance level, if that effect and the model assumptions were correct. The 90% column uses a higher target chance of detection and therefore generally calls for more participants. Smaller assumed effects require more participants per arm because they are harder to distinguish from variability. These figures are not observed power and do not describe the probability that the trial’s findings are correct.

The planning illustrations use observed effects as a starting point, so they can appear optimistic if the observed effects are larger than effects expected in a future study. They also rely on equal allocation for the planning exercise, whereas the trial’s posted group sizes are not equal. The figures are best read as sensitivity illustrations about how required group size changes when the assumed effect is reduced, not as a reconstruction of the trial’s design.

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 = 549.000; pooled standard deviation σ = 528.119.
2. n = 2 × (1.96 + 0.842)² × 528.119² / 549.000² = 14.5, rounded up to 15 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 <- 549                      # observed difference in means
sigma <- 528.119399149164                     # 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(507026)
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 = 15 (80% power), 20 (90% power)
 75% of observed effect: n per arm = 26 (80% power), 35 (90% power)
 50% of observed effect: n per arm = 59 (80% power), 78 (90% power)
power.t.test: 15.5 
Simulated power at n = 15 per arm: 0.790 (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. Analyses Reported by the Sponsor

The registry’s formal results report ANCOVA superiority analyses for diclofenac versus placebo and ketorolac versus placebo. Both are reported with p < 0.001. The posted record does not include the corresponding adjusted estimates or confidence intervals, nor does it describe the ANCOVA covariates.

The recalculated comparisons are different in scope: they are simple, unadjusted contrasts of the posted means. They give mean differences of 549.0 for diclofenac versus placebo and 535.2 for ketorolac versus placebo, with 95% confidence intervals of 413.3 to 684.7 and 356.9 to 713.5. The recalculated p-values are <0.001 for both. Because they do not reproduce the ANCOVA model, these estimates should be read as supplementary checks based on group summaries, not as replacements for the sponsor’s reported analysis.

The registry indicates no discrepancy between the reported analysis and the recalculation. Even so, the two approaches are not identical: one is the sponsor’s covariate-adjusted method, and the other is an unadjusted calculation. Their p-values also have different displayed thresholds. Neither result is a direct diclofenac-versus-ketorolac test. The registry record does not report enough model detail to independently reconstruct the ANCOVA results.

14. Multiplicity, Missing Data and Analysis Populations

The registry lists two primary superiority comparisons: diclofenac versus placebo and ketorolac versus placebo. It also lists 15 outcome measures and 10 analyses, with secondary outcomes spanning pain intensity over time, pain relief, rescue-medication use, and participant global evaluation. When several hypotheses are tested, the chance of at least one false-positive result can increase if each test is interpreted separately. In general, multiplicity procedures can control that risk by adjusting the testing or interpretation strategy.

The registry record does not report whether a multiplicity adjustment was used, how the primary comparisons were ordered, or which secondary outcomes were designated confirmatory. The two reported ANCOVA p-values of < 0.001 should therefore be understood as the p-values for the stated comparisons, rather than as evidence of a particular family-wise error control across every outcome or analysis. The record also does not report an analysis hierarchy or rules for interpreting the secondary results.

The primary population is described as the ITT population, including all randomized participants who received at least 1 dose of study medication. This is not an all-randomized population as literally described in some uses of the term ITT. The registry record does not report the number excluded from this outcome population, the extent or pattern of missing outcome data, the method used to handle missing pain assessments, or sensitivity analyses for missing data. It also does not report whether participants were analyzed according to assigned group after treatment discontinuation.

Those omissions matter because the endpoint combines repeated measurements over 24 hours. Missing assessments can affect an individual’s calculated SPID, and the treatment of incomplete records can influence group estimates. Without the missing-data methods and relevant analysis details, the robustness of the primary estimate to alternative assumptions cannot be assessed from the registry record. These uncertainties do not invalidate the posted results, but they limit how completely the analysis can be evaluated.

15. Safety Results

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

Other (non-serious) adverse events

Participants with at least one non-serious event at or above the registry's 0% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: deep vein thrombosis (2.1%), cardiac failure congestive (1.7%), staphylococcal infection (1.7%), urinary tract infection (1.7%), femur fracture (1.7%).

The registry reports serious adverse events affecting 7 participants in the diclofenac group, 4 in the ketorolac group, and 3 in the placebo group. The corresponding groups at risk were 145, 60, and 72. These are counts of participants with serious adverse events as recorded in the registry, not counts of all serious-event episodes. The record does not provide deaths for these groups.

Other adverse events affected 109 participants in the diclofenac group, 46 in the ketorolac group, and 60 in the placebo group, with the same respective group sizes at risk. Serious and other adverse-event categories should be read separately; the registry’s counts alone do not explain event severity beyond the category labels or establish whether an event was caused by treatment.

The listed serious-event terms and percentages include deep vein thrombosis at 2.1%, cardiac failure congestive at 1.7%, staphylococcal infection at 1.7%, urinary tract infection at 1.7%, femur fracture at 1.7%, and blindness unilateral at 1.4%. The record does not give a group-specific breakdown for these terms in the summary available. Counts across groups should not be directly compared without considering the different group sizes, and these figures do not establish relative risk or causality.

The safety information is descriptive. The registry record does not report event timing, exposure duration, attribution, or a detailed statistical comparison of adverse events. The counts should therefore be interpreted as recorded trial events, not as a complete comparative safety analysis.

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

The posted results show higher mean 24-hour SPID in both active-treatment groups than in the placebo group. Since the endpoint defines higher values as greater pain reduction, the group summaries and sponsor-reported superiority analyses indicate a clear statistical separation from placebo on this outcome. The recalculated unadjusted differences and confidence intervals provide an additional description of the magnitude and precision of those group contrasts.

That conclusion is specific to the trial’s endpoint, population, and comparisons. SPID summarizes changes in pain intensity over time; it does not describe every time point separately or every participant’s experience. The registry does not report a threshold for a clinically meaningful SPID difference. Statistical evidence against a null difference is not, on its own, a complete assessment of clinical importance.

The reported primary analyses compare each active group with placebo. They do not establish that diclofenac and ketorolac differ from each other, and similar observed means do not establish equivalence. The record does not report a noninferiority or equivalence analysis. It also does not report enough information about the ANCOVA model, missing-data handling, or multiplicity strategy to assess those features fully.

The findings do not guarantee an outcome for any individual, nor do they establish results in populations or settings beyond the study description. They should be considered alongside the registry’s safety counts, which are descriptive and do not provide a complete causal or comparative safety assessment. The available information supports a statistical reading of the posted trial results, not medical advice or a recommendation about treatment.

17. Why This Trial Matters Statistically

This trial is statistically informative because it puts an active-treatment comparison beside a placebo comparison and uses a repeated-measure-derived endpoint. The SPID measure summarizes both direction and duration of pain-intensity changes over 24 hours, rather than reducing the outcome to a single assessment. That makes the endpoint interpretable as an aggregate of the post-dose trajectory, while also making the result dependent on how repeated assessments contribute to the summary.

The posted group means and standard deviations show why effect estimation and precision should be considered together. The active groups have higher mean SPID than placebo, while participant outcomes vary substantially within each group. The unadjusted confidence intervals quantify uncertainty around the mean contrasts and help distinguish the estimated magnitude from its precision. They do not substitute for the sponsor’s adjusted ANCOVA analysis, whose model details and adjusted estimates are not reported.

The registry also illustrates an important reporting distinction: a formal sponsor analysis and a transparent recalculation from group summaries can address the same contrast using different methods. The ANCOVA p-values and the recalculated mean differences are related evidence, but they are not identical statistical quantities. Readers should keep the analysis method attached to each estimate rather than treating all reported values as one calculation.

Finally, the existence of multiple outcome measures and analyses raises interpretive questions about multiplicity and missing data. The registry does not report enough detail to resolve them. Recognizing what the record establishes, and what it leaves unreported, is central to interpreting the apparent separation from placebo without overstating the breadth or certainty of the evidence.

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.

Several trial-specific limitations arise from the information not reported in the registry record. The ANCOVA method is named, but its covariates, model specification, adjusted estimates, and confidence intervals are not reported. Consequently, the sponsor’s formal effect estimates cannot be independently reconstructed from the posted group summaries, and the recalculated contrasts remain unadjusted.

The primary endpoint aggregates repeated pain assessments over 24 hours. The registry gives the formula and possible range but does not report the detailed assessment schedule or the handling of missing post-dose measurements. It therefore is not possible to assess whether incomplete assessment patterns could have influenced SPID or how robust the findings are under alternative missing-data assumptions. The record also does not provide a clinical-importance threshold for interpreting the size of a SPID difference.

The registry reports two primary comparisons and numerous outcome measures and analyses, but does not report the multiplicity strategy or identify an analysis hierarchy in the information. This limits interpretation of the p-values across the full set of analyses. Nor does the reported primary analysis directly compare diclofenac with ketorolac; the similar group means do not amount to evidence of equivalence.

The safety summary supplies counts and selected serious-event terms but does not give event timing, attribution, exposure details, or a full group-specific breakdown for the listed terms. The posted information therefore does not support a comprehensive comparative safety assessment. The registry record also does not report the original sample-size justification, randomization procedure, or detailed treatment schedule. These limitations concern the detail available for evaluating the record and should not be mistaken for evidence that a particular method was or was not used.

19. Related Tutorials

20. Related Statistical Calculators

21. Sources

• ClinicalTrials.gov: Efficacy and Safety of IV Diclofenac (DIV075V)for Pain After Elective Orthopedic Surgery — NCT00507026.
• Registry record last updated 2021-10-29; status last verified 2021-09.
• Results first posted 2021-10-29.
• 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

NCT00507026 is a completed randomized, quadruple-masked, parallel-group phase 3 study of intravenous diclofenac (DIC075V), intravenous ketorolac, and placebo for acute postoperative pain after elective orthopedic surgery. Enrollment was 277. The primary endpoint was mean SPID over 24 hours after the first dose in the ITT population, defined as randomized participants who received at least 1 dose.

The posted means were 577.0 for diclofenac, 563.2 for ketorolac, and 28.0 for placebo. The registry reports ANCOVA superiority results of p < 0.001 for each active-treatment comparison with placebo. Unadjusted recalculations give mean differences of 549.0 and 535.2, with 95% confidence intervals of 413.3 to 684.7 and 356.9 to 713.5. The registry record does not report a direct primary comparison of diclofenac with ketorolac, ANCOVA estimates, or the original sample-size justification. Safety findings are descriptive counts and do not establish causality.