A Study of Subcutaneous Mircera for the Treatment of Anemia in Dialysis Patients.: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for A Study of Subcutaneous Mircera for the Treatment of Anemia in Dialysis Patients. (NCT00077623), a randomized phase 3 study. The primary outcome is Mean Change in Hemoglobin Concentration From Baseline to Evaluation Periods.

2. Clinical Question

3. Trial Design

This phase 3 study was randomized, controlled, open-label and parallel-group. It is registered as NCT00077623, with Hoffmann-La Roche listed as sponsor. The registry record gives enrollment as 572 and status as completed. The recorded start was 2004-03 and primary completion was 2005-09.

Participants were assigned to one of three groups: RO0503821 (Mircera) subcutaneously once every two weeks, RO0503821 subcutaneously once every four weeks, or an epoetin reference regimen. The two Mircera schedules were described as lasting 52 weeks. The reference group continued ongoing subcutaneous epoetin alfa or beta for 52 weeks, administered one, two or three times weekly. The summary says participants had previously received subcutaneous epoetin and describes the treatment as maintenance therapy for renal anemia in chronic kidney disease patients on dialysis.

Masking was none, so the study was open-label. The registry lists nine outcome measures and two analyses. The primary outcome was a continuous measure: mean change in hemoglobin concentration from baseline to the evaluation period. Its primary analysis population was Per Protocol, rather than all randomized participants. The registry description lists some reasons for exclusion from that population, but ends before the full eligibility description is complete.

The summary anticipated 1-2 years on study treatment and described a target sample size of 100-500 individuals. Those are summary statements; the registry record does not report the original design assumptions or sample-size justification. The posted outcome analysis uses the Week 29 to Week 36 evaluation period, while the arm descriptions state a 52-week treatment period.

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 572 participants were randomized, of which 1 participant did not received the study drug.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: The Per Protocol population included all randomized participants except those not meeting inclusion criterion related to stable baseline Hb values, inadequate iron status, hemoglobinopathies/hemolysis, RBC transfusion/blood loss, with \<5 recorded Hb values during the evaluation period, with missing administrations of the study drug/ reference drug

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 analysis was done on safety population which included all participants who received at least one dose of study drug.

6. Primary Endpoint

The registry lists the primary outcome as Mean Change in Hemoglobin Concentration From Baseline to Evaluation Periods, measured over Baseline (Week -4 to Week -1) and Evaluation period (Week 29 to Week 36).

The primary endpoint measures the average change in hemoglobin concentration, in g/dL, between a baseline period and a later evaluation period. Baseline was defined as Week -4 to Week -1, and the evaluation period as Week 29 to Week 36. A time-adjusted average was calculated separately for each period using an area-under-the-curve approach. The change was then obtained by subtracting the average baseline value from the average evaluation-period value.

Hemoglobin is the measured outcome in this anemia maintenance study. Comparing changes from baseline focuses the analysis on movement over the study periods, rather than on evaluation-period levels alone. The endpoint's period averages are intended to reflect measurements across specified windows; the registry says blood samples were taken before study-drug administration.

The record says last observation carried forward (LOCF) was used for missing hemoglobin values to address early dropouts. The primary analysis was restricted to the Per Protocol population. Its description lists exclusions related to stable baseline hemoglobin, iron status, hemoglobinopathies or hemolysis, red blood cell transfusion or blood loss, and having fewer than 5 recorded hemoglobin values during the evaluation period. The remaining text describing exclusions is incomplete in the registry record.

General: a mean change condenses participant-level measurements into an average, so it does not show the full distribution of responses. The posted standard deviations indicate variability around group means, while the between-group contrast and its confidence interval describe the estimated difference and its precision.

7. Results Posted in the Registry

• RO0503821 (1x/2 Weeks): mean -0.00 versus -0.12 on control; difference 0.12 (95% CI -0.10 to 0.34), standardized effect d = 0.12; z-test p = 0.282.
• RO0503821 (1x/4 Weeks): mean -0.11 versus -0.12 on control; difference 0.01 (95% CI -0.21 to 0.23), standardized effect d = 0.01; z-test p = 0.929.

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.

Number of Participants Maintaining Average Hb Concentration During the Evaluation Period Within +-1 g/dL of Their Average Baseline Hb Concentration

• Endpoint type: Binary: the number of participants with the event, compared as proportions.
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • RO0503821 (1x/2 Weeks): 75.6% (124/164) versus 72.2% (127/176) on control; risk difference 3.5 percentage points (95% CI -5.9 to 12.8), risk ratio 1.05 (0.92–1.19); unadjusted two-proportion z-test p = 0.470.
  • RO0503821 (1x/4 Weeks): 66.1% (111/168) versus 72.2% (127/176) on control; risk difference -6.1 percentage points (95% CI -15.8 to 3.7), risk ratio 0.92 (0.79–1.06); unadjusted two-proportion z-test p = 0.222.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between RO0503821 (1x/2 Weeks) and the control. The interval gives the range of effects that remain compatible with the data.
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between RO0503821 (1x/4 Weeks) and the control. The interval gives the range of effects that remain compatible with the data.

Number of Participants With Red Blood Cell Transfusions

• Endpoint type: Binary: the number of participants with the event, compared as proportions.
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • RO0503821 (1x/2 Weeks): 6.3% (12/190) versus 9.9% (19/191) on control; risk difference -3.6 percentage points (95% CI -9.1 to 1.8), risk ratio 0.63 (0.32–1.27); unadjusted two-proportion z-test p = 0.195.
  • RO0503821 (1x/4 Weeks): 10.5% (20/190) versus 9.9% (19/191) on control; risk difference 0.6 percentage points (95% CI -5.5 to 6.7), risk ratio 1.06 (0.58–1.92); unadjusted two-proportion z-test p = 0.852.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between RO0503821 (1x/2 Weeks) and the control. The interval gives the range of effects that remain compatible with the data.
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between RO0503821 (1x/4 Weeks) and the control. The interval gives the range of effects that remain compatible with the data.

The posted secondary outcomes cover hemoglobin maintenance within the stated range, red blood cell transfusions, safety events, laboratory abnormalities, and blood pressure and pulse measures in dialysis subgroups. The registry lists nine outcome measures overall, while eight secondary outcomes are posted in the results summary. These outcomes broaden the description of treatment experience beyond the primary hemoglobin-change measure, but they do not all address the same clinical or statistical question.

The posted recalculated comparisons for the two available binary outcomes are unadjusted, and their intervals are broad enough to include differences in either direction. The record does not give sponsor-reported analyses for those two secondary outcomes. The other listed outcomes do not have numerical results in the registry record, so their direction, magnitude and precision cannot be assessed from the posted information.

In relation to the primary result, these secondary outcomes provide additional context rather than a direct resolution of the discrepancy between the sponsor ANCOVA and the unadjusted recalculation. A binary event such as transfusion is not interchangeable with a mean change in hemoglobin, and physiological or safety outcomes address separate domains. Consistency across outcomes would require prespecified analysis and careful consideration of their different definitions and populations; that full information is not reported in the registry record.

Unadjusted secondary results carry less weight because multiple outcomes and comparisons create more opportunities for chance findings, and no multiplicity adjustment is reported. They are also not a substitute for the primary endpoint or for a complete safety analysis. The record does not report a full hierarchy of secondary hypotheses or the planned approach to multiplicity, so isolated secondary p-values or apparent patterns should be interpreted cautiously.

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

• Number of Participants Maintaining Average Hb Concentration During the Evaluation Period Within +-1 g/dL of Their Average Baseline Hb Concentration
• Number of Participants With Red Blood Cell Transfusions
• Number of Participants With Any Adverse Events, Any Serious Adverse Events, and Deaths
• Number of Participants With Marked Laboratory Abnormalities
• Change From Baseline in Systolic and Diastolic Blood Pressure - at Weeks 36 and 52 in Hemodialysis Participants
• Change From Baseline in Pulse Rate at Weeks 36 and 52 in Hemodialysis Participants
• Change From Baseline in Systolic and Diastolic Blood Pressure at Weeks 36 and 52 in Peritoneal Dialysis Participants
• Change From Baseline in Pulse Rate - Peritoneal Dialysis Participants

9. Statistical Interpretation

10. Statistical Methods Explained

General: the primary outcome is continuous because hemoglobin concentration is measured on a numeric scale and summarized as a mean. A mean difference compares the average change in one group with the average change in the reference group. It is expressed in the same units as the endpoint, here g/dL. A difference near zero means the group averages are close on that scale; uncertainty around the estimate matters for judging how precisely the difference is known.

How was the primary outcome constructed?

The registry describes a time-adjusted average based on an area-under-the-curve approach. The approach uses hemoglobin measurements across each defined period to calculate an average for baseline and an average for evaluation, then subtracts the baseline average from the evaluation average. The periods were Week -4 to Week -1 and Week 29 to Week 36. Samples were collected before study-drug administration. This construction uses repeated measurements within the windows rather than relying on a single measurement from a particular visit.

The analysis used LOCF for missing hemoglobin values, described as a way to correct for the impact of early dropouts. In general, LOCF carries a participant's last recorded value forward to fill a later missing value. This relies on an assumption about what the unobserved values would have been after the last measurement. The registry does not report sensitivity analyses that test alternative missing-data assumptions, so the robustness of the result to those assumptions is not reported in the registry record.

Who was included in the primary analysis?

The primary analysis population was Per Protocol. The record says it included randomized participants except those failing specified criteria related to stable baseline hemoglobin, adequate iron status, hemoglobinopathies or hemolysis, red blood cell transfusion or blood loss, and having fewer than 5 recorded hemoglobin values during the evaluation period. The population description is cut off, so the complete list of exclusions is not reported in the registry record. A Per Protocol analysis is not identical to an analysis of all randomized participants; participants can be omitted based on post-randomization information. The record does not provide a corresponding intention-to-treat primary result in the posted analysis details.

What does ANCOVA contribute?

The sponsor reported an analysis using ANCOVA and confidence intervals for the difference between groups. In general, analysis of covariance estimates a group contrast while accounting for one or more covariates. The covariates used in this analysis are not reported in the registry record. The sponsor results are described as non-inferiority or equivalence analyses, but the margin and complete decision rules are not reported. Those details are important because non-inferiority asks whether the data exclude an unacceptable loss relative to a prespecified boundary; a p-value from that test does not necessarily answer the same question as a conventional test of whether the difference equals zero.

The recalculated comparisons use the posted means, standard deviations and group sizes to form simple, unadjusted contrasts. They do not include covariate adjustment and do not reconstruct the sponsor's analysis. The record explicitly flags a discrepancy, so the sponsor's own analysis takes precedence when describing the formal study analysis. Recalculations are useful as a check on the scale and uncertainty of a simple group comparison, not as a replacement for the analysis plan or the sponsor's analysis.

Why do confidence intervals matter?

General: a confidence interval gives a range of values compatible with the data and the method used to calculate it. Its width reflects precision: wider intervals leave more uncertainty about the size and direction of the contrast. For a conventional difference-from-zero interpretation, an interval that includes zero does not rule out no difference at the confidence level used. Non-inferiority interpretation requires the interval to be judged against the non-inferiority margin, which is not reported in the registry record. The sponsor's intervals and recalculated intervals therefore answer different analytical questions only to the extent that their methods and target quantities differ; the registry does not provide enough detail to fully reconcile them.

Why can variability affect an analysis?

General: standard deviations describe how dispersed participant measurements are around a group mean. Greater variability makes a group mean less precise when other factors are similar. The primary arm standard deviations are close in size, but the registry does not report the full participant-level distribution, so the means and standard deviations cannot describe all variation in response. More complex issues such as clustering or over-dispersion are not identified as features of the primary continuous analysis in the registry record.

The registry does not report the full statistical analysis plan, covariate list, original sample-size justification, interim analyses, or a complete missing-data strategy beyond the stated LOCF method. Those omissions limit independent reconstruction of the formal analysis and its assumptions.

11. Confidence Intervals and Precision

The sponsor-reported primary comparisons use 97.5% confidence intervals. For RO0503821 every two weeks versus epoetin, the reported mean difference is 0.141, with an interval from -0.098 to 0.380. For RO0503821 every four weeks versus epoetin, the mean difference is -0.022, with an interval from -0.262 to 0.217. Each sponsor interval spans values on both sides of zero. For a conventional difference-from-zero question, that indicates uncertainty about the direction of the contrast; for a non-inferiority question, the relevant comparison is with the margin, which is not reported in the registry record.

The unadjusted recalculations have 95% confidence intervals. The every-two-weeks comparison is 0.12, with an interval from -0.099 to 0.339. The every-four-weeks comparison is 0.01, with an interval from -0.210 to 0.230. These intervals also span zero and cover differences in either direction. Their widths show that the posted summaries do not pinpoint a narrow range around either recalculated estimate.

General: a confidence interval is not a guarantee that the true effect lies within the displayed endpoints. It quantifies uncertainty under the specified statistical method and assumptions. An interval's interpretation depends on the target quantity, analysis population and method. Here, the sponsor used ANCOVA and a Per Protocol population is described for the primary endpoint, while the simple recalculations use unadjusted summaries. The discrepancy between the sponsor results and recalculations means the intervals should not be conflated. The registry does not provide enough detail about the model or non-inferiority margin to resolve the difference independently.

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 figures are participants per arm for these comparisons; they are not total event counts. They are illustrations based on the observed effect, not the original trial's design. The stated assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, a normal approximation, and an effect to detect equal to the stated fraction of the observed effect. The original design assumptions and sample-size justification are not reported in the registry record.

The 80% power column gives the illustrative number of participants per arm for a design with an 80% chance of detecting the selected effect under those assumptions, if that effect is the true effect and the model assumptions hold. The 90% power column gives the corresponding planning number for a higher target chance of detection. These are design calculations; this is not observed power.

Rows that plan for 100% of the observed effect use the full recalculated contrast as the target effect. Rows planning for 75% or 50% of the observed effect use smaller target effects. Smaller effects need more participants to distinguish them from random variation with the same planned power and significance level. This is especially apparent where the observed contrast is very small: the resulting illustrative requirements can become very large.

The values are not forecasts of what a future study will achieve. They depend on the observed effect, variability, equal allocation and the other stated assumptions. Because the primary results have a flagged discrepancy between the sponsor analysis and the simple recalculation, the planning illustrations should not be taken as a substitute for a study-specific design based on a prespecified clinically relevant effect and a justified analysis plan.

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

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 = 0.120; pooled standard deviation σ = 1.002.
2. n = 2 × (1.96 + 0.842)² × 1.002² / 0.120² = 1,095.4, rounded up to 1,096 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 <- 0.12                      # observed difference in means
sigma <- 1.00242714851892                     # 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(77623)
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 = 1,096 (80% power), 1,467 (90% power)
 75% of observed effect: n per arm = 1,948 (80% power), 2,608 (90% power)
 50% of observed effect: n per arm = 4,382 (80% power), 5,866 (90% power)
power.t.test: 1096.4 
Simulated power at n = 1,096 per arm: 0.797 (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.

14. Analyses Reported by the Sponsor

The registry reports sponsor ANCOVA results for both Mircera schedules against the epoetin reference. For the every-two-weeks schedule, the sponsor reports a mean difference of 0.141 and a 97.5% confidence interval from -0.098 to 0.380, with p < 0.001. For the every-four-weeks schedule, the reported difference is -0.022 and the interval is -0.262 to 0.217, also with p < 0.001. Both analyses are labeled non-inferiority or equivalence.

The unadjusted recalculations from the posted group summaries yield different estimates and p-values: 0.12 with p = 0.282 for every two weeks, and 0.01 with p = 0.929 for every four weeks. The discrepancy flag indicates that these simple comparisons do not reproduce the sponsor's own analysis. They may not target the same adjusted estimand or use the same analysis details, but the registry does not give enough information to establish the precise reason for the discrepancy.

Accordingly, the sponsor's own analysis takes precedence as the trial's reported analysis. That does not remove the need to interpret it in context: the non-inferiority margin, covariates, and full decision rule are not reported in the registry record. The recalculations are best viewed as transparent, unadjusted checks using posted summaries, not as a correction to or replacement for the sponsor's analysis.

15. Multiplicity, Missing Data and Analysis Populations

The registry lists a primary endpoint and multiple secondary outcomes, with two primary group comparisons against the epoetin reference. Testing several outcomes and contrasts can increase the chance of a nominally notable result occurring by chance. General: multiplicity procedures can control that risk by adjusting significance thresholds or by specifying a testing sequence. The registry record does not report a multiplicity strategy, a hierarchy among secondary outcomes, or whether the two primary comparisons were adjusted for multiplicity.

The sponsor analyses are labeled non-inferiority or equivalence, but the non-inferiority margin and detailed decision rule are not reported in the registry record. Without the margin, the reported confidence intervals cannot independently establish how the sponsor's criterion was applied. The p-values are both reported as p < 0.001, but their exact role in the decision rule is not described. The original statistical analysis plan is not reported in the registry record.

Missing-data handling for the primary endpoint is partly described: LOCF was used for missing hemoglobin values to address early dropouts. The registry does not report the amount or pattern of missing primary data, sensitivity analyses, or whether alternative assumptions were examined. LOCF can be sensitive to changes that occur after the last observed value, so an analysis under different assumptions could help assess robustness; whether such analyses were performed is not reported in the registry record.

The primary analysis population was Per Protocol, with several exclusions described but the full list truncated. The registry record does not report a corresponding intention-to-treat primary analysis in the posted results summary. Differences between analysis populations can matter in non-inferiority settings, making the incomplete population and analysis details particularly relevant to interpretation.

Interim analyses, stopping rules, covariate selection, and the original sample-size justification are not reported in the registry record. These gaps limit independent assessment of the study's prespecification and the robustness of the reported formal analyses.

16. Safety Results

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

Other (non-serious) adverse events

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

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: myocardial infarction (4.2%), pneumonia (4.2%), cardiac failure congestive (2.6%), bronchopneumonia (2.1%), cardiac arrest (2.1%).

The safety figures list participants at risk and participants with serious adverse events or other adverse events by group. For serious adverse events, the figures are 70 of 190 in the every-two-weeks Mircera group, 73 of 190 in the every-four-weeks group, and 85 of 191 in the epoetin reference group. For other adverse events, the corresponding figures are 113 of 190, 131 of 190 and 111 of 191. These are counts within the recorded safety groups; they should be read with their displayed denominators and not treated as a complete description of event timing or severity.

The most frequently listed serious-event terms include myocardial infarction and pneumonia, each reported as 4.2%, followed by cardiac failure congestive at 2.6%, bronchopneumonia at 2.1%, cardiac arrest at 2.1%, and arteriovenous fistula thrombosis at 2.1%. The registry record does not report death counts for these groups. It also does not provide enough detail to assess exposure-adjusted event rates, event duration, causal attribution, or comparative statistical uncertainty. Counts alone do not establish whether a difference between groups is meaningful.

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

The registry record describes a randomized, open-label comparison of two subcutaneous Mircera schedules with ongoing epoetin in dialysis patients with chronic kidney disease and anemia. Its primary measure was mean change in hemoglobin from baseline to the Week 29 to Week 36 evaluation period. The posted group means were close to zero, and the simple recalculated comparisons have confidence intervals spanning zero. The sponsor's ANCOVA results are different and are labeled as non-inferiority or equivalence analyses.

Those findings do not support a single unqualified interpretation from the registry record alone. The sponsor analysis takes precedence over the recalculation because the record flags a discrepancy, but the non-inferiority margin and full analysis details are absent. The recalculated intervals describe uncertainty for simple unadjusted contrasts and should not be substituted for the sponsor's adjusted analysis. Neither analysis can be fully assessed against the study's formal decision criterion without the margin and decision rule.

General: a non-significant conventional test is not proof that two treatments are identical, and an interval that includes zero can still be compatible with differences in either direction. Likewise, a non-inferiority conclusion depends on a prespecified margin and appropriate analysis, not merely on a small p-value. The registry record does not provide the details needed to independently verify that conclusion.

The secondary outcomes offer information about hemoglobin maintenance, transfusions, safety and physiological measures, but many listed outcomes have no numerical results posted, and the available recalculated comparisons are unadjusted. Multiplicity handling is not reported. The safety figures describe event counts, not a complete causal or comparative assessment, and deaths are not reported.

These results do not establish a broad claim about treatment benefit or harm. They describe the recorded outcomes, estimates and uncertainty for this trial, with important limitations in the posted analysis detail.

18. Why This Trial Matters Statistically

This trial illustrates why the analysis method and target question matter as much as the group averages. The primary outcome is a change in hemoglobin averaged across specified periods, while the formal sponsor analysis is ANCOVA and is labeled non-inferiority or equivalence. A simple difference calculated from arm summaries is not necessarily the same estimand as an adjusted analysis, and the registry flags a discrepancy between them.

It also illustrates that a p-value cannot be interpreted without knowing the hypothesis being tested. The sponsor reports p < 0.001 alongside confidence intervals that span zero. Those results may reflect a non-inferiority or equivalence test rather than a conventional test of difference from zero, but the margin and precise decision rule are not reported. In this setting, identifying the test and its margin is essential for a statistically sound interpretation.

The primary endpoint uses repeated hemoglobin measurements and LOCF for missing values, while the primary population is Per Protocol. Both choices shape which data contribute to the result. Without a complete description of the analysis population, missing-data sensitivity analyses and model covariates, readers cannot fully reproduce or stress-test the formal finding.

Finally, the planning illustrations show how sample-size needs depend on the effect chosen for detection. Smaller assumed effects lead to larger illustrative participant requirements, but those calculations use the observed recalculated effect and are not the original design justification. The record therefore offers a useful example of the distinction between descriptive data, formal inference and planning assumptions.

19. 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 constrain interpretation of the posted results. The study was open-label, so participants and study personnel were not masked to assigned regimen. The primary analysis population was Per Protocol, and the registry's description of its exclusions is incomplete. The posted summary does not include an intention-to-treat primary result, making it difficult to compare conclusions across analysis populations.

The sponsor's reported ANCOVA contrasts do not match simple unadjusted recalculations from the posted summaries, and the record explicitly flags a discrepancy. The sponsor analysis takes precedence, but the registry does not report its covariates or enough methodological detail to reconcile the estimates. The non-inferiority margin and full decision rule are also absent, limiting independent evaluation of the formal claim.

Missing hemoglobin values were handled with LOCF, but the extent and pattern of missingness and sensitivity analyses are not reported in the registry record. The record also does not report an adjustment strategy for multiple outcomes or comparisons. These omissions matter because the primary and secondary findings depend on analysis choices and because multiple unadjusted comparisons can be misleading.

Results are not posted for every listed secondary outcome. Safety reporting includes serious and other adverse-event counts and selected serious-event terms, but deaths are not reported, and event timing, exposure-adjusted rates and causal attribution are not available in this summary. The record does not report the original sample-size justification, interim analyses or a full statistical analysis plan. These are limitations of the available registry information and do not, by themselves, establish that any particular analysis was or was not conducted.

20. Related Tutorials

21. Related Statistical Calculators

22. Sources

• ClinicalTrials.gov: A Study of Subcutaneous Mircera for the Treatment of Anemia in Dialysis Patients. — NCT00077623.
• Registry record last updated 2016-05-25; status last verified 2016-04.
• Results first posted 2016-04-26.
• 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.

23. Record Summary

NCT00077623 was a completed phase 3 randomized, open-label, parallel-group study sponsored by Hoffmann-La Roche. It compared subcutaneous RO0503821 (Mircera) every two weeks or every four weeks with ongoing subcutaneous epoetin alfa or beta in dialysis patients with chronic kidney disease and anemia. The primary endpoint was mean change in hemoglobin from baseline to the Week 29 to Week 36 evaluation period.

The sponsor reported ANCOVA non-inferiority or equivalence analyses, while unadjusted recalculations from the posted summaries differed; the registry flags this discrepancy, so the sponsor's own analysis takes precedence. The non-inferiority margin and several details needed to assess the analysis are not reported in the registry record. Secondary outcomes and safety counts provide additional context, but multiplicity handling, some outcome results and death counts are not reported.