Study of Roxadustat in Non-Dialysis Chronic Kidney Disease Participants With Anemia: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Study of Roxadustat in Non-Dialysis Chronic Kidney Disease Participants With Anemia (NCT01244763), a randomized phase 2 study. The primary outcome is Number (%) of Participants With an Hb Response by Week 17.

2. Clinical Question

The study addressed anemia correction in people with chronic kidney disease who were not receiving dialysis. Its primary endpoint combined an absolute hemoglobin threshold with a minimum increase from baseline. This definition asks whether participants both reached a specified hemoglobin level and improved by the specified amount, rather than measuring change alone.

The registry record describes a randomized, open-label, parallel study and lists six roxadustat cohorts. The primary outcome is a count of participants meeting the response definition, not a continuous average change in hemoglobin. The posted primary outcome timeframe is up to Week 17, while other listed outcomes assess hemoglobin response and levels at several time points and over different windows.

The central statistical question is therefore descriptive and comparative: how often was the response criterion met in each cohort, and how precisely do the posted estimates distinguish each cohort from Cohort A? Because every cohort received roxadustat, these results do not compare roxadustat with no treatment or another drug. They describe response counts across the dosing cohorts recorded in this study.

3. Trial Design

This was a Phase 2, randomized, open-label, parallel study of roxadustat in participants with anemia and non-dialysis chronic kidney disease. The registry record lists an enrollment of 145 and a completed status. The recorded study start was 2010-10-29, and primary completion was 2012-06-13. The sponsor is Kyntra Bio.

Six cohorts received different roxadustat regimens. Cohort A used tiered, weight-based dosing three times weekly for 16 weeks. Cohort B used tiered, weight-based dosing three times weekly and then twice weekly for 16 weeks. Cohort C received 50 mg three times weekly for 24 weeks, and Cohort D received 100 mg three times weekly for 24 weeks. Cohort E used tiered, weight-based dosing twice weekly and then once weekly for 24 weeks. Cohort F received 70 mg twice weekly and then once weekly for 24 weeks. The registry descriptions also mention dose adjustments for some regimens; the complete protocol details are not reported in the registry record.

For the primary endpoint, the posted cohort counts were 19 of 23 in Cohort A, 24 of 24 in Cohort B, 21 of 23 in Cohort C, 23 of 24 in Cohort D, 23 of 24 in Cohort E, and 21 of 25 in Cohort F. These are the efficacy-evaluable denominators recorded for that endpoint, not the overall enrollment. The registry record does not report a separate non-roxadustat arm in these results.

The study was open-label, so treatment masking was absent. The registry record does not report how randomization was generated, whether allocation was concealed, or how participants were assigned among the six cohorts. It also does not report the rationale for the cohort sizes or a protocol-based sample-size justification. Those omissions limit what can be concluded about the design beyond the registered characteristics.

4. Participant Flow and Accounting

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

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: EE Population included all participants who received at least 2 weeks of study drug with a Hb BL and a valid post-BL value at 2 weeks into study.

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: Safety Population included all participants who received any dose of study drug.

Only the first 6 of 7 groups are shown.

6. Primary Endpoint

The registry lists the primary outcome as Number (%) of Participants With an Hb Response by Week 17, measured over Up to Week 17.

The primary endpoint is the number and percentage of participants with a hemoglobin response by Week 17. A response required a hemoglobin level of at least 11 g/dL and an increase from baseline of at least 1 g/dL. The registry record specifies the timeframe as up to Week 17 and defines the efficacy-evaluable population as participants who received at least 2 weeks of study drug, had a baseline hemoglobin value, and had a valid post-baseline value at 2 weeks into the study.

This is a binary endpoint: each participant either meets the response definition or does not. Reporting the count makes the number of observed responses explicit; the denominator indicates how many participants in that cohort contributed to this endpoint. The endpoint is clinically interpretable as a combined threshold-and-change criterion, although the statistical results alone do not establish the broader clinical meaning of reaching that criterion.

Using both a minimum absolute hemoglobin level and a minimum increase from baseline distinguishes a participant who reaches the target level after an improvement from one who merely has a high value without the specified change. Conversely, a participant may improve but not meet the absolute threshold. The endpoint therefore captures both elements at once, and its result should not be read as a measure of average hemoglobin change.

The registry lists additional hemoglobin outcomes, including response at other time points, changes from baseline, and counts in stated hemoglobin ranges. The primary result remains specific to the Week 17 response definition and the efficacy-evaluable population. Details of how missing values or participants outside that population were handled are not reported in the registry record.

7. Results Posted in the Registry

• Cohort B: Roxadustat Tiered, Weight Based Dosing TIW Then BIW: 100.0% (24/24) versus 82.6% (19/23) on control; risk difference 17.4 percentage points (95% CI 1.9 to 32.9), risk ratio 1.21 (1.00–1.46); unadjusted two-proportion z-test p = 0.033.
• Cohort C: Roxadustat at 50 mg TIW: 91.3% (21/23) versus 82.6% (19/23) on control; risk difference 8.7 percentage points (95% CI -10.6 to 28.0), risk ratio 1.11 (0.88–1.39); unadjusted two-proportion z-test p = 0.381.
• Cohort D: Roxadustat at 100 mg TIW: 95.8% (23/24) versus 82.6% (19/23) on control; risk difference 13.2 percentage points (95% CI -4.2 to 30.7), risk ratio 1.16 (0.94–1.42); unadjusted two-proportion z-test p = 0.142.
• Cohort E: Roxadustat Tiered, Weight Based Dosing BIW Then QW: 95.8% (23/24) versus 82.6% (19/23) on control; risk difference 13.2 percentage points (95% CI -4.2 to 30.7), risk ratio 1.16 (0.94–1.42); unadjusted two-proportion z-test p = 0.142.

8. Statistical Interpretation

9. Statistical Methods Explained

What does the primary outcome count?

The primary outcome is binary at the participant level. A participant is counted as a responder only if both parts of the registry definition are met: hemoglobin is at least 11 g/dL and the increase from baseline is at least 1 g/dL. The reported counts are paired with cohort denominators from the efficacy-evaluable population. This makes the result a count of participants meeting the criterion, rather than a mean or median hemoglobin value.

In general, binary outcomes are often compared using a difference between the proportions in two groups. A risk difference is the proportion in one group minus the proportion in the reference group. A value above zero indicates a higher observed response proportion in the first-named cohort; a value below zero indicates a lower one. The posted estimates compare each listed cohort with Cohort A.

How should the p-values be read?

The posted p-values are unadjusted. In general, a p-value measures how compatible the observed data are with a specified null hypothesis under the analysis assumptions; it is not the probability that the null hypothesis is true, nor does it measure the size or clinical importance of an effect. The unadjusted p-value for Cohort B is 0.033. The registry record does not report a multiplicity adjustment or a primary comparison hierarchy. When multiple comparisons are examined, the chance of at least one apparently unusual result can increase; the statistical plan for addressing that issue is not reported in the registry record.

What do confidence intervals add?

In general, a confidence interval expresses the uncertainty around an estimate using a method with stated coverage properties. It provides a range of values compatible with the data and analysis assumptions, rather than a guarantee that the true effect lies in that particular interval. Here, the intervals are for risk differences. Intervals spanning zero include both a positive and a negative difference, as well as no difference, under the displayed scale. The interval for Cohort B does not span zero, while those for Cohorts C, D, E, and F do.

The cohort denominators are small, and some response counts are close to their denominators. Estimates from small groups can be sensitive to the outcome of a small number of participants. This is one reason to read the interval widths as well as the point estimates and p-values. The registered result does not report adjusted estimates that might account for baseline characteristics or other covariates.

What do the planning illustrations mean?

Planning figures are illustrations based on the observed effect, not the trial's own sample-size justification and not observed power. Their stated assumptions are a two-sided significance level of 0.05, equal allocation between two arms, a normal approximation, and an effect to detect set at the stated fraction of the observed effect. The displayed figures refer to participants per arm for 80% or 90% power.

For Cohort C versus Cohort A, planning for 100% of the observed effect gives 235 participants per arm for 80% power and 314 for 90% power. Planning for 75% of the observed effect gives 447 participants per arm for 80% power and 598 for 90% power. Planning for 50% of the observed effect gives 1,071 participants per arm for 80% power and 1,433 for 90% power.

For Cohort D and Cohort E versus Cohort A, the displayed planning figures are 86 and 114 participants per arm for 80% and 90% power when planning for 100% of the observed effect; 173 and 231 when planning for 75%; and 434 and 580 when planning for 50%. For Cohort F versus Cohort A, the corresponding figures are 11,278 and 15,098; 20,217 and 27,064; and 45,862 and 61,396 participants per arm. These are hypothetical planning illustrations. They do not describe the number enrolled or demonstrate the power of this completed study.

The planning figures vary substantially because the observed differences used as inputs vary. In general, smaller target effects require larger samples to detect under otherwise similar assumptions. This general principle does not validate the observed effects as reliable estimates for planning; the current cohort estimates are uncertain and may not represent effects that would recur in a larger study.

What is not reported about analysis?

The registry record does not report the statistical model or test used for the sponsor's primary analysis, the method used to construct the posted confidence intervals, any covariate adjustment, or how missing endpoint data were handled. It also does not report interim analyses, a sample-size justification, or a multiplicity strategy. The posted comparisons should therefore be interpreted as the estimates and unadjusted p-values recorded, without assuming additional features of an analysis plan.

10. Confidence Intervals and Precision

The precision of a cohort estimate depends on both the observed count and the number of participants contributing to it. The primary endpoint denominators range from 23 to 25 across the six cohorts. With cohorts of this size, the point estimates can be sensitive to a small number of participants, and the confidence intervals provide important context beyond the counts alone.

The risk-difference interval for Cohort B versus Cohort A extends from 0.019 to 0.329. It is entirely above zero, but it remains broad enough to include a range of positive differences. For Cohort C, the interval extends from -0.106 to 0.28. The intervals for Cohorts D and E extend from -0.042 to 0.307, and the interval for Cohort F extends from -0.197 to 0.225. These latter intervals include zero and span both directions of difference.

In general, a confidence interval reflects sampling uncertainty under the method used; it does not account automatically for every possible source of uncertainty. The posted intervals do not, by themselves, address unreported issues such as bias from missing data, differences in participant characteristics, or the effect of examining multiple comparisons. The registry record does not report whether the comparison estimates were adjusted for such factors.

The small denominators and interval widths mean that point estimates should not be read as precise rankings of regimens. In particular, a numerical difference between two estimates is not necessarily evidence of a dependable difference. The confidence intervals describe uncertainty around each comparison with Cohort A, but the registry record does not report intervals for direct comparisons between the other cohorts.

11. Multiplicity, Missing Data and Analysis Populations

The primary endpoint is defined for an efficacy-evaluable population: participants had to receive at least 2 weeks of study drug, have a baseline hemoglobin value, and have a valid post-baseline value at 2 weeks into the study. The posted cohort denominators for this endpoint are 23, 24, 23, 24, 24, and 25. The overall enrollment was 145, but the registry record does not explain the relationship between enrollment and the endpoint population for each cohort or describe reasons participants were not included in the primary analysis.

The registry record does not report how missing hemoglobin measurements or other missing outcome data were handled. It is not reported in the registry record whether missing values were treated as non-response, excluded, imputed, or addressed using another method. It also does not report a sensitivity analysis for missing data. These details matter because the endpoint depends on having both baseline and follow-up information.

The record lists one primary endpoint and multiple secondary hemoglobin outcomes, including response at several time points, hemoglobin change, and counts within stated hemoglobin ranges. The posted comparisons also include several cohorts contrasted with Cohort A. The registry record does not report whether statistical testing was adjusted for multiple outcomes or comparisons, whether testing followed a prespecified hierarchy, or whether any outcomes were designated as exploratory. The listed p-values are unadjusted.

In general, multiple testing can make a nominally small p-value less persuasive as evidence when many analyses are considered. Without a reported plan for multiplicity, the Cohort B p-value should be viewed as an unadjusted result rather than a multiplicity-controlled finding. The registry record does not report interim analyses or stopping rules, and no assumptions about them can be made from the posted endpoint results.

12. 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: cardiac failure congestive (12.0%), cellulitis (8.0%), cardio-respiratory arrest (4.2%), acute myocardial infarction (4.2%), renal failure acute (4.2%).

The registry reports serious adverse event counts by cohort among participants at risk. The number affected was 3 of 24 in Cohort A, 4 of 24 in Cohort B, 4 of 24 in Cohort C, 4 of 24 in Cohort D, 10 of 24 in Cohort E, and 10 of 25 in Cohort F. Other adverse events affected 15 of 24, 15 of 24, 10 of 24, 10 of 24, 11 of 24, and 10 of 25 participants, respectively. These are separate summaries of participants affected; they do not show event rates over time or establish that events were caused by treatment.

The registry also lists serious event terms and percentages: Cardiac Failure Congestive, 12.0%; Cellulitis, 8.0%; Cardio-Respiratory Arrest, 4.2%; Acute Myocardial Infarction, 4.2%; Renal Failure Acute, 4.2%; and Renal Failure Chronic, 4.2%. The record does not provide additional context in these figures about event timing, recurrence, severity beyond the serious-event classification, or attribution.

Death counts are not reported in the registry record. The cohort counts should be read descriptively: the denominators are small, and the figures do not establish comparative safety. This information alone cannot determine whether an event was caused by roxadustat or how risks compare with another treatment.

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

The posted primary results show that participants meeting the defined hemoglobin response criterion were recorded in every cohort. The counts range from 19 of 23 in Cohort A to 24 of 24 in Cohort B, with the other cohort counts and denominators as stated in the primary endpoint results. These findings describe response within the efficacy-evaluable groups; they do not establish the effect of roxadustat compared with no treatment, placebo, or another drug.

Relative to Cohort A, Cohort B has a positive posted risk-difference estimate, and its unadjusted confidence interval does not include zero. The remaining posted intervals include zero. Because the cohorts are small and the record does not report multiplicity adjustment, these comparisons do not support a definitive ranking of all regimens. The outcomes also reflect different dosing schedules and, for some cohorts, different study durations. They cannot isolate the effect of dose amount alone.

The primary response definition is a specific combined criterion, not a complete assessment of anemia, kidney disease, participant experience, or long-term outcomes. The registered primary timeframe is up to Week 17. Other hemoglobin measures are listed, but the record does not report all corresponding results in the information summarized for this endpoint. Long-term effects and outcomes beyond the registered study periods are not reported in the registry record.

In general, statistical uncertainty and design limitations should temper interpretation of an observed difference. An unadjusted p-value does not establish clinical importance, and a confidence interval does not resolve possible missing-data or multiplicity concerns. The results can be described as cohort-level response counts and comparisons, but they do not establish that any particular regimen is preferable or make a treatment recommendation.

14. Why This Trial Matters Statistically

This trial illustrates how a binary endpoint can make a treatment response operational: a participant must satisfy both a hemoglobin threshold and a minimum increase from baseline. That definition is more specific than a simple change score, but it also means the result depends on both baseline and follow-up measurements and on the chosen cutoff.

It also demonstrates the importance of interpreting estimates alongside their precision. The cohort response counts are high in relation to their denominators, but the groups are small, and the risk-difference intervals vary in width. The Cohort B comparison has a positive estimate and an interval above zero, while the other posted intervals include zero. These contrasts show why point estimates alone are not enough to assess the strength of evidence.

The study has multiple cohorts and multiple registered outcomes, yet the registry record does not report a multiplicity plan or a sample-size justification. This makes it important to distinguish unadjusted comparisons from confirmatory evidence. The planning illustrations also show how assumed effect size influences the participant counts needed for a future study; they are planning scenarios, not observed power or evidence that the observed effect is stable.

Finally, the open-label design, lack of a non-roxadustat comparison in the posted results, and distinct cohort schedules constrain causal interpretation. The registry record offers useful detail about endpoint definitions and observed counts, while leaving important analysis and design details unreported. Recognizing both aspects is central to responsible statistical reading.

15. Limitations

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

Several limitations are specific to these registered results. The primary endpoint uses an efficacy-evaluable population rather than all enrolled participants. The inclusion criteria require treatment exposure and valid hemoglobin measurements, while the registry record does not explain exclusions from the primary analysis or report how missing data were handled. Consequently, the posted counts cannot be assumed to represent all enrolled participants.

The six cohorts are small, with primary endpoint denominators from 23 to 25. The resulting comparisons have considerable uncertainty, and the reported confidence intervals for Cohorts C, D, E, and F include zero. The Cohort B interval is also broad, despite excluding zero. Multiple cohort comparisons and multiple registered secondary outcomes are present, but multiplicity handling is not reported.

The study is open-label, and the registry record does not describe allocation concealment or the randomization procedure. All posted cohorts received roxadustat, so the primary comparisons are between regimens, not against an untreated or active comparator. Cohort schedules also differ in frequency and duration, which makes it difficult to attribute differences specifically to one dosing feature.

The registry record does not report the primary analysis method, covariate adjustment, interim analyses, or the rationale for the sample size. It also lists safety counts without reporting death counts. These missing details limit assessment of bias, analysis robustness, and comparative safety. The study's Phase 2 status and observed response counts should not be treated as proof of a definitive regimen ranking.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Study of Roxadustat in Non-Dialysis Chronic Kidney Disease Participants With Anemia — NCT01244763.
• Registry record last updated 2022-02-10; status last verified 2022-01.
• Results first posted 2022-02-10.
• 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.

19. Record Summary

NCT01244763 is a completed Phase 2, randomized, open-label, parallel study of roxadustat in participants with anemia and non-dialysis chronic kidney disease. The registry lists 145 enrolled participants and six dosing cohorts. The primary endpoint was the number of participants with a hemoglobin response by Week 17, defined as hemoglobin of at least 11 g/dL and an increase from baseline of at least 1 g/dL.

Posted primary endpoint counts ranged from 19 of 23 in Cohort A to 24 of 24 in Cohort B. Unadjusted risk-difference comparisons use Cohort A as the reference. The Cohort B interval was above zero; intervals for the other compared cohorts included zero. The registry record does not report a multiplicity strategy, missing-data method, primary analysis details, or protocol-based sample-size justification. Because every cohort received roxadustat and the cohort schedules differed, these data describe responses across dosing regimens rather than treatment effects against a control.