Anemia Studies in Chronic Kidney Disease (CKD): Erythropoiesis Via a Novel Prolyl Hydroxylase Inhibitor (PHI) Daprodustat in Non-Dialysis Subjects Evaluating Hemoglobin (Hgb) and Quality of Life (ASCEND-NHQ): Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Anemia Studies in Chronic Kidney Disease (CKD): Erythropoiesis Via a Novel Prolyl Hydroxylase Inhibitor (PHI) Daprodustat in Non-Dialysis Subjects Evaluating Hemoglobin (Hgb) and Quality of Life (ASCEND-NHQ) (NCT03409107), a randomized phase 3 study. The primary outcome is Mean Change in Hemoglobin From Baseline and Over the Evaluation Period (Mean Over Week 24 and 28).

2. Clinical Question

Quality of life and safety were also part of the study’s stated purpose. The registry lists further outcomes concerning hemoglobin response, time in a target range, rescue criteria, and questionnaire scores. The primary comparison, however, is the adjusted between-group estimate for hemoglobin change over the evaluation period.

3. Trial Design

This was a Phase 3, randomized, double-blind, placebo-controlled, parallel-group, multicenter study. Its official title describes a 28-week study in recombinant human erythropoietin-naïve, non-dialysis participants with anemia associated with chronic kidney disease. The registry lists anemia as the condition. The study status is completed.

Participants were assigned to once-daily oral daprodustat or once-daily oral placebo. The registry records enrollment of 614. The arm results for the primary endpoint list 307 participants in each group. The primary analysis used the Intent-to-Treat population, which included all randomized participants regardless of study-drug use. The registry does not report further details about the randomization procedure, concealment, or how masking was maintained.

The primary outcome was mean change in hemoglobin from baseline, assessed as the average of post-randomization measurements during Week 24 to Week 28, inclusive. Baseline was the latest non-missing pre-dose assessment on or before randomization. The study’s description says blood samples were collected at given time points. The registry lists 22 outcome measures and 27 analyses. It identifies secondary outcomes concerning hemoglobin response and target range, the SF-36 vitality domain, rescue criteria, and the Chronic Kidney Disease–Anemia Questionnaire.

The registry lists GlaxoSmithKline as sponsor. The recorded study start was 2018-03-05 and primary completion was 2020-10-07. The original sample-size justification, the planned allocation ratio, and further details of the protocol are not reported in the registry record. The presence of an enrolled sample and completed status does not itself explain how the original target was chosen.

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 1336 participants were screened, of which 722 were screen failures. A total of 614 participants were enrolled in the study.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Intent-to-Treat (ITT) Population comprised all randomized participants regardless of whether they took study 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.

6. Primary Endpoint

The registry lists the primary outcome as Mean Change in Hemoglobin From Baseline and Over the Evaluation Period (Mean Over Week 24 and 28), measured over Baseline (Day 1) and Week 24 to Week 28.

The primary endpoint is mean change in hemoglobin, measured in grams per deciliter, from baseline to the evaluation period. The evaluation-period value is defined as the mean of all available post-randomization hemoglobin values, whether measured on or off treatment, during Week 24 to Week 28, inclusive. Change is that average minus the baseline value. Baseline is the latest non-missing pre-dose assessment on or before randomization.

Hemoglobin is the measured quantity central to the trial’s stated evaluation of anemia treatment. Using an average across the evaluation period describes the level over that period rather than relying on a single post-randomization measurement. The registry also lists secondary outcomes based on hemoglobin response and time within a specified hemoglobin range, but those are distinct questions from the primary mean-change endpoint.

For the primary analysis, missing post-baseline hemoglobin values were imputed using prespecified multiple imputations. The analysis used an ANCOVA model with terms for treatment, baseline hemoglobin, and region. This model estimates a treatment-group contrast while accounting for the listed baseline measure and region. The registry’s primary population was the Intent-to-Treat population, including all randomized participants regardless of whether they took study drug. These definitions matter because the estimate describes a model-based comparison of average change over the evaluation period, not an unadjusted comparison of one visit or a result limited to participants who completed treatment.

7. Results Posted in the Registry

• Daprodustat: mean 1.58 versus 0.19 on control; difference 1.39 (95% CI 1.22 to 1.56), standardized effect d = 1.29; z-test p = <0.001.

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.

Percentage of Participants With Hemoglobin Increase of >=1.0 Grams Per Deciliter From Baseline to Evaluation Period

• Endpoint type: Binary: the percentage of participants with the event (counts rebuilt from the posted percentages and group sizes).
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • Daprodustat: 76.9% (236/307) versus 17.9% (55/307) on control; risk difference 59.0 percentage points (95% CI 52.6 to 65.3), risk ratio 4.29 (3.35–5.49); unadjusted two-proportion z-test p = <0.001.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the proportion with the event is statistically higher for Daprodustat than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

Change From Baseline in Short Form-36 (SF-36) Questionnaire Vitality Domain Score by Traditional Scoring at Week 28

• Endpoint type: Continuous: a mean value per participant, compared as a difference in means.
• Test: Unadjusted two-sample z-test for the difference in means, using the posted standard errors.
• Result:
  • Daprodustat: mean 7.29 versus 1.93 on control; difference 5.36 (95% CI 2.20 to 8.52), standardized effect d = 0.27; z-test p = <0.001.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the mean is statistically higher for Daprodustat than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

Percentage of Participants With Hgb Response (Hgb in the 11-12 Grams/Deciliter Range) During Evaluation Period (Week 24 to Week 28 Inclusive)

• Endpoint type: Binary: the percentage of participants with the event (counts rebuilt from the posted percentages and group sizes).
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • Daprodustat: 52.1% (160/307) versus 8.1% (25/307) on control; risk difference 44.0 percentage points (95% CI 37.6 to 50.3), risk ratio 6.40 (4.33–9.46); unadjusted two-proportion z-test p = <0.001.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the proportion with the event is statistically higher for Daprodustat than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

The reported secondary results broadly align in direction with the primary hemoglobin finding: the posted analyses favor daprodustat on hemoglobin response measures, and the SF-36 vitality estimate is also higher in the daprodustat group. These endpoints examine different aspects of outcome than the primary mean change, and they should not be treated as interchangeable confirmations of the same measure.

The registry lists additional secondary outcomes, including time within a hemoglobin target range, post-randomization hemoglobin change at Week 28, rescue-treatment criteria, and questionnaire domains. Numerical results for those outcomes are not reported in the registry record in the information available for this analysis. Their absence from the reported results should not be read as evidence of either benefit or no effect.

Secondary findings generally carry less inferential weight than a prespecified primary endpoint. A study may examine many outcomes, and testing more hypotheses increases the chance that some results appear notable by chance. Adjustment for multiple comparisons can address that issue, but multiplicity procedures for the secondary outcomes are not reported in the registry record. The reported secondary p-values are therefore best read as endpoint-specific, unadjusted evidence rather than a definitive set of independent confirmations.

Different outcome definitions, assessment times, and analysis methods also limit direct comparisons across endpoints. A hemoglobin response threshold, an average vitality score change, and time in a target range each represent distinct questions. Taken together, the reported secondary pattern is supportive in direction, while incomplete numerical reporting and unreported multiplicity handling limit how strongly it can be interpreted.

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

• Percentage of Participants With Hemoglobin Increase of >=1.0 Grams Per Deciliter From Baseline to Evaluation Period
• Change From Baseline in Short Form-36 (SF-36) Questionnaire Vitality Domain Score by Traditional Scoring at Week 28
• Percentage of Participants With Hgb Response (Hgb in the 11-12 Grams/Deciliter Range) During Evaluation Period (Week 24 to Week 28 Inclusive)
• Percentage of Time With Hgb Within the Target Range (11-12 Grams Per Deciliter) During Evaluation Period (Week 24 to Week 28 Inclusive) (Hodges-Lehmann Estimate)
• Percentage of Time With Hgb Within the Target Range (11-12 Grams Per Deciliter) During Evaluation Period (Week 24 to Week 28 Inclusive) (Mann-Whitney Estimate)
• Change From Baseline in Post-randomization Hgb at Week 28
• Rate of Participants Permanently Stopping Randomized Treatment Due to Meeting Rescue Criteria
• Change From Baseline by Domain and Single Item Scores on the Chronic Kidney Disease -Anemia Questionnaire (CKD-AQ) Symptom Questionnaire

9. Statistical Interpretation

10. Statistical Methods Explained

The primary analysis used analysis of covariance, or ANCOVA. In general, ANCOVA models an outcome while including treatment group and selected covariates. Here, the registry specifies treatment, baseline hemoglobin, and region. Including baseline hemoglobin can account for differences in starting measurements when estimating the treatment contrast; region is another model term stated in the registry. The reported least-squares mean difference is a model-adjusted contrast, rather than simply the raw difference between two group averages.

What does the primary estimate represent?

The endpoint is the change from baseline in the mean of available post-randomization hemoglobin values during the evaluation period. The reported 1.40 grams per deciliter is the estimated adjusted difference between treatment groups for that endpoint. A positive contrast in this coding corresponds to a higher estimated mean change for daprodustat than placebo. It is a comparison of group-level estimates; it does not imply that each person’s measurement changed by this amount.

The registry gives group least-squares means of 1.58 for daprodustat and 0.19 for placebo, each with a standard error. A standard error describes uncertainty in an estimated mean under the analysis, rather than the variation among individual observations. It should not be interpreted as a standard deviation. The record does not provide the participant-level distribution of hemoglobin change in the registry record used.

Why is an adjusted analysis used?

In a randomized trial, randomization is intended to support a fair group comparison. Adjustment for baseline hemoglobin and region can improve the precision of the estimated contrast and accounts for those specified model terms. The registry does not report additional covariates, model diagnostics, or details such as the precise coding of region. Those aspects should not be inferred from the ANCOVA label alone.

The registry’s primary analysis includes all randomized participants in the Intent-to-Treat population, regardless of whether they took study drug. This preserves the randomized comparison as defined in the record. It does not mean all measurements were observed: the registry explicitly says missing post-baseline hemoglobin measurements were imputed using prespecified multiple imputations. The imputation method aims to incorporate uncertainty about missing values, but its validity depends on assumptions about the missing data and the imputation model. Details of those assumptions and model specifications are not reported in the registry record.

What does the confidence interval convey?

A confidence interval gives a range of estimates compatible with the data and the statistical model at the stated confidence level. In repeated use of the same procedure under its assumptions, intervals constructed this way would cover the underlying parameter at the stated rate. It is not a range containing a stated proportion of individual responses, nor a probability statement about the fixed parameter after observing this particular interval. The primary 95% confidence interval is 1.23 to 1.56 grams per deciliter. Its width indicates the estimate is relatively precise on the scale of the endpoint, while uncertainty remains.

The p-value is a measure of how incompatible the observed data are with a specified null hypothesis under the model and its assumptions. The primary analysis reports p < 0.001. It is not the probability that the null hypothesis is true, and it does not measure the size or clinical importance of the treatment contrast. The registry’s superiority label is consistent with its reported hypothesis test, but interpretation should also consider the estimate, confidence interval, endpoint definition, and limitations.

Why do the simple recalculations differ?

The recalculated comparison uses posted summary statistics and is explicitly unadjusted. It yields 1.39 grams per deciliter with a 95% confidence interval of 1.22 to 1.56 and p < 0.001. The sponsor’s reported analysis uses ANCOVA and gives 1.40 with a 95% confidence interval of 1.23 to 1.56 and p < 0.001. These estimates are similar, but they are not interchangeable: the adjusted analysis uses the registry’s stated model terms, while the recalculation does not reconstruct that model.

For a continuous outcome, the uncertainty of a group mean depends on the outcome’s variability and the number of observations contributing to it. For a binary outcome, the uncertainty depends on the event frequency and group size. In either setting, a confidence interval is more informative about precision than a point estimate alone. The registry’s secondary outcomes use different measures and methods, so their effect estimates and intervals should be interpreted on their own scales.

Why can over-dispersion matter?

As a general statistical point, over-dispersion occurs when observed variation exceeds what a chosen probability model expects. If a model assumes less variation than the data contain, standard errors may be too small and confidence intervals too narrow. Other model choices may be needed to represent the data adequately. The registry entries available for this analysis do not state that over-dispersion was assessed or that it affected any analysis; this is a general consideration, not a finding about this trial.

The registry also reports binary secondary analyses using the Cochran-Mantel-Haenszel method and continuous secondary analyses using ANCOVA. These methods are designed for different outcome types and answer distinct questions. The Cochran-Mantel-Haenszel approach estimates a response-rate contrast while accounting for strata, but the strata used are not identified in the registry record. The registry does not report a complete statistical analysis plan, diagnostics, or all model details. Interpretation therefore rests on the analysis descriptions and results the registry does report.

11. Confidence Intervals and Precision

The primary reported estimate is a least-squares mean difference of 1.40 grams per deciliter, with a 95% confidence interval from 1.23 to 1.56. The interval quantifies uncertainty in the adjusted average contrast under the stated ANCOVA model. Because both ends are above zero, the reported interval is consistent with a positive average difference in the direction of daprodustat relative to placebo under that model.

Precision concerns how narrowly an effect is estimated, not whether an effect is important or applies uniformly to participants. The interval does not describe individual-level variability, and it does not resolve questions about missing-data assumptions or outcomes outside the measured period. The reported p-value, < 0.001, addresses a null-hypothesis test and does not replace the interval as a description of estimate uncertainty.

The unadjusted recalculation is 1.39, with a 95% confidence interval from 1.22 to 1.56. Its close proximity to the reported adjusted interval provides a similar broad account of the primary contrast, but the recalculation does not reproduce the model adjustment. The registry’s ANCOVA estimate remains the appropriate result to cite as the sponsor-reported primary analysis.

Secondary outcomes have their own uncertainty and scales. For instance, the posted SF-36 estimate has a reported 95% confidence interval of 2.17 to 8.56 in the sponsor analysis. Binary response outcomes are expressed as differences in response rate in their reported analyses. These intervals address their respective endpoints; their widths should not be compared as if the outcome scales or analysis methods were the same. The registry record does not report confidence intervals for every outcome listed.

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 illustrations based on the observed effect, not the trial’s original design calculation. For these non-time-to-event outcomes, the figures mean participants per arm. The stated planning assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation; each illustration considers detecting the stated fraction of the observed effect. The trial’s own original design assumptions and sample-size justification are not reported in the registry record.

The 80% power column gives the illustrated participants per arm for a design that would have an 80% chance of detecting the assumed effect under the planning assumptions, if that effect and the statistical model assumptions were true. The 90% power column expresses the corresponding planning target for a 90% chance. These are design-planning concepts, not guarantees that a future study will detect an effect. The figures are not observed power.

Rows that plan for 100% of the observed effect use the full effect estimated in this trial as the hypothetical target. Rows for 75% and 50% use smaller hypothetical effects. Smaller effects are harder to distinguish from random variation, so they require more participants per arm to reach the same planned power under otherwise fixed assumptions. The figures are not evidence that the original trial was designed around these effects, and the record does not establish that the assumptions match a future study.

Where a planning curve accompanies the figures, it should be read as a visualization of the relationship between assumed effect size and required participants under the stated planning assumptions. It is not a display of actual enrollment outcomes or observed power. A sample-size illustration based on an effect observed in the same trial can be optimistic if that observed estimate overstates the underlying effect; the registry’s planning assumptions do not quantify that uncertainty.

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 = 1.390; pooled standard deviation σ = 1.078.
2. n = 2 × (1.96 + 0.842)² × 1.078² / 1.390² = 9.4, rounded up to 10 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 <- 1.39                      # observed difference in means
sigma <- 1.07760266332262                     # 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(3409107)
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 = 10 (80% power), 13 (90% power)
 75% of observed effect: n per arm = 17 (80% power), 23 (90% power)
 50% of observed effect: n per arm = 38 (80% power), 51 (90% power)
power.t.test: 10.5 
Simulated power at n = 10 per arm: 0.775 (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 sponsor-reported primary result is an ANCOVA least-squares mean difference of 1.40 grams per deciliter, with a 95% confidence interval of 1.23 to 1.56 and p < 0.001. The simple recalculation from posted summary statistics gives an unadjusted mean difference of 1.39, with a 95% confidence interval of 1.22 to 1.56 and p < 0.001. The estimates are close, but the methods differ.

The registry’s ANCOVA includes treatment, baseline hemoglobin, and region. The recalculation uses summary statistics and does not adjust for those model terms. It is a useful independent arithmetic check on the broad magnitude and direction, but it is not a substitute for the reported model analysis. The registry indicates no discrepancy between the recalculation and the reported analysis, and the sponsor’s analysis remains the result to cite for the trial’s primary endpoint.

Related differences also appear in secondary analyses: the registry reports sponsor analyses for those outcomes using the stated endpoint-specific methods. Recalculations from summary statistics should not be expected to reproduce methods that use adjustments or other analysis details not captured by the summary data. The comparison should therefore distinguish a simple unadjusted recalculation from the registry’s reported inferential analysis rather than treating small numerical differences as contradictions.

15. Multiplicity, Missing Data and Analysis Populations

The primary analysis description specifies multiple imputation for missing post-baseline hemoglobin values. The registry says the imputation was prespecified, but the imputation model, assumptions about why measurements were missing, number of imputed datasets, and sensitivity analyses are not reported in the registry record. This matters because conclusions from imputed data depend on whether the model reasonably represents the observed data and the missingness process.

The primary analysis population is Intent-to-Treat, including all randomized participants regardless of whether they took study drug. The registry’s primary endpoint definition includes available post-randomization measurements on and off treatment, with missing post-baseline values handled through multiple imputation. Further details on treatment discontinuation, data collection after discontinuation, and the handling of other missing outcomes are not reported in the registry record.

The registry reports many outcome measures and analyses, including secondary outcomes. Testing multiple outcomes creates a multiplicity issue: even when no underlying effect exists for any one of several questions, the chance of at least one apparently notable result can increase as more tests are considered. Statistical procedures can control that risk, but the multiplicity strategy for the secondary outcomes is not reported in the registry record. The reported secondary p-values should therefore not be interpreted as if they were necessarily adjusted for all analyses.

The registry records a superiority analysis for the primary endpoint and reports a p-value and confidence interval. Further details of the complete statistical analysis plan, including any interim analysis plans, are not reported in the registry record. No interim-analysis procedure or stopping rule should be inferred from the completed status. These gaps do not erase the reported estimate, but they constrain assessment of how all analysis decisions were planned and how robust the findings are to missing-data or multiplicity choices.

16. Safety Results

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

Deaths

Other (non-serious) adverse events

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

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: anaemia (2.6%), renal failure (2.0%), acute kidney injury (1.6%), cardiac failure acute (1.3%), urinary tract infection (1.3%).

The registry reports serious adverse events by group: the placebo group had 68 affected participants among 306 at risk, and the daprodustat group had 62 affected participants among 308 at risk. Deaths are listed as 16 for placebo and 10 for daprodustat. The at-risk denominators in this safety record differ from the 307 participants listed per arm for the primary outcome; the registry’s safety figures should be read with their own stated denominators.

The most frequently listed serious adverse event terms include anemia at 2.6%, renal failure at 2.0%, acute kidney injury at 1.6%, cardiac failure acute at 1.3%, urinary tract infection at 1.3%, and cardiac failure at 1.3%. These are registry-reported terms and percentages. No causal attribution should be inferred from the counts alone. The registry record, as, does not provide enough detail to assess event timing, severity distributions, exposure-adjusted rates, or comparative risk estimates. These figures are descriptive safety information, not a complete assessment of safety.

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

The primary result describes a statistically adjusted difference in average hemoglobin change between daprodustat and placebo during the specified evaluation period among the Intent-to-Treat population. The sponsor-reported estimate is 1.40 grams per deciliter, with a 95% confidence interval of 1.23 to 1.56. The interval and reported p-value support a difference under the specified analysis model; they do not determine how every participant responded.

The result does not establish that the same change would occur in every population, at every time point, or beyond the study timeframe. It does not establish a particular effect on outcomes not reported numerically, nor does it make the reported average a measure of individual benefit. The secondary measures address separate questions and have their own uncertainty. Missing-data assumptions and multiplicity handling also affect how broadly the full set of results can be interpreted.

The registry provides serious adverse event counts and selected event terms, but those descriptive figures alone cannot establish the cause of events or fully characterize comparative safety. The information available does not provide exposure-adjusted event comparisons or all details needed to assess safety over time. Efficacy and safety should be considered as distinct dimensions of the trial record.

Statistical significance is not the same as clinical importance. The registry reports the measured endpoint and analysis, but a judgment about the practical importance of the estimate requires context not reported, including what magnitude of hemoglobin change would be meaningful to participants and how outcomes relate to their experience. This statistical description does not offer medical advice or establish a treatment decision.

18. Why This Trial Matters Statistically

This trial illustrates how the choice of endpoint definition shapes the question being answered. Its primary outcome uses average hemoglobin measurements over an evaluation period rather than a single assessment. That can characterize the group’s average level during the period, while still leaving questions about individual trajectories and responses outside the endpoint definition.

It also illustrates the distinction between an adjusted sponsor analysis and a simple recalculation from summary statistics. The reported ANCOVA estimate accounts for baseline hemoglobin and region, while the recalculation does not. Their close results offer a useful consistency check, but the model-based estimate is the primary inferential result described by the registry.

The record provides a clear example of why confidence intervals matter alongside p-values. The estimate’s interval conveys the precision of the average contrast under the model, while the p-value addresses evidence against a null hypothesis. Neither answers questions about individual variation, clinical importance, or unreported outcomes.

Finally, the trial highlights issues that matter when reading a broad outcome program: missing endpoint measurements, numerous secondary questions, and safety information all require careful interpretation. Multiple imputation is reported for missing primary hemoglobin values, but details of its assumptions are absent. Multiplicity handling for secondary outcomes is also not reported. These are not reasons to discard the primary result; they are reasons to distinguish the well-described primary comparison from less fully specified parts of the record.

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.

The registry record does not provide the complete statistical analysis plan, detailed randomization procedures, or information about how masking was maintained. It also does not report the original sample-size justification, the trial’s original design assumptions, or any interim-analysis plan. These omissions limit assessment of how the reported analysis compares with all planned procedures.

Although multiple imputation is specified for missing primary hemoglobin measurements, details of the imputation model and assumptions are not reported. Sensitivity analyses for departures from those assumptions are also not reported in the registry record. The available information does not describe the extent or pattern of missingness, nor does it provide detailed handling of missing values for all secondary outcomes.

Many secondary measures are listed, but numerical results are not available for all of them in the information. The registry does not report the multiplicity strategy for secondary analyses. As a result, the secondary findings should be interpreted cautiously and not treated as a set of independently confirmed effects.

The serious adverse event information is descriptive and uses stated safety denominators that differ from those listed for the primary endpoint. The record does not provide exposure-adjusted comparisons, detailed event timing, or causal assessments in the registry record for this analysis. The trial’s stated timeframe and evaluation period also bound what can be concluded from the primary endpoint; findings beyond those periods are not reported in the registry record.

20. Related Tutorials

21. Related Statistical Calculators

22. Sources

• ClinicalTrials.gov: Anemia Studies in Chronic Kidney Disease (CKD): Erythropoiesis Via a Novel Prolyl Hydroxylase Inhibitor (PHI) Daprodustat in Non-Dialysis Subjects Evaluating Hemoglobin (Hgb) and Quality of Life (ASCEND-NHQ) — NCT03409107.
• Registry record last updated 2024-04-02; status last verified 2024-03.
• Results first posted 2021-11-03.
• 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

NCT03409107 is a completed Phase 3 randomized, double-blind, placebo-controlled, parallel-group multicenter study of once-daily oral daprodustat versus once-daily oral placebo in non-dialysis participants with chronic kidney disease-associated anemia who were naïve to recombinant human erythropoietin. The record lists enrollment of 614 and identifies GlaxoSmithKline as sponsor.

The primary outcome was mean hemoglobin change from baseline over Week 24 to Week 28, inclusive. The sponsor-reported ANCOVA estimate was a least-squares mean difference of 1.40 grams per deciliter, with a 95% confidence interval of 1.23 to 1.56 and p < 0.001. Missing post-baseline primary hemoglobin values were imputed using prespecified multiple imputations. The analysis population was Intent-to-Treat.

Reported secondary findings broadly favor daprodustat on the posted hemoglobin-response and vitality measures. Serious adverse events and deaths are reported descriptively. Details of the original sample-size rationale, full multiplicity strategy, and imputation assumptions are not reported in the registry record.