Study to Evaluate the Effect of GBT440 on TCD in Pediatrics With Sickle Cell Disease: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Study to Evaluate the Effect of GBT440 on TCD in Pediatrics With Sickle Cell Disease (NCT04218084), a randomized phase 3 study. The primary outcome is Change From Baseline in Time-Averaged Maximum of Mean Velocity (TAMMV) Arterial Cerebral Blood Flow at Week 24.

2. Clinical Question

3. Trial Design

This was a Phase 3 randomized, double-blind, placebo-controlled, parallel-group study of voxelotor in pediatric participants with sickle cell disease. The masking designation in the registry is quadruple. The study compared daily voxelotor with matching placebo, and the voxelotor formulation could be a tablet, dispersible tablet, or powder for oral suspension. The registry lists Pfizer as the sponsor.

The planned enrollment was 236 participants. The registry gives a study start date of 2020-11-11 and a primary completion date of 2024-01-18. Its current status is terminated. The reason for termination and the timing or circumstances of that decision are not reported in the registry record. The record also does not report whether the planned enrollment was reached or provide details about the duration of treatment for each participant.

The study population was described as children aged 2 years and older but under 15 years. Participants were eligible for the stated TCD flow category: conditional velocities of 170 to < 200 cm/sec. The registry distinguishes this category from normal flow, defined as < 170 cm/sec, and abnormal flow, defined as at least 200 cm/sec.

The primary outcome was change from baseline in TAMMV at Week 24. The baseline value was the value at screening. The primary analysis population is described as all randomized participants, while the record clarifies that the analyzed number for this outcome represents participants evaluable for it. The registry posts 114 evaluable participants in the voxelotor arm and 108 in the placebo arm. The record does not report the protocol's sample-size justification, allocation ratio details beyond the observed counts, or further operational details of randomization and masking.

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 236 participants were assigned to the study treatment. Study was terminated based on sponsor decision.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: ITT analysis population included all randomized participants. Here, "Overall Number of Participants Analyzed'' signifies number of participants evaluable for this outcome measure.

5. Baseline Characteristics

The table reproduces the baseline characteristics posted in the registry, which describe the participants analysed at baseline and help show whether the randomized groups were comparable.

Population: Intent-to-Treat (ITT) population included all randomized participants.

6. Primary Endpoint

The registry lists the primary outcome as Change From Baseline in Time-Averaged Maximum of Mean Velocity (TAMMV) Arterial Cerebral Blood Flow at Week 24, measured over Baseline (value at screening), Week 24.

TAMMV is the time-averaged maximum of mean velocity in an artery, measured using transcranial Doppler ultrasound. The registry describes it as an ultrasound measure used to assess blood flow in cerebral arteries. The primary endpoint compares the change from the screening baseline to Week 24 between the voxelotor and placebo groups.

TCD velocity is used in this study to classify cerebral arterial flow. The record defines normal flow as < 170 cm/sec, conditional flow as 170 to < 200 cm/sec, and abnormal flow as at least 200 cm/sec. The eligible population had conditional flow velocities. Measuring change in TAMMV therefore asks whether the average measured velocity changed differently between randomized groups over the stated follow-up period.

The endpoint is continuous: each participant's measurement can take a range of values rather than falling into only a small set of categories. The posted arm estimates are least-squares means, and the primary contrast is a difference in those means from a repeated-measures model. A negative change indicates a lower Week 24 value relative to baseline on the scale used by the analysis. The between-group estimate describes an average contrast, not the result for every participant. It also does not directly report how many participants moved between the registry's normal, conditional, and abnormal categories.

General: continuous measurements can retain information about the size and direction of change, while a category-based outcome instead records whether a threshold was crossed. The primary endpoint should therefore be interpreted as a change in measured velocity, not as a direct count of clinical events.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The primary outcome is a continuous change-from-baseline measure, and the registry reports least-squares means for each arm. The primary comparison uses a mixed model for repeated measures (MMRM), with the reported contrast expressed as a difference in least-squares means. This approach is designed for outcomes measured repeatedly over time and estimates group means while accounting for the repeated measurement structure specified in the model. The registry record does not report the full model specification, including its covariates, covariance structure, or handling of missing measurements.

What does the least-squares mean represent?

General: a least-squares mean is a model-based estimate of an outcome mean. Depending on the model, it can represent a mean adjusted for included terms rather than a simple arithmetic average of all observed values. The posted primary result is therefore best read as the model's estimated between-arm difference, not as a difference reconstructed from individual participant data. The registry does not give enough detail to reproduce the model independently.

The posted voxelotor least-squares mean change is -12.06 cm/sec, and the placebo estimate is -4.29 cm/sec. The sponsor-reported model contrast is -7.77 cm/sec. The contrast is reported with a 95% confidence interval from -13.18 to -2.37 and p = 0.005. For interpreting the primary question, the contrast and its interval are more informative than comparing the overlap of the separate arm-level intervals. Separate confidence intervals are not a substitute for the confidence interval calculated directly for the between-arm contrast.

How should the confidence interval and p-value be read?

General: a confidence interval gives an estimate's uncertainty under the analysis assumptions. It conveys both direction and precision, while a p-value summarizes how compatible the data are with a null hypothesis under those assumptions. Neither quantity measures the probability that a treatment is effective, and neither alone establishes clinical importance. The primary interval excludes zero, while the reported p-value is 0.005. These are two related summaries of the same model-based comparison, not independent pieces of evidence.

The registry specifies a 95% confidence interval for the primary contrast. The primary result is a difference on the cm/sec scale, so its magnitude is expressed in the same units as the endpoint. The analysis population is described as all randomized participants, but the registry also says that the analyzed number represents those evaluable for the outcome. This distinction matters: the intended population description does not mean every randomized participant necessarily contributed an evaluable Week 24 outcome.

Why do repeated measurements and over-dispersion matter?

General: repeated measurements on the same participant are usually related rather than independent. A repeated-measures model accounts for this dependence according to its specified structure; treating all measurements as independent can misstate uncertainty. Over-dispersion is a separate concern that arises when the variation in data exceeds what a chosen model expects. Accounting for extra variation can affect standard errors and confidence intervals. The registry does not report whether over-dispersion was assessed or how the repeated-measures covariance was specified, so no conclusion about those choices can be drawn from the record.

Missing measurements can also affect a repeated-measures analysis. Such models commonly rely on assumptions about the relationship between missingness and observed information. The exact assumptions and missing-data procedures used for this trial are not reported in the registry record. The count evaluable for the primary endpoint is not, by itself, enough to establish why other randomized participants were not included in that outcome analysis or how missing values were treated.

How do the secondary time-to-event methods differ?

The secondary outcome for time to conversion to abnormal TCD flow is analyzed using a log-rank method, and the posted comparison is a hazard ratio. General: a hazard ratio compares the event rates over time between groups under the analysis framework; it is not a ratio of median times. The median estimates are not reported as estimable for either arm. The reported hazard ratio is 0.62, with a 95% confidence interval from 0.35 to 1.09 and p = 0.098. Because the interval includes 1, the estimate is uncertain with respect to a difference in event rates under the conventional interpretation.

The record also includes planning illustrations based on the observed effect. These are hypothetical calculations, not observed trial outcomes and not observed power. For time-to-event outcomes, the planning quantities refer to total events; for other outcomes, the illustrations refer to participants per arm. The assumptions listed for these illustrations are a two-sided significance level of 0.05, equal allocation, and a normal approximation. The registry explicitly distinguishes these planning calculations from the trial's own design. The trial's sample-size justification and design assumptions are not reported in the registry record.

What can be reproduced from the registry?

The posted arm summaries and reported model estimate allow the result to be described, but they are not enough to refit the MMRM or verify its assumptions. The record does not include participant-level data, full model settings, or a detailed statistical analysis plan. General: a numerical estimate can be reported without every analytic detail needed for independent reproduction. Interpretation should therefore remain tied to the stated analysis and acknowledge the information that is not reported.

10. Analyses Reported by the Sponsor

The sponsor's reported primary analysis is a mixed model for repeated measures. It estimates a difference in least-squares mean change of -7.77 cm/sec, with a 95% confidence interval from -13.18 to -2.37 and p = 0.005. This is the comparative estimate to use when describing the sponsor's primary result.

The registry also gives arm-level least-squares mean changes: -12.06 cm/sec for voxelotor and -4.29 cm/sec for placebo. Arm-level summaries and a model-based contrast are related, but they are not necessarily interchangeable. The contrast comes from the stated repeated-measures analysis, while a simple comparison of posted summary values would not reproduce the full model or its uncertainty. The registry does not provide enough model detail to reconstruct that analysis independently.

The record does not flag a discrepancy between the sponsor analysis and the recalculation. No separately reported recalculated estimate is included in the record, so there is no independent numerical recalculation to compare with the sponsor's model result. The sponsor analysis should therefore be described on its own terms, without treating the arm summaries as a replacement for the reported contrast.

11. Multiplicity, Missing Data and Analysis Populations

The registry reports 11 outcome measures and 21 analyses. Assessing multiple outcomes or comparisons raises the issue of multiplicity: when many statistical tests are conducted, the chance of observing at least one small p-value by chance can increase. General: procedures such as an adjustment across tests or a prespecified hierarchy can help control the interpretation of multiple comparisons. The registry record does not report whether such a procedure was planned or used, how outcomes were ordered, or which analyses were intended to be confirmatory.

The primary result is reported as a mixed-model contrast with a 95% confidence interval and p-value. Without the statistical analysis plan, the record does not establish how the primary endpoint was protected from multiplicity across the other outcomes and analyses. This does not invalidate the posted estimate, but it limits assessment of the broader family of findings. Secondary results should therefore be interpreted cautiously, especially when full numerical results and the multiplicity strategy are unavailable.

The primary population is described as all randomized participants. The registry also states that the number analyzed for the primary endpoint represents participants evaluable for that outcome, and it reports 114 analyzed in the voxelotor arm and 108 in the placebo arm. These descriptions distinguish the analysis population definition from the number with evaluable outcome data. They do not explain why any randomized participants were not evaluable.

The handling of missing outcome data, including imputation, sensitivity analyses, and assumptions about why data were missing, is not reported in the registry record. The record does not report whether missingness was related to treatment assignment or outcome values. No conclusion about the adequacy of a missing-data method can be drawn from the posted counts alone. Details of interim analyses, stopping rules, and any adjustments associated with stopping are also not reported. Since the study status is terminated, those unreported details would be relevant context, but the reason for termination and its statistical implications cannot be inferred from status alone.

12. 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 0% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: sickle cell anaemia with crisis (36.7%), malaria (16.7%), pyrexia (15.8%), anaemia (11.2%), acute chest syndrome (6.9%).

The registry reports serious adverse events and other adverse events by arm, with the number at risk specified for each category. Serious adverse events were reported for 63 of 120 participants at risk in the voxelotor arm and 43 of 116 in the placebo arm. Other adverse events were reported for 81 of 120 and 83 of 116, respectively. The registry also records 8 deaths in the voxelotor group and 2 in the placebo group.

These are descriptive counts from the registry. They do not, by themselves, establish that an event was caused by study treatment, nor do they adjust for exposure duration, participant characteristics, or other factors. The registry lists sickle cell anaemia with crisis, malaria, pyrexia, anaemia, acute chest syndrome, and sepsis among the top serious adverse event terms, with the listed percentages of 36.7%, 16.7%, 15.8%, 11.2%, 6.9%, and 6.7%, respectively. The record does not clarify the denominator for these term percentages in the information, so they should not be read as rates among all randomized participants.

Safety counts are not efficacy estimates. Differences in the reported counts cannot establish causality or comparative risk without fuller context on event definitions, follow-up, exposure, and analysis. Those details are not reported in the registry record.

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

The primary result describes a difference between randomized groups in mean change in TAMMV at Week 24 under the sponsor's repeated-measures model. Its estimate is -7.77 cm/sec, with a 95% confidence interval from -13.18 to -2.37 and p = 0.005. The interval indicates uncertainty around the estimated group contrast, and the estimate is expressed on the measured velocity scale.

This does not mean that every participant experienced the same change, or that an individual participant's result can be predicted from the group average. It also does not directly state how many participants moved from conditional to normal or abnormal TCD categories. The primary endpoint measures a continuous change; category conversion is a separate question.

The result does not by itself establish a change in clinical events, a mechanism of action, or the clinical importance of the observed difference. Those questions require evidence beyond a statistical contrast in the measured endpoint. The registry's additional outcomes may have been intended to address related questions, but their complete numerical results and the full analysis plan are not reported in the registry record.

The secondary time-to-conversion estimate is not conclusive on its own: its confidence interval includes 1, and the median times are not reported as estimable. It should not be combined informally with the primary continuous outcome as though both measured the same quantity. The study's terminated status is also relevant background, but the record does not report why it was terminated or what consequences that had for the intended analyses.

General: statistical evidence, measurement relevance, and clinical importance are distinct considerations. A confidence interval and p-value characterize the statistical estimate under model assumptions; they do not determine whether a difference matters to participants. The registry result should be read as a finding about the posted outcome analysis, with uncertainty and missing design details kept in view. It is not a treatment recommendation.

14. Why This Trial Matters Statistically

This trial is statistically informative because it tests a continuous physiological measure in a randomized, placebo-controlled design. Randomization provides a framework for comparing groups, while the primary endpoint records change in TCD-measured arterial flow velocity rather than only whether a threshold was crossed. The reported model-based contrast and confidence interval make both estimated direction and precision available for interpretation.

The endpoint also illustrates why the estimand matters. A mean change in TAMMV at Week 24 is not the same question as time to conversion to abnormal flow, time to reversion to normal flow, or the proportion crossing a specified change threshold. These outcomes can offer complementary perspectives, but they need separate interpretations and complete reporting.

The registry's account also highlights practical limits in evaluating a trial from posted results. Although the primary analysis method and headline estimate are reported, the full model specification, missing-data procedures, multiplicity strategy, and reason for termination are not. These omissions limit independent assessment of assumptions and the relationship among outcomes. The listed planning illustrations are not observed power and are not a substitute for the trial's own design justification, which is not reported in the registry record.

For biostatisticians and clinical researchers, the case underscores the value of separating a model-based primary contrast from arm-level summaries, and of distinguishing statistical evidence from clinical interpretation. It also shows why outcome completeness and analysis-plan details matter when many measures are registered. The primary estimate can be described precisely, while broader conclusions must remain proportionate to the available record.

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.

The study is listed as terminated, but the reason, timing, and consequences of termination are not reported in the registry record. The planned enrollment was 236; the primary outcome has 114 evaluable participants in the voxelotor arm and 108 in the placebo arm. The difference between planned enrollment and evaluable counts cannot, by itself, explain participant disposition or missing outcome data.

The full repeated-measures model specification is not reported. In particular, the record does not describe included covariates, the covariance structure, or the approach to missing measurements. The analysis population is defined as all randomized participants, but the outcome-specific counts indicate that not all randomized participants are represented in the evaluable primary outcome numbers. Reasons and missing-data assumptions are not reported.

The registry lists many outcomes and analyses, but the multiplicity strategy and the status of secondary outcomes as confirmatory or exploratory are not reported. Numerical results for all secondary outcomes are not available in the record's detailed outcome information, limiting assessment of consistency across measures.

The record does not report a sample-size justification, interim-analysis details, stopping rules, or a full statistical analysis plan. It therefore cannot establish whether the termination affected the planned inferential approach or whether the posted analysis followed every prespecified procedure. The serious adverse event counts and deaths are descriptive; detailed exposure time, event adjudication, and causal attribution are not reported in the information available. These limitations constrain interpretation without changing the numerical result reported for the primary analysis.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Study to Evaluate the Effect of GBT440 on TCD in Pediatrics With Sickle Cell Disease — NCT04218084.
• Registry record last updated 2026-03-10; status last verified 2026-03.
• Results first posted 2025-06-13.
• 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

The registry describes a Phase 3 randomized, double-blind, placebo-controlled, parallel-group study of voxelotor versus matching placebo in pediatric participants with sickle cell disease. The eligible age range was 2 years and older but under 15 years, with conditional TCD flow velocities. Planned enrollment was 236, and the study status is terminated.

The primary outcome was change from baseline in TAMMV at Week 24. The sponsor's mixed-model estimate of the difference in least-squares mean change was -7.77 cm/sec, with a 95% confidence interval from -13.18 to -2.37 and p = 0.005. The record also includes a secondary time-to-conversion analysis, additional outcome listings, and safety counts.

The reason for termination, complete secondary numerical results, detailed missing-data methods, multiplicity procedures, and the full statistical analysis plan are not reported in the registry record. These gaps limit conclusions beyond the posted primary estimate and the descriptive results recorded.