This page provides an independent statistical analysis and educational interpretation of publicly reported results. ClinicalTrials.gov provides the official trial registry record.
1. Trial at a Glance
BROADWAY was a randomized, parallel, phase 3 trial evaluating whether adding obicetrapib 10 mg to maximum tolerated lipid-modifying therapies changed LDL-C and other lipid measures compared with placebo. The primary endpoint was the least-squares mean percent change in LDL-C from baseline to Day 84, with LDL-C measured by preparative ultracentrifugation.
| Feature | BROADWAY |
|---|---|
| Phase | Phase 3 |
| Status | Completed |
| Therapeutic area | Cardiovascular |
| Conditions | Dyslipidemias; High Cholesterol; Hypercholesterolemia; Familial Hypercholesterolemia; Atherosclerotic Cardiovascular Disease |
| Design | Randomized, parallel-group, quadruple-masked |
| Allocation | Randomized |
| Primary purpose | Treatment |
| Enrollment | 2530 |
| Interventions | Placebo; Obicetrapib |
| Primary endpoint | Percent Change in Low-Density Lipoprotein Cholesterol (LDL-C) From Baseline to Day 84 [PUC] |
| Primary endpoint type | Binary |
| Hypothesis type | Superiority |
| Primary statistical method | ANCOVA |
| Lead sponsor | NewAmsterdam Pharma |
| Start | 2021-12-15 |
| Primary completion | 2024-09-26 |
| ClinicalTrials.gov | NCT05142722 |
2. Clinical Question
The central question was whether adding obicetrapib 10 mg to maximum tolerated lipid-modifying therapy produced a greater change in LDL-C from baseline to Day 84 than placebo.
Population
Participants enrolled in a phase 3 study of dyslipidemias, high cholesterol, hypercholesterolemia, familial hypercholesterolemia, and atherosclerotic cardiovascular disease.
Intervention
Obicetrapib 10 mg.
Comparator
Placebo.
Primary question
Does obicetrapib produce a different LS mean percent change in LDL-C from baseline to Day 84 compared with placebo?
3. Trial Design
Placebo
- Placebo was the randomized comparator.
- Primary and secondary analyses compared this group with obicetrapib 10 mg.
Obicetrapib 10mg
- Obicetrapib 10 mg was the randomized intervention.
- Primary and secondary analyses compared this group with placebo.
Analysis population
The registry defines the Baseline Population as all participants who were randomized in the study. Baseline values were defined as the last measurement prior to the first dose of study drug. This definition anchors the change-from-baseline analysis to randomized participants rather than restricting the primary efficacy population to participants who completed a particular follow-up visit.
4. Endpoints
| Endpoint | Time frame | Registry definition |
|---|---|---|
| Percent Change in Low-Density Lipoprotein Cholesterol (LDL-C) From Baseline to Day 84 [PUC] | 84 Days | LS mean percent change from baseline to Day 84 in LDL-C in the obicetrapib group compared to the placebo group. LDL-C level was measured by preparative ultracentrifugation (PUC). |
| Percent Change in Low Density Lipoprotein-Cholesterol (LDL-C) From Baseline to Day 180 [Martin/Hopkins] | 180 Days | Percent change from baseline in LDL-C comparing placebo and obicetrapib 10mg. |
| Percent Change in Low Density Lipoprotein-Cholesterol (LDL-C) From Baseline to Day 365 [PUC] | 365 Days | Percent change from baseline in LDL-C comparing placebo and obicetrapib 10mg. |
| Percent Change in Apolipoprotein B (ApoB) From Baseline to Day 84 | 84 Days | Percent change from baseline in ApoB comparing placebo and obicetrapib 10mg. |
| Percent Change in Apolipoprotein B (ApoB) From Baseline to Day 180 | 180 Days | Percent change from baseline in ApoB comparing placebo and obicetrapib 10mg. |
| Percent Change in Apolipoprotein B (ApoB) From Baseline to Day 365 | 365 Days | Percent change from baseline in ApoB comparing placebo and obicetrapib 10mg. |
| Percent Change in Non-HDL-C From Baseline to Day 84 | 84 Days | Percent change from baseline in non-HDL-C comparing placebo and obicetrapib 10mg. |
| Percent Change in Non-HDL-C From Baseline to Day 180 | 180 Days | Percent change from baseline in non-HDL-C comparing placebo and obicetrapib 10mg. |
| Percent Change in Non-HDL-C From Baseline to Day 365 | 365 Days | Percent change from baseline in non-HDL-C comparing placebo and obicetrapib 10mg. |
| Percent Change in HDL-C From Baseline to Day 84 | 84 Days | Percent change from baseline in HDL-C comparing placebo and obicetrapib 10mg. |
| Percent Change in HDL-C From Baseline to Day 180 | 180 Days | Percent change from baseline in HDL-C comparing placebo and obicetrapib 10mg. |
| Percent Change in HDL-C From Baseline to Day 365 | 365 Days | Percent change from baseline in HDL-C comparing placebo and obicetrapib 10mg. |
| Percent Change in Lp(a) From Baseline to Day 84 | 84 Days | Percent change from baseline in Lp(a) comparing placebo and obicetrapib 10mg. |
The registry reports 21 outcome measures in total, with 13 statistical analyses posted and one primary-endpoint analysis. The results reported here correspond to the statistical analyses posted for the primary endpoint and the secondary lipid endpoints listed above.
5. Statistical Methodology
ANCOVA for change from baseline
The primary and posted secondary analyses used analysis of covariance (ANCOVA). The reported effect measure was the least-squares (LS) mean, normalized here as a mean difference between the randomized treatment groups.
For a change-from-baseline endpoint, ANCOVA provides a model-based comparison of treatment groups while accounting for baseline information. The resulting LS means represent adjusted group means from the fitted model rather than simple arithmetic means calculated directly from the observed follow-up values.
Least-squares mean difference
The primary effect estimate was -32.65. Because the outcome is expressed as percent change from baseline, the estimate represents the modeled difference between the obicetrapib and placebo groups in percent change. A negative value indicates a more negative change in the obicetrapib group relative to placebo.
The corresponding 95% two-sided confidence interval was -35.79 to -29.50, with a superiority P-value of <.0001.
Multiple imputation for missing data
Missing values were imputed using a multiple imputation model assuming that the data were missing not at random. Missing measurements of non-retrieved dropouts were modeled using known measurements from retrieved dropouts in the same treatment group.
This is an important feature of the analysis because the reported LS mean comparison is not simply a comparison of participants with observed measurements at the endpoint. The statistical result incorporates an explicit model for missing measurements.
Baseline population definition
The registry defines the Baseline Population as all participants who were randomized. Baseline values were the last measurements obtained before the first dose of study drug. This creates a clearly defined reference point for the percent-change endpoints.
Superiority framework
The registered hypothesis type was superiority. The inferential question is therefore whether the modeled treatment-group difference is inconsistent with the null comparison of no treatment difference, rather than whether obicetrapib falls within a prespecified non-inferiority margin.
6. Primary Result: LDL-C at Day 84
The primary endpoint was percent change in LDL-C from baseline to Day 84, with LDL-C measured by preparative ultracentrifugation. The registry reports an ANCOVA comparison of placebo versus obicetrapib 10 mg using the randomized Baseline Population and multiple imputation for missing values.
Primary treatment effect
LS mean difference in percent change from baseline
95% CI: -35.79 to -29.50 · P <.0001 · Superiority
| Primary endpoint | Comparison | Method | Estimate | 95% CI | P-value |
|---|---|---|---|---|---|
| LDL-C percent change from baseline to Day 84 [PUC] | Placebo vs Obicetrapib 10mg | ANCOVA | -32.65 | -35.79 to -29.50 | <.0001 |
What the estimate means: The LS mean difference of -32.65 means that the modeled percent change in LDL-C was 32.65 percentage points more negative with obicetrapib than with placebo at Day 84. The direction of the estimate therefore indicates a greater reduction in LDL-C in the obicetrapib group.
What it does not mean: It does not mean that every participant experienced exactly a 32.65% reduction, nor does it mean that the individual-level treatment effect was identical across participants. It is an adjusted group-level estimate from an ANCOVA model.
What the confidence interval says: The 95% two-sided confidence interval extends from -35.79 to -29.50. It describes statistical uncertainty around the estimated between-group difference under the analysis framework. It is not a range containing 95% of individual treatment effects.
Why the P-value does not measure effect size: The P-value of <.0001 addresses the strength of evidence against the null comparison used for the superiority test. It does not say that the treatment effect is "<.0001 large" or quantify the clinical magnitude of the effect. The estimate and its confidence interval provide the magnitude and precision.
Missing-data caution: The analysis used multiple imputation under a missing-not-at-random assumption. Consequently, interpretation depends in part on the plausibility of that missing-data model and on the information used to model non-retrieved dropouts.
7. Secondary Results: LDL-C
LDL-C at Day 180
Placebo-adjusted LS mean difference
95% CI: -37.07 to -30.49 · P <.0001
At Day 180, the registry reports an ANCOVA estimate of -33.78 for the difference between the placebo and obicetrapib 10 mg groups in percent change from baseline in LDL-C measured using the Martin/Hopkins method. The 95% two-sided confidence interval was -37.07 to -30.49.
LDL-C at Day 365
Placebo-adjusted LS mean difference
95% CI: -27.87 to -20.09 · P <.0001
At Day 365, the ANCOVA estimate was -23.98 for percent change from baseline in LDL-C measured by preparative ultracentrifugation. The 95% two-sided confidence interval was -27.87 to -20.09.
| Endpoint | Time frame | Estimate | 95% CI | P-value |
|---|---|---|---|---|
| LDL-C [PUC] | Day 84 | -32.65 | -35.79 to -29.50 | <.0001 |
| LDL-C [Martin/Hopkins] | Day 180 | -33.78 | -37.07 to -30.49 | <.0001 |
| LDL-C [PUC] | Day 365 | -23.98 | -27.87 to -20.09 | <.0001 |
The estimates are all negative, indicating a more negative modeled percent change from baseline in the obicetrapib group relative to placebo. The Day 84 and Day 180 estimates are not directly interchangeable with the Day 365 estimate: they correspond to different follow-up times, and the Day 180 endpoint uses the Martin/Hopkins method whereas the Day 84 and Day 365 endpoints use PUC.
8. Secondary Results: Apolipoprotein B
The registry reports three ANCOVA analyses of percent change in ApoB from baseline. All three comparisons use placebo versus obicetrapib 10 mg and incorporate multiple imputation under the same missing-not-at-random framework.
| Endpoint | Time frame | Estimate | 95% CI | P-value |
|---|---|---|---|---|
| ApoB percent change from baseline | Day 84 | -18.92 | -20.76 to -17.09 | <.0001 |
| ApoB percent change from baseline | Day 180 | -18.31 | -20.38 to -16.23 | <.0001 |
| ApoB percent change from baseline | Day 365 | -13.80 | -16.20 to -11.41 | <.0001 |
How to interpret the ApoB estimates
The Day 84 estimate of -18.92 indicates that the modeled percent change in ApoB was 18.92 percentage points more negative with obicetrapib than with placebo. The corresponding 95% confidence interval, -20.76 to -17.09, is entirely below zero.
The same interpretation applies to the Day 180 estimate of -18.31 and the Day 365 estimate of -13.80. These are separate time-specific treatment comparisons rather than a single longitudinal effect estimate.
9. Secondary Results: Non-HDL-C
| Endpoint | Time frame | Estimate | 95% CI | P-value |
|---|---|---|---|---|
| Non-HDL-C percent change from baseline | Day 84 | -29.44 | -31.89 to -26.99 | <.0001 |
| Non-HDL-C percent change from baseline | Day 180 | -28.32 | -30.94 to -25.69 | <.0001 |
| Non-HDL-C percent change from baseline | Day 365 | -23.02 | -26.09 to -19.95 | <.0001 |
All three posted analyses use ANCOVA and report LS means as the effect measure. The estimated differences are negative at each time point, with two-sided 95% confidence intervals that remain below zero and P-values of <.0001.
The pattern illustrates an important statistical distinction between direction and magnitude. All three estimates point in the same direction, but the numerical estimates differ across follow-up times. A comparison across these values should therefore be described as a comparison of time-specific estimates rather than as evidence of a particular rate of change over time.
10. Secondary Results: HDL-C
| Endpoint | Time frame | Estimate | 95% CI | P-value |
|---|---|---|---|---|
| HDL-C percent change from baseline | Day 84 | 136.26 | 132.46 to 140.07 | <.0001 |
| HDL-C percent change from baseline | Day 180 | 134.43 | 129.84 to 139.03 | <.0001 |
| HDL-C percent change from baseline | Day 365 | 122.04 | 117.53 to 126.55 | <.0001 |
The HDL-C results differ in direction from the LDL-C, ApoB, and non-HDL-C results. The positive estimates indicate a greater positive modeled percent change from baseline in HDL-C in the obicetrapib group relative to placebo.
The estimate is a difference in percent change from baseline, not a 136.26 percentage-point difference in the absolute HDL-C concentration. The unit of the endpoint is therefore essential to its interpretation.
The Day 84 confidence interval of 132.46 to 140.07, the Day 180 interval of 129.84 to 139.03, and the Day 365 interval of 117.53 to 126.55 all lie above zero. Each corresponding P-value is <.0001.
11. Secondary Result: Lp(a)
Lp(a) at Day 84
95% CI: -53.68 to 25.44 · P = 0.4841
The Day 84 ANCOVA estimate for percent change in Lp(a) was -14.12. The 95% two-sided confidence interval ranged from -53.68 to 25.44, and the P-value was 0.4841.
The point estimate is negative, but the confidence interval spans zero by a wide margin. The data therefore provide substantially more uncertainty about the treatment-group difference in Lp(a) than the posted analyses for LDL-C, ApoB, non-HDL-C, and HDL-C.
The P-value of 0.4841 should not be interpreted as the probability that the null hypothesis is true, nor as evidence that the treatment groups are exactly equivalent. It is an inferential quantity calculated under the statistical testing framework. The confidence interval is particularly useful here because it shows the range of treatment differences compatible with the estimated uncertainty.
12. Consolidated Statistical Results
| Endpoint | Time | Method | Effect estimate | 95% CI | P-value |
|---|---|---|---|---|---|
| LDL-C [PUC] | Day 84 | ANCOVA | -32.65 | -35.79 to -29.50 | <.0001 |
| LDL-C [Martin/Hopkins] | Day 180 | ANCOVA | -33.78 | -37.07 to -30.49 | <.0001 |
| LDL-C [PUC] | Day 365 | ANCOVA | -23.98 | -27.87 to -20.09 | <.0001 |
| ApoB | Day 84 | ANCOVA | -18.92 | -20.76 to -17.09 | <.0001 |
| ApoB | Day 180 | ANCOVA | -18.31 | -20.38 to -16.23 | <.0001 |
| ApoB | Day 365 | ANCOVA | -13.80 | -16.20 to -11.41 | <.0001 |
| Non-HDL-C | Day 84 | ANCOVA | -29.44 | -31.89 to -26.99 | <.0001 |
| Non-HDL-C | Day 180 | ANCOVA | -28.32 | -30.94 to -25.69 | <.0001 |
| Non-HDL-C | Day 365 | ANCOVA | -23.02 | -26.09 to -19.95 | <.0001 |
| HDL-C | Day 84 | ANCOVA | 136.26 | 132.46 to 140.07 | <.0001 |
| HDL-C | Day 180 | ANCOVA | 134.43 | 129.84 to 139.03 | <.0001 |
| HDL-C | Day 365 | ANCOVA | 122.04 | 117.53 to 126.55 | <.0001 |
| Lp(a) | Day 84 | ANCOVA | -14.12 | -53.68 to 25.44 | 0.4841 |
The table highlights the overall statistical structure of the posted results: twelve secondary analyses and one primary analysis use the same broad ANCOVA framework, while the estimated effects differ substantially according to the lipid biomarker and time point.
13. Statistical Methods Explained
Why was ANCOVA used?
ANCOVA is appropriate for comparing treatment groups on a continuous outcome while incorporating baseline information. In this trial, the endpoints were expressed as percent changes from baseline, and the registry reports ANCOVA as the method for the primary and secondary statistical analyses. The resulting LS means provide model-based adjusted treatment-group estimates.
What does an LS mean difference of -32.65 mean?
It means the modeled percent change from baseline in LDL-C was 32.65 percentage points more negative in the obicetrapib group than in the placebo group at Day 84. Because the endpoint is a percent change, the estimate is expressed in percentage points of change, not as a ratio or hazard ratio.
Why does the confidence interval matter?
The confidence interval describes the statistical precision of the estimated treatment difference. For the primary endpoint, the interval from -35.79 to -29.50 remains below zero, so the direction of the estimated difference is stable within the reported interval. A narrower interval generally represents greater precision than a wider interval, although precision and clinical importance are separate concepts.
Why does the P-value not measure the size of the treatment effect?
A P-value addresses compatibility of the observed result with the null hypothesis under the statistical model. It does not quantify the magnitude of the treatment effect. For example, the primary P-value is <.0001, but the treatment magnitude is communicated by the estimate of -32.65 and its confidence interval of -35.79 to -29.50.
Why was multiple imputation important?
The registry states that missing values were imputed using a multiple imputation model assuming the data were missing not at random. This matters because a complete-case analysis would use only participants with observed measurements, whereas the posted analysis incorporates modeled information about missing measurements. The resulting inference therefore depends partly on the assumptions of the imputation model.
What does the Lp(a) confidence interval tell us?
The Lp(a) estimate of -14.12 has a 95% confidence interval from -53.68 to 25.44. Because the interval includes zero, the estimate is compatible with both a negative and a positive treatment-group difference under the stated uncertainty framework. The P-value of 0.4841 provides the corresponding hypothesis-testing result.
Why should the Day 84, Day 180, and Day 365 estimates not be treated as one effect?
Each estimate is tied to a specific endpoint definition and time frame. In addition, the LDL-C Day 180 endpoint specifies the Martin/Hopkins measurement method, whereas the Day 84 and Day 365 LDL-C endpoints specify PUC. The numerical estimates therefore should be interpreted within their respective endpoint definitions rather than combined into a single summary effect.
14. Multiple Comparisons and Endpoint Interpretation
The registry reports one primary endpoint and multiple secondary endpoints, with 13 statistical analyses posted in total. The primary hypothesis is explicitly classified as a superiority hypothesis. The posted secondary analyses likewise report superiority as their hypothesis type.
| Feature | Statistical implication |
|---|---|
| One registered primary endpoint | The primary inference is centered on the Day 84 LDL-C [PUC] comparison. |
| Multiple secondary endpoints | Secondary P-values should be interpreted in the context of multiple statistical questions. |
| Multiple time points | Day 84, Day 180, and Day 365 represent distinct estimands and follow-up windows. |
| Superiority hypothesis | The inferential framework tests for a treatment difference rather than non-inferiority against a margin. |
| ANCOVA across analyses | The broad modeling framework is consistent across the posted primary and secondary analyses. |
Because the registry reports many endpoint-specific P-values, a small P-value should not automatically be interpreted as establishing an independently confirmatory claim for every endpoint. The inferential status of each endpoint depends on the prespecified multiplicity strategy and endpoint hierarchy. The posted data identify the endpoint roles and analyses but do not provide a separate multiplicity-adjustment description in the statistical-analysis information presented here.
15. Missing Data and Multiple Imputation
Missing-data handling is one of the most consequential methodological features of the BROADWAY analyses. The registry states that missing values were imputed using a multiple imputation model assuming the data were missing not at random.
Randomized baseline population
The Baseline Population includes all participants who were randomized, and baseline is defined as the last measurement before the first dose.
Missing values
Missing measurements were addressed through multiple imputation rather than simply excluding participants with missing endpoint observations.
Missing not at random
The imputation model assumes that missingness may depend on information not fully captured by the observed data under a missing-at-random framework.
Non-retrieved dropouts
Missing measurements from non-retrieved dropouts were modeled using known measurements from retrieved dropouts in the same treatment group.
The key interpretive point is that the final estimate is model-dependent. Multiple imputation does not eliminate uncertainty caused by missing data; rather, it propagates uncertainty under an explicit statistical model. The missing-not-at-random assumption is particularly important because it represents a stronger assumption about the mechanism generating missing measurements.
16. Interpreting the Lipid Biomarker Results Together
The posted analyses cover LDL-C, ApoB, non-HDL-C, HDL-C, and Lp(a). The direction of the estimated treatment differences is not uniform because these biomarkers represent different quantities and the treatment effect is expressed as a percent change from baseline.
| Biomarker | Earliest posted time point | Estimate | Interpretive direction |
|---|---|---|---|
| LDL-C [PUC] | Day 84 | -32.65 | More negative percent change with obicetrapib relative to placebo |
| ApoB | Day 84 | -18.92 | More negative percent change with obicetrapib relative to placebo |
| Non-HDL-C | Day 84 | -29.44 | More negative percent change with obicetrapib relative to placebo |
| HDL-C | Day 84 | 136.26 | More positive percent change with obicetrapib relative to placebo |
| Lp(a) | Day 84 | -14.12 | Point estimate is negative, but the confidence interval includes zero |
This is an example of why statistical interpretation should begin with the endpoint definition. A positive estimate is not inherently favorable or unfavorable, and a negative estimate is not inherently favorable or unfavorable. The meaning depends on what was measured and how the endpoint was defined.
17. Safety Results
The registry reports serious adverse events by randomized arm as affected participants divided by participants at risk.
| Safety measure | Affected / at risk |
|---|---|
| Placebo | 117 / 843 |
| Obicetrapib 10mg | 211 / 1685 |
Placebo
117 participants with serious adverse events among 843 at risk.
Obicetrapib 10mg
211 participants with serious adverse events among 1685 at risk.
These counts should be interpreted as the reported serious-adverse-event counts and denominators rather than converted into a new comparative statistic. The ClinicalTrials.gov record does not report a confidence interval, P-value, relative risk, odds ratio, or formal hypothesis test for this safety comparison.
The reported safety denominators also differ from the overall enrollment figure of 2530. The appropriate interpretation is therefore to retain the registry's stated affected/at-risk counts rather than infer a new safety population definition from the overall enrollment number.
18. Trial Timeline
Study start
The BROADWAY study began on December 15, 2021.
Randomized treatment comparison
The phase 3 trial used randomized parallel-group allocation with quadruple masking and two treatment arms.
Primary efficacy assessment
The primary endpoint assessed percent change in LDL-C from baseline to Day 84 using the PUC measurement method.
Intermediate secondary assessments
Secondary analyses evaluated LDL-C, ApoB, non-HDL-C, and HDL-C percent change from baseline at Day 180.
Longer-term lipid assessment
Secondary analyses evaluated LDL-C, ApoB, non-HDL-C, and HDL-C percent change from baseline at Day 365.
Primary completion
The registry lists September 26, 2024 as the primary completion date.
19. What the Primary Estimate Does — and Does Not — Mean
The primary estimate of -32.65 is a modeled difference in percent change from baseline between the obicetrapib 10 mg and placebo groups at Day 84. It describes a group-level treatment contrast under the ANCOVA model.
It does not mean that each individual participant had exactly a -32.65% treatment effect. Individual responses can vary, while the LS mean is a population-level model-based quantity.
The 95% confidence interval of -35.79 to -29.50 communicates the statistical uncertainty around the estimated treatment difference. It is substantially narrower than the corresponding Lp(a) interval of -53.68 to 25.44, illustrating how endpoint-specific precision can differ even when the same broad ANCOVA method is used.
The primary P-value of <.0001 provides evidence against the null hypothesis under the stated superiority testing framework. It does not itself quantify the magnitude, clinical relevance, or individual variability of the effect.
Because the endpoint is percent change from baseline, the estimate must be interpreted in percentage-point units of change. This distinction becomes especially important for HDL-C, where the reported treatment-group difference is positive and numerically large.
20. Clinical Interpretation vs Statistical Interpretation
Statistical interpretation
The primary ANCOVA comparison produced an LS mean difference of -32.65 with a 95% confidence interval of -35.79 to -29.50 and a P-value of <.0001. The posted secondary analyses show similarly negative estimates for LDL-C, ApoB, and non-HDL-C and positive estimates for HDL-C.
Clinical interpretation
The statistical findings describe differences in lipid biomarkers between randomized treatment groups. The registry information presented here does not establish a clinical outcome benefit from these biomarker changes.
This distinction is essential. A statistically precise biomarker effect and a clinical outcome effect are different estimands. The primary endpoint in BROADWAY is a laboratory-based lipid measure rather than a patient-centered event endpoint.
21. Important Limitations and Interpretation Issues
- Endpoint is a biomarker: the primary endpoint is percent change in LDL-C rather than a clinical cardiovascular event or survival endpoint. A statistical difference in LDL-C should therefore not be interpreted as a direct estimate of a clinical event effect.
- Measurement method matters: the primary endpoint and Day 365 LDL-C endpoint specify PUC, while the Day 180 LDL-C endpoint specifies the Martin/Hopkins method. These measurements should not be treated as identical endpoint definitions.
- Multiple endpoints: the registry reports one primary endpoint and numerous secondary endpoints. The interpretation of secondary P-values depends on the prespecified multiplicity strategy.
- Multiple time points: the Day 84, Day 180, and Day 365 analyses answer distinct time-specific questions and should not automatically be interpreted as a single longitudinal treatment effect.
- Missing data: multiple imputation was performed under a missing-not-at-random assumption. This makes the missing-data model an important component of inference.
- Lp(a) precision: the Day 84 Lp(a) estimate has a wide confidence interval spanning zero, indicating considerably greater uncertainty for that endpoint than for several other posted lipid analyses.
- Safety denominator: the reported serious-adverse-event denominators are 843 for placebo and 1685 for obicetrapib 10mg. They should not be replaced by the overall enrollment figure.
- Unreported characteristics: ClinicalTrials.gov does not report a baseline-characteristics table in the statistical-analysis information presented here, so treatment-group balance cannot be quantitatively assessed from the available registry fields.
- No survival analysis: the posted analyses use ANCOVA and mean-difference/LS-mean methodology; no Kaplan-Meier or Cox model result is reported in the statistical analyses presented here.
22. Why This Trial Matters Statistically
BROADWAY is useful as a teaching case because it demonstrates how a large randomized trial can use a relatively straightforward linear-model framework to answer a precise biomarker question while still requiring careful interpretation of endpoint definitions, missing data, confidence intervals, and multiple comparisons.
| Concept | How it appears in BROADWAY |
|---|---|
| Randomization | Participants were randomized to placebo or obicetrapib 10 mg in a parallel-group phase 3 trial. |
| Masking | The study was quadruple-masked. |
| ANCOVA | The primary and posted secondary analyses used ANCOVA. |
| Least-squares means | The reported effect measure was LS means, represented as a mean difference for the treatment comparison. |
| Confidence intervals | Posted estimates include two-sided 95% confidence intervals. |
| P-values | The primary and most secondary analyses report P-values of <.0001, while the Lp(a) analysis reports 0.4841. |
| Missing data | Multiple imputation was used under a missing-not-at-random assumption. |
| Endpoint scale | Most posted outcomes are percent change from baseline, requiring interpretation in percentage-point units. |
| Multiple endpoints | The registry reports one primary endpoint and multiple secondary analyses across biomarkers and time points. |
| Safety analysis | Serious adverse events are reported by arm as affected participants divided by participants at risk. |
23. Related Tutorials
Learn more about the methods used in this trial:
24. Related Calculators
25. Sources
- ClinicalTrials.gov: NCT05142722.
- Linked publication: PMID 40337982.
- Linked publication: PMID 41109840.
Continue through Clinical Biostats
Explore statistical tutorials, calculators, and additional clinical trial analyses covering the methods used throughout randomized clinical research.
26. Record Summary
BROADWAY provides a clear example of statistical analysis for a randomized biomarker endpoint. The phase 3 trial used randomized parallel-group allocation and quadruple masking, with ANCOVA as the primary and posted secondary analytical method. The primary Day 84 LDL-C analysis produced an LS mean difference of -32.65, with a two-sided 95% confidence interval of -35.79 to -29.50 and a P-value of <.0001. Secondary analyses reported negative treatment-group differences for LDL-C, ApoB, and non-HDL-C, positive differences for HDL-C, and a less precise Lp(a) estimate whose confidence interval included zero.
The statistical story is therefore broader than any individual P-value. The primary estimate describes the magnitude of the modeled between-group difference; the confidence interval describes its precision; ANCOVA defines the modeling framework; and multiple imputation determines how missing observations contribute to the inference. Because the endpoints are lipid biomarkers measured as percent changes from baseline, the results should be interpreted on that scale rather than translated automatically into clinical event effects.