Multi-trait analysis of gene-by-environment interactions in large-scale genetic studies

Identifying genotype-by-environment interaction (GEI) is challenging because the GEI analysis generally has low power. Large-scale consortium-based studies are ultimately needed to achieve adequate power for identifying GEI. We introduce Multi-Trait Analysis of Gene-Environment Interactions (MTAGEI), a powerful, robust, and computationally efficient framework to test gene-environment interactions on multiple traits in large data sets, such as the UK Biobank (UKB). To facilitate the meta-analysis of GEI studies in a consortium, MTAGEI efficiently generates summary statistics of genetic associations for multiple traits under different environmental conditions and integrates the summary statistics for GEI analysis. MTAGEI enhances the power of GEI analysis by aggregating GEI signals across multiple traits and variants that would otherwise be difficult to detect individually. MTAGEI achieves robustness by combining complementary tests under a wide spectrum of genetic architectures. We demonstrate the advantages of MTAGEI over existing single-trait-based GEI tests through extensive simulation studies and the analysis of the whole exome sequencing data from the UKB.

Is inadequate risk stratification diluting hazard ratio estimates in randomized clinical trials?

In randomized clinical trials, analyses of time-to-event data without risk stratification, or with stratification based on pre-selected factors revealed at the end of the trial to be at most weakly associated with risk, are quite common. We caution that such analyses are likely delivering hazard ratio estimates that unwittingly dilute the evidence of benefit for the test relative to the control treatment. To make our case, first, we use a hypothetical scenario to contrast risk-unstratified and risk-stratified hazard ratios. Thereafter, we draw attention to the previously published 5-step stratified testing and amalgamation routine (5-STAR) approach in which a pre-specified treatment-blinded algorithm is applied to survival times from the trial to partition patients into well-separated risk strata using baseline covariates determined to be jointly strongly prognostic for event risk. After treatment unblinding, a treatment comparison is done within each risk stratum and stratum-level results are averaged for overall inference. For illustration, we use 5-STAR to reanalyze data for the primary and key secondary time-to-event endpoints from three published cardiovascular outcomes trials. The results show that the 5-STAR estimate is typically smaller (i.e. more in favor of the test treatment) than the originally reported (traditional) estimate. This is not surprising because 5-STAR mitigates the presumed dilution bias in the traditional hazard ratio estimate caused by no or inadequate risk stratification, as evidenced by two detailed examples. Pre-selection of stratification factors at the trial design stage to achieve adequate risk stratification for the analysis will often be challenging. In such settings, an objective risk stratification approach such as 5-STAR, which is partly aligned with guidance from the US Food and Drug Administration on covariate-adjustment in clinical trials, is worthy of consideration.

Applying polygenic risk score methods to pharmacogenomics GWAS: challenges and opportunities

Polygenic risk scores (PRSs) have emerged as promising tools for the prediction of human diseases and complex traits in disease genome-wide association studies (GWAS). Applying PRSs to pharmacogenomics (PGx) studies has begun to show great potential for improving patient stratification and drug response prediction. However, there are unique challenges that arise when applying PRSs to PGx GWAS beyond those typically encountered in disease GWAS (e.g. Eurocentric or trans-ethnic bias). These challenges include: (i) the lack of knowledge about whether PGx or disease GWAS/variants should be used in the base cohort (BC); (ii) the small sample sizes in PGx GWAS with corresponding low power and (iii) the more complex PRS statistical modeling required for handling both prognostic and predictive effects simultaneously. To gain insights in this landscape about the general trends, challenges and possible solutions, we first conduct a systematic review of both PRS applications and PRS method development in PGx GWAS. To further address the challenges, we propose (i) a novel PRS application strategy by leveraging both PGx and disease GWAS summary statistics in the BC for PRS construction and (ii) a new Bayesian method (PRS-PGx-Bayesx) to reduce Eurocentric or cross-population PRS prediction bias. Extensive simulations are conducted to demonstrate their advantages over existing PRS methods applied in PGx GWAS. Our systematic review and methodology research work not only highlights current gaps and key considerations while applying PRS methods to PGx GWAS, but also provides possible solutions for better PGx PRS applications and future research.

Calculation of Phase 2 dose-finding study sample size for reliable Phase 3 dose selection

Sample sizes of Phase 2 dose-finding studies, usually determined based on a power requirement to detect a significant dose-response relationship, will generally not provide adequate precision for Phase 3 target dose selection. We propose to calculate the sample size of a dose-finding study based on the probability of successfully identifying the target dose within an acceptable range (e.g., 80%-120% of the target) using the multiple comparison and modeling procedure (MCP-Mod). With the proposed approach, different design options for the Phase 2 dose-finding study can also be compared. Due to inherent uncertainty around an assumed true dose-response relationship, sensitivity analyses to assess the robustness of the sample size calculations to deviations from modeling assumptions are recommended. Planning for a hypothetical Phase 2 dose-finding study is used to illustrate the main points. Codes for the proposed approach is available at https://github.com/happysundae/posMCPMod.

Statistical assessment of biomarker replicability using MAJAR method

In the era of precision medicine, many biomarkers have been discovered to be associated with drug efficacy and safety responses, which can be used for patient stratification and drug response prediction. Due to the small sample size and limited power of randomized clinical studies, meta-analysis is usually conducted to aggregate all available studies to maximize the power for identifying prognostic and predictive biomarkers. However, it is often challenging to find an independent study to replicate the discoveries from the meta-analysis (e.g. meta-analysis of pharmacogenomics genome-wide association studies (PGx GWAS)), which seriously limits the potential impacts of the discovered biomarkers. To overcome this challenge, we develop a novel statistical framework, MAJAR (meta-analysis of joint effect associations for biomarker replicability assessment), to jointly test prognostic and predictive effects and assess the replicability of identified biomarkers by implementing an enhanced expectation-maximization algorithm and calculating their posterior-probability-of-replicabilities and Bayesian false discovery rates (Fdr). Extensive simulation studies were conducted to compare the performance of MAJAR and existing methods in terms of Fdr, power, and computational efficiency. The simulation results showed improved statistical power with well-controlled Fdr of MAJAR over existing methods and robustness to outliers under different data generation processes. We further demonstrated the advantages of MAJAR over existing methods by applying MAJAR to the PGx GWAS summary statistics data from a large cardiovascular randomized clinical trial. Compared to testing main effects only, MAJAR identified 12 novel variants associated with the treatment-related low-density lipoprotein cholesterol reduction from baseline.

Integrating multiple traits for improving polygenic risk prediction in disease and pharmacogenomics GWAS

Polygenic risk score (PRS) has been recently developed for predicting complex traits and drug responses. It remains unknown whether multi-trait PRS (mtPRS) methods, by integrating information from multiple genetically correlated traits, can improve prediction accuracy and power for PRS analysis compared with single-trait PRS (stPRS) methods. In this paper, we first review commonly used mtPRS methods and find that they do not directly model the underlying genetic correlations among traits, which has been shown to be useful in guiding multi-trait association analysis in the literature. To overcome this limitation, we propose a mtPRS-PCA method to combine PRSs from multiple traits with weights obtained from performing principal component analysis (PCA) on the genetic correlation matrix. To accommodate various genetic architectures covering different effect directions, signal sparseness and across-trait correlation structures, we further propose an omnibus mtPRS method (mtPRS-O) by combining P values from mtPRS-PCA, mtPRS-ML (mtPRS based on machine learning) and stPRSs using Cauchy Combination Test. Our extensive simulation studies show that mtPRS-PCA outperforms other mtPRS methods in both disease and pharmacogenomics (PGx) genome-wide association studies (GWAS) contexts when traits are similarly correlated, with dense signal effects and in similar effect directions, and mtPRS-O is consistently superior to most other methods due to its robustness under various genetic architectures. We further apply mtPRS-PCA, mtPRS-O and other methods to PGx GWAS data from a randomized clinical trial in the cardiovascular domain and demonstrate performance improvement of mtPRS-PCA in both prediction accuracy and patient stratification as well as the robustness of mtPRS-O in PRS association test.

AWOT and CWOT for genotype and genotype-by-treatment interaction joint analysis in pharmacogenetics GWAS

MOTIVATION

Pharmacogenomics (PGx) research holds the promise for detecting association between genetic variants and drug responses in randomized clinical trials, but it is limited by small populations and thus has low power to detect signals. It is critical to increase the power of PGx genome-wide association studies (GWAS) with small sample sizes so that variant-drug-response association discoveries are not limited to common variants with extremely large effect.

RESULTS

In this article, we first discuss the challenges of PGx GWAS studies and then propose the adaptively weighted joint test (AWOT) and Cauchy Weighted jOint Test (CWOT), which are two flexible and robust joint tests of the single nucleotide polymorphism main effect and genotype-by-treatment interaction effect for continuous and binary endpoints. Two analytic procedures are proposed to accurately calculate the joint test P-value. We evaluate AWOT and CWOT through extensive simulations under various scenarios. The results show that the proposed AWOT and CWOT control type I error well and outperform existing methods in detecting the most interesting signal patterns in PGx settings (i.e. with strong genotype-by-treatment interaction effects, but weak genotype main effects). We demonstrate the value of AWOT and CWOT by applying them to the PGx GWAS from the Bezlotoxumab Clostridium difficile MODIFY I/II Phase 3 trials.

AVAILABILITY AND IMPLEMENTATION

The R package COWT is publicly available on CRAN https://cran.r-project.org/web/packages/cwot/index.html.

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

A fast and powerful linear mixed model approach for genotype-environment interaction tests in large-scale GWAS

Genotype-by-environment interaction (GEI or GxE) plays an important role in understanding complex human traits. However, it is usually challenging to detect GEI signals efficiently and accurately while adjusting for population stratification and sample relatedness in large-scale genome-wide association studies (GWAS). Here we propose a fast and powerful linear mixed model-based approach, fastGWA-GE, to test for GEI effect and G + GxE joint effect. Our extensive simulations show that fastGWA-GE outperforms other existing GEI test methods by controlling genomic inflation better, providing larger power and running hundreds to thousands of times faster. We performed a fastGWA-GE analysis of ~7.27 million variants on 452 249 individuals of European ancestry for 13 quantitative traits and five environment variables in the UK Biobank GWAS data and identified 96 significant signals (72 variants across 57 loci) with GEI test P-values < 1 × 10-9, including 27 novel GEI associations, which highlights the effectiveness of fastGWA-GE in GEI signal discovery in large-scale GWAS.

Pharmacogenomics polygenic risk score for drug response prediction using PRS-PGx methods

Polygenic risk scores (PRS) have been successfully developed for the prediction of human diseases and complex traits in the past years. For drug response prediction in randomized clinical trials, a common practice is to apply PRS built from a disease genome-wide association study (GWAS) directly to a corresponding pharmacogenomics (PGx) setting. Here, we show that such an approach relies on stringent assumptions about the prognostic and predictive effects of the selected genetic variants. We propose a shift from disease PRS to PGx PRS approaches by simultaneously modeling both the prognostic and predictive effects and further make this shift possible by developing a series of PRS-PGx methods, including a novel Bayesian regression approach (PRS-PGx-Bayes). Simulation studies show that PRS-PGx methods generally outperform the disease PRS methods and PRS-PGx-Bayes is superior to all other PRS-PGx methods. We further apply the PRS-PGx methods to PGx GWAS data from a large cardiovascular randomized clinical trial (IMPROVE-IT) to predict treatment related LDL cholesterol reduction. The results demonstrate substantial improvement of PRS-PGx-Bayes in both prediction accuracy and the capability of capturing the treatment-specific predictive effects while compared with the disease PRS approaches.

To try to predict an individual’s drug response using genetic data, most studies have used traditional polygenic risk score (PRS) methods. Here, the authors develop a pharmacogenomics-specific PRS method, which can improve drug response prediction and patient stratification in pharmacogenomics studies.

CAPITAL: Optimal subgroup identification via constrained policy tree search

Personalized medicine, a paradigm of medicine tailored to a patient's characteristics, is an increasingly attractive field in health care. An important goal of personalized medicine is to identify a subgroup of patients, based on baseline covariates, that benefits more from the targeted treatment than other comparative treatments. Most of the current subgroup identification methods only focus on obtaining a subgroup with an enhanced treatment effect without paying attention to subgroup size. Yet, a clinically meaningful subgroup learning approach should identify the maximum number of patients who can benefit from the better treatment. In this article, we present an optimal subgroup selection rule (SSR) that maximizes the number of selected patients, and in the meantime, achieves the pre‐specified clinically meaningful mean outcome, such as the average treatment effect. We derive two equivalent theoretical forms of the optimal SSR based on the contrast function that describes the treatment‐covariates interaction in the outcome. We further propose a constrained policy tree search algorithm (CAPITAL) to find the optimal SSR within the interpretable decision tree class. The proposed method is flexible to handle multiple constraints that penalize the inclusion of patients with negative treatment effects, and to address time to event data using the restricted mean survival time as the clinically interesting mean outcome. Extensive simulations, comparison studies, and real data applications are conducted to demonstrate the validity and utility of our method.

A statistical perspective on baseline adjustment in pharmacogenomic genome-wide association studies of quantitative change

In pharmacogenetic (PGx) studies, drug response phenotypes are often measured in the form of change in a quantitative trait before and after treatment. There is some debate in recent literature regarding baseline adjustment, or inclusion of pre-treatment or baseline value as a covariate, in PGx genome-wide association studies (GWAS) analysis. Here, we provide a clear statistical perspective on this baseline adjustment issue by running extensive simulations based on nine statistical models to evaluate the influence of baseline adjustment on type I error and power. We then apply these nine models to analyzing the change in low-density lipoprotein cholesterol (LDL-C) levels with ezetimibe + simvastatin combination therapy compared with simvastatin monotherapy therapy in the 5661 participants of the IMPROVE-IT (IMProved Reduction of Outcomes: Vytroin Efficacy International Trial) PGx GWAS, supporting the conclusions drawn from our simulations. Both simulations and GWAS analyses consistently show that baseline-unadjusted models inflate type I error for the variants associated with baseline value if the baseline value is also associated with change from baseline (e.g., when baseline value is a mediator between a variant and change from baseline), while baseline-adjusted models can control type I error in various scenarios. We thus recommend performing baseline-adjusted analyses in PGx GWASs of quantitative change.

GWAS of longitudinal trajectories at biobank scale

Biobanks linked to massive, longitudinal electronic health record (EHR) data make numerous new genetic research questions feasible. One among these is the study of biomarker trajectories. For example, high blood pressure measurements over visits strongly predict stroke onset, and consistently high fasting glucose and Hb1Ac levels define diabetes. Recent research reveals that not only the mean level of biomarker trajectories but also their fluctuations, or within-subject (WS) variability, are risk factors for many diseases. Glycemic variation, for instance, is recently considered an important clinical metric in diabetes management. It is crucial to identify the genetic factors that shift the mean or alter the WS variability of a biomarker trajectory. Compared to traditional cross-sectional studies, trajectory analysis utilizes more data points and captures a complete picture of the impact of time-varying factors, including medication history and lifestyle. Currently, there are no efficient tools for genome-wide association studies (GWASs) of biomarker trajectories at the biobank scale, even for just mean effects. We propose TrajGWAS, a linear mixed effect model-based method for testing genetic effects that shift the mean or alter the WS variability of a biomarker trajectory. It is scalable to biobank data with 100,000 to 1,000,000 individuals and many longitudinal measurements and robust to distributional assumptions. Simulation studies corroborate that TrajGWAS controls the type I error rate and is powerful. Analysis of eleven biomarkers measured longitudinally and extracted from UK Biobank primary care data for more than 150,000 participants with 1,800,000 observations reveals loci that significantly alter the mean or WS variability.

Germline HLA landscape does not predict efficacy of pembrolizumab monotherapy across solid tumor types

We evaluated the impact of class I and class II human leukocyte antigen (HLA) genotypes, heterozygosity, and diversity on the efficacy of pembrolizumab. Seventeen pembrolizumab clinical trials across eight tumor types and one basket trial in patients with advanced solid tumors were included (n > 3,500 analyzed). Germline DNA was genotyped using a custom genotyping array. HLA diversity (measured by heterozygosity and evolutionary divergence) across class I loci was not associated with improved response to pembrolizumab, either within each tumor type evaluated or across all patients. Similarly, HLA heterozygosity at each class I and class II gene was not associated with response to pembrolizumab after accounting for the number of tests conducted. No conclusive association between HLA genotype and response to pembrolizumab was identified in this dataset. Germline HLA genotype or diversity alone is not an important independent determinant of response to pembrolizumab and should not be used for clinical decision-making in patients treated with pembrolizumab.

Evaluating the Efficacy of Coronavirus Disease 2019 Vaccines

A large number of studies are being conducted to evaluate the efficacy and safety of candidate vaccines against coronavirus disease 2019 (COVID-19). Most phase 3 trials have adopted virologically confirmed symptomatic COVID-19 as the primary efficacy end point, although laboratory-confirmed severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) is also of interest. In addition, it is important to evaluate the effect of vaccination on disease severity. To provide a full picture of vaccine efficacy and make efficient use of available data, we propose using SARS-CoV-2 infection, symptomatic COVID-19, and severe COVID-19 as dual or triple primary end points. We demonstrate the advantages of this strategy through realistic simulation studies. Finally, we show how this approach can provide rigorous interim monitoring of the trials and efficient assessment of the durability of vaccine efficacy.

CauchyCP: A powerful test under non-proportional hazards using Cauchy combination of change-point Cox regressions

Non-proportional hazards data are routinely encountered in randomized clinical trials. In such cases, classic Cox proportional hazards model can suffer from severe power loss, with difficulty in interpretation of the estimated hazard ratio since the treatment effect varies over time. We propose CauchyCP, an omnibus test of change-point Cox regression models, to overcome both challenges while detecting signals of non-proportional hazards patterns. Extensive simulation studies demonstrate that, compared to existing treatment comparison tests under non-proportional hazards, the proposed CauchyCP test (a) controls the type I error better at small α levels (<0.01); (b) increases the power of detecting time-varying effects; and (c) is more computationally efficient than popular methods like MaxCombo for large-scale data analysis. The superior performance of CauchyCP is further illustrated using retrospective analyses of two randomized clinical trial datasets and a pharmacogenetic biomarker study dataset. The R package CauchyCP is publicly available on CRAN.

A Deferred-Vaccination Design to Assess Durability of COVID-19 Vaccine Effect After the Placebo Group Is Vaccinated

Multiple candidate vaccines to prevent COVID-19 have entered large-scale phase 3 placebo-controlled randomized clinical trials, and several have demonstrated substantial short-term efficacy. At some point after demonstration of substantial efficacy, placebo recipients should be offered the efficacious vaccine from their trial, which will occur before longer-term efficacy and safety are known. The absence of a placebo group could compromise assessment of longer-term vaccine effects. However, by continuing follow-up after vaccination of the placebo group, this study shows that placebo-controlled vaccine efficacy can be mathematically derived by assuming that the benefit of vaccination over time has the same profile for the original vaccine recipients and the original placebo recipients after their vaccination. Although this derivation provides less precise estimates than would be obtained by a standard trial where the placebo group remains unvaccinated, this proposed approach allows estimation of longer-term effect, including durability of vaccine efficacy and whether the vaccine eventually becomes harmful for some. Deferred vaccination, if done open-label, may lead to riskier behavior in the unblinded original vaccine group, confounding estimates of long-term vaccine efficacy. Hence, deferred vaccination via blinded crossover, where the vaccine group receives placebo and vice versa, would be the preferred way to assess vaccine durability and potential delayed harm. Deferred vaccination allows placebo recipients timely access to the vaccine when it would no longer be proper to maintain them on placebo, yet still allows important insights about immunologic and clinical effectiveness over time.

Principal stratum strategy: Potential role in drug development

A randomized trial allows estimation of the causal effect of an intervention compared to a control in the overall population and in subpopulations defined by baseline characteristics. Often, however, clinical questions also arise regarding the treatment effect in subpopulations of patients, which would experience clinical or disease related events post-randomization. Events that occur after treatment initiation and potentially affect the interpretation or the existence of the measurements are called intercurrent events in the ICH E9(R1) guideline. If the intercurrent event is a consequence of treatment, randomization alone is no longer sufficient to meaningfully estimate the treatment effect. Analyses comparing the subgroups of patients without the intercurrent events for intervention and control will not estimate a causal effect. This is well known, but post-hoc analyses of this kind are commonly performed in drug development. An alternative approach is the principal stratum strategy, which classifies subjects according to their potential occurrence of an intercurrent event on both study arms. We illustrate with examples that questions formulated through principal strata occur naturally in drug development and argue that approaching these questions with the ICH E9(R1) estimand framework has the potential to lead to more transparent assumptions as well as more adequate analyses and conclusions. In addition, we provide an overview of assumptions required for estimation of effects in principal strata. Most of these assumptions are unverifiable and should hence be based on solid scientific understanding. Sensitivity analyses are needed to assess robustness of conclusions.

Clinical Endpoints for Evaluating Efficacy in COVID-19 Vaccine Trials

Several vaccine candidates to protect against severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection or coronavirus disease 2019 (COVID-19) have entered or will soon enter large-scale, phase 3, placebo-controlled randomized clinical trials. To facilitate harmonized evaluation and comparison of the efficacy of these vaccines, a general set of clinical endpoints is proposed, along with considerations to guide the selection of the primary endpoints on the basis of clinical and statistical reasoning. The plausibility that vaccine protection against symptomatic COVID-19 could be accompanied by a shift toward more SARS-CoV-2 infections that are asymptomatic is highlighted, as well as the potential implications of such a shift.

Survival analysis using a 5-step stratified testing and amalgamation routine (5-STAR) in randomized clinical trials

Randomized clinical trials are often designed to assess whether a test treatment prolongs survival relative to a control treatment. Increased patient heterogeneity, while desirable for generalizability of results, can weaken the ability of common statistical approaches to detect treatment differences, potentially hampering the regulatory approval of safe and efficacious therapies. A novel solution to this problem is proposed. A list of baseline covariates that have the potential to be prognostic for survival under either treatment is pre-specified in the analysis plan. At the analysis stage, using all observed survival times but blinded to patient-level treatment assignment, "noise" covariates are removed with elastic net Cox regression. The shortened covariate list is used by a conditional inference tree algorithm to segment the heterogeneous trial population into subpopulations of prognostically homogeneous patients (risk strata). After patient-level treatment unblinding, a treatment comparison is done within each formed risk stratum and stratum-level results are combined for overall statistical inference. The impressive power-boosting performance of our proposed 5-step stratified testing and amalgamation routine (5-STAR), relative to that of the logrank test and other common approaches that do not leverage inherently structured patient heterogeneity, is illustrated using a hypothetical and two real datasets along with simulation results. Furthermore, the importance of reporting stratum-level comparative treatment effects (time ratios from accelerated failure time model fits in conjunction with model averaging and, as needed, hazard ratios from Cox proportional hazard model fits) is highlighted as a potential enabler of personalized medicine. An R package is available at https://github.com/rmarceauwest/fiveSTAR.

Assessing Durability of Vaccine Effect Following Blinded Crossover in COVID-19 Vaccine Efficacy Trials

BACKGROUND

Several candidate vaccines to prevent COVID-19 disease have entered large-scale phase 3 placebo-controlled randomized clinical trials and some have demonstrated substantial short-term efficacy. Efficacious vaccines should, at some point, be offered to placebo participants, which will occur before long-term efficacy and safety are known.

METHODS

Following vaccination of the placebo group, we show that placebo-controlled vaccine efficacy can be derived by assuming the benefit of vaccination over time has the same profile for the original vaccine recipients and the placebo crossovers. This reconstruction allows estimation of both vaccine durability and potential vaccine-associated enhanced disease.

RESULTS

Post-crossover estimates of vaccine efficacy can provide insights about durability, identify waning efficacy, and identify late enhancement of disease, but are less reliable estimates than those obtained by a standard trial where the placebo cohort is maintained. As vaccine efficacy estimates for post-crossover periods depend on prior vaccine efficacy estimates, longer pre-crossover periods with higher case counts provide better estimates of late vaccine efficacy. Further, open-label crossover may lead to riskier behavior in the immediate crossover period for the unblinded vaccine arm, confounding vaccine efficacy estimates for all post-crossover periods.

CONCLUSIONS

We advocate blinded crossover and continued follow-up of trial participants to best assess vaccine durability and potential delayed enhancement of disease. This approach allows placebo recipients timely access to the vaccine when it would no longer be proper to maintain participants on placebo, yet still allows important insights about immunological and clinical effectiveness over time.

Multi-trait analysis of rare-variant association summary statistics using MTAR

Integrating association evidence across multiple traits can improve the power of gene discovery and reveal pleiotropy. Most multi-trait analysis methods focus on individual common variants in genome-wide association studies. Here, we introduce multi-trait analysis of rare-variant associations (MTAR), a framework for joint analysis of association summary statistics between multiple rare variants and different traits. MTAR achieves substantial power gain by leveraging the genome-wide genetic correlation measure to inform the degree of gene-level effect heterogeneity across traits. We apply MTAR to rare-variant summary statistics for three lipid traits in the Global Lipids Genetics Consortium. 99 genome-wide significant genes were identified in the single-trait-based tests, and MTAR increases this to 139. Among the 11 novel lipid-associated genes discovered by MTAR, 7 are replicated in an independent UK Biobank GWAS analysis. Our study demonstrates that MTAR is substantially more powerful than single-trait-based tests and highlights the value of MTAR for novel gene discovery.

The pharmacogenetics of OATP1B1 variants and their impact on the pharmacokinetics and efficacy of elbasvir/grazoprevir

Aim: To evaluate the effect of SLCO1B1 genetic variants on grazoprevir pharmacokinetics and efficacy. Methods: A retrospective analysis of 1578 hepatitis C virus-infected participants from ten Phase II/III clinical trials. Results: Relative to noncarriers of the risk allele, geometric mean ratios (95% CI) of grazoprevir area under curve (AUC)_0-24 were: rs4149056 (risk allele C), one copy, 1.13 (1.06-1.21), two copies, 1.43 (1.16-1.77); and rs11045819 (risk allele A), one copy, 0.93 (0.87-1.00); two copies, 0.78 (0.61-1.00). The rs2306283 variant was not associated with grazoprevir exposure. None of the SLCO1B1 variants were associated with sustained virologic response. Conclusion: Genetic variants in SLCO1B1 were associated with modest changes in grazoprevir pharmacokinetics, but not with meaningful differences in efficacy.

Composite Kernel Association Test (CKAT) for SNP-set joint assessment of genotype and genotype-by-treatment interaction in Pharmacogenetics studies

MOTIVATION

It is of substantial interest to discover novel genetic markers that influence drug response in order to develop personalized treatment strategies that maximize therapeutic efficacy and safety. To help enable such discoveries, we focus on testing the association between the cumulative effect of multiple single nucleotide polymorphisms (SNPs) in a particular genomic region and a drug response of interest. However, the currently existing methods are either computational inefficient or not able to control type I error and provide decent power for whole exome or genome analysis in Pharmacogenetics (PGx) studies with small sample sizes.

RESULTS

In this article, we propose the Composite Kernel Association Test (CKAT), a flexible and robust kernel machine-based approach to jointly test the genetic main effect and SNP-treatment interaction effect for SNP-sets in Pharmacogenetics (PGx) assessments embedded within randomized clinical trials. An analytic procedure is developed to accurately calculate the P-value so that computationally extensive procedures (e.g. permutation or perturbation) can be avoided. We evaluate CKAT through extensive simulation studies and application to the gene-level association test of the reduction in Clostridium difficile infection recurrence in patients treated with bezlotoxumab. The results demonstrate that the proposed CKAT controls type I error well for PGx studies, is efficient for whole exome/genome association analysis and provides better power performance than existing methods across multiple scenarios.

AVAILABILITY AND IMPLEMENTATION

The R package CKAT is publicly available on CRAN https://cran.r-project.org/web/packages/CKAT/.

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

A note on the draft International Council for Harmonisation guidance on estimands and sensitivity analysis

In the second half of 2014, the Steering Committee of the International Council for Harmonisation endorsed the formation of an expert working group to develop an addendum to the International Council for Harmonisation E9 guideline ( Statistical Principles for Clinical Trials). The addendum was to focus on two clinical trial topics: estimands and sensitivity analysis. A draft of the addendum, referred to as E9/R1, was developed by the expert working group and made available for public comments across the International Council for Harmonisation regions in the second half of 2017. A structured framework for clinical trial design and analysis proposed in the draft addendum are briefly described, including four key inputs for developing objective-driven estimands and strategies for tackling one of the inputs (‘intercurrent events’). The proposed framework aligns each clinical trial objective with the corresponding statistical target of estimation (estimand), trial design and data to be collected, main method of estimation/inference, and sensitivity analysis to pressure test key analytic assumption(s) in the main analysis. A case study from the diabetes therapeutic area illustrates how the framework can be implemented in practice. International Council for Harmonisation E9/R1 is expected to enable better planning, conduct, analysis, and interpretation of randomised clinical trials. This will facilitate improvements in new drug applications and strengthen understanding of decision making by regulatory authorities and advisory committees.

Hazard ratio inference in stratified clinical trials with time-to-event endpoints and limited sample size

The stratified Cox model is commonly used for stratified clinical trials with time-to-event endpoints. The estimated log hazard ratio is approximately a weighted average of corresponding stratum-specific Cox model estimates using inverse-variance weights; the latter are optimal only under the (often implausible) assumption of a constant hazard ratio across strata. Focusing on trials with limited sample sizes (50-200 subjects per treatment), we propose an alternative approach in which stratum-specific estimates are obtained using a refined generalized logrank (RGLR) approach and then combined using either sample size or minimum risk weights for overall inference. Our proposal extends the work of Mehrotra et al, to incorporate the RGLR statistic, which outperforms the Cox model in the setting of proportional hazards and small samples. This work also entails development of a remarkably accurate plug-in formula for the variance of RGLR-based estimated log hazard ratios. We demonstrate using simulations that our proposed two-step RGLR analysis delivers notably better results through smaller estimation bias and mean squared error and larger power than the stratified Cox model analysis when there is a treatment-by-stratum interaction, with similar performance when there is no interaction. Additionally, our method controls the type I error rate while the stratified Cox model does not in small samples. We illustrate our method using data from a clinical trial comparing two treatments for colon cancer.

Hazard ratio estimation and inference in clinical trials with many tied event times

The medical literature contains numerous examples of randomized clinical trials with time-to-event endpoints in which large numbers of events accrued over relatively short follow-up periods, resulting in many tied event times. A generally common feature across such examples was that the logrank test was used for hypothesis testing and the Cox proportional hazards model was used for hazard ratio estimation. We caution that this common practice is particularly risky in the setting of many tied event times for two reasons. First, the estimator of the hazard ratio can be severely biased if the Breslow tie-handling approximation for the Cox model (the default in SAS and Stata software) is used. Second, the 95% confidence interval for the hazard ratio can include one even when the corresponding logrank test p-value is less than 0.05. To help establish a better practice, with applicability for both superiority and noninferiority trials, we use theory and simulations to contrast Wald and score tests based on well-known tie-handling approximations for the Cox model. Our recommendation is to report the Wald test p-value and corresponding confidence interval based on the Efron approximation. The recommended test is essentially as powerful as the logrank test, the accompanying point and interval estimates of the hazard ratio have excellent statistical properties even in settings with many tied event times, inferential alignment between the p-value and confidence interval is guaranteed, and implementation is straightforward using commonly used software.

Incorporating baseline measurements into the analysis of crossover trials with time-to-event endpoints

Two-period two-treatment (2×2) crossover designs are commonly used in clinical trials. For continuous endpoints, it has been shown that baseline (pretreatment) measurements collected before the start of each treatment period can be useful in improving the power of the analysis. Methods to achieve a corresponding gain for censored time-to-event endpoints have not been adequately studied. We propose a method in which censored values are treated as missing data and multiply imputed using prespecified parametric event time models. The event times in each imputed data set are then log-transformed and analyzed using a linear model suitable for a 2×2 crossover design with continuous endpoints, with the difference in period-specific baselines included as a covariate. Results obtained from the imputed data sets are synthesized for point and confidence interval estimation of the treatment ratio of geometric mean event times using model averaging in conjunction with Rubin's combination rule. We use simulations to illustrate the favorable operating characteristics of our method relative to two other methods for crossover trials with censored time-to-event data, ie, a hierarchical rank test that ignores the baselines and a stratified Cox model that uses each study subject as a stratum and includes period-specific baselines as a covariate. Application to a real data example is provided.

Abstract 4759: Designing clinical trials in tumor indications with a positive signal in phase 1

Increasingly, the conventional proof-of-concept randomized Phase 2 study is skipped in favor of directly going to Phase 3 after an experimental oncology drug has demonstrated promising anti-tumor activity in Phase 1 with a small sample size. This shift in the balance between certainty and speed is especially evident in the immune-oncology space where the tremendous success of immune checkpoint inhibitors has brought unprecedented competition in the field. The aggressive approach can be very risky no matter how promising an experimental drug appears in Phase 1. In this presentation, we introduce a novel adaptive design that mitigates the risk of late-stage programs. The proposed approach starts with a Phase 2 trial and adds an option in the design that allows the expansion of the Phase 2 trial into Phase 3 if the interim result based on the Phase 2 endpoint is promising. If the decision is to not expand, the study is kept as a Phase 2 trial and the primary analysis is conducted at the end of Phase 2. Otherwise, the study is expanded into a Phase 3 trial and the primary analysis of the study is conducted at the end of Phase 3, utilizing data from all enrolled patients including those already used for the decision making in the ongoing trial. The proposed approach is more efficient than the conventional approach that conducts and analyzes Phase 2 and Phase 3 trials sequentially, and is less risky than the contemporary approach of skipping Phase 2. Importantly, we will show that this design controls overall Type I error regardless of the expansion criterion. As a result, the study can still be considered positive even without expansion to Phase 3.

Citation Format: Cong Chen, Keaven Anderson, Devan V. Mehrotra, Eric H. Rubin, Archie Tse. Designing clinical trials in tumor indications with a positive signal in phase 1 [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2018; 2018 Apr 14-18; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2018;78(13 Suppl):Abstract nr 4759.

Efficient baseline utilization for incomplete block crossover clinical trials

Incomplete block crossover trials with period-specific baseline and post-baseline (outcome) measures for each subject are often used in clinical drug development; without loss of generality, we focus on the three-treatment two-period ( 3×2 ) crossover. Data from such trials are commonly analyzed using a mixed effects model with indicator terms for treatment and period, and an unstructured covariance matrix for the vector of intra-subject measurements. It is well-known that treatment effect estimates from this analysis are complex functions of both within-subject and between-subject treatment contrasts. We caution that the associated type I error rate and power for hypothesis testing can be non-trivially influenced by how the baselines are utilized. Specifically, the mixed effects analysis which uses change from baseline as the dependent variable is shown to consistently underperform corresponding analyses in which the outcome is the dependent variable and linear combinations of the baselines are used as period-specific and/or period-invariant covariates. A simpler fixed effects analysis of covariance involving only within-subject contrasts is also described for small sample situations in which the mixed effects analyses can suffer from increased type I error rates. Theoretical insights, simulation results and an illustrative example with real data are used to develop the main points.

Genome Wide Analysis Reveals Host Genetic Variants that Associate with Reduction in Clostridium difficile Infection Recurrence (rCDI) in Patients Treated with Bezlotoxumab

Background

Bezlotoxumab (BEZ) and actoxumab (ACT) are monoclonal antibodies against C. difficile toxins B and A, respectively. Patients receiving a single infusion of BEZ alone or with ACT in the MODIFY I/II trials showed a consistent reduction in the rate of rCDI over a 12-week period compared with a placebo (PBO) infusion. Exploratory genome wide analyses were conducted to determine whether genetic variants across the genome were associated with treatment response (rCDI).

Methods

DNA was extracted from blood obtained from patients who consented to genetic analysis (PGx population). Genetic data were generated on a commercial Axiom array platform (Affymetrix). Genotype imputation was performed using the 1000 Genomes Phase 3 reference data and Impute2 software after genetic quality control. Data from BEZ and ACT+BEZ arms were combined to provide increased power. The logistic regression with likelihood ratio test was used to search for single nucleotide polymorphisms (SNPs) that were strongly associated with a treatment effect on rCDI.

Results

An SNP rs2516513 located in the extended major histocompatibility complex (xMHC), region with a minor allele frequency of 25% in the general population, was associated with rCDI (P = 3.04E-08) (Figure 1). rCDI rates for the PGx population and in subgroups at high/low risk for rCDI stratified by SNP rs2516513 are shown in Table 1. Carriers of the T allele of SNP rs2516513 were associated with a statistically significant reduction in rCDI in BEZ-treated patients but not in PBO-treated patients (DrCDI = -21.5%). The magnitude of the effect of the T allele on rCDI is most prominent in patients who have ≥1 risk factor for rCDI (DrCDI = -24.6%), but is also present in patients without risk factors (DrCDI = -10.6%). In CC homozygous patients, rCDI rates are similar in both treatment groups and in patients at high and low risk of rCDI.

Conclusion

An SNP variant rs2516513 is associated with a lower rate of rCDI recurrence in patients treated with BEZ. The location of the associated genetic variant on chromosome 6 within xMHC, suggests that a host driven, immunological mechanism may play a role in rCDI and may predict patients most likely to respond to BEZ. As this is an exploratory finding, the results should be replicated in an independent validation study.

Disclosures

P. Shaw, Merck & Co., Inc.: Employee, May own stock/hold stock options in Company; J. Shen, Merck & Co., Inc.: Employee, may hold stock/hold stock options in the Company; M. B. Dorr, Merck & Co., Inc.: Employee and Shareholder, may own stock/hold stock options in the Company; J. Li, BGI-Shenzhen: Employee, Salary; R. Mogg, Merck & Co., Inc.: Employee, May hold stock/stock options in the Company; D. V. Mehrotra, Merck & Co., Inc.: Employee, may own stock/hold stock options in the Company; R. L. Blanchard, Merck & Co., Inc.: Employee, may own stock/hold stock options in the Company

A 2-in-1 adaptive phase 2/3 design for expedited oncology drug development

We propose an adaptive design that allows us to expand an ongoing Phase 2 trial into a Phase 3 trial to expedite a drug development program with fewer patients. Rather than the usual practice of increasing sample size with a less positive interim outcome, here we propose maintaining sample size with such a result and wait for fully mature data. The final Phase 2 data may be negative, may warrant a larger Phase 3 trial, or, in the extreme, could provide a definitively positive outcome. If the interim outcome is more positive, the trial continues to an originally planned larger sample size for a definitive Phase 3 evaluation. All patients from the study are used for inference regardless of the interim expansion decision. We show that no penalty needs to be paid in order to control the overall Type I error of the study, under a mild assumption that is expected to generally hold in practice. The proposed design may be considered an alternative approach to sample size adjustment for ongoing trials. As such, the use of an intermediate endpoint for adaptive decision is a unique feature of the design. A hypothetical example is provided for illustration purpose.

Current Practices in Choosing Estimands and Sensitivity Analyses in Clinical Trials: Results of the ICH E9 Survey

BACKGROUND

An addendum to the International Conference on Harmonisation E9 (ICH E9) guidance document (Statistical Principles for Clinical Trials) is currently under development. The aim of the addendum is to promote harmonized standards on the choice of estimand (a well-defined measure of the treatment effect that is being estimated) in clinical trials and to describe a consensual framework for planning, conducting, and interpreting sensitivity analyses of clinical trial data.

METHODS

In order to help understand current practices relating to the choice of estimands and sensitivity analyses for clinical trials, the ICH E9 working group developing the addendum conducted a survey with a primary focus on clinical trials involving drugs, vaccines, and biologics. The survey was distributed electronically between May 19, 2015, and June 11, 2015, to various stakeholder groups within ICH, including industry, regulatory, and academic communities. A total of 1305 respondents participated.

RESULTS

Of the 1305 respondents 547 (42%), 344 (26%) and 283 (22%) were from Europe, USA and Japan respectively. Over half of the respondents work in pharmaceutical companies, and approximately a quarter of respondents noted oncology as the primary therapeutic area they work in. Over half of the respondents (595, 55%) noted the treatment effect being estimated was 'in the entire target population of patients regardless of whether they will take treatment as instructed'. The most common methods for handling missing data in primary analyses were mixed-models repeated measures (555, 56% respondents) and last observation carried forward (549, 55% respondents). The majority of respondents (816, 83%) noted they conducted sensitivity analyses to estimate treatment effects in different ways compared to the primary analysis by using alternative assumptions (627, 78%) and/or using alternative statistical methods (616, 76%).

CONCLUSIONS

The survey results have provided useful information to the ICH E9 working group on current practices on the choice of primary estimands for measuring treatment effects in confirmatory clinical trials, and approaches used to select sensitivity analyses.

Use of the false discovery rate for evaluating clinical safety data

Clinical adverse experience (AE) data are routinely evaluated using between group P values for every AE encountered within each of several body systems. If the P values are reported and interpreted without multiplicity considerations, there is a potential for an excess of false positive findings. Procedures based on confidence interval estimates of treatment effects have the same potential for false positive findings as P value methods. Excess false positive findings can needlessly complicate the safety profile of a safe drug or vaccine. Accordingly, we propose a novel method for addressing multiplicity in the evaluation of adverse experience data arising in clinical trial settings. The method involves a two-step application of adjusted P values based on the Benjamini and Hochberg1 false discovery rate (FDR). Data from three moderate to large vaccine trials are used to illustrate our proposed ‘Double FDR’ approach, and to reinforce the potential impact of failing to account for multiplicity. This work was in collaboration with the late Professor John W. Tukey who coined the term ‘Double FDR’.

Seeking harmony: estimands and sensitivity analyses for confirmatory clinical trials

In October 2014, the Steering Committee of the International Conference on Harmonization endorsed the formation of an expert working group to develop an addendum to the International Conference on Harmonization E9 guideline (“Statistical Principles for Clinical Trials”). The addendum will focus on two topics involving randomized confirmatory clinical trials: estimands and sensitivity analyses. Both topics are motivated, in part, by the need to improve the precision with which scientific questions of interest are formulated and addressed by clinical trialists and regulators, specifically in the context of post-randomization events such as use of rescue medication or missing data resulting from dropouts. Given the importance of these topics for the statistical and medical community, we articulate the reasons for the planned addendum. The resulting “ICH E9/R1” guideline will include a framework for improved trial planning, conduct, analysis, and interpretation; a draft is expected to be ready for public comment in the second half of 2016.

Effect of rAd5-Vector HIV-1 Preventive Vaccines on HIV-1 Acquisition: A Participant-Level Meta-Analysis of Randomized Trials

Background

Three phase 2b, double-blind, placebo-controlled, randomized efficacy trials have tested recombinant Adenovirus serotype-5 (rAd5)-vector preventive HIV-1 vaccines: MRKAd5 HIV-1 gag/pol/nef in Step and Phambili, and DNA/rAd5 HIV-1 env/gag/pol in HVTN505. Due to efficacy futility observed at the first interim analysis in Step and HVTN505, participants of all three studies were unblinded to their vaccination assignments during the study but continued follow–up. Rigorous meta-analysis can provide crucial information to advise the future utility of rAd5-vector vaccines.

Methods

We included participant-level data from all three efficacy trials, and three Phase 1–2 trials evaluating the HVTN505 vaccine regimen. We predefined two co-primary analysis cohorts for assessing the vaccine effect on HIV-1 acquisition. The modified-intention-to-treat (MITT) cohort included all randomly assigned participants HIV-1 uninfected at study entry, who received at least the first vaccine/placebo, and the Ad5 cohort included MITT participants who received at least one dose of rAd5-HIV vaccine or rAd5-placebo. Multivariable Cox regression models were used to estimate hazard ratios (HRs) of HIV-1 infection (vaccine vs. placebo) and evaluate HR variation across vaccine regimens, time since vaccination, and subgroups using interaction tests.

Findings

Results are similar for the MITT and Ad5 cohorts; we summarize MITT cohort results. Pooled across the efficacy trials, over all follow-up time 403 (n = 224 vaccine; n = 179 placebo) of 6266 MITT participants acquired HIV-1, with a non-significantly higher incidence in vaccine recipients (HR 1.21, 95% CI 0.99–1.48, P = 0.06). The HRs significantly differed by vaccine regimen (interaction P = 0.03; MRKAd5 HR 1.41, 95% CI 1.11–1.78, P = 0.005 vs. DNA/rAd5 HR 0.88, 95% CI 0.61–1.26, P = 0.48). Results were similar when including the Phase 1–2 trials. Exploratory analyses based on the efficacy trials supported that the MRKAd5 vaccine-increased risk was concentrated in Ad5-positive or uncircumcised men early in follow-up, and in Ad5-negative or circumcised men later. Overall, MRKAd5 vaccine-increased risk was evident across subgroups except in circumcised Ad5-negative men (HR 0.97, 95% CI 0.58−1.63, P = 0.91); there was little evidence that the DNA/rAd5 vaccine, that was tested in this subgroup, increased risk (HR 0.88, 95% CI 0.61–1.26, P = 0.48). When restricting the analysis of Step and Phambili to follow-up time before unblinding, 114 (n = 65 vaccine; n = 49 placebo) of 3770 MITT participants acquired HIV-1, with a non-significantly higher incidence in MRKAd5 vaccine recipients (HR 1.30, 95% CI 0.89–1.14, P = 0.18).

Interpretation and Significance

The data support increased risk of HIV-1 infection by MRKAd5 over all follow-up time, but do not support increased risk of HIV-1 infection by DNA/rAd5. This study provides a rationale for including monitoring plans enabling detection of increased susceptibility to infection in HIV-1 at-risk populations.

Safety Profile of the Merck Human Immunodeficiency Virus-1 Clade B gag DNA Plasmid Vaccine With and Without Adjuvants

The immunogenicity results from 3 phase I trials of the Merck DNA human immunodeficiency virus (HIV) vaccine have previously been reported. Because preventive DNA vaccine strategies continue to be leveraged for diverse infections, the safety and tolerability results from these studies can inform the field moving forward, particularly regarding adverse reactions and adjuvants. No serious vaccine-related adverse events were reported during the 3-dose priming phase. Pain at the injection site was more common with adjuvanted formulations than with the phosphate-buffered saline diluent alone. Febrile reactions were usually low grade. Although the AlPO₄ or CRL1005 adjuvants used in these studies did not significantly enhance the immunogenicity of the DNA vaccine, adverse events were numerically more common with adjuvanted formulations than without adjuvants.

Rank-based principal stratum sensitivity analyses

We describe rank-based approaches to assess principal stratification treatment effects in studies where the outcome of interest is only well-defined in a subgroup selected after randomization. Our methods are sensitivity analyses, in that estimands are identified by fixing a parameter and then we investigate the sensitivity of results by varying this parameter over a range of plausible values. We present three rank-based test statistics and compare their performance through simulations, and provide recommendations. We also study three different bootstrap approaches for determining levels of significance. Finally, we apply our methods to two studies: an HIV vaccine trial and a prostate cancer prevention trial.

Analysis of longitudinal clinical trials with missing data using multiple imputation in conjunction with robust regression

In a typical randomized clinical trial, a continuous variable of interest (e.g., bone density) is measured at baseline and fixed postbaseline time points. The resulting longitudinal data, often incomplete due to dropouts and other reasons, are commonly analyzed using parametric likelihood-based methods that assume multivariate normality of the response vector. If the normality assumption is deemed untenable, then semiparametric methods such as (weighted) generalized estimating equations are considered. We propose an alternate approach in which the missing data problem is tackled using multiple imputation, and each imputed dataset is analyzed using robust regression (M-estimation; Huber, 1973, Annals of Statistics 1, 799-821.) to protect against potential non-normality/outliers in the original or imputed dataset. The robust analysis results from each imputed dataset are combined for overall estimation and inference using either the simple Rubin (1987, Multiple Imputation for Nonresponse in Surveys, New York: Wiley) method, or the more complex but potentially more accurate Robins and Wang (2000, Biometrika 87, 113-124.) method. We use simulations to show that our proposed approach performs at least as well as the standard methods under normality, but is notably better under both elliptically symmetric and asymmetric non-normal distributions. A clinical trial example is used for illustration.