Lurdes Y. T. Inoue

Implantable Cardioverter-Defibrillators for Primary Prevention of Sudden Cardiac Death in CKD: A Meta-analysis of Patient-Level Data From 3 Randomized Trials

BACKGROUND The benefit of a primary prevention implantable cardioverter-defibrillator (ICD) among patients with chronic kidney disease is uncertain. STUDY DESIGN Meta-analysis of patient-level data from randomized controlled trials. SETTING & POPULATION Patients with symptomatic heart failure and left ventricular ejection fraction<35%. SELECTION CRITERIA FOR STUDIES From 7 available randomized controlled studies with patient-level data, we selected studies with available data for important covariates. Studies without patient-level data for baseline estimated glomerular filtration rate (eGFR) were excluded. INTERVENTION Primary prevention ICD versus usual care effect modification by eGFR. OUTCOMES Mortality, rehospitalizations, and effect modification by eGFR. RESULTS We included data from the Multicenter Automatic Defibrillator Implantation Trial I (MADIT-I), MADIT-II, and the Sudden Cardiac Death in Heart Failure Trial (SCD-HeFT). 2,867 patients were included; 36.3% had eGFR<60 mL/min/1.73m2. Kaplan-Meier estimate of the probability of death during follow-up was 43.3% for 1,334 patients receiving usual care and 35.8% for 1,533 ICD recipients. After adjustment for baseline differences, there was evidence that the survival benefit of ICDs in comparison to usual care depends on eGFR (posterior probability for null interaction P<0.001). The ICD was associated with survival benefit for patients with eGFR≥60 mL/min/1.73 m2 (adjusted HR, 0.49; 95% posterior credible interval, 0.24-0.95), but not for patients with eGFR<60 mL/min/1.73 m2 (adjusted HR, 0.80; 95% posterior credible interval, 0.40-1.53). eGFR did not modify the association between the ICD and rehospitalizations. LIMITATIONS Few patients with eGFR<30 mL/min/1.73 m2 were available. Differences in trial-to-trial measurement techniques may lead to residual confounding. CONCLUSIONS Reductions in baseline eGFR decrease the survival benefit associated with the ICD. These findings should be confirmed by additional studies specifically targeting patients with varying eGFRs.

Survival Benefit of the Primary Prevention Implantable Cardioverter-Defibrillator Among Older Patients Does Age Matter? An Analysis of Pooled Data From 5 Clinical Trials

Background—The impact of patient age on the risks of death or rehospitalization after primary prevention implantable cardioverter-defibrillator (ICD) placement is uncertain. Methods and Results—Data from 5 major ICD trials were merged: the Multicenter Automatic Defibrillator Implantation Trial I (MADIT-I), the Multicenter UnSustained Tachycardia Trial (MUSTT), the Multicenter Automatic Defibrillator Implantation Trial II (MADIT-II), the Defibrillators in Nonischemic Cardiomyopathy Treatment Evaluation Trial (DEFINITE), and the Sudden Cardiac Death in Heart Failure Trial (SCD-HeFT). Median age at enrollment was 62 (interquartile range 53–70) years. Compared with their younger counterparts, older patients had a greater burden of comorbid illness. In unadjusted exploratory analyses, ICD recipients were less likely to die than nonrecipients in all age groups: among patients aged <55 years: hazard ratio 0.48, 95% posterior credible interval 0.33 to 0.69; among patients aged 55 to 64 years: hazard ratio 0.69, 95% posterior credible interval 0.53 to 0.90; among patients aged 65 to 74 years: hazard ratio 0.67, 95% posterior credible interval, 0.53 to 0.85; and among patients aged ≥75 years: hazard ratio 0.54, 95% posterior credible interval 0.37 to 0.78. Sample sizes were limited among patients aged ≥75 years. In adjusted Bayesian–Weibull modeling, point estimates indicate ICD efficacy persists but is attenuated with increasing age. There was evidence of an interaction between age and ICD treatment on survival (two-sided posterior tail probability of no interaction <0.01). Using an adjusted Bayesian logistic regression model, there was no evidence of an interaction between age and ICD treatment on rehospitalization (two-sided posterior tail probability of no interaction 0.44). Conclusions—In this analysis, the survival benefit of the ICD exists but is attenuated with increasing age. The latter finding may be because of the higher burden of comorbid illness, competing causes of death, or limited sample size of older patients. There was no evidence that age modifies the association between ICD treatment and rehospitalization.

Outcomes of implantable cardioverter-defibrillator use in patients with comorbidities: results from a combined analysis of 4 randomized clinical trials.

OBJECTIVES The aim of this study was to determine if the benefit of implantable cardioverter-defibrillators (ICDs) is modulated by medical comorbidity. BACKGROUND Primary prevention ICDs improve survival in patients at risk for sudden cardiac death. Their benefit in patients with significant comorbid illness has not been demonstrated. METHODS Original, patient-level datasets from MADIT I (Multicenter Automatic Defibrillator Implantation Trial I), MADIT II, DEFINITE (Defibrillators in Non-Ischemic Cardiomyopathy Treatment Evaluation), and SCD-HeFT (Sudden Cardiac Death in Heart Failure Trial) were combined. Patients in the combined population (N = 3,348) were assessed with respect to the following comorbidities: smoking, pulmonary disease, diabetes, peripheral vascular disease, atrial fibrillation, ischemic heart disease, and chronic kidney disease. The primary outcome was overall mortality, using the hazard ratio (HR) of time to death for patients receiving an ICD versus no ICD by extent of medical comorbidity, and adjusted for age, sex, race, left ventricular ejection fraction, use of antiarrhythmic drugs, beta-blockers, and angiotensin-converting enzyme inhibitors. RESULTS Overall, 25% of patients (n = 830) had <2 comorbid conditions versus 75% (n = 2,518) with significant comorbidity (≥2). The unadjusted hazard of death for patients with an ICD versus no ICD was significantly lower, but this effect was less for patients with ≥2 comorbidities (unadjusted HR: 0.71; 95% confidence interval: 0.61 to 0.84) compared with those with <2 comorbidities (unadjusted HR: 0.59; 95% confidence interval: 0.40 to 0.87). After adjustment, the benefit of an ICD decreased with increasing number of comorbidities (p = 0.004). CONCLUSIONS Patients with extensive comorbid medical illnesses may experience less benefit from primary prevention ICDs than those with less comorbidity; implantation should be carefully considered in sick patients. Further study of ICDs in medically complex patients is warranted.

What If I Don't Treat My PSA-Detected Prostate Cancer? Answers from Three Natural History Models

Background: Making an informed decision about treating a prostate cancer detected after a routine prostate-specific antigen (PSA) test requires knowledge about disease natural history, such as the chances that it would have been clinically diagnosed in the absence of screening and that it would metastasize or lead to death in the absence of treatment. Methods: We use three independently developed models of prostate cancer natural history to project risks of clinical progression events and disease-specific deaths for PSA-detected cases assuming they receive no primary treatment. Results: The three models project that 20%–33% of men have preclinical onset; of these 38%–50% would be clinically diagnosed and 12%–25% would die of the disease in the absence of screening and primary treatment. The risk that men age less than 60 at PSA detection with Gleason score 2–7 would be clinically diagnosed in the absence of screening is 67%–93% and would die of the disease in the absence of primary treatment is 23%–34%. For Gleason score 8 to 10 these risks are 90%–96% and 63%–83%. Conclusions: Risks of disease progression among untreated PSA-detected cases can be nontrivial, particularly for younger men and men with high Gleason scores. Model projections can be useful for informing decisions about treatment. Impact: This is the first study to project population-based natural history summaries in the absence of screening or primary treatment and risks of clinical progression events following PSA detection in the absence of primary treatment. Cancer Epidemiol Biomarkers Prev; 20(5); 740–50. ©2011 AACR.

Bayesian Hierarchical Curve Registration

Functional data often exhibit a common shape, but with variations in amplitude and phase across curves. The analysis often proceeds by synchronization of the data through curve registration. In this article we propose a Bayesian hierarchical model for curve registration. Our hierarchical model provides a formal account of amplitude and phase variability while borrowing strength from the data across curves in the estimation of the model parameters. We discuss extensions of the model by using penalized B-splines in the representation of the shape and time-transformation functions, and by allowing temporal misalignment of the curves. We discuss applications of our model to simulated data, as well as to two data sets. In particular, we use our model in a nonstandard analysis aimed at investigating regulatory network in time course microarray data.

Relationship between bayesian and frequentist sample size determination

Sample size determination is among the most commonly encountered tasks in statistical practice. A broad range of frequentist and Bayesian methods for sample size determination can be described as choosing the smallest sample that is sufficient to achieve some set of goals. An example for the frequentist is seeking the smallest sample size that is sufficient to achieve a desired power at a specified significance level. An example for the Bayesian is seeking the smallest sample size necessary to obtain, in expectation, a desired rate of correct classification of the hypothesis as true or false. This article explores parallels between Bayesian and frequentist methods for determining sample size. We provide a simple but general and pragmatic framework for investigating the relationship between the two approaches, based on identifying mappings to connect the Bayesian and frequentist inputs necessary to obtain the same sample size. We illustrate this mapping with examples, highlighting a somewhat surprising “approximate functional correspondence” between power-based and information-based optimal sample sizes.

Individualized Estimates of Overdiagnosis in Screen-Detected Prostate Cancer

BACKGROUND The chance that a prostate cancer detected by screening is overdiagnosed (ie, it would not have been detected in the absence of screening) can vary widely depending on the patients age and tumor characteristics. The purpose of this study is to use age, Gleason score, and prostate-specific antigen (PSA) level to help inform patients with screen-detected prostate cancers about the chances their cancers were overdiagnosed. METHODS A computer microsimulation model of prostate cancer natural history was used to generate virtual life histories in the presence and absence of PSA screening, including an indicator of whether screen-detected cancers are overdiagnosed. A logistic regression model was fit to nonmetastatic patients diagnosed by screening with PSA less than 10 ng/mL, and a nomogram was created to predict the individualized risk of overdiagnosis given age, Gleason score, and PSA at diagnosis. RESULTS The calibrated microsimulation model closely reproduces observed incidence trends in the Surveillance, Epidemiology, and End Results registries by age, stage, and Gleason score. The fitted logistic regression predicts risks of overdiagnosis among PSA-detected patients with an area under the curve of 0.75. Chances of overdiagnosis range from 2.9% to 88.1%. CONCLUSIONS The chances of overdiagnosis vary considerably by age, Gleason score, and PSA at diagnosis. The overdiagnosis nomogram presents tailored estimates of these risks based on patient and tumor information known at diagnosis and can be used to inform decisions about treating PSA-detected prostate cancers.

Calibrating disease progression models using population data: a critical precursor to policy development in cancer control

There are many more strategies for early detection of cancer than can be evaluated with randomized trials. Consequently, model-projected outcomes under different strategies can be useful for developing cancer control policy provided that the projections are representative of the population. To project population-representative disease progression outcomes and to demonstrate their value in assessing competing early detection strategies, we implement a model linking prostate-specific antigen (PSA) levels and prostate cancer progression and calibrate it to disease incidence in the US population. PSA growth is linear on the logarithmic scale with a higher slope after disease onset and with random effects on intercepts and slopes; parameters are estimated using data from the Prostate Cancer Prevention Trial. Disease onset, metastatic spread, and clinical detection are governed by hazard functions that depend on age or PSA levels; parameters are estimated by comparing projected incidence under observed screening and biopsy patterns with incidence observed in the Surveillance, Epidemiology, and End Results registries. We demonstrate implications of the model for policy development by projecting early detections, overdiagnosis, and mean lead times for PSA cutoffs 4.0 and 2.5 ng/mL and for screening ages 50-74 or 50-84. The calibrated model validates well, quantifies the tradeoffs involved across policies, and indicates that PSA screening with cutoff 4.0 ng/mL and screening ages 50-74 performs best in terms of overdiagnosis per early detection. The model produces representative outcomes for selected PSA screening policies and is shown to be useful for informing the development of sound cancer control policy.

Primary Prevention Implantable Cardioverter Defibrillators in Patients With Nonischemic Cardiomyopathy: A Meta-analysis

Importance Conflicting data have emerged on the efficacy of implantable cardioverter defibrillators (ICDs) for primary prevention of sudden cardiac death (primary prevention ICDs) in patients with nonischemic cardiomyopathy. Objective To investigate the association of primary prevention ICDs with all-cause mortality in patients with nonischemic cardiomyopathy. Data Sources PubMed was searched from January 1, 2000, through October 31, 2016, for the terms implantable defibrillator OR implantable cardioverter defibrillator AND non-ischemic cardiomyopathy. Additional references were identified from bibliographies of pertinent articles and queries to experts in this field. Study Selection Inclusion criteria consisted of a randomized clinical trial design and comparison of the ICD with medical therapy (control) in at least 100 patients with nonischemic cardiomyopathy. In addition, studies had to report on all-cause mortality during a follow-up period of at least 12 months and be published in English. The search yielded 10 studies, of which only 1 met the inclusion criteria. A search of bibliographies of pertinent articles and queries of experts in this field led to 3 additional studies. Data Extraction and Synthesis The PRISMA guidelines were used to abstract data and assess data quality and validity. Data were pooled using fixed- and random-effects models. Main Outcomes and Measures The primary end point was all-cause mortality. Before data collection started, primary prevention ICDs were hypothesized to reduce all-cause mortality among patients with nonischemic cardiomyopathy. Results Four randomized clinical trials met the selection criteria and included 1874 unique patients; 937 were in the ICD group and 937 in the control group. Pooling data from these trials showed a significant reduction in all-cause mortality with an ICD (hazard ratio, 0.75; 95% CI, 0.61-0.93; P = .008; P = .87 for heterogeneity). Conclusions and Relevance Primary prevention ICDs are efficacious at reducing all-cause mortality among patients with nonischemic cardiomyopathy. These findings support professional guidelines that recommend the use of ICDs in such patients.

Modeling grade progression in an active surveillance study

Prostate cancer grade, assessed with the Gleason score, describes how abnormal the tumor tissue and cells appear, and it is an important prognostic indicator of disease progression. Whether prostate tumors change grade is a question that has implications for screening and treatment. Empirical data on tumor grade over time have become available from men biopsied regularly as part of active surveillance (AS). However, biopsy (BX) grade is subject to misclassification. In this article, we develop a model that allows for estimation of the time of grade change while accounting for the misclassification error from BX grade. We use misclassification rates from studies of prostate cancer BXs followed by radical prostatectomy. Estimation of the transition times from true low-grade to high-grade disease is conducted within a Bayesian framework. We apply our model to serial observations on BX grade among 627 cases enrolled in a cohort of AS patients at Johns Hopkins University who were biopsied annually and referred to treatment if there was any evidence of disease progression on BX. We consider different prior distributions for the time to true grade progression. The estimated likelihood of grade progression within 10 years of study entry ranges from 12% to 24% depending on the prior. We conclude that knowledge of rates of grade misclassification allows for determination of true grade progression rates among men with serial BXs on AS. Although our results are sensitive to prior specifications, they indicate that in a nontrivial fraction of the patient population, tumor grade can progress.