Recognizing the challenge in optimizing evidence-based interventions in complex patients with multiple comorbidities, we propose to evaluate optimal utilization of the implantable cardiac defibrillator (ICD) in individuals with heart failure (HF) and chronic kidney disease (CKD). The proposed project is of critical relevance to the US healthcare system given the increasing prevalence, and common coexistence, of HF and CKD, and the paucity of data from clinical trials and practice guidelines to specifically guide therapy in such complex patients. Of particular importance is the association between CKD and increased risk of competing modes of death, other than sudden cardiac death (SCD), which can alter the overall effectiveness of the ICD in patients with these two chronic conditions. Therefore, the aims of this modeling study are three-fold. First, we will use primary data from a HF population without an ICD enrolled in the Study of Left Ventricular Dysfunction (SOLVD) trials to estimate how variation in severity of HF and CKD affect the risk of cause-specific mortality (SCD, non-arrhythmic cardiac death, and non-cardiac mortality), and estimate the rates of CKD and HF disease progression. This will be achieved using competing risks proportional hazards regressions to model the three cause-specific modes of death across several categories of HF/CKD severity and age, and proportional hazards regression models to estimate HF/CKD progression. Second, we will validate and expand on our estimates from the primary data modeling based on a systematic literature review/data synthesis focusing on risks/benefits of ICD therapy, and rates of cause-specific mortality and CKD/HF disease progression in patients of varying age, and HF/CKD severity. Third, based on the transition probabilities obtained from the primary data modeling and literature review, we will estimate the lifetime benefits of ICD therapy in patients of varying age, HF/CKD severity using a Markov model. This will be accomplished by initially constructing a Markov model that will estimate the outcomes in patients without ICDs, and calibrating this non-ICD model with the SOLVD data and the control arm of an ICD trial (the Multicenter Automatic Defibrillator Trial, MADIT-II). We will then incorporate the effect of ICD treatment into the Markov model, and calibrate the full ICD versus no-ICD model on the MADIT-II data. The final and calibrated Markov model will then be used to estimate life-expectancy benefit of ICD therapy across several categories of HF/CKD severity and age, including in patients with severe CKD/HF who were under-represented in ICD trials.
| Trikalinos, Thomas A; Segal, Jodi B; Boyd, Cynthia M (2014) Addressing multimorbidity in evidence integration and synthesis. J Gen Intern Med 29:661-9 |
| Weiss, Carlos O; Varadhan, Ravi; Puhan, Milo A et al. (2014) Multimorbidity and evidence generation. J Gen Intern Med 29:653-60 |
| Uhlig, Katrin; Leff, Bruce; Kent, David et al. (2014) A framework for crafting clinical practice guidelines that are relevant to the care and management of people with multimorbidity. J Gen Intern Med 29:670-9 |
| Boyd, Cynthia M; Kent, David M (2014) Evidence-based medicine and the hard problem of multimorbidity. J Gen Intern Med 29:552-3 |
| Alsheikh-Ali, Alawi A; Trikalinos, Thomas A; Ruthazer, Robin et al. (2011) Risk of arrhythmic and nonarrhythmic death in patients with heart failure and chronic kidney disease. Am Heart J 161:204-209.e1 |
