National Death Index Search for Participants in SADHART

Columbia University Health Sciences, New York NY

Investigators

Linked publications & trials

Abstract

DESCRIPTION (provided by applicant): It is clear that major depressive disorder (MDD) following MI is associated with a marked increase in death. The immediate objective of this project is to establish which MDD patients are at highest risk of death and long-term to establish if antidepressant treatment would reduce this risk. Originally, no evidence existed that any antidepressant drug was safe in the immediate post-Mi period. SADHART tested the safety of sertraline and found efficacy and no evidence of harm. Although 369 patients were not adequate to test mortality reduction, there was a surprising 23% reduction in life-threatening events over the 6 months among patients receiving sertraline (just less than a trend). A recent NHLBI study comparing psychotherapy with usual cardiac care found no benefit of psychotherapy on mortality but a subset of severe depressions who received an SSRI had a statistically significant 42% reduction in mortality. Although this was neither randomized nor controlled it does reinforce the trend suggested by SADHART. Baseline depression severity, usually measured on the Beck Depression Inventory (BDI) does predict the risk of death. We will search the National Death Index (NDI) to identify SADHART enrollees who have died since the trial was completed. NDI is a computer index of death records established by the Government only for research purposes. This is expected to yield about 90 deaths and would give sufficient power to detect a 35% increase in relative risk of several other characteristics of depression that are hypothesized to be predictive of death and to compare these predictors with BDI. The characteristics to be examined are duration of present episode (chronicity) and number of previous episodes of MDD. In addition, data exist suggesting that the extra mortality that occurs in depressed patients is primarily sudden arrhythmic death and can be predicted by the number of extra ventricular beats in the post-Mi period. This longer follow-up would replicate this hypothesis and if this interaction exists, test if it is independent of severity, chronicity, or prior episodes. This additional information would identify patients at extremely high risk for death and in urgent need of treatment, and would aid in designing a definitive trial testing if treatment reduces mortality.