Objectives:

To determine the efficacy, effectiveness, and safety of cardiac resynchronization therapy (CRT) and/or implantable cardioverter defibrillators (ICD) in patients with left ventricular systolic dysfunction (LVSD).

Data Sources:

A systematic and comprehensive literature search was conducted to identify randomized controlled trials (RCTs) evaluating efficacy and observational studies evaluating effectiveness or safety of CRT and/or ICD in patients with LVSD.

Review Methods:

Study selection, quality assessment, and data extraction were completed by several investigators in duplicate and independently. Random-effects models were used for analyses.

Results:

From 11,340 citations, we identified 14 RCTs (4,420 patients) for the CRT efficacy review, 106 studies (9,209 patients) for the CRT effectiveness review, 89 studies (9,677 patients) for the CRT safety review, 12 RCTs (8,516 patients) for the ICD efficacy review, 48 studies (15,097 patients) for the ICD effectiveness review, and 49 studies (12,592 patients) for the ICD safety review—all studies enrolled only patients with LVSD. An additional 12 studies (68,848 patients) were included for an analysis of peri-implant outcomes for all patients with ICD (i.e., not only LVSD patients).

All patients in the CRT studies had LVSD (mean LVEF from 21 to 30 percent) and prolonged QRS duration (mean from 155 to 209 msec), and 91 percent had New York Heart Association (NYHA) class III or IV symptoms. In patients with LVSD and heart failure symptoms, CRT improved ejection fraction (weighted mean difference 3.0 percent [95% CI, 0.9 to 5.1]), quality of life (weighted mean reduction in Minnesota Living with Heart Failure Questionnaire 8.0 points [95% CI, 5.6 to 10.4 points]), and function (59 percent of CRT recipients vs. 37 percent of controls improved by at least one NYHA class in the RCTs and between 63 percent and 82 percent of CRT recipients improved by at least one NYHA class in observational studies). The proportion of patients hospitalized for HF was reduced by 37 percent (95% CI, 7 to 57 percent) and all-cause mortality was reduced by 22 percent (95% CI, 9 to 33 percent; NNT=29 over 6 months). Implant success rate was 93 percent, 0.3 percent of patients with LVSD died during implantation. Over a median 11-month followup, 6.6 percent of CRT devices exhibited lead problems and 5 percent malfunctioned.

In patients with LVSD, ICD reduced all-cause mortality by 20 percent (95% CI, 10 to 29 percent; NNT=20 over 35 months). ICD implant success rate was 99 percent and peri-implant deaths occurred in 1.2 percent of LVSD patients and 1.3 percent of all implantees. The frequency of post-implantation complications in LVSD patients per 100 patient years included 1.4 (95% CI, 1.2 to 1.6) device malfunctions, 1.5 (95% CI, 1.3 to 1.8) lead problems, 0.6 (95% CI, 0.5 to 0.8) implant site infections, and 19.1 (95% CI, 16.5 to 22.0) inappropriate discharges in RCT participants and 4.7 (95% CI, 4.3 to 5.1) inappropriate discharges in patients enrolled in observational studies.

Conclusions:

ICD and CRT reduce all-cause mortality in patients with LVSD meeting RCT entry criteria. The incremental benefit of CRT plus ICD over CRT alone in patients with LVSD remains uncertain. None of the trials reported differences in the efficacy of CRT or ICD across patient subgroups, nor did our meta-regression detect any subgroup effects; however, subgroup analyses and meta-regression using aggregate trial data are post-hoc analyses and were underpowered to detect such effects. Examination of individual patient trial data is urgently needed to define which clinical subgroups are most likely to benefit from these devices.