Nonthoracotomy- versus thoracotomy-implantable defibrillators. Intention-to-treat comparison of clinical outcomes.

BACKGROUND Nonthoracotomy-implantable cardioverter/defibrillator (ICD) systems may represent a significant advance in the treatment of patients with life-threatening ventricular arrhythmias, but their merits relative to those of the well-established thoracotomy systems remain largely unknown. The objective of this study was to compare the short- and long-term clinical outcomes after attempted ICD implantation via a nonthoracotomy versus thoracotomy approach in similar groups of patients. METHODS AND RESULTS Between September 1990 and December 1992, 212 consecutive patients underwent attempted ICD system implantation without concomitant cardiac surgery at a single institution. Approach selection was not randomized but rather was based primarily on hardware availability. Primary comparisons of short- and long-term outcome were performed according to the "intention-to-treat" principle. Implantation was attempted via a nonthoracotomy approach in 120 patients (57%) and via a thoracotomy approach in 92 patients (43%). Prior cardiac surgery was more prevalent in the nonthoracotomy patients; otherwise, groups did not differ significantly in terms of prognostically relevant clinical characteristics. Nonthoracotomy implantation was successful in 101 patients (84%). After crossover to thoracotomy implantation (14 patients), the eventual success rate for ICD system implantation was 96% in the nonthoracotomy group. Thoracotomy implantation was successful in 89 patients (97%). Operative mortality was 3.3% in the nonthoracotomy and 4.3% in the thoracotomy groups (P = .73). Nonthoracotomy group patients were less likely to experience postoperative congestive heart failure (6% versus 16%; P = .02) or supraventricular arrhythmia (6% versus 18%; P = .004) and had significantly shorter postoperative intensive care and total hospitalization. Total hospital costs were significantly lower in the nonthoracotomy group ($32,205 versus $37,265; P = .001). After a follow-up of 16 +/- 9 months, there were 17 deaths in the nonthoracotomy group (none sudden) and 12 deaths in the thoracotomy group (1 sudden). One- and 2-year Kaplan-Meier survival probabilities were .87 (95% CI, .78 to .91) and .80 (95% CI, .68 to .88) in the nonthoracotomy group and .90 (95% CI, .82 to .95) and .87 (95% CI, .77 to .93) in the thoracotomy group (P = .56; log-rank test). CONCLUSIONS Nonthoracotomy ICD implantation is associated with reduced surgical morbidity, postoperative hospital care requirement, and hospital costs and has similar efficacy in preventing sudden death relative to the thoracotomy approach. From these nonrandomized data, it appears that a nonthoracotomy approach should be considered preferable in most patients requiring ICD therapy.