Impact of hospital volume and type on outcomes of open and endovascular repair of descending thoracic aneurysms in the United States Medicare population.

OBJECTIVE: Favorable outcomes of thoracic endovascular aortic repair (TEVAR) compared with open repair for descending thoracic aortic aneurysms (DTAs) have led to increasing TEVAR use. We evaluated the effect of case volume and hospital teaching status on clinical outcomes of intact DTA repair.
METHODS: The Medicare Provider Analysis and Review (MEDPAR) data set (2004 to 2007) was queried to identify open repair or TEVAR for DTA. Hospitals were stratified by DTA volume into high volume (HV; ≥ 8 cases/y) or low volume (LV; <8 cases/y) and teaching or nonteaching. The effect of hospital variables on the primary study end point of 30-day mortality and secondary end points of 30-day complications and long-term survival after open repair and TEVAR DTA repair were studied using univariate testing, multivariable regression modeling, Kaplan-Meier survival analysis, and Cox proportional hazards regression modeling.
RESULTS: We identified 763 hospitals performing 3554 open repairs and 3517 TEVARs. Overall DTA repair increased (P < .01) from 1375 in 2004 to 1987 in 2007. The proportion of hospitals performing open repair significantly decreased from 95% in 2004 to 57% in 2007 (P < .01), whereas those performing TEVAR increased (P < .01) from 24% to 76%. Overall repair type shifted from open (74% in 2004, the year before initial commercial availability of TEVAR) to TEVAR (39% open in 2007; P < .01). The fraction of open repairs at LV hospitals decreased from 56% in 2004 to 44% in 2007 (P < .01), whereas TEVAR increased from 24% in 2004 to 51% in 2007 (P < .01). Overall mortality during the study interval for open repair was 15% at LV hospitals vs 11% at HV hospitals (P < .01), whereas TEVAR mortality was similar, at 3.9% in LV vs 5.5% in HV hospitals (P = .43). LV was independently associated with increased mortality after open repair (odds ratio, 1.4; 95% confidence interval, 1.1-1.8; P < .01) but not after TEVAR. There was no independent effect of hospital teaching status on mortality or complications after open repair or TEVAR repair.
CONCLUSIONS: The total number of DTA repairs has significantly increased. Operative mortality for TEVAR is independent of hospital volume and type, whereas mortality after open surgery is lower at HV hospitals, suggesting that TEVAR can be safely performed across a spectrum of hospitals, whereas open surgery should be performed only at HV hospitals.