Risk analysis for thoracoscopic lung volume reduction: a multi-institutional experience

Objective: Most reports of thoracoscopic lung volume reduction (TLVR) are r elatively small and early experiences from a single institution, factors wh ich limit both the statistical validity and the applicability to the popula tion at large. In order to address these shortcomings we undertook an analy sis of the TLVR experience at five separate institutions to assess operativ e morbidity and identify predictors of mortality. Methods: Questionnaires w ere sent to four groups of surgical investigators at five institutions acti vely performing TLVR. Data was requested regarding preoperative, operative and postoperative parameters. Twenty-five potential predictors of mortality were analyzed and seven proved to be at least marginally significant (P < 0.10). These parameters were entered into a stepwise logistic regression an alysis to identify independent predictors. Results: The 682 patients (415 m ales, 267 females, mean age 64.0 years) underwent unilateral (410) or bilat eral (272) TLVRs. Overall, operative mortality was 6% with half of the deat hs resulting from respiratory causes. The remaining patients were discharge d to home (88%), a rehabilitation facility (4%) or a ventilator facility (2 %). There were 25 perioperative factors chosen representing clinically impo rtant indices such as spirometry, oxygenation, functional status, clinical and demographic variables. Univariate analysis identified seven variables a s predictors of mortality (P < 0.10) and these were entered into a stepwise logistic regression analysis. Only age, 6-min walk, gender (male 8%, femal e 3% mortality) and the procedure performed (unilateral 4.6%, bilateral 8%) were independent predictors while preoperative steroid therapy, preoperati ve oxygen administration, and time since smoking cessation dropped out of t he model. The specific institution, learning curve (early vs. late experien ce), type of lung disease, spirometric indices and predicted maximum VO2 we re not significant predictors. Conclusion: This experience suggests that un ilateral and bilateral lung volume reduction procedure can be performed wit h acceptable morbidity and mortality. Although age, gender, exercise capaci ty and the procedure performed are all independent predictors of mortality, the risk of operative death did not appear excessive in this fragile patie nt subset. (C) 2000 Elsevier Science B.V. All rights reserved.