Predictors of early mortality after radical nephrectomy with renal vein or inferior vena cava thrombectomy - A population-based study

Study Type - Prognosis (cohort) Level of Evidence 2a What's known on the subject? and What does the study add? Surgical volume has been well established as a predictor of outcomes for several complex surgical procedures, yet few studies have evaluated this relationship with regards to radical nephrectomy with either renal vein or inferior vena cava thrombectomy. In addition, most published literature consists of single-institution series from centres of excellence. We performed a population-level analysis and identified surgeon volume as a significant predictor of short-term mortality for this procedure. Such findings have potential implications regarding future policy and regionalization of care. Objective: To study the short-term mortality associated with radical nephrectomy with renal vein or inferior vena cava thrombectomy and the variables associated with this adverse outcome. Methods: Using the Ontario Cancer Registry, we identified 433 patients in the province of Ontario, Canada undergoing radical nephrectomy with venous thrombectomy between 1995 and 2004. We determined mortality rates at postoperative days 30 and 90. Other variables analysed include pathological tumour characteristics, surgeon graduation year, hospital/surgeon academic status, surgery year and hospital/surgeon volume. We used multivariable logistic regression to assess outcomes. Results: Overall mortality was 2.8% (30-day) and 5.8% (90-day). Surgeons performing a single nephrectomy with venous thrombectomy performed 14% of the cases and had the highest 30-day (6.7%) and 90-day (10%) mortality. The mortality rate for surgeons performing more than one surgery was 2.1% (30-day) and 5.1% (90-day). In recent years, this procedure was performed more commonly by the highest volume surgeons - 67% of cases in 2004 vs 40% in 1995. Significant predictors of 30-day mortality included procedure year and low surgeon volume. Significant predictors of 90-day mortality included procedure year, low surgeon volume, left-sided tumour and increasing hospital volume. Conclusions: For radical nephrectomy with venous thrombectomy, surgeon volume predicts short-term mortality, emphasizing the importance of experience in patient outcome. Despite a shift towards high-volume surgeons, 13.8% of cases continued to be performed by low-volume providers. If these results are confirmed in other jurisdictions, radical nephrectomy with venous thrombectomy should be regionalized and performed by surgeons who manage these cases regularly.