Relationship of the number of removed lymph nodes to bladder cancer and competing mortality after radical cystectomy

Abstract

The extent of lymph node dissection in radical cystectomy is a subject of controversy. A more extended dissection has been reported to be associated with superior survival. We analyzed the relationship between the lymph node count and different causes of death in a sample of 735 patients who underwent radical cystectomy for recurrent or muscle-invasive urothelial or undifferentiated carcinoma of the bladder. The median follow-up was 7.8 yr. The median lymph node count was 17, and the median age was 67 yr. Although there was a clear association between lymph node count and overall survival (≥21 vs. <10 lymph nodes: 10-yr rates: 59% vs. 32%, respectively; hazard ratio: 0.63; 95% confidence interval, 0.46-0.87; log-rank test: p=0.0056), there was no detectable relationship between bladder cancer mortality and lymph node count (narrowly congruent cumulative mortality curves, Pepe-Mori test, p values ranging between 0.40 and 0.93). The differences were virtually entirely attributable to differences in competing mortality. These observations indicate that serious bias may occur when the lymph node count is used to stratify patients undergoing radical cystectomy. The results of the ongoing randomized trials should be awaited to reliably answer the question of the degree to which more extensive dissection may improve outcome.

Patient summary: Survival differences in patients stratified by lymph node count may be attributed to competing mortality. The results of ongoing randomized trials should be awaited to answer the question of the degree to which more extensive lymph node dissection may improve outcome.

Trial registration: ClinicalTrials.gov NCT01215071 NCT01224665.

Keywords: Adjuvant; Bladder; Chemotherapy; Cystectomy; Lymph nodes; Proportional hazards model; Urologic neoplasms.