Background The objective of the study was to determine if initial surgery (IS) or initial chemotherapy (IC) affects rates of optimal surgery and survival in a population with significant medical comorbidities.

Methods Data of all patients with stage III-IV ovarian, peritoneal, and fallopian tube cancers diagnosed from 1995 to 2008 were reviewed. Clinical and pathologic data were abstracted.

Results There were 551 cases for review: 255 (46.3%) received IS, and 296 (53.7%) received IC. Patients who received IC had higher stage (P < 0.001), higher-grade cancers (P < 0.001), higher mean CA-125 (P = 0.015), higher rates of diabetes (P = 0.006), hypertension (P = 0.008), and presurgical embolism (P < 0.022) and were older (P = 0.043). There was no difference with respect to body mass index, albumin, extent of surgery, or intensive care use. Rates of optimal cytoreduction were higher with IC compared with IS (72.7% vs 56.1%, P < 0.001). IS was associated with more blood loss (P = 0.005) and higher rates of postsurgical venous thrombosis (P < 0.001). Optimal cytoreduction predicted survival in both groups. Among optimal patients, IS improved median survival: progression-free survival of 14 months (IS) versus 12 months (IC), P = 0.004; overall survival of 58 months (IS) versus 34 months (IC), P = 0.002. Factors influencing this difference were receipt of IC and history of diabetes; both predictors of mortality: hazard ratios, 1.9 (95% confidence interval, 1.3–2.8; P < 0.001) and 1.8 (95% confidence interval, 1.02–3.1; P = 0.042), respectively.

Conclusions The achievement of optimal cytoreduction continues to be a significant predictor of survival, regardless of treatment approach. Patients selected for IS and in whom optimal cytoreduction was achieved had improvements in both progression-free survival and overall survival. However, the differences could not be explained by surgical effort alone as diabetes was independently associated with mortality.