Predictors of Distant Metastasis and Mortality in Patients with Stage II Colorectal Cancer

The aim of our study was to determine clinical factors that predicted distant metastasis and mortality in patients with stage II colorectal cancer (CRC). A retrospective review of all patients admitted to the Dallas Veteran's Affairs Medical Center from 1998 to 2002 with stage II CRC was performed. Factors associated with distant metastasis and mortality were assessed by univariate analysis. Independent predictors of mortality and distant metastasis were assessed by multivariate analysis. Fifty-five patients with stage II CRC were identified (96% men, age 65 +/- 1.2 years old). Univariate analysis demonstrated that patients with distant metastasis (n = 13) had a history of alcohol intake (54% vs 20%; P = 0.029), less history of angiotensin-converting enzyme inhibitor use for the management of hypertension (31% vs 67%; P = 0.029), greater incidence of a rectal location for cancer (54% vs 12%; P = 0.004), an abdominoperineal resection (APR) for surgical management of their cancers (23% vs 0%; P = 0.011), and less negative lymph nodes examined during surgical resection (9 +/- 2.2 vs 15 +/- 1.2, P = 0.022). Follow-up of at least 3 years revealed 14 mortalities. Survivors had a less preoperative serum carcinoembryonic antigen level (5.9 +/- 0.9 vs 19.1 +/- 8.6; P = 0.031), an average less tumor depth involvement (T stage: 3.0 +/- 0.03 vs 3.2 +/- 0.11; P = 0.065), more negative lymph nodes examined at the time of surgical resection (14.0 +/- 1.3 vs 10.0 +/- 1.7; P = 0.058), less incidence of rectal cancer (21% vs 43%; P = 0.060), and none had undergone an APR (0% vs. 21%; P = 0.010). Multivariate analysis revealed that alcohol intake (P = 0.014; odds ration OR = 5.3), patients undergoing an APR (P = 0.011; OR = 25), and less frequency of angiotensin-converting enzyme inhibitor use (P = 0.007; OR = 4.5) independently predicted distant metastasis, whereas preoperative carcinoembryonic antigen (P = 0.038; OR = 2.8) and patients undergoing an APR (P = 0.019; OR = 25) independently predicted mortality.