A comparative study on the prognostic impact of concurrent smoking and alcohol drinking on colon and rectal cancers: A frailty competing risks survival analysis

Aim: This study aimed to design a model and to compare the prognostic impact of concurrent using of tobacco and alcohol in colon and rectal cancers via a competing risks approach. Background: Many authors have confirmed both alcohol and tobacco smoking as the risk factors of CRC. The effect of concurrent using has been explored for an association with CRC and a comparison between sub-sites found in few studies. Patients and methods: 1219 patients with CRC diagnosis according to the pathology report of Research Institute For Gastroenterology And Liver Diseases (RIGLD) cancer registry, from 1 January 2002 to 1 October 2007, were entered into the study. Separately and concurrently, tobacco smoking and alcohol drinking were analyzed using competing risk parametric survival analysis with frailty parameter adjustment utilizing STATA statistical software.Results: In separate evaluations, tobacco smoking and alcohol use were significantly related to the survival only in patients with colon cancer (Hazard Ratio (HR) =1.61 and 95% Confidence Interval (CI) = (1.16-2.23) for tobacco and HR=1.93 and 95% CI= (1.22-3.06) for alcohol). In addition, these factors were significantly different between two subsites of colon and rectum (HR=1.78, (95% CI= (1.12-2.83) for tobacco and HR=4.44, 95% CI= (1.74-11.37) for alcohol). Also, results of concurrent analysis showed that only current or past tobacco- current or past alcohol category had significant relationship to the survival in patients with colon cancer (HR=2.17 and 95% CI= (1.27-3.71)) and thiswas significantly different between two sub-sites (HR(C/R) =5.16 and 95% CI= (1.65-16.12)). In total, survival probability of colonic patients was lower than that of rectum cancer patients. Conclusion: Concurrent using of tobacco and alcohol might be a prognostic factor of survival in patients with colon cancer. These results could be beneficial for prognosis and treatment application planning screening programs and its possible modifications.