BACKGROUND:Most studies investigate the association between the baseline Romano-Charlson comorbidity index (CCI) and survival on hemodialysis (HD). Few consider the effect of progression in the CCI score (CCIp) on patient survival before HD initiation. That is CCIp = CCI-1 - CCI-3, where CCI(-1) is the CCI score in the first year before HD initiation, and CCI-3 is the CCI score in the third year before HD initiation. The present study investigated whether CCIp affects the survival of incident HD patients. METHODS:Using the National Health Insurance (NHI) Research Database of Taiwan, we recruited 7391 adult incident HD patients in the year 2006 for this historical cohort study. We followed the cohort until the end of 2007. Using the Romano-Charlson method, each comorbidity was assigned a score of 1, 2, 3 or 6. The scores were then summed to produce a total score (CCI), which predicts mortality. The log-rank test and a Cox regression model were used to analyze the association between CCIp and survival, and the risk markers of survival. RESULTS:Diabetes, ulcers, congestive heart failure, chronic pulmonary disease and cerebrovascular disease were the most common comorbid conditions. The median CCI-3 was 2 (interquartile range 0-3). The overall survival rate in 1 year was 82.8%. In incident patients with a CCI-3 score of <3, the rate was 85.1%, and in patients with a CCI-3 score of ≥ 3, the rate was 76.8%. Each increase of one point in the CCI-3 score (HR = 1.69, 95% CI 1.42-2.01) and the CCIp (HR = 1.22, 95% CI 1.17-1.27) affected survival in HD patients. CONCLUSIONS:The CCI-3 and CCIp before HD initiation are valuable predictors of survival in incident patients.