Sepsis is a major cause of morbidity and mortality worldwide. With the advance of medical care, the mortality of sepsis has decreased in the past decades. Many treatments and diagnostic tools still lack supporting evidence. We conducted a retrospective population-based cohort study with propensity score matched subcohorts based on a prospectively collected national longitudinal health insurance database in Taiwan. Severe sepsis-associated hospital admissions from 2000 to 2011 based on International Classification of Diseases, Ninth Revision, Clinical Modification codes of infections and acute organ dysfunction were identified. To compare the effectiveness of treatment and diagnostic tool, propensity scores were generated to match the comparable control groups. During the 12-year period, 33 375 patients and 50 465 hospitalizations of severe sepsis were identified. The age-standardized 28-day in-hospital mortality decreased significantly from 21% in 2008 to 15% in 2011 with increasingly implemented treatment and diagnostic tool. After propensity score matching, procalcitonin (odds ratio [OR]: 0.70, 95% confidence interval [95% CI]: 0.61-0.81) and lactate testing (OR: 0.90, 95% CI: 0.84-0.97, respectively), transfusion of packed red blood cell (OR: 0.60, 95% CI: 0.52-0.69), albumin (OR: 0.72, 95% CI: 0.55-0.93), balanced crystalloid (OR: 0.29, 95% CI: 0.20-0.41), and use of dopamine (OR: 0.44, 95% CI: 0.39-0.49) were found to be significantly associated with lower mortality rate. However, inconsistent findings need to be further validated.
Keywords: albumin; balanced crystalloid; blood transfusion; cohort; lactate; mortality; procalcitonin; propensity score matching; sepsis; vasoactive agents.