Objective: Coagulopathy occurs frequently in critically ill patients, but its epidemiology, current treatment, and relation to patient outcome are poorly understood. We described the prevalence, risk factors, and treatment of prolongation of the prothrombin time in critically ill patients using the international normalized ratio to standardize data and explored its association with intensive care unit survival.
Design: Prospective multiple center observational cohort study.
Setting: Twenty-nine adult intensive care units in the United Kingdom.
Patients: All sequentially admitted patients over an 8-wk period.
Measurements and Main Results: Prospective daily data were collected concerning prevalence, predefined risk factors, and treatment of coagulopathy throughout intensive care unit admission. Of 1923 intensive care unit admissions, 30% developed abnormal international normalized ratio values (defined as an international normalized ratio >1.5). Most international normalized ratio abnormalities were minor and short-lived (73% of worst international normalized ratio values 1.6–2.5). Male sex, chronic liver disease, sepsis, warfarin therapy, increments in Acute Physiology and Chronic Health Evaluation II score, severity of renal and hepatic dysfunction, and red cell transfusions were all independent risk factors for international normalized ratio abnormalities (all p < .001). In all regression models, there was a strong independent association between abnormal international normalized ratio values and greater intensive care unit mortality (p < .0001), particularly when international normalized ratio increased after intensive care unit admission. Among patients with abnormal international normalized ratios, 33% received fresh-frozen plasma transfusions during their intensive care unit stay, but the pretransfusion international normalized ratio value varied widely. Fifty-one percent of fresh-frozen plasma treatments were to nonbleeding patients and 40% to nonbleeding patients whose international normalized ratio was normal or only modestly deranged (≤2.5). The dose of fresh-frozen plasma administered was highly variable (median dose 10.8 mL/kg−1 (first, third quartile 7.2, 14.4; range, 2.4–41.1 mL/kg−1).
Conclusions: Prothrombin time prolongation is prevalent in critically ill patients and is independently associated with greater intensive care unit mortality. Wide variation in fresh-frozen plasma treatment exists suggesting clinical uncertainty regarding best practice, particularly as a prophylactic treatment.