Thrombocytopenia in Intensive Care Patients: A Comprehensive Analysis of Risk Factors in 314 Patients

Abstract

Objective: To define the incidence and severity of thrombocytopenia in a mixed medical-surgical population of critically ill patients and to examine factors that may be related to the development of thrombocytopenia. Design: Retrospective chart review of 314 critically ill patients requiring at least 3 days of critical care. Setting: A 17-bed combined medical-surgical intensive care unit (ICU) in a 560-bed tertiary care community hospital. Patients: Medical and surgical patients admitted to the ICU. Interventions: All medical records over the duration of the ICU stay were reviewed. All scheduled medications, including dosage and start/stop dates, were recorded. All platelet counts, placement of pulmonary artery catheters, liver function test results, and admission serum creatinine concentrations were collected. Measurement and Main Results: Thrombocytopenia (platelet count less than 200 × 109L) was observed frequently, but rarely reached a severe stage (7 patients). No single diagnostic category was significantly associated with thrombocytopenia alone, although the combination of sepsis syndrome/septic shock and respiratory failure was strongly correlated (p < 0.0001) with thrombocytopenia. Liver function abnormalities were correlated strongly with thrombocytopenia, and the majority of patients (5 of 7) with severe thrombocytopenia (less than 20 × 109/L) were found to have concurrent severe alterations in liver function test results. Pulmonary artery catheter placement and heparin exposure were associated strongly with thrombocytopenia (p < 0.0001). Drug therapies that were correlated with thrombocytopenia included heparin and vancomycin (p < 0.05). Hemodynamic instability was correlated strongly with the presence and severity of thrombocytopenia. In a stepwise linear regression model, the admission platelet count accounted for the largest proportion of the variance (43%), followed by hemodynamic instability (8%) and the requirement for inotropic agents (2%). Conclusions: Thrombocytopenia in the critically ill occurs frequently, rarely reaches severely depressed concentrations, and primarily represents a manifestation of disease processes initiated prior to admission. Hemodynamic instability and/or heparin exposure appear to be the strongest identifiable correlates with thrombocytopenia. Although these may cause infrequent isolated cases, other specific drug causes of thrombocytopenia are not responsible for the majority of cases of thrombocytopenia in the critically ill.