Thrombocytopenia is a common complication of critical illness and an independent risk factor for bleeding and death in the intensive care unit (ICU). Platelet transfusions are commonly used to improve platelet counts; however, the expected platelet increment from a transfusion in this setting has not been established. The objective of this study was to describe the frequency of platelet transfusion administration and their effect on platelet count increments in a large cohort of non-oncology critically ill adults.


We performed an analysis of a registry database, which was developed to capture clinical and laboratory data on all blood transfusions administered in 3 academic hospitals in Hamilton, Ontario, Canada. We included all patients ≥18 years who received one or more platelet transfusion during an ICU admission. Data validation was done by integrity checks with medical records and laboratory information system performed by a biostatistician. Non-transfused ICU patients were used as controls. The absolute increment in platelet count was calculated for each single platelet transfusion using the closest platelet count taken within 24 hours before the transfusion and 4-24 hours after the transfusion.


Between April 2006 and October 2012, 33,222 patients were admitted to ICU, including 29,511 (88.8%) who did not have a diagnosis of cancer. Of those, 4,502 (15.3%) received one or more platelet transfusion during any ICU admission (n=4,690); 31.9% were female and median age at the time of first admission was 69 years (IQR 59-77). Among the 25,009 non-transfused patients admitted to ICU during the same period, 38.1% were female and the median age was 65 years (IQR 52–76). Median pre-transfusion platelet count was 87 x109/L (IQR 59-131) and a single platelet transfusion resulted in a median platelet count increment of 21 x109/L (IQR 6-40) as measured 6.7 hours (IQR 5.1-9.8) after the transfusion. There were 277 (25.4%) transfusions that yielded a platelet count increment of 5 x109/L or less. ICU mortality was 562/4,690 (12.4%) for patients who received a platelet transfusion, compared with 2,251/33,033(6.8%) for patients who were not transfused during their ICU stay.


Among this large cohort of non-oncology ICU patients, platelet transfusions were commonly administered for thrombocytopenia that was generally mild. In this setting one platelet transfusion resulted in a median platelet count rise of 21 x109/L. Many transfusion episodes yielded no appreciable increase in platelet count. Further studies are needed to determine the clinical effects of platelet transfusion in this setting controlling for confounding.


Heddle:CIHR: Research Funding; Canadian Blood Services: Membership on an entity’s Board of Directors or advisory committees; Health Canada: Research Funding; Macopharma: Consultancy; ASH: Honoraria.

Author notes


Asterisk with author names denotes non-ASH members.