Summary.  Background:  It is uncertain whether reactive thrombocytosis is associated with an increased risk of venous thromboembolism. This study assessed the incidence of reactive thrombocytosis, defined as platelet count ≥ 500 × 109 L−1, at intensive care unit discharge and its association with subsequent venous thromboembolism. Methods and Results:  This cohort study involved linkage of routinely collected intensive care unit, laboratory, radiology and death registry data of critically ill patients admitted to the intensive care unit between January 2009 and March 2010. The census date for survival and radiologically confirmed venous thromboembolism was 31 October 2011. Of the 1446 patients who survived to intensive care unit discharge, 139 patients had reactive thrombocytosis (9.6%, 95% confidence interval [CI] 8.2–11.2%). Twenty‐nine patients developed venous thromboembolism after discharge (2%, 95% CI 1.4–2.9%; 67 per 100 person‐years, 95% CI 45–97) and the median time to develop venous thromboembolism was 25 days (interquartile range 8–148). Reactive thrombocytosis was associated with an increased risk of subsequent venous thromboembolism (hazard ratio 5.3, 95% CI 1.7–16.4), after adjusting for other covariates. Platelet counts explained about 34% of the variability in the risk of venous thromboembolism and had a relatively linear relationship with the risk of venous thromboembolism when the platelet counts were > 400 × 109 L−1. Venous thromboembolism after intensive care unit discharge was associated with an increased risk of mortality (hazard ratio 2.0, 95% CI 1.1–3.9), after adjusting for reactive thrombocytosis. Conclusions:  Reactive thrombocytosis during the recovery phase of critical illness was associated with an increased risk of subsequent venous thromboembolism.

No abstract

Objectives: Critically ill patients with deranged conventional coagulation tests are often perceived to have an increased bleeding risk. Whether anticoagulant prophylaxis for these patients should be withheld is contentious. This study assessed the ability of using in vitro clot strength, as measured by thromboelastography, to predict thromboembolism in patients with abnormal coagulation profiles. Design: Prospective cohort study. Setting: A tertiary ICU. Patients: Two-hundred and fifteen critically ill coagulopathic patients with thrombocytopenia and/or a derangement in at least one conventional coagulation test (international normalized ratio or activated partial thromboplastin time) within 48 hours of ICU admission. Interventions: None. Measurements and Main Results: Thromboelastography was performed for all study patients, and plasma thrombotic biomarkers were measured in a nested cohort (n = 40). Of the 215 patients included, 34 patients (16%) developed subsequent thromboembolism—predominantly among those with a normal (maximum amplitude, 54–72 mm) or increased (maximum amplitude, > 72 mm) in vitro clot strength on thromboelastography (91%; area under the receiver-operating characteristic curve, 0.74; 95% CI, 0.64–0.83). The ability of the maximum amplitude to predict thromboembolism was comparable to plasma P-selectin concentrations (thromboembolism, 78.3 ng/mL vs no thromboembolism, 59.5 ng/mL; p = 0.031; area under the receiver-operating characteristic curve, 0.73; 95% CI, 0.52–0.95). In addition, patients with an increased maximum amplitude were also less likely to receive blood product transfusions within 24 hours of testing compared with those with a subnormal maximum amplitude (12.8% vs 69.2%, respectively; area under the receiver-operating characteristic curve, 0.74; 95% CI, 0.67–0.80). Conclusions: In patients with abnormal coagulation profiles, an increased in vitro clot strength on thromboelastography was associated with an increased risk of thromboembolism, and a reduced risk of requiring transfusion compared with those with a normal or reduced in vitro clot strength.

It is uncertain whether reactive thrombocytosis is associated with an increased risk of thrombosis. This prospective case-control study assessed the in vitro thrombotic tendency of patients with reactive thrombocytosis. Forty-eight patients with reactive thrombocytosis, defined by platelet count >500×10 9 /l and 55 similar, randomly selected critically ill patients who did not have reactive thrombocytosis were considered. In vitro thrombotic tendency in both groups of patients was assessed using the maximal amplitude (normal range 54 to 72 mm) and alpha angle (normal range 47 to 74°) on the thromboelastograph. The associations between reactive thrombocytosis and C-reactive protein, the coagulation profile and Sequential Organ Failure Assessment score were also evaluated. Patients with reactive thrombocytosis had an associated increased in vitro thrombotic tendency (maximal amplitude 77 vs 69 mm, mean difference 8 mm, 95% confidence interval 4.9 to 10.9, P=0.001), a higher fibrinogen concentration (7.2 vs 5.8 g/l, P=0.003), and a higher incidence of infection requiring antibiotics (50 vs 27%, P=0.025) compared to patients without thrombocytosis. Platelet count had a relatively linear relationship with the maximal amplitude and the alpha angle of the thrombo-elastograph tracing (Pearson correlation coefficient: 0.53, P=0.001). In the multivariate analysis, only reactive thrombocytosis (odds ratio 5.9, 95% confidence interval 1.3 to 27.8, P=0.025) and activated partial thromboplastin time (odds ratio 0.93 per second increment, 95% confidence interval 0.87 to 0.99, P=0.016) were significantly associated with a strong in vitro thrombotic tendency. In summary, reactive thrombocytosis was associated with infection requiring antibiotics and evidence of increased in vitro thrombotic tendency in critically ill patients.

Trauma patients are at a high risk of both bleeding and thromboembolism. This study assessed whether conventional coagulation blood tests were reliable predictors of an increased in vitro thrombotic and bleeding tendency of trauma and non-trauma patients. Conventional coagulation blood tests and thromboelastographs of 63 trauma and 63 randomly selected, critically ill non-trauma patients were compared. Increased in vitro thrombotic and bleeding tendencies were defined by a maximum amplitude >72 mm or an angle >74° on the thromboelastograph and a maximum amplitude < 54 mm or an angle <47°, respectively. In vitro thrombotic tendency was more common than bleeding tendency and this was not different between the critically ill trauma and non-trauma patients (59% versus 67% with thrombotic tendency, P=0.461; 11% versus 10% with bleeding tendency, respectively, P=0.999). Thrombocytopenia (<150 × 109 /l) and low fibrinogen concentrations (<2 g/l) were the only two factors associated with an increased in vitro bleeding tendency (both P=0.001) and thrombocytopenia was the only factor associated with a lower risk of in vitro thrombotic tendency (21% versus 75%, P=0.001). Platelet counts (Pearson's correlation coefficient [r]: 0.59, P=0.001) and fibrinogen concentrations (r 0.61, P=0.001) both had a relatively linear association with maximum amplitude of the thromboelastograph. Prolonged International Normalized Ratio (>1.5) and activated Partial Thromboplastin Time (>40 seconds) were, however, not significantly associated with an increased in vitro thrombotic or bleeding tendency. In conclusion, in vitro thrombotic tendency was more common than bleeding tendency in critically ill trauma and non-trauma patients. Platelet counts and fibrinogen concentrations were better predictors of increased in vitro thrombotic and bleeding risks than International Normalized Ratio or activated Partial Thromboplastin Time.