Journal of Intensive Care volume 1, Article number: 9 [2013] Cite this article
22k Accesses
124 Citations
6 Altmetric
Metrics details
Abstract
Background
Sepsis is a major risk factor for the development of thrombocytopenia, but few studies have specifically evaluated prognostic importance of thrombocytopenia in patients with sepsis. We investigated the incidence, risk factors, and prognostic importance of thrombocytopenia in adult patients admitted to the intensive care unit [ICU] with sepsis.
Methods
A retrospective analysis of patients admitted with severe sepsis/septic shock from December 2007 to January 2009 to a 24-bed medical ICU was done.
Results
A total of 304 patients were included in the study. The patients' mean [±SD] age was 68.8 [±15.8] years. The majority [93.7%] had septic shock, and pneumonia was the most common infection [38.8%]. Thrombocytopenia developed in 145 patients [47.6%]: 77 [25.3%] at ICU admission and 68 [22.3%] during their hospital course. The median [IQR] duration of thrombocytopenia was 4.4 [1.9–6.9] days. Patients who developed thrombocytopenia had more episodes of major bleeding [14.4% vs. 3.7%, P < 0.01] and received more transfusions. Patients with thrombocytopenia had a higher incidence of acute kidney injury [44.1% vs. 29.5%, P < 0.01], prolonged vasopressor support [median [IQR]: 37 [17–76] vs. 23 [13–46] h, P < 0.01], and longer ICU stay [median [IQR]: 3.1 [1.6–7.8] vs. 2.1 [1.2–4.4] days, P < 0.01]. The 28-day mortality was similar between patients with and without thrombocytopenia [32.4% vs. 24.5%, P = 0.12]. However, while 15 of 86 patients [17.4%] who resolved their thrombocytopenia died, 32 of 59 patients [54.2%] whose thrombocytopenia did not resolve died [P < 0.01]. The association between non-resolution of thrombocytopenia and mortality remained significant after adjusting for age, APACHE III score and compliance with a sepsis resuscitation bundle [P < 0.01].
Conclusions
Thrombocytopenia is common in patients who are admitted to the ICU with severe sepsis and septic shock. Patients with thrombocytopenia had more episodes of major bleeding, increased incidence of acute kidney injury, and prolonged ICU stay. Non-resolution of thrombocytopenia, but not thrombocytopenia itself, was associated with increased 28-day mortality.
Background
Thrombocytopenia [platelet count < 150,000/μl] is common in critically ill patients, with an estimated incidence of 20%–40% at some point during the intensive care unit [ICU] stay [1]. Thrombocytopenia is recognized as an independent risk factor for mortality in ICU patients [2, 3]. Both the nadir platelet count and a large drop in platelet count predict a poor outcome in adult ICU patients [3]. Prolonged thrombocytopenia and absence of relative increase in the platelet count were also associated with a greater risk of mortality [4]. Many studies tried to identify consequential risk factors for the development of thrombocytopenia in the ICU. Sepsis was found to be the most common risk factor in several studies [2, 5–7]. Increased severity of illness [as suggested by high Acute Physiology, Age, and Chronic Health Evaluation [APACHE] II and Sequential Organ Failure Assessment [SOFA] scores] [2, 8] and drugs [heparin, beta lactam antibiotics, and vancomycin] [5, 6] were also suggested to be risk factors for thrombocytopenia; however, these findings have not been consistent among various studies.
To date, most studies focus on the incidence and risk factors for the development of thrombocytopenia, as well as its association with clinical outcomes in general ICU populations. The data are limited about incidence of thrombocytopenia and its association with clinical outcomes in patients with severe sepsis and septic shock. There are also limited data available about the incidence of secondary consumptive thrombocytopenia, like disseminated intravascular coagulation [DIC] and thrombotic thrombocytopenic purpura [TTP] in patients with sepsis. The incidence of DIC in severe sepsis/septic shock patients is estimated from randomized control trial [RCT] data evaluating the therapeutic role of antithrombin III or immunomodulatory drugs [9–12]. However, the data from these RCTs may be of limited use to make epidemiological inferences because these studies represent a highly selective population group. There are also inconsistencies among the various studies evaluating the risk factors for development of thrombocytopenia [5, 7]. A recent study evaluating the incidence and risk factors for thrombocytopenia in septic shock patients identified a higher SOFA score [mean SOFA score of 12.7] as an independent risk factor for the development of thrombocytopenia [8]. This finding has limited utility because one of the factors used to calculate the SOFA score is actually the platelet count.
The objective of this study was to evaluate the incidence, risk factors, and prognostic importance of thrombocytopenia in patients admitted to an ICU with severe sepsis or septic shock. We also investigated the relationship between thrombocytopenia and major bleeding episodes and transfusion requirements.
Methods
This is a retrospective cohort study conducted at the Mayo Clinic Medical Center, Rochester, MN. The study was approved by the Institutional Review Board. Informed consent was waived because of the anonymous and observational nature of this study. Adult patients [age ≥ 18 years] admitted to the 24-bed medical ICU at the St. Marys Hospital, Rochester, MN with a diagnosis of severe sepsis or septic shock between December 2007 and January 2009 were included in the study. Study patients were identified from the existing Sepsis Quality Initiative database, which contains data of patients admitted to the medical ICU with sepsis at the Mayo Medical Center. Only the first admission was included for patients who had multiple ICU admissions. Patients who have not authorized their medical records to be reviewed for research were excluded. Other exclusion criteria were history of platelet disorders [e.g., idiopathic thrombocytopenia purpura, congenital thrombocytopenia, hypersplenism], hematologic malignancies, use of chemotherapy [in the last 30 days prior to admission], mechanical heart valves, alcohol abuse, and hepatic cirrhosis.
The following data were abstracted from the patients’ electronic medical records: age, ethnicity, gender, co-morbidities, stage of sepsis, compliance with institutional sepsis resuscitation bundle, source of infection, microbial organism, development of acute lung injury [ALI] or acute respiratory distress syndrome [ARDS], duration of mechanical ventilation, vasopressor use, major bleeding episodes, transfusion requirements, duration of ICU stay, and 28-day mortality. The APACHE III score and predicted mortality were abstracted from the ICU Data Mart of our institution [13]. For patients who developed acute kidney injury, worst risk of renal dysfunction, injury to the kidney, failure of kidney function, loss of kidney function, and end-stage kidney disease [RIFLE] stage was determined [14]. The following laboratory values were collected for all patients when available: hemoglobin, creatinine, lactate, albumin, daily platelet count [up to 14 days or death whichever happened earlier], prothrombin time, activated partial thromboplastin time, D-dimer assay, fibrinogen level, peripheral blood smear evidence of hemolysis, serum lactate dehydrogenase, serum hepatic function panel, haptoglobin, and anti-platelet factor 4-heparin antibody by enzyme-linked immunosorbent assay [ELISA]. Data about the usage of medications that are commonly associated with thrombocytopenia in critically ill patients was recorded [15]. We reviewed whether the study patients received beta-lactam antibiotics, vancomycin, linezolid, trimethoprim-sulfamethoxazole, H2 receptor antagonists, or heparin products [unfractionated and low-molecular weight heparin]. One-time orders and medications prescribed ‘as needed’ were excluded.
Definitions
Thrombocytopenia was defined as having a platelet count of 14 drinks/week [25].
Statistical analysis
Data were summarized as mean [standard deviation [SD]], median [interquartile range [IQR]] or percentages. We used unpaired Student’s t test to compare continuous variables with normal distribution and Mann-Whitney U test for skewed distribution. For comparison of categorical variables, we used chi-square test if the number of elements in each cell were 5 or higher and Fisher’s exact test, otherwise. Multivariate logistic regression analysis was performed to determine independent risk factors associated with the development of thrombocytopenia. Univariate analysis was done to identify the candidate variables for multiple logistic regression analysis. Colinear variable with a low P value [≤0.1] or variables that were thought to have strong biologic associations were included in the model. To determine the impact of thrombocytopenia on mortality, we performed a multivariate logistic regression analysis using the 28-day mortality as the dependent factor. Variables considered for the multivariate modeling included age, APACHE III score, and compliance with sepsis resuscitation bundle. The performance of the multiple logistic regression models were assessed by the area under the receiver operating characteristic curve statistics. When appropriate, the odds ratio [OR] and 95% confidence intervals [CI] were calculated. A P value of