Andrea T. Obi1, Rafael Alvarez1, Christopher Pannucci1, Shipra Arya1, Andrew Nackashi2, Newaj Abdullah3, Vinita Bahl1, Lena Napolitano1, Thomas W. Wakefield1, Peter K. Henke1
1Surgery, University of Michigan, Ann Arbor, MI; 2West Virginia College of Osteopathic Medicine, Lewisburg, WV; 3Wayne State University School of Medicine, Detroit, MI.
OBJECTIVES: This study aims to retrospectively validate a VTE risk score and compare with physician entered scores in a previously unstudied high risk cohort: the critically ill.
METHODS: VTE risk factors from the ACCP guidelines via Caprini risk scoring and events were identified for 4,856 patients admitted to our Surgical Intensive Care Unit from 2007-2012, of which the majority had both clinician-entered VTE and retrospectively calculated scores. Logistic regression was used to calculate odds ratio for each level of VTE risk.
RESULTS: Overall incidence of VTE was 7.4% and increased with risk level: 0% in low risk patients, 3.5% in moderate risk patients, 5.5% in high risk patients and 8.3% in highest risk group. The difference between highest and high-risk patients was statistically significant (p=0.005). Incidence of acquired VTE accelerated in the highest risk group according to cumulative risk score, which was significant in the 7-8 (8.4%, p=0.0217) and 8+ (11.5%, p=0.0015) cohorts. The four risk levels were significantly associated with development of VTE during hospitalization. Clinician-entered scores were within 1 point of the retrospectively calculated score in 74% of cases. However, in 26% of cases, there was significant disagreement by ≥2 points between the scores. The clinician-entered score significantly underestimated risk in 80% of these cases. In patients misclassified from the highest risk group downward, the risk of VTE was higher (p=0.0007) than the physician score predicted.
CONCLUSIONS: The VTE risk assessment score is valid in the critically ill and supports the use of individualized risk assessment upon admittance to the ICU. Clinician-entered scores significantly diverge from retrospectively calculated scores in one quarter of cases and result in poor discrimination of VTE risk. Discrepancies were usually the result of under-scoring. The impact of misclassification on thromboprophylaxis prescribing regimen and incidence of VTE requires further study.
AUTHOR DISCLOSURES: N. Abdullah: Nothing to disclose; R. Alvarez: Nothing to disclose; S. Arya: Nothing to disclose; V. Bahl: Nothing to disclose; P. K. Henke: Nothing to disclose; A. Nackashi: Nothing to disclose; L. Napolitano: Nothing to disclose; A. T. Obi: Nothing to disclose; C. Pannucci: Nothing to disclose; T. W. Wakefield: Nothing to disclose.
Posted April 2013