Abstract

Introduction

The American College of Chest Physicians and government agencies recommend risk stratification and pharmacologic prophylaxis to prevent venous thromboembolism (VT) in hospitalized individuals. While risk factor assessment is often completed at admission, clinical changes during hospitalization may influence VT risk. Based on clinical observation of coincidence of C. difficile infection and VT, we sought to determine if C. diff or clinical suspicion of C. diff are risk factors for hospital acquired VT in medical inpatients.

Methods

Derivation and validation cohorts for the Medical Inpatient Venous Thrombosis Risk Assessment Score were used. Patients were admitted to the medical services at Fletcher Allen Health Care, a 500 bed teaching hospital for the University of Vermont. For derivation, all cases of hospital-acquired VT between 2002-2009 were identified using ICD-9 codes and verified by chart review; cases were matched by admission year and medical service to controls; VT risk factors and C. diff testing and Results were confirmed by chart review. For validation, hospital-acquired VT between 2009-2012 were captured by ICD-9 codes with confirmatory imaging; testing and Results for C. diff were ascertained from the microbiology lab database; VT risk factors were ascertained by ICD-9 codes, electronic problem lists, vital sign data, and lab values. Logistic regression, accounting for VT risk factors (Table), was used to determine whether 1) testing for C. diff or 2) a positive confirmation for C. diff was associated with hospital-acquired VT.

Table

Association of Clostridium difficile testing and infection with risk of hospital-acquired VT*

Derivation Cohort (299 VT cases in 64,334 admissions)Validation Cohort (120 VT cases in 20,946 admissions)
nOR* (95% CI) for VTnOR* (95% CI) for VT
C. diff test performed 4,793 2.14 (1.28-3.61) 1,708 4.25 (2.76-6.55) 
C. diff test positive 478 3.23 (1.00-10.45) 260 2.79 (1.15-6.73) 
C. diff test negative 4,314 1.95 (1.09-3.46) 1,448 3.96 (2.53-6.18) 
Derivation Cohort (299 VT cases in 64,334 admissions)Validation Cohort (120 VT cases in 20,946 admissions)
nOR* (95% CI) for VTnOR* (95% CI) for VT
C. diff test performed 4,793 2.14 (1.28-3.61) 1,708 4.25 (2.76-6.55) 
C. diff test positive 478 3.23 (1.00-10.45) 260 2.79 (1.15-6.73) 
C. diff test negative 4,314 1.95 (1.09-3.46) 1,448 3.96 (2.53-6.18) 
*

Adjusted for past history of heart failure, rheumatic or inflammatory disease; fracture within 3 months; past history of VT; cancer in prior 12 months; heart rate >100, oxygen saturation <90%, intubation, white cell count >11, or platelet count >350 on admission; use of pharmacologic VT prophylaxis or anticoagulation.

Results

In the derivation analysis 299 cases of hospital-acquired VT were identified from 64,334 medical admissions and matched to 601 controls. In the validation analysis 120 hospital-acquired VT were identified among 20,946 admissions. In the derivation cohort there were 4,793 tests for C. diff and 478 confirmed cases; and in the validation cohort there were 1,708 tests for C. diff and 260 confirmed cases. After accounting for other VT risk factors in the derivation cohort, testing for C. diff was associated with an OR of 2.14 (95% CI 1.28, 3.61) for VT, positive C. diff with an OR of 3.23 (95% CI 1.00, 10.45) for VT, and negative C. diff with an OR of 1.95 (95% CI 1.09-3.46) for VT. These associations were confirmed in the validation cohort (see Table).

Conclusions

Both testing for, and a diagnosis of C. diff were associated with hospital-acquired VT in medical inpatients. This relationship could be a surrogate of antimicrobial therapy indicating another active infection or a general marker for systemic illness, direct causation appears less likely. These data suggest that VT risk is dynamic during hospitalization and that further studies incorporating dynamic VT risk assessment are warranted in medical inpatients.

Disclosures:

No relevant conflicts of interest to declare.

Author notes

*

Asterisk with author names denotes non-ASH members.