Infection and venous thromboembolism in patients undergoing colorectal surgery: what is the relationship?

Abstract

Background: There is evidence demonstrating an association between infection and venous thromboembolism. We recently identified this association in the postoperative setting; however, the temporal relationship between infection and venous thromboembolism is not well defined

Objective: We sought to determine the temporal relationship between venous thromboembolism and postoperative infectious complications in patients undergoing colorectal surgery.

Design, setting, and patients: A retrospective cohort analysis was performed using data for patients undergoing colorectal surgery in the National Surgical Quality Improvement Project 2010 database.

Main outcome measures: The primary outcome measures were the rate and timing of venous thromboembolism and postoperative infection among patients undergoing colorectal surgery during 30 postoperative days.

Results: Of 39,831 patients who underwent colorectal surgery, the overall rate of venous thromboembolism was 2.4% (n = 948); 729 (1.8%) patients were diagnosed with deep vein thrombosis, and 307 (0.77%) patients were diagnosed with pulmonary embolism. Eighty-eight (0.22%) patients were reported as developing both deep vein thrombosis and pulmonary embolism. Following colorectal surgery, the development of a urinary tract infection, pneumonia, organ space surgical site infection, or deep surgical site infection was associated with a significantly increased risk for venous thromboembolism. The majority (52%-85%) of venous thromboembolisms in this population occurred the same day or a median of 3.5 to 8 days following the diagnosis of infection. The approximate relative risk for developing any venous thromboembolism increased each day following the development of each type of infection (range, 0.40%-1.0%) in comparison with patients not developing an infection.

Limitations: We are unable to account for differences in data collection, prophylaxis, and venous thromboembolism surveillance between hospitals in the database. Additionally, there is limited patient follow-up.

Conclusions: These findings of a temporal association between infection and venous thromboembolism suggest a potential early indicator for using certain postoperative infectious complications as clinical warning signs that a patient is more likely to develop venous thromboembolism. Further studies into best practices for prevention are warranted.

FIGURE 1

Figures demonstrating the temporal relationship between identification of postoperative DVT (A) and PE (B) and postoperative infectious complications. These graphs examine the temporal relationship between infection and VTE by using the categories infection following VTE, infection the same day as VTE, and infection prior to VTE. Ten percent to 30% of patients developing PE had a postoperative infection identified the same day. DVT = deep vein thrombosis; PE = pulmonary embolism; UTI = urinary tract infection; SSI = surgical site infection; VTE, venous thromboembolic event.

FIGURE 2

This chart demonstrates the number of days between the diagnosis of DVT and discharge. Negative numbers represent DVT diagnosis that occurred before hospital discharge, whereas positive numbers represent DVT diagnosis postdischarge. Twenty-eight percent of patients were diagnosed with DVT postdischarge. The number of days between diagnosis and discharge was not calculated for 24 patients with missing data. DVT = deep vein thrombosis

FIGURE 3

This chart demonstrates the number of days between diagnosis of PE and discharge. Negative numbers represent PE diagnosis that occurred before hospital discharge, whereas positive numbers represent PE diagnosis postdischarge. Forty-one percent of patients were diagnosed with PE postdischarge. The number of days between diagnosis and discharge was not calculated for 8 patients with missing data. PE = pulmonary embolism.