Abstract
Background :
Despite early ambulation, use of anticoagulation and several other strategies, postoperative venous thromboembolism (VTE) remains a major cause of morbidity and mortality. Therefore, the search for modifiable preoperative risk factors is crucial. Few reports in the literature mention an increased risk of VTE with exogenous steroid use and endogenous hypercortisolism. To date however, the direct relationship between steroid use and VTE remains unexplored.
Objective:
To assess if an association between preoperative steroid use and postoperative venous thromboembolism exists.
Methods:
Using data from the American College of Surgeons’ National Surgical Quality Improvement Program (NSQIP), which is a prospective validated outcomes registry including around 548 hospitals in the United States and around the world, we identified patients with no history of thrombotic events. We excluded patients with a history of one or more deep venous thrombosis (DVT), and patients with sepsis. After analyzing patient characteristics, we used multivariate logistic regression to assess the crude and adjusted effect of steroids on VTE, our primary outcome.
Results:
We obtained data for 1 921 901 patients, 58 667 of whom were on steroids for at least 30 days before operative intervention. Amongst those, 1241 patients had a postoperative VTE. After adjustment for different variables, VTE was higher in patients on steroids with an adjusted odds ratio of 1.54, 95% confidence interval (CI) 1.45-1.64. Secondary assessed outcomes included mortality, urinary tract infection (UTI), wound occurrences, sepsis, cardiac and respiratory adverse events. Adjusted odds ratios for the latter were 1.42 (CI 1.35-1.49), 1.40 (CI 1.30-1.50), 1.58 (CI 1.51-1.66), 1.51 (CI 1.42-1.60), 1.19(CI 1.11-1.29) and 1.302 (CI 1.301-1.303) respectively.
Discussion:
Our results suggest that surgical patients with prolonged preoperative steroid intake are at a higher risk of developing postoperative VTE. This is an important finding since preoperative steroid use is a modifiable factor. Furthermore, our data also shows that the exposure of interest is associated with a number of secondary outcomes considered including: all-cause mortality, UTI, sepsis, wound occurrences, cardiac and respiratory adverse events. Association with wound occurrences, sepsis and UTI is concordant with the well-known inhibitory effect of prolonged steroid use on wound healing and immune functions. The association with mortality, cardiac and respiratory events might in fact be due to the association between preoperative steroid use and our main outcome
Future perspectives include investigating the mechanism by which steroid use might lead to an increased risk of postoperative VTE with a particular emphasis on the coagulation cascade and potential interference with its normal physiologic function.
Abstract 2865. Table 1. Various outcomes and steroid use Steroid intake No Steroid ORcrude CI p-value ORadj CI P-value Venous Thromboembolic Event 1241 14544 2.74 2.59-2.91 <.0001 1.54 1.45-1.64 <.0001 Mortality 2525 21868 3.78 3.63-3.94 <.0001 1.42 1.35-1.49 <.0001 Cardiac 852 11665 2.33 2.18-2.50 <.0001 1.19 1.11-1.29 <.0001 Respiratory 4094 44220 3.08 2.98-3.18 <.0001 1.302 1.301-1.303 <.0001 Urinary Tract 937 10752 2.79 2.61-2.99 <.0001 1.40 1.30-1.50 <.0001 Wound 2135 29221 2.37 2.26-2.47 <.0001 1.58 1.51-1.66 <.0001 Sepsis 1515 13286 3.69 3.49-3.89 <.0001 1.51 1.42-1.60 <.0001
Abbreviations: OR: Odds ratio, CI: confidence interval, adj: adjusted.
Disclosures
No relevant conflicts of interest to declare.
Risk factors for major bleeding in the SEATTLE II trial
