A study from Vanderbilt University Medical Center details the apparent risks associated with bleeding and venous thromboembolism following noncardiac surgery.
Robert Freundlich, MD
New research is seeking to draw a firm conclusion on a common question: does postoperative bleeding or postoperative venous thromboembolism pose a greater risk to patient mortality?
An analysis of a national database from the American College of Surgeon’s, results indicate postoperative bleeding is a more consistent source of mortality risk among patients undergoing noncardiac surgery.
“We wanted to do a head-to-head comparison of those two in a large surgical population and look at the tradeoff in terms of mortality. We found, in every single year, a consistent signal for bleeding,” said lead investigator Robert Freundlich, MD, assistant professor of Anesthesiology and Biomedical Informatics at Vanderbilt University Medical Center, in a statement. “Bleeding had very, very high attributable mortality, meaning death that you could say statistically was related to the occurrence of bleeding. Whereas, in every single year, we didn’t see that same signal for blood clots.”
With bleeding and thromboembolic events representing 2 of the greatest postoperative concerns for patients and clinicians, Freundlich and a team of colleagues sought to determine the risk and attributable mortality related to both of these event types among patients undergoing noncardiac surgery in the US. Using the American College of Surgeons’ National Surgical Quality Improvement Program (ACS-NSQIP) database, which records data related to bleeding events for first 72 hours after surgery and records data related to blood clots are tracked for up to 30 days, investigators designed their study using data from surgeries between 2006-2017.
Information from the ACS-NSQIP database also included data related to preoperative risk factors, intraoperative variables, 30-day postoperative mortality and morbidity outcomes, and more than 150 variables. For the purpose of the analysis, investigators defined bleeding and venous thromboembolic events according to the American College of Surgeons’ definitions.
Of note, ACS-NSQIP changed the definition of bleeding in 2010, reducing the number of transfused units required and adding transfusion during the intraoperative period to the definition. The main goal of the investigators’ analyses was to calculate the odds ratios and attributable mortality for both of the aforementioned event types.
During the study period, a total of 6,637,635 were identified for inclusion—ranging from 152,287 patients from 2006 to 1,028,731 from 2017. Due to the change in definition, the rate of bleeding prior to January 2010 was 0.4%-0.6% and this number grew to 4.5%-8.4% following the change in definition. In comparison, the rate of venous thromboembolic events ranged from 0.8%-1.0% during the years examined.
After adjusting for confounders, results of the analysis indicated bleeding had a more significant association with mortality than venous thromboembolic events in all 12 years. Prior to 2010, the odds ratios for mortality ranged between 1.62-3.44 and, following the definition change, ranged from 1.19-1.42 (all P values <.001). For venous thromboembolic events, the odds ratio for mortality ranged from 1.21-1.49 beginning after 2012 (all P values <.001). According to study results, there was no evidence VTE was associated with mortality before 2012 except in 2008, when the odds ratio was 1.32 (95% CI, 1.07-1.64; P=.01).
When examining attributable mortality, investigators found attributable mortality from bleeding ranged from 6-17 per 100,000 patients prior to the 2010 definition change. Following the definition change, the attributable mortality ranged from 10-52 per 100,000. For venous thromboembolic events, attributable mortality peaked in 2008 reaching a rate of 6 per 100,000 patients. Following 2008, the rate of attributable mortality ranged from 2-5 per 100,000 patients.
“I think people are doing a really good job of addressing blood clots after surgery, making sure that even when blood clots do happen they aren’t bad enough to lead to a patient dying,” Freundlich said. “And I would like to think that current medical therapy is probably working well, which is what we see in the database. Really just very, very undetectable mortality.”
This study, “The Attributable Mortality of Postoperative Bleeding Exceeds the Attributable Mortality of Postoperative Venous Thromboembolism,” was published in Anesthesia & Analgesia.