PerspectiveIn the Journals Plus

Venous thromboembolic events persist despite in-hospital prophylaxis

Show Citation

August 11, 2017

John A. Heit

Near-universal in-hospital prophylaxis did not lead to declines in hospital-related venous thromboembolic events, according to a study published in Blood.

More than half of the nearly 500,000 VTEs that occur in the United States each year were related to current or recent hospitalization, but nearly 50% were unrelated.

“The incidence of VTE did not change significantly from 1981 to 2010, despite widely disseminated guidelines on the prevention of VTE,” John A. Heit, MD, hematologic researcher at Mayo Clinic in Rochester, Minnesota, told HemOnc Today. “Hospitalization with or without surgery accounts for about one-half of all VTE events, and we wished to determine the effect of near-universal in-hospital VTE prophylaxis on the occurrence of hospitalization-related VTE.”

VTE — comprised of deep vein thrombosis and pulmonary embolism — has an overall age- and sex-adjusted incidence rate of 123 (95% CI, 118-127) per 100,000 persons, which has remained relatively constant for the past 30 years.

Researchers used Rochester Epidemiology Project resources to identify all Olmstead County, Minnesota residents who had incident or recurrent VTE from 2005 to 2010.

Over the 6-year period, 855 residents (median age, 62 years; 52.5% women) developed a first-ever VTE. Median age for first VTE onset was 64 years for women and 60 years for men.

DVT alone accounted for 51.4% of all events, followed by 28.2% PEs, and 20.5% DVTs and PEs. Another 281 residents reported 345 recurrent VTE events (57.4% DVT; 26.4% PE; 16.2% DVT and PE).

Researchers calculated an average annual VTE attack rates related to hospitalization — defined as events occurring in the hospital or within 92 days after discharge — of 282 (95% CI, 257-308) per 10,000 person-years. Rates unrelated to hospitalization were 8.1 (95% CI, 7.5-8.7) per 10,000.

Attack rates for PE with or without DVT were 133 (95% CI, 117-150) per 10,000 person-years for current or recently hospitalized patients, and 3.8 (95% CI, 3.4-4.2) per 10,000 persons not recently hospitalized.

Attack rates for leg DVT alone were 149 (95% CI, 130-168) per 10,000 person-years for those hospitalized and 4.3 (95% CI, 3.9-4.8) per 10,000 person-years for those not hospitalized.

Researchers applied these data to the number of U.S. hospital bed-days per year and to census data to estimate the total number of VTE events among U.S. residents.

Overall, researchers estimated 495,669 VTE events occurred each year in the United States, 52% of which were related to hospitalization.


The Olmstead County in-hospital and non-hospitalized VTE attack rates did not change significantly from 2005 to 2010.

However, during that time, researchers reported the in-hospital rate of either receiving VTE prophylaxis or having an indication that prophylaxis was unnecessary increased from between 15% and 40% in 2005 to about 90% in 2010, mostly driven by an increase in prophylaxis.

Researchers then evaluated detailed electronic VTE prophylaxis data from 2008 to 2010, during which time 15,533 unique Olmsted County residents were hospitalized for 25,617 hospitalizations.

Median duration of anticoagulant-based and/or sequential compression device prophylaxis was 70 hours and median duration of hospitalization was 3 days.

Researchers reported prophylaxis failure rates of 1.11% (95% CI, 0.77-1.55) for warfarin, 1.17% (95% CI, 0.95-1.43) for unfractionated heparin, 1.35% (95% CI, 0.97-1.83) for low-molecular-weight heparin and 0.72% (95% CI, 0.48-1.03) for sequential compression devices.

VTE rate was 0.06% (95% CI, 0.02-0.15) in patients where prophylaxis was judged not indicated.

“Achievement of near-universal in-hospital prophylaxis failed to reduce the occurrence of hospitalization-related VTE, possibly due to the very short duration of in-hospital prophylaxis,” Heit said. “About 75% of hospitalization-related VTE events occurred after hospital discharge.”

Additionally, during the 6-year duration of the study, 108 unique patients received an inferior vena cava filter as VTE prophylaxis. Ten of those patients (9.3%) developed VTE (leg DVT, n = 8; PE, n = 2).

Extending the duration of VTE prophylaxis beyond hospital discharge could reduce the occurrence of hospitalization-related VTE, but at the expense of increased bleeding, Heit said.

“About one-half of VTE events are unrelated to hospitalization,” Heit said. “Thus, even if we were able to develop a completely effective means of preventing hospitalization-related VTE, only about 50% of the burden of VTE disease in the U.S. would be prevented. We need better risk assessment tools for identifying the nonhospitalized individual (eg, active cancer patients) at high risk for VTE, and targeting VTE prophylaxis to that individual.” – by Chuck Gormley

For more information:

John A. Heit , MD, can be reached at Stabile 6-Hematology Research, May Clinic, 200 First St., SW, Rochester, MN 55905; email:

Disclosures: NHLBI and the Mayo Foundation funded this study. The researchers report no relevant financial disclosures.

itj+ Perspective

Author Name

Hospital-associated venous thromboembolism has been highlighted as an important safety problem and has been the focus of numerous national, hospital-based quality measures. The goal of most measures has been to improve VTE risk assessment and delivery of pharmacologic prophylaxis to patients at risk. Unfortunately, several studies have shown that higher performance on VTE prophylaxis measures is not associated with lower rates of VTE. Many of these studies, however, have had limitations such as focus on a narrowly defined patient population, an inability to address surveillance bias — the more you look, the more you find — or the concern that current performance measures only address one narrow window of an overall hospitalization (eg, day 1 or 2 of admission.)

Heit and colleagues overcame many of these limitations utilizing the Rochester Epidemiology Project data to estimate annual incident and recurrent VTE rates, the proportion attributed to hospitalization, and the impact of near-universal hospital-based prophylaxis at Mayo Clinic hospitals on VTE rates. Of total VTE events, 52% were related to current or recent (within prior 3 months) hospitalization. Interestingly, over a period from 2005 to 2010 rates of “appropriate prophylaxis” at Mayo Clinic hospitals increased from between 15% and 40% to 90%, yet the hospital-associated VTE rate did not change. The researchers attribute the lack of effect of VTE prophylaxis to the short duration over which it was delivered, as the median length of hospital stay was only 3 days and virtually no patients were discharged with pharmacologic prophylaxis.

Heit and colleagues are not alone in calling for better identification of high-risk patients and, thus, more targeted delivery of extended out-of-hospital prophylaxis to this cohort. But, until we have such results, what is the approach we should take? There are numerous VTE risk-stratification tools that effectively classify patients into “at-risk” or “not-at-risk” categories. But, it should be recognized that, although these tools identify populations with a higher relative risk for VTE, they have a more limited ability to identify populations that benefit from pharmacologic prophylaxis. Some estimates suggest that 500 medical patients need to be treated to prevent one VTE. 

We can assume the Mayo Clinic approach included use of such a tool and yet found no benefit in delivering higher rates of pharmacologic prophylaxis to at-risk patients. Perhaps an equally important focus should be on identifying low-risk patients and avoiding use of pharmacologic prophylaxis in this group. In studies of noncritically ill medical patients, use of a variety of risk scores showed less than 20% of patients to be at increased risk for VTE. The 80% classified as low-risk had a rate of VTE at 90 days consistently less than 0.8% to 0.9%. In hospitals using a “one-size-fits-all” approach, many of these low-risk patients receive pharmacologic prophylaxis. These patients are, thus, needlessly exposed to adverse drug events, including bleeding, discomfort associated with subcutaneous injections, and increased costs. Until we have a better approach for the high-risk patients, reducing overuse of unnecessary prophylaxis in low-risk patients and implementing effective VTE prevention strategies, such as hospital mobility enhancement programs, may be the safest strategy.


Flanders SA, et al. JAMA Intern Med. 2014;doi:10.1001/jamainternmed.2014.3384.

Greene MT, et al. Am J Med. 2016;doi:10.1016/j.amjmed.2016.03.031.


Scott A. Flanders, MD

Michigan Medicine

University of Michigan

Disclosure: Flanders reports no relevant financial disclosures.