Association between intermittent pneumatic compression use and deep vein thrombosis in hospitalized adult medical patients: a retrospective single-center case-cohort study in Japan.

Abstract

Background: Clinical practice guidelines of venous thromboembolism prophylaxis for acutely ill hospitalized medical patients recommend intermittent pneumatic compression (IPC) devices as a mechanical thromboprophylaxis option for those at high risk for bleeding. However, supporting evidence focusing on this patient population is limited.

Methods: This case-cohort study used a university hospital-based retrospective cohort to assess the association between IPC use and lower-extremity deep vein thrombosis (DVT) development within 30 days of admission among inpatients in Japan. The parental cohort included 3876 consecutive acutely ill patients aged 15 years and older admitted to a general internal medicine department between July 2016 and July 2021. Weighted multivariable survival models and analyses that additionally accounted for time-varying covariate effects and competing risks were employed.

Results: In total, 52 patients who developed lower-extremity DVT within 30 days (9/52 [17%] with proximal lower-extremity DVT) and 144 who did not were finally included as cases and non-cases, respectively. Patients highly at risk for developing DVT, as classified by the proposed risk prediction scores, used IPC devices more frequently (p for trend range = 0.004-0.015). However, the association between pharmacologic prophylaxis frequency and the risk group demonstrated no difference. In the multivariable weighted Cox model, the adjusted point estimate indicated a lower risk of developing VTE associated with IPC use, although this result was not significant (hazard ratio [HR] = 0.23; 95% confidence interval [CI]: 0.03-1.56; p = 0.22). Regarding sensitivity, multivariable models accounted for both nonproportional hazards and deaths from non-DVT causes as competing risks showed that IPC use was significantly associated with reduced VTE development risk (HR = 0.16; 95% CI: 0.03-0.94; p = 0.04); however, the CI was wide.

Conclusions: Evidence on the association between IPC use versus nonuse and reduced VTE risk in general medical inpatients remains inconclusive. Given that conducting randomized clinical trials requires substantial costs and resources, further research should focus on multicenter, practice-based, prospective studies with larger sample sizes.