Receipt of anticoagulation after venous thromboembolism diagnoses during delivery hospitalizations

Obstetric venous thromboembolism (VTE) is a leading cause of maternal mortality. While hospital discharge data provide a readily accessible means of studying this relatively rare outcome, diagnosis codes are of limited validity. Prior studies have demonstrated that VTE billing codes may be subject to misclassification and false positives and overestimate obstetric VTE risk. Given the public health significance of accurately estimating obstetric VTE, the purpose of this study was to determine to what degree patients received anticoagulants after discharge from a delivery hospitalization associated with an acute VTE diagnosis as pharmacy claims may more accurately assess the incidence of obstetric VTE.

A retrospective cohort study using the MarketScan database was performed using 2008-2014 claims data. We identified women 15-54 years of age diagnosed with acute VTE during a delivery hospitalization. We determined the proportion of women with VTE that received anticoagulants within 60 days of delivery discharge. Only women with ≥60 days of pharmacy benefits after discharge were included. Receipt of low molecular weight and unfractionated heparin, warfarin, and Xa inhibitors was ascertained. Receipt of anticoagulants was analyzed individually based on diagnoses for deep vein thrombosis (DVT), pulmonary embolism (PE), or both. The Chi-square test was performed for categorical comparisons.

Of 2,664,951 delivery hospitalizations, 2112 women had a diagnosis of VTE (0.08%) including 236 women with PE alone, 1760 women with DVT alone, and 116 women with both DVT and PE. Of these women, 51.3% (95% CI 49.2-53.4%) received an anticoagulant including 49.5% of women with DVT (95% CI 47.2-51.8%), 50.0% of women with PE (95% CI 43.7-56.3%), and 81.9% of women with both DVT and PE (95% CI 73.9-87.9%).

This analysis of pharmacy claims found that estimates for the proportion of deliveries with acute VTE diagnoses that subsequently received anticoagulants was similar to chart-confirmed VTE, albeit in a large population. In addition to previous studies comparing database claims to chart review that showed that the prevalence of VTE was grossly overestimated, these findings support that the proportion of cases with VTE during delivery hospitalization may be approximately half that ascertained with billing codes.