Analyzing the Rates of Heavy Menstrual Bleeding By Anticoagulant Type in Menstruating Individuals: A Large National Database Study

Background

Menstruating individuals on anticoagulation often underreport heavy menstrual bleeding (HMB), making direct comparisons between drugs challenging. While several studies have noted high HMB rates with anticoagulants, particularly directly acting oral anticoagulants (DOACs), data from large national datasets are lacking. This study aimed to assess and compare the prevalence of HMB among patients on different anticoagulants using a national electronic medical record database. This study aimed to assess and compare the prevalence of HMB among patients on different anticoagulants utilizing a national electronic medical record database.

Methods

Population-level data was collected from TriNETX, LLC, an administrative database containing comprehensive inpatient and outpatient data from 56 tertiary care hospitals in the United States. Data on females aged 15-45 years, from January 1, 2010, to December 30, 2023, who received a new prescription for an anticoagulant (identified by pharmaceutical name and RX Norm code), were analyzed. HMB was defined as the presence of one or more ICD-9 or ICD-10 codes for heavy menstrual bleeding diagnosis within three months after the anticoagulant prescription. Exclusion criteria included prior history of heavy menstrual bleeding, menopause, premature ovarian failure, pregnancy and known bleeding disorders (hemophilia, von Willebrand disease, thrombocytopenia, and qualitative platelet disorders). Data was summarized descriptively, and prevalence rates were calculated. Propensity score matching was completed based on age at the index event and race. The relative risk of HMB was also calculated and compared between the anticoagulants, with a p-value <0.05 considered significant.

Results

A total of 433,872 females on anticoagulant therapy between the ages of 15 to 45 met our inclusion criteria. This included 341,068 patients on Lovenox, 26,075 patients on Warfarin, 26,782 patients on Rivaroxaban, 38,311 patients on Apixaban and 1636 patients on Dabigatran. Patients on Betrixaban and Edoxaban were not analyzed due to their small sample size. Our results showed that prevalence of HMB was lowest among patients on Lovenox compared to warfarin, rivaroxaban, apixaban and dabigatran (p<0.05). Patients on warfarin had the highest prevalence of HMB (6.1%). Among patients on DOACs, those on rivaroxaban had the highest prevalence of HMB (5.6%). Comparative analysis between DOACs showed the following: rivaroxaban vs. apixaban (0.051 vs. 0.048; p=0.003), rivaroxaban vs. dabigatran (0.056 vs. 0.033; p=0.003), and apixaban vs. dabigatran (0.037 vs. 0.033; p=0.533). Further stratified analysis based on race, ethnicity and age is ongoing.

Conclusion:

Our study demonstrates that patients on anticoagulants have a significant risk of HMB, emphasizing the importance of careful patient history and follow-up for HMB assessment. If HMB is present, anticoagulants with lower rates of HMB should be preferred.

No relevant conflicts of interest to declare.