Introduction: Patients with multiple myeloma (MM) have a high risk of venous thromboembolism (VTE) due to unique disease-related (i.e., vascular inflammation, hyperviscosity) and treatment-related factors (i.e., immunomodulatory drugs [IMiDs], high-dose dexamethasone). Thromboprophylaxis is recommended for most patients starting MM therapy. Anticoagulants are preferred over aspirin except in patients at low risk for VTE. However, compliance with prophylaxis guidelines is low. The risk of bleeding from thromboprophylaxis is not well described in these patients so an excessive perceived risk may influence clinicians' practice. We conducted a real-world analysis of bleeding complications from VTE prophylaxis in MM. Methods: The IBM MarketScan Commercial Claims Database was used to identify MM cases between 2013 and 2021. Cases required MM diagnosis from 1 inpatient or 2 outpatient claims, based on ICD-9 or ICD-10 codes, and MM-specific therapy started between 7 days before and 90 days after index diagnosis code, based on HCPCS codes and prescription records. Those with preexisting VTE, atrial fibrillation or stroke were excluded. Thromboprophylaxis was defined as at least 1 prescription for warfarin, low-molecular weight heparin (LMWH) or direct oral anticoagulants (DOAC) within 30 days of diagnosis. Major bleeding was defined as bleeding requiring hospitalization using inpatient ICD codes previously validated (Cunningham algorithm). Patients were followed from end of prophylaxis exposure window (index MM + 30 days) to first major bleeding event, disenrollment (due to death or other reasons) or December 31, 2021. Descriptive statistics were used to calculate rates of bleeding. Cox regression was used to assess the risk of bleeding from thromboprophylaxis and define risk factors associated with major bleeding. Results: We included 6,656 patients diagnosed and treated for MM, with mean age 62.7 years and 55.3% males. Although over 50% of the population would have had a high thrombotic risk based on consensus guidelines, concurrent anticoagulant prophylaxis was prescribed in only 6.6% (436). Prophylaxis agent of choice was warfarin in 23% (102), LMWH in 39% (169), and DOAC in 38% (165), respectively. There was a trend over time favoring DOAC over other agents (25% increase in DOAC use vs. other agents per year on average, p<0.01). Patients on prophylaxis had a higher rate of IMiD-based therapy (63.8% vs. 46.7%, p<0.01) and lower concomitant anti-platelet use (2.1% vs. 4.7%, p=0.01) than those not receiving prophylaxis. There were no differences in age/sex distribution, Charlson comorbidity index (CCI) or history of bleeding between groups. With a median 1.3 years follow-up, the incidence of major bleeding among MM patients receiving VTE prophylaxis was 1.4% or 7.8 bleeding events per 1,000 person-years. This was numerically but not statistically lower than the incident bleeding rate in patients not receiving prophylaxis: 1.8% or 10.1 events per 1,000 person-years. Table 1 shows risk factors for bleeding in this MM population, adjusting for age and sex, including higher age (per 10-year increase, HR 1.38), higher CCI (per SD increment, HR 1.18), history of bleeding (HR 1.54), hypertension (HR 1.87), and renal disease (HR 1.56). IMiD-based MM therapy was associated with decreased risk of bleeding (HR 0.69). Table 2 shows the estimated bleeding risk associated with prophylaxis across several models. No statistically significant association was identified between major bleeding and anticoagulant prophylaxis. Discussion: In a real-world analysis of commercial claims data, the risk of major bleeding associated with VTE prophylaxis in patients with MM undergoing cancer therapy is low and no significant increase in bleeding risk was identified for concurrent anticoagulant use. The rate of major bleeding events appears to be comparable between patients who receive VTE prophylaxis and those who do not. Nevertheless, underutilization of VTE prophylaxis is frequent. Higher CCI, bleeding history, hypertension, and renal disease were clinical factors associated bleeding risk, which may be included in models to help clinicians assess individual patient's risk and guide clinical decisions.
Read full abstract