Background
Adequate risk-stratification of venous thromboembolism (VTE) in multiple myeloma (MM) is an unmet need.
Methods
This is a retrospective study of 320 newly diagnosed MM (NDMM) patients treated from 2011 to 2023. All gave their informed consent. Required minimal follow-up (FU) length was 12 months or lower if VTE or death occurred. We aimed to evaluate VTE incidence at 6 and 12 months of first line treatment (L1) initiation, identify risk predictors and assess VTE impact on overall survival (OS). Comorbidities, baseline demographic (excluding race), clinical, biological (including FISH) and imaging data, treatment exposures, response profile and OS were all analyzed. Total scores for 4 existing risk assessment models (RAM) (IMPEDE-VTE, SAVED, PRISM, IFM/DFCI-related) were also calculated from L1 start. VTE events (VTEe) were adjudicated to exclude superficial ones and restrict analysis to documented deep vein thrombosis (DVT) and pulmonary embolism (PE). Predictors validated by Cox model were weighted based on their hazard ratios (HR).
Results
The cohort comprised 147 (45.9%) women and 111 (34.7%) transplant-eligible patients. Overall median age was 70.1 (36.6-98.4) years and 205 (64.1%) subjects were >65. Two hundred and twenty-five (70.3%) received standard-of-care regimens containing immunomodulatory drugs (IMID) (Thalidomide in 98: 30.6%; Revlimid in 127: 39.7%), while 130 (40.6%) received high-dose (≥160 mg) Dexamethasone (HDD) per L1 cycle. VTE prophylaxis was omitted in 32 (10%) or consisted of NSAID in 97 (30.3%), prophylactic low molecular weight heparin (LMWHp) in 115 (36%) and direct oral anticoagulant (DOA) in 55 (17.2%), while 10 (3.1%), 9 (2.8%) and 2 (0.6%) were previously receiving curative LMWH (LMWHc), vitamin K antagonist (VKA) or a mixed regimen, respectively, for prior VTE or other indications.
In 33 (10.3%) patients, 34 VTEe occurred (31 [91.2%] within the first 6 months): 4 (11.8%) as both PE and DVT; 3 (8.8%) as PE; 27 (79.4%) as isolated proximal (12 [35.3%]) or distal (15 [44.1%]) DVT of the upper (2 [5.9%]) or lower (25 [73.5%]) limbs, including 3 (8.8%) with both proximal and distal and 3 (8.8%) close to central venous catheter.
Based on Cox univariate analysis, VTEe were not related to usual parameters such as past VTE history, IMID, HDD or EPO exposure, prophylaxis omission or modality or recent surgery. A 6-parameter RAM, named Irwazh score, was built from multivariate Cox model, including: age <55 years (HR 0.34; 95% confidence interval [CI]: 0.0-0.8; p=0.014), prior diabetes (HR 3.41; 95% CI: 1.48-7.85; p=0.004), creatinine clearance (Cr-Cl) using MDRD <40 ml/mn (HR 2.64; 95% CI: 1.26-5.51; p=0.009), CRP ≥10 mg/dl (HR 2.65; 95% CI: 1.27-5.48; p=0.009), ≥50% 1p32 deletion (HR 3.44; 95% CI: 1.17-10.09; p=0.024) and minimal response (MR) to L1 (HR 4.67; 95% CI: 1.71-12.73; p=0.003).
Area under the curve (AUC) was 0.718 (95% CI: 0.624-0.811; p<0.0001) at 12 months and 0.731 (95% CI: 0.636-0.825; p<0.0001) at 6. In comparison, AUC was 0.649 (95% CI: 0.546-0.751; p=0.004) for IMPEDE-VTE, 0.597 (95% CI: 0.505-0.689; p=0.039) for PRISM, 0.584 (95% CI: 0.486-0.682: p=0.091) for IFM/DFCI-related and 0.580 (95% CI: 0.473-0.687; p=0.141) for SAVED (Figure 1). Based on points attributed to each parameter (age <55 years: 1; prior diabetes: 9; Cr-Cl<40 ml/mn: 6; CRP≥10 mg/dl: 6; ≥50% 1p32 deletion: 9; MR: 11), patients were stratified into low (<6 points), intermediate (6-15) and high risk (≥16), with 12-month cumulative VTE incidence of 4,51% (6/133), 10,45% (16/153) and 35,29% (12/34), respectively. Of low-risk patients, 2 were receiving NSAID and 4 LMWHp; of intermediate-risk ones, 2 had no prophylaxis, 9 were under NSAID and 5 LMWHp; of high-risk ones, 2 had no prophylaxis, 5 were receiving NSAID, 2 LMWHp, 1 VKA, 1 LMWHc and 1 DOA. Using Kaplan-Meier analysis for VTE-free survival, statistical difference was found significant between low- and intermediate-risk ( p=0.043), low- and high-risk ( p<0.0001) and intermediate- and high-risk groups ( p=0.000) (Figure 2). Median total FU was 33.2 (0.2-147) months; OS assessment using Cox model did not find a significant statistical impact of VTE neither at 6 ( p=0.491) and 12 ( p=0.199) months nor during the whole FU length ( p=0.438).
Conclusion
The Irwazh score, with an unprecedented FISH feature, a statistically significant risk-stratifying power and a better AUC than existing VTE RAMs is worth validating in a larger clinical study.
Disclosures
No relevant conflicts of interest to declare.
