Clinical background
Patients with multiple myeloma have up to a ninefold increased risk of venous thromboembolism (VTE) compared with the general population, and VTE is a leading cause of death in this group [1]. The risk is amplified by disease-specific treatment, particularly immunomodulatory drugs (IMiDs) and high-dose dexamethasone, but also by patient-related factors such as a high BMI, previous VTE and a central venous catheter. Despite thromboprophylaxis, the VTE incidence remains above 10% in many cohorts [3].
Before IMPEDE VTE there was no validated quantitative risk model for myeloma. The IMWG and NCCN published qualitative guidance based on the presence of individual risk factors, but these identified only about 55% of the patients who developed VTE as being at high risk [1]. In the derivation cohort, the IMWG/NCCN guidance achieved a c-statistic of 0.55, barely better than chance [1]. IMPEDE VTE was developed to replace this qualitative assessment with a points-based model that better separates low- from high-risk patients and can thereby guide who should receive pharmacological thromboprophylaxis.
Calculating IMPEDE VTE
The score is the sum of ten variables, some contributing positive and some negative points:
The variables and their point values were derived from a multivariable competing-risk model (Fine and Gray) in which the parameter estimates were multiplied by 5 and rounded to the nearest whole number [1]. The variables included are:
- Immunomodulatory drug (thalidomide, lenalidomide or pomalidomide): +4
- BMI ≥25 kg/m²: +1
- Recent pathological pelvic, hip or femoral fracture: +4
- Erythropoiesis-stimulating agent: +1
- Doxorubicin as part of current treatment: +3
- Dexamethasone dose: low dose (≤160 mg/month) +2, high dose (>160 mg/month) +4
- Asian or Pacific Islander ethnicity: −3
- Previous VTE: +5
- Tunnelled catheter or central venous catheter: +2
- Current thromboprophylaxis: prophylactic acetylsalicylic acid −3, therapeutic anticoagulation (LMWH, warfarin or a DOAC) −4
The derivation cohort consisted of 4,446 patients with newly diagnosed multiple myeloma who started chemotherapy within 6 months of diagnosis, identified in the Veterans Administration Central Cancer Registry (VACCR) between 1999 and 2014 [1]. Follow-up was 180 days from the start of chemotherapy. VTE occurred in 259 patients (5.8%). The cohort was dominated by men, reflecting the composition of the VA population. The modelling accounted for the competing risk of death without VTE and used time-dependent variables for intermittent treatments to minimise immortal time bias [1].
External validation was performed in the SEER-Medicare cohort of 4,256 patients diagnosed between 2007 and 2013, all aged ≥65 years and continuously insured through Medicare parts A, B and D [1]. Here VTE occurred in 221 patients (5.2%).
Interpretation in practice
The risk classification used by the calculator is:
| Score | Risk category | Clinical action |
|---|---|---|
| ≤3 | Low | Clinical surveillance; pharmacological prophylaxis usually not justified unless other factors argue for it |
| 4–7 | Intermediate | Individual judgement; prophylaxis should be considered, particularly during IMiD treatment |
| ≥8 | High | Pharmacological thromboprophylaxis strongly indicated |
A negative total score is possible and reflects a patient in whom protective factors, above all established therapeutic anticoagulation, outweigh the risk factors. This is not an error in the calculation but a consequence of ongoing prophylaxis being modelled as a variable that lowers the predicted risk [1].
The clinical value lies mainly in separating low-risk patients, in whom prophylaxis can be avoided, from high-risk patients, in whom it should be started. In the derivation cohort, VTE risk rose significantly with increasing score (hazard ratio 1.20 per point, p < 0.0001) [1].
Validation and performance
In the derivation cohort, IMPEDE VTE achieved a c-statistic of 0.66, unchanged whether point values or raw beta coefficients were used [1]. Bootstrap validation with 500 resamples gave a mean c-statistic of 0.66 (95% CI 0.63–0.70) [1]. Calibration was satisfactory by the Hosmer-Lemeshow test (p = 0.41) [1].
In the external SEER-Medicare cohort the c-statistic fell marginally to 0.64 [1]. By comparison, the IMWG/NCCN guidance achieved a c-statistic of 0.55 in the same cohort, confirming that the quantitative model is superior to the qualitative approach [1].
A French external validation was published by Chalayer et al. in a cohort of newly diagnosed myeloma patients treated with IMiDs, but the full text was not available for detailed review [2].
A Turkish multicentre study of 455 myeloma patients (2019–2023) reported a c-statistic of 0.618 for IMPEDE VTE, compared with 0.633 for SAVED and 0.701 for IMPEDED VTE, a variant in which the D-dimer has been added as a biomarker [4]. The VTE incidence in this cohort was 10.7%, considerably higher than in the US cohorts, reflecting a different treatment mix and a higher proportion of patients on IMiDs. The study included no patients of Asian or Pacific Islander ethnicity, so the negative point variable could not be evaluated [4].
A systematic review and meta-analysis from 2026 included 14 studies and seven risk models for VTE in myeloma [5]. The pooled AUC for all the models was below 0.70, with IMPEDE VTE and IMPEDED VTE performing best. The authors noted that all included studies were judged to be at high risk of bias by PROBAST, above all in the outcome and analysis domains. The pooled VTE incidence was 7.9% (95% CI 6.2–10.1%) [5].
Limitations
IMPEDE VTE was derived and validated exclusively in US cohorts: the VA (predominantly male military veterans) and SEER-Medicare (people aged ≥65 years). Its generalisability to younger patients, to women and to non-US populations has not been well tested. The Turkish validation suggests lower discrimination in a different population mix [4].
The model assumes that the patient has recently started chemotherapy for multiple myeloma. It does not apply to patients with relapsed or refractory disease in later lines, where the treatment landscape has changed with anti-CD38 antibodies, bispecific antibodies and CAR-T cells and where the risk profile may differ. Nor does it apply to patients undergoing stem cell transplantation, since these were excluded from the derivation [1].
The variable Asian or Pacific Islander ethnicity is assigned negative points on the basis of the observation in the VA cohort that this group had a lower VTE incidence. The mechanism is unclear and the finding has not been reproduced in populations lacking this ethnic group [4]. In populations where this group is uncommon, the variable is rarely relevant and may cause the score to underestimate the risk if it is left unanswered.
The dexamethasone dose threshold of 160 mg/month reflects treatment patterns from 1999 to 2014. With current standard regimens, in which lower dexamethasone doses are often used particularly in older and frail patients, this variable may be difficult to establish retrospectively.
DOACs (direct oral anticoagulants) were not included in the derivation because they were not available during the study period. The calculator nonetheless lists them as therapeutic anticoagulation scoring −4, but this rests on extrapolation from LMWH and warfarin, not on empirical data from the derivation cohort.
All the available risk models for VTE in myeloma, IMPEDE VTE included, have a pooled AUC below 0.70, indicating moderate discrimination [5]. No prospective randomised trial has yet shown that treatment decisions based on IMPEDE VTE lead to a lower VTE incidence without an increase in bleeding [3]. The score should therefore be seen as an aid to clinical judgement, not as a definitive answer.
References
- Sanfilippo KM, Luo S, Wang TF, et al. Predicting venous thromboembolism in multiple myeloma: development and validation of the IMPEDE VTE score. Am J Hematol. 2019;94(11):1176-1184. PMID: 31379000
- Chalayer E, Teste A, Guyotat D, et al. Predicting the risk of venous thromboembolism in newly diagnosed myeloma with immunomodulatory drugs: External validation of the IMPEDE VTE score. Am J Hematol. 2020;95(1):E18-E20. PMID: 31617240
- Covut F, Sanfilippo KM. Mitigating the risk of venous thromboembolism in patients with multiple myeloma receiving immunomodulatory-based therapy. Hematology Am Soc Hematol Educ Program. 2022;2022(1):363-367. PMID: 36485142
- Gursoy V, Baysal M, Sadri S, et al. Comparative analysis and validation of the IMPEDED VTE, IMPEDE VTE, and SAVED risk models in predicting venous thromboembolism in multiple myeloma patients: A retrospective study in Türkiye. Diagnostics (Basel). 2025;15(5):633. PMID: 40075881
- Yang J, Pu Y, Li Y, et al. Risk prediction models for venous thromboembolism among patients with multiple myeloma: a systematic review and meta-analysis. PeerJ. 2026;14:e21322. PMID: 42222482
