Background: The HCT-CI is a recently developed comorbidity score which has been adapted to hematopoietic stem cell transplantation, and identified 3 risk groups with increased non-relapse mortality (NRM) and lower overall survival (OS) (
Methods: The analysis included all patients undergoing MUD transplant from 1996–2006 who received GVHD prophylaxis with cyclosporine 2 mg/kg iv BID from day −2, methotrexate 15 mg/m2 iv on day +1 and 10 mg/m2 iv on days +3 and +6, and methylprednisolone 0.25 mg/kg iv BID beginning on day +7 and tapering from day +28. Patients were stratified by disease risk per CIBMTR classification. The comorbidities were obtained by retrospective chart review and scored according to the HCT-CI score.
Results: 150 patients (median age 40) received the 3 drug-regimen, including 38% with low-, 34% with intermediate- and 28% with high-risk disease. Diagnoses included acute leukemia in 50%, MDS in 12%, CML in 15%, lymphoma in 18%, and multiple myeloma in 3.0%. Conditioning regimens included Cy-TBI in 64% and Bu-Cy in 21%. Source of stem cells was PBSC in 47.3% and marrow in 50.7%. HCT-CI scores of 0, 1–2 or ≥3 were found in 17%, 30% and 53% of patients evaluated. The majority of comorbidities were pulmonary (72%). With a median follow-up of 46 weeks, day 100 and 5-year OS were 82.7 and 33%, with a 23% and 50.4% cumulative incidence of NRM. Five year relapse-related mortality was 15.8%. Although higher HCT-CI scores were associated with increased NRM and decreased OS, no statistically significant differences were detected when using the published HCT-CI grouping of 0, 1–2 and ≥3. Unadjusted hazard ratio (HR) for inferior survival were 0.9 (CI 0.47–1.85, P=.79) and 1.65 (CI 0.885–3.090, P=.11) for scores 1–2 and ≥3, respectively. We then determined an alternate prognostic model based on 2 groups. Statistical modeling separated patients with a score of 0–3 (n=97, 64%) and ≥4 (n=53, 35.6%), with a 3 month and 5 year OS of 84% and 45% versus 52% and 10%, respectively (P<.0001). Cumulative incidence of day 100 and 5-year NRM was 16% and 38% versus 43% and 73%, respectively. Unadjusted HR for inferior survival was 2.77 (CI 1.816–4.225, P<.0001) for a score of ≥4. By multivariate analysis, only the HCT-CI score (P<.0001) and the disease risk per CIBMTR (P=.0058) were predictive of OS and NRM, but not age, CMV positivity, sex- or HLA-mismatch, or regimen.
Conclusions: While our data confirm that the HCT-CI score is predictive of NRM and OS in a high-risk MUD transplant cohort, we were unable to detect statistically significant differences between the 3 risk groups defined in the original score. A modified 2-group scoring system readily stratified the patient population into low-risk and high-risk risk groups with scores of 0–3 and ≥4, respectively, that was predictive of OS and NRM. This simplified, 2-tiered scoring system will have utility in clinical decision-making and in defining patient populations eligible for clinical trials. Additional single and multi-institutional analyses will ultimately determine the optimal applications of the HCT-CI score.
Disclosure: No relevant conflicts of interest to declare.