Comorbidity Index in the Context of Cord Blood Transplantation: An Accurate Predictor of Treatment Related Mortality?

Background: The hematopoietic cell transplantation comorbidity index (HCT-CI) has been demonstrated to be a useful tool for evaluating the risk of treatment related mortality (TRM) in HLA-matched sibling (MRD) and matched unrelated donor (URD) bone marrow (BM) and peripheral blood stem cell transplant (PBSCT) patients (Sorror et al 2005). In recent years, umbilical cord blood transplant (CBT) has emerged as a common treatment for hematologic malignancy patients without a suitably matched conventional BM or PBSCT donor and patients who require a rapidly available donor. However, the initial analysis evaluating the use of the HCT-CI did not include CBT patients. Therefore, the value of this risk assessment tool as it applies to risk of TRM in the CBT setting is not known. Reduced intensity conditioning (RIC) CBT is often reserved for older adults and those with high comorbidities due to concern for excessive TRM with myeloablative regimens. In this retrospective analysis, we sought to evaluate the association of HCT-CI scores with TRM in patients undergoing RIC UCBT.

Methods: Charts were reviewed and HCT-CI scores tabulated for 143 patients (80 male, 63 female), median age 54 years (5-73), who were sequentially enrolled on two multicenter CBT studies coordinated by the Fred Hutchinson Cancer Research Center (FHCRC) between 2006 - 2013. Patients were transplanted for high risk (n = 68) and standard risk (n = 75) AML/MDS (n =91), ALL (n = 22), Non-Hodgkin Lymphomas (n = 12), CML (n = 3), Hodgkin Lymphoma (n =3), and other hematologic malignancies (n = 12). 77 patients received RIC with cyclophosphamide 50mg/kg, fludarabine 200mg/m², and either 200 or 300 cGy TBI, while the remaining 66 received a “midi” regimen of treosulfan 42g/m², fludarabine 150-200mg/m², and 200 cGy TBI. All patients received cyclosporine and mycophenolate mofeteil for GVHD prophylaxis. Only seven patients received a single CBT, and the remaining 136 received a double CBT. The incidence of TRM was estimated by using cumulative incidence methods, with relapse considered a competing risk.

Results: Median HCT-CI for the whole group (n =143) was 2 (range 0 - 8) with distribution as follows: (HCT-CI = 0 (n = 25), HCT-CI = 1 (n = 15), HCT-CI = 2 (n = 34), HCT-CI = 3 (n = 34), HCT-CI = 4 (n = 16), HCT-CI = 5 (n = 9), HCT-CI = 6 (n = 8), HCT-CI = 7 (n = 2), and HCT-CI = 8 (n = 1). TRM for the group as a whole was 7% (95% CI: 3-12) at Day 100 and 32% (95% CI: 24-42) at 1 year. When divided into 2 groups by HCT-CI score ≤ 3 (n = 107) and >3 (n = 36), demographics were similar with the exception of younger age in the HCT-CI ≤ 3 group (52 vs. 57 years; p = 0.02) and more CMV seropositive patients in the HCT-CI ≤ 3 group (73 vs. 50%; p = 0.01). The incidence of TRM at 1-year was 25% (CI 95%: 16-35) vs. 58% (CI 95%: 35-75) in the HCT-CI ≤ 3 vs > 3 groups respectively (p=0.001). The hazard of TRM in the HCT-CI > 3 group was higher than that in the HCT-CI ≤ 3 group after adjusting for age, CMV serostatus, gender, disease risk, and conditioning regimen (HR=2.31; 95% CI 1.13-4.7; p=0.021). At Day 100 there was a trend for higher incidence of TRM in patients with HCT-CI > 3 = 14% (CI 95%: 5-27) vs HCT-CI ≤ 3 = 5% (CI 95%: 2-9) (p=0.08). Successful engraftment occurred in 92% of patients (n = 131) at a median of 19 (6-89) days. Of the 126 patients evaluable, 17 developed grade III-IV acute GVHD. OS for the whole group at 1 year was 55% (CI 95%: 45-63): patients with HCT-CI ≤ 3 = 62% (52-72) vs. > 3 = 32% (16-48) (p = 0.0016).

Conclusions: This analysis demonstrates that higher HCT-CI score is associated with increased risk of TRM at 1 year post-transplant in a series of patients undergoing RIC UCBT. Due to the significant risk of TRM in patients with HCT-CI > 3 compared to those with HCT-CI ≤ 3, patients with a HCT-CI score >3 should be counseled regarding their heightened risk of TRM prior to undergoing CBT.