Centers that perform stem cell transplants have patient selection criteria for treatment, typically including patient performance status, acceptable cardiopulmonary function, normal renal function, absence of active infection, and underlying disease likely to respond to treatment. Despite these selection criteria, transplant-related morbidity and mortality are high, particularly in allogeneic transplants. We sought to find a method to better predict outcome after SCT. The Charlson Comorbidity Index (CCI) is a tool to assess comorbidities and outcome in a varitety of diseases, but its utility in SCT is unclear. This retrospective study used chart review from 187 procedures (109 autologous and 78 allogeneic; 43 myeloablative and 35 non-myeloablative) performed between 2002 and 2004 to assign CCI scores. Transplant outcomes, including 100-day and 1-year survival and transplant related toxicity, were grouped according to CCI and analyzed using regression analysis, T-test, Fisher Exact and ANOVA. Median age at transplant was 55 years for autologous and 48 years for allogeneic (41 years for myeloablative and 55 years for non-myeloablative). The CCI was not predictive of survival or toxicity in autologous transplant patients. In allogeneic transplants, the CCI was predictive of only long-term survival when a score of zero was compared to all other scores, but was not clinically useful since 76% of patients had a CCI score of zero and values over one had similar outcomes. We thus sought a better tool to predict outcomes and modified the CCI using clinical parameters that were anticipated to impact outcome. These included pre-transplant lab values (AST, ALT, creatinine and bilirubin), scores for prior organ dysfunction (CNS, cardiovascular, diabetes, lung, liver and kidney), pulmonary function, alcohol use and smoking history. This new SCT Comorbidity Index, SCI, was used to evaluate outcomes using the same methodology. The modification of the original CCI expanded the range of comorbidities to 27 from the 19 medical conditions assessed in the original CCI. By this expansion, a new maximum score of 38 was created with an individual patient high SCI score of 8. The autologous and allogeneic SCT populations had comparable comorbidity scores, while non-myeloablative patients had higher comorbidity scores, on average, than myeloablative. For short term 100-day survival, the SCI proved to be a better predictor than the original CCI only in autologous SCTs (99% survival with SCI scores of 0 or 1 vs. 83% survival with other SCI scores p=0.004). Long-term survival 100 days to 1 year was opposite of anticipated in autologous SCTs (67% survival with SCI score of 0 vs. 91% survival with other SCI scores p=0.007) but appeared to be predictive in allogenic transplants (87% survival with SCI score 0–2 vs. 50% for higher SCI scores p=.002). The SCI did not predict long-term survival in the non-myeloablative allogeneic SCT subset of patients. In addition, neither index predicted overall toxicity (d0 to d30), and neither predicted grade III–IV toxicity (d0 to d30) in any patient population. Thus neither the Charlson Comorbidity Index nor a modified index was found to be a helpful in the consistent prediction of stem cell transplant outcomes. This likely reflects the selection criteria used for determining candidacy that limits the degree of comorbidity.

Author notes

Corresponding author