Residual β-Cell Function Predicts Clinical Response After Autologous Hematopoietic Stem Cell Transplantation

New strategies of autologous hematopoietic stem cell transplantation (auto-HSCT) have gained much interest for the treatment of type 1 diabetes mellitus. However, assessing the clinical response and residual beta-cell function still has limitations. The aim of the study was to select the optimal quantitative index to assess pre-existing beta-cell function and to explore its predictive function for clinical response after auto-HSCT therapy. In this study, all of the patients who had undergone auto-HSCT were clustered into a responder group (Delta beta-score > 0) and a nonresponder group (Delta beta-score <= 0). We compared their quantitative metabolic indexes at baseline and performed receiver-operating characteristic (ROC) analysis to analyze the correlations between the indexes and clinical response. Kaplan-Meier analysis was conducted to compare the cumulative response durations in each quartile of the selected indexes. In an average of 15.13 +/- 6.15 months of follow-up, 44 of 112 patients achieved a clinical response. The responder group had lower levels of fasting plasma glucose and quantitative insulin sensitivity check index (QUICKI) but higher levels of fasting C-peptide, fasting insulin, and homeostasis model assessments for insulin resistance (HOMA-IR). ROC analysis showed that HOMA-IR had the largest area under the curve (0.756), which was similar to that of QUICKI. Kaplan-Meier analysis further confirmed that the third quartile (1.3371-1.7018) of HOMA-IR or the second quartile (0.3523-0.3657) of QUICKI was preferential for a prolonged response. In conclusion, HOMA-IR and QUICKI could be optimal measurements for beta-cell reserves, and they were predictive for the clinical response after auto-HSCT.