New‐onset posttransplant diabetes mellitus after haploidentical hematopoietic cell transplantation with posttransplant cyclophosphamide

Abstract Haploidentical hematopoietic cell transplantation (haplo‐HCT) with posttransplant cyclophosphamide (PTCY) is utilized for patients with hematological disorders but without conventional donors. The effects of new‐onset posttransplant diabetes mellitus (PTDM) following haplo‐HCT are unknown. We examined PTDM incidence and outcomes after haplo‐HCT with PTCY. Patients without diabetes receiving haplo‐HCT (n = 64) were analyzed for PTDM diagnosis (defined as blood glucose ≥ 200 mg/dL). By day 100, 14 (22%) patients developed PTDM (median, 18 days). Hyperglycemia (blood glucose ≥ 200 mg/dL) preceded corticosteroids in 11 (79%) individuals. PTDM patients had increased death/relapse (P = .029). PTDM occurs frequently, precedes corticosteroids, and leads to inferior outcomes following haplo‐HCT. PTDM prophylaxis/treatment may improve HCT survival.

By increasing or decreasing immune activation, HLA mismatch or PTCY could either promote or suppress PTDM development after haplo-HCT, respectively. As previously discussed, PTDM in HLA-matched HCTs has a negative effect on survival, though a lack of data exists for the distinct haplo-HCT population. Therefore, we examined the incidence, risk factors, and outcomes for PTDM after haplo-HCT with PTCY. We hypothesized that hyperglycemia would occur despite PTCY and that PTDM would affect survival negatively following haplo-HCT.

METHODS
A single-center, retrospective study was conducted at Vanderbilt University Medical Center (VUMC) to assess the incidence, clinical risk fac-  [11]. The statistical analysis plan is discussed in the Supporting Information Methods.

DISCUSSION
PTDM developed in about one quarter of patients undergoing haplo-HCT and increased the risk of death or relapse four times. These data are novel since previous PTDM research focused on conventional donor HCT and the incidence of hyperglycemia after PTCY is not well studied. As in other research, hyperglycemia generally preceded the initiation of corticosteroids [12]. This is important because it negates the commonly held belief that PTDM is merely a side effect of immunosuppression. Although it is possible that corticosteroids exacerbate clinical hyperglycemia, our data indicate that glucose metabolism is already altered by the time these medications are started and other causes for PTDM need to be identified.
From previous trials, the incidence of PTDM has ranged from about 20 to 60% in the nonhaplo-HCT population [1][2][3][4]. The 22% development of PTDM in this study falls within that range, but on the lower side. The lower incidence after haplo-HCT could be due to the lack of routine fasting labs. In this retrospective study, episodes of fasting hyperglycemia could have been missed leading to an underestimate of the true rate of PTDM. Another hypothesis to account for this lower incidence may be the conditioning regimen and PTCY used in the haplo-HCT population. Cyclophosphamide selectively depletes alloreactive T cells sparing T cells responsible for immune reconstitution and regulation [9]. Post-HCT cyclophosphamide given on days 3 and 4 after haplo-HCT has been shown to be an effective method for GVHD prophylaxis.
haplo-HCT. Although further research is required to validate, our proposed model of inflammation-induced PTDM is outlined in Figure S2.
Older age, nonwhite ethnicity, corticosteroids, parenteral nutrition, unrelated donor, ablative conditioning, impaired fasting glucose, and elevated C-peptide levels have been reported as risk factors for PTDM following HCT with conventional HLA-matched donors. [3,4] In our study, ablative chemotherapy and steroid exposure were more common in the PTDM group, however this did not reach statistical significance. We were unable to definitively identify any clinical predictors of PTDM after haplo-HCT. It is possible that previously reported risk factors for PTDM are not as relevant following PTCY. As previously mentioned, our research was limited by its retrospective nature, lack of F I G U R E 1 Haplo-identical hematopoietic cell transplant (haplo-HCT) outcomes stratified for development of posttransplant diabetes mellitus (PTDM). Overall survival and cumulative incidence of acute graft-versus-host-disease (GVHD) for the entire cohort (panels A and B, respectively). Disease-free survival and cumulative incidence of relapse for patients with malignancy only (panels C and D, respectively) routine fasting blood glucose measurements, and modest cohort size that may have influenced the results. It is likely that we did not have enough statistical power to detect differences between the PTDM and non-PTDM groups. In addition, the multivariate regression model may have been over fit for the cohort size.
Diabetes is known to complicate cancer care. The HCT comorbidity index has shown that pre-existing diabetes mellitus increases mortality after HCT [13]. In this context, we were not surprised by the inferior outcomes for patients developing new-onset PTDM after haplo-HCT.
Interestingly, we also noted a trend for excess relapse in PTDM  (Table S2) [1,2,5,6,12,16]. With the clinical importance now established, it is time to shift focus from examining the incidence of PTDM, and begin to study the mechanistic causes for hyperglycemia and potential therapeutic options to prevent or treat metabolic complications after HCT. Understanding the actionable targets that lead to PTDM remains an unmet need and should be actively investigated to optimize patient outcomes.

ACKNOWLEDGMENTS
This work was supported by the National Institutes of Health/ National Heart, Lung, and Blood Institute Grant R01-HL141943 (to B.G.E).