Characteristics of chemotherapy-induced diabetes mellitus in acute lymphoblastic leukemia patients

Abstract

Acute lymphocytic leukemia (ALL) is one of the most common malignancies, especially in young people. Combination chemotherapy for ALL typically includes corticosteroids (Kantarjian et al., 2000). Hyperglycemia is a well-recognized complication of corticosteroids, and chemotherapy-induced diabetes (CID) is not uncommon (27.5%–37.0%) during the treatment of ALL (Hsu et al., 2002; Weiser et al., 2004; Alves et al., 2007). Besides the effect of corticosteroids, potential factors triggering hyperglycemia in ALL also include direct infiltration of the pancreas by leukemia cells and β cell dysfunction induced by chemotherapeutic agents such as L-asparagine (Mohn et al., 2004).

Acute lymphocytic leukemia (ALL) is one of the most common malignancies, especially in young people. Combination chemotherapy for ALL typically includes corticosteroids (Kantarjian et al., 2000). Hyperglycemia is a well-recognized complication of corticosteroids, and chemotherapy-induced diabetes (CID) is not uncommon (27.5%–37.0%) during the treatment of ALL (Hsu et al., 2002; Weiser et al., 2004; Alves et al., 2007). Besides the effect of corticosteroids, potential factors triggering hyperglycemia in ALL also include direct infiltration of the pancreas by leukemia cells and β cell dysfunction induced by chemotherapeutic agents such as L-asparagine (Mohn et al., 2004).

Several studies have noted alteration in glucose metabolism during the treatment of ALL (Koltin et al., 2012; Gifford et al., 2013; Banihashem et al., 2014), but few studies have investigated the clinical significance of CID. Weiser et al. (2004) reported that patients with hyperglycemia (37%) had a shorter median complete remission duration and a shorter median survival. However, their study focused on selected populations (patients treated with the hyperfractionated cyclophosphamide, vincristine, doxorubicin, and dexamethasone (Hyper-CVAD) regimen). Since patients using the Hyper-CVAD regimen represent only a small part of the patient population in clinical setting, it is not clear whether the prognostic value of CID has a rationale in all ALL patients. To further address this issue, we retrospectively investigated an unselected ALL population with several treatment regimens. We focused on an assessment of its characteristics, risk factors, and its prognostic value on survival.

We collected data from 177 patients with newly diagnosed ALL in the database of the First Affiliated Hospital, Zhejiang University School of Medicine (Hangzhou, China) from January 2011 to October 2013. Twelve patients who were treated at other institutions after being initially diagnosed in our hospital and were then lost to follow-up were excluded. A total of 165 patients had completed medical records. Among them, three patients had diabetes mellitus previously and six patients were newly diagnosed as having diabetes mellitus before induction chemotherapy. These patients were not included in the analysis. The remaining 156 patients were included in the current analysis.

The results were statistically analyzed using SPSS statistical software version 20.0 (SPSS Inc., Chicago, IL, USA). The Chi-square test or Fisher’s exact test was used to compare categorical variables. Student’s t-tests and Mann-Whitney U-test were used to compare continuous variables. Survival analysis was assessed using the Kaplan-Meier method. All P values were two-sided, with a P value of <0.05 indicating statistical significance.

Of the 156 subjects included, 87 (55.8%) were male. The median follow-up period was 24.7 months (range 0.5–48.3 months), and the median age at diagnosis was 34.5 years (range 15.0–73.0 years). In 118 patients (75.6%) ALL originated from B-cell precursor cells, while in 30 patients (19.2%) ALL originated from T cell or natural killer (NK)-T cell precursor cells. Breakpoint cluster region-Abelson (BCR-ABL) fusion gene transcript was detected in 46 patients (29.5%). All 156 patients received VP (vincristine+prednisone)-based induction chemotherapy. Two thirds (111 patients, 71.2%) received the VDCP (or VICP) regimen, which includes vincristine, daunorubicin (or idarubicin), cyclophosphamide, and prednisone. Fifteen patients (9.6%) used L-asparaginase in addition to the VDCP regimen. Fourteen patients (9.0%) underwent the Hyper-CVAD regimen which is comprised of cyclophosphamide twice daily on Days 1–3, doxorubicin on Day 4, vincristine on Days 4 and 11, and dexamethasone on Days 1–4 and 11–14.

Since all patients had fasting glucose level measurements as part of routine blood work two times a week during their hospital admission in our center, hyperglycemia in this study was defined using fasting glucose levels. Patients with a plasma fasting glucose level greater than or equal to 7.0 mmol/L during two or more continuous determinations were defined as having CID. This is in accordance with the American Diabetes Association (2018) definition. Thirty-three patients (21.2%) met the criteria of hyperglycemia during chemotherapy. Of them, 24 patients experienced transient hyperglycemia (lasting one or two chemotherapy cycles and then they recovered) and 9 patients developed persistent hyperglycemia. Seventeen patients (51.5%) were diagnosed with hyperglycemia during induction chemotherapy and the others in following chemotherapies. The median number of chemotherapy cycles developing hyperglycemia was 2.82.

Baseline characteristics according to hyperglycemia status are shown in Table S1. The two groups with or without hyperglycemia did not differ with respect to sex, incidence of hypertension, body mass index (BMI), immunophenotype, or the induction chemotherapy regimen. However, patients in the CID group were significantly older (median age 44 years vs. 31 years, P=0.017).

We then analyzed associations between CID and characteristics such as age, sex, BMI, hypertension, low-density lipoprotein (LDL), triglyceride (TG), and immunophenotype (Table 1). Among the factors selected, univariate analysis showed that CID was significantly influenced by age (P=0.031). Characteristics with P<0.5 were selected into the multivariate analysis. In the multivariate analysis, age greater than or equal to 35 years was significantly associated with the development of CID (P=0.026).

Table 1.

Risk factors of chemotherapy-induced diabetes (logistic analysis)

Variable	Univariate	Multivariate
	P value	HR (95% CI)		P value
Sex (male vs. female)	0.343		0.830
Age (≥35 years vs. <35 years)	0.031	2.824 (1.133–7.034)	0.026
Hypertension (yes vs. no)	1.000
WBC (>30×109 L−1 vs. ≤30×109 L−1)	0.615
HB (≥90 g/L vs. <90 g/L)	0.839
PLT (≥100×109 L−1 vs. <100×109 L−1)	0.571
BMI (≥25 kg/m2 vs. <25 kg/m2)	0.155		0.333
LDL (>3.29 mmol/L vs. ≤3.29 mmol/L)	0.679
TG (>1.70 mmol/L vs. ≤1.70 mmol/L)	0.482		0.801
Fusion gene (BCR-ABL+ vs. BCR-ABL−)	0.588
Immunophenotype (B vs. non-B)	0.117		0.707
Induction chemotherapy	0.189		0.318

WBC, white blood cell; HB, hemoglobin; PLT, platelet; BMI, body mass index; LDL, low-density lipoprotein; TG, triglyceride; BCR-ABL, breakpoint cluster region-Abelson; HR, hazard ratio; CI, confidence interval

Next, we assessed the prognosis value of CID in ALL patients. In order to eliminate the confounding effect of age on prognosis analysis, we divided all the patients into two groups. The first group was comprised of those younger than 35 years and the second of the rest. Table S2 shows both demographic characteristics and a comparison of patients between the hyperglycemia and the normal blood glucose subgroups organized separately into young adult patients and older ALL patients. In both the young adult patients and the older patients, there were no differences with respect to age, gender, incidence of hypertension, BMI, immunophenotype, or the induction chemotherapy regimen between the hyperglycemia and the normal blood glucose subgroups.

Survival results from the young adult patient group are shown in Fig. 1. Patients with hyperglycemia were found to have a shorter median overall survival (OS, 11.3 months vs. >48.0 months; P=0.002) and a shorter median event-free survival (EFS, 9.1 months vs. 15.9 months; P=0.005). Outcomes for older adult ALL patients are shown in Fig. 2. Using Kaplan-Meier analysis, the median OS was found to be 8.7 months for the hyperglycemia group and 11.8 months for the normal blood glucose group (P=0.989, Fig. 2a). The median EFS rates were 5.7 months and 6.4 months, respectively (P=0.635, Fig. 2b). There was no significant difference between the two groups in terms of OS and EFS in older adult patients.

Fig. 1.

Kaplan-Meier estimates of overall survival (a) and event-free survival (b) in young adult patients

Fig. 2.

Kaplan-Meier estimates of overall survival (a) and event-free survival (b) in older adult patients

A number of studies have noted disorders of glucose metabolism during the treatment of ALL (Mohn et al., 2004; Koltin et al., 2012; Gifford et al., 2013; Banihashem et al., 2014). However, the impact of CID on ALL patients is not yet well established. In this study, we demonstrate for the first time the negative impact of hyperglycemia on outcomes for young adult ALL patients. Interestingly, there were no differences among older adult patients. Weiser et al. (2004) found that patients with hyperglycemia during induction chemotherapy had a shorter median complete remission duration (P<0.001) and a shorter median survival (P<0.001). However, the authors did not further analyze the hybrid effect of age, which is an independent negative prognostic factor for both hyperglycemia and ALL (Giovannucci et al., 2010). To eliminate the confounding effect of age, we conducted stratified analysis by classifying age into two groups and we found that an age of 35 years was the best cutoff according to receiver operating characteristic (ROC) analysis. This result still needs validation in other independent cohorts of ALL patients.

The relation between CID and poor outcomes in ALL patients is multifactorial. One important factor may be associated with the increased incidence of infection during intensive chemotherapy. Weiser et al. (2004) showed that patients who developed hyperglycemia during induction chemotherapy for ALL were more likely to develop sepsis (P=0.03) or complicated infections (P=0.016) compared with patients without hyperglycemia. On the other hand, hyperinsulinemia, hyperglycemia, and inflammatory cytokines have been shown to be a direct link between diabetes and cancer, which promote the neoplastic process. All these complex factors lead to poor clinical outcomes in CID patients (Giovannucci et al., 2010). Abnormal glucose metabolism may also play a critical role. Studies have shown that acute myeloid leukemia (AML) patients present an altered glucose metabolism signature with poor clinical outcomes (Wang et al., 2013; Chen et al., 2014). Glucose metabolism may also be abnormal in ALL patients, especially CID patients, which would further affect the prognosis of the patients. More research is needed to confirm this.

We also studied the management of hyperglycemia in these CID patients. However, in our study, only eight CID patients (8/33, 24.2%) received glucose-lowing treatments. All these eight patients had undergone insulin therapy and none took oral hypoglycemic drugs, such as metformin or sulfonylureas. Since only one patient in the younger group received treatment for diabetes, we could not analyze whether glucose-lowering treatments could improve the prognosis for these CID patients, which is a limitation of this study. It should be noted that metformin, a widely used anti-diabetic drug, has recently attracted strong interest as a possible new anti-cancer molecule (Rosilio et al., 2013; Zi et al., 2015; Biondani and Peyron, 2018). Rosilio et al. (2013) have reported that metformin could interfere with the growth and survival of human T-ALL cancer cells, and it potentiates the anti-leukemia effects of dexamethasone. Additionally, differences in gut microbial composition were found between ALL children and healthy controls before, during, and even after cessation of chemotherapy (Chua et al., 2020; Thomas et al., 2020). Metformin has been found to benefit microbiota composition, promote gut barrier integrity, and improve metabolic function, which may have a role in the long-term wellbeing in ALL survivors (Ouyang et al., 2020a, 2020b). All these studies showed that metformin may play a unique role as a glucose-lowing therapy in CID patients. Further studies are needed to explore the value of glucose-lowering treatments, especially metformin, in CID patients.

Collectively, our results show that CID is not uncommon in ALL patients and age is an independent prognostic factor for CID development. Furthermore, we demonstrate that CID is an independent prognostic factor for inferior survival in young adult ALL patients.

List of electronic supplementary materials

Table S1

Baseline characteristics by hyperglycemia status of 156 patients studied

Table S2

Baseline characteristics by hyperglycemia status in patients aged <35 years and ≥35 years