Abstract
BACKGROUND
Over half of children treated for acute lymphoblastic leukemia (ALL) develop one or more treatment-related medical condition in their lifetime, many are known risk factors for diabetes mellitus. We evaluated the prevalence and risk factors for diabetes mellitus among clinically-assessed adult survivors of childhood ALL
METHODS
We performed a retrospective evaluation of data from ALL survivors and community controls enrolled in the St. Jude Lifetime Cohort Study between October 1, 2007 and June 30, 2016. Participants were adult ≥10-year survivors of childhood ALL and community controls who completed clinical and laboratory evaluations. Data for this analysis was abstracted from medical records. Exposures evaluated in this analysis included: chemotherapy and radiation exposures; and medical history, including drug-induced diabetes mellitus.
RESULTS
Of 1,360 eligible adults ≥10-year survivors of childhood ALL, 1,044 (mean age 33.97±9.14 years, 50.86% male) completed evaluations. The 368 controls (45.65% male) had a mean age of 35.33±10.21 years. Type 2 diabetes mellitus (T2DM) was found in 7.47% of survivors and 3.80% of controls (odds ratio [OR] 2.07, 95% confidence interval [CI] 1.11–3.87). Among survivors, in adjusted models, older age (odds ratio [OR] 1.05 for each additional year, 95% CI 1.02–1.08), BMI ≥30kg/m2 (OR 7.40, 95% CI 2.61–20.97) and drug-induced diabetes during ALL therapy (OR 4.67, 95% CI 2.53–8.61) were associated with T2DM.
CONCLUSION
Adult survivors of childhood ALL are at increased risk for T2DM. Adult survivors of childhood ALL who are of older age, overweight or obese BMI and/or developed drug-induced diabetes mellitus during treatment should be closely monitored for T2DM during long-term follow-up.
Keywords: Acute lymphoblastic leukemia, childhood, diabetes, survivor
PRECIS FOR USE IN TABLE OF CONTENTS:
This study enumerates a high prevalence of clinically ascertained diabetes in adult survivors of childhood ALL. These data identify an association between the development of drug-induced diabetes during therapy and the development of future diabetes.
Introduction
Improvements in risk-directed therapy for childhood acute lymphoblastic leukemia (ALL) have resulted in current 5-year survival rates of 90%.1–3 However, treatment-related adverse outcomes are common, with over half of survivors suffering from one or more severe, disabling or life-threatening chronic health conditions in their lifetime.3,4 Data from the Childhood Cancer Survivor Study (CCSS) show that adult survivors of childhood ALL, when compared to siblings, are at an increased risk for diabetes mellitus, which increases the risk for cardiovascular disease and early mortality beyond that conferred by anti-leukemic treatment.5,6 Thus, identification of survivors at greatest risk for diabetes mellitus is important for counseling and follow-up care.
Evidence indicates that host and lifestyle factors contribute to development of metabolic syndrome, a potential prediabetic state, in childhood ALL survivors.7,8 However, contributions of host and lifestyle factors to the development of overt diabetes in this population are not documented. In addition, although total body or abdominal radiation exposure have been identified as risk factors for diabetes mellitus among childhood cancer survivors, specific chemotherapy agents have not been implicated.5,6 Because drug-induced diabetes mellitus (DIDM), typically transient, occurring during the induction phase of ALL therapy when high doses of glucocorticoids and asparaginase are administered, is reported in approximately 10 to 20% of patients receiving treatment for ALL,9–11 investigation of the association between these agents, DIDM and later development of diabetes mellitus may identify the most vulnerable long-term survivors. One group of investigators,12 in a small study of 90 ALL survivors, provide some suggestive data that indicate that adolescents with a history of DIDM, when compared to those without this history, appear to be at greater risk for glucose intolerance later in life (13.3% vs. 1.1%, p=0.07). However, their small sample size precluded a comprehensive examination of associations between host, lifestyle and treatment factors and diabetes mellitus.
The aims of this study were to evaluate, in a clinically-assessed cohort of childhood ALL survivors, the contributions of host, lifestyle and treatment-related risk factors for diabetes mellitus, and to examine if development of DIDM during treatment increased risk for type 2 diabetes mellitus (T2DM) later in life. The availability the St. Jude Lifetime (SJLIFE) cohort, which includes a large number of ALL survivors with comprehensive demographic and lifestyle data, treatment exposures, medically validated health conditions, laboratory evaluations, and longitudinal clinical follow-up during adulthood, offers this opportunity.
Methods
Participants
SJLIFE is a cohort study designed to characterize health outcomes in childhood cancer survivors as they age. The study population and procedures have been described previously.13–16 These analyses included adult (≥18 years) survivors of childhood ALL who were treated at St. Jude Children’s Research Hospital (SJCRH), and were ≥10 years post diagnosis at the time of study participation. Medical records were abstracted, and participants completed detailed demographic, health and lifestyle questionnaires, and an on-campus comprehensive clinical assessment. Because SJCRH provides care for children (and thus follows survivors) from across the United States, age-, sex- and race-frequency matched community comparison group members (N=368) were recruited from among friends and non-first-degree family members of current SJCRH patients rather than from among members of the greater Memphis Community. When SJLIFE participants were enrolled, they were asked it they would like to invite a friend or non-first degree relative to serve as a community control. Current SJCRH family members and friends were recruited via on-campus advertisement.
Outcomes
Diabetes and prediabetes were classified, per American Diabetes Association (ADA) guidelines, using measures of plasma glucose and glycosylated hemoglobin (HgbA1c) after an overnight fast.17 Prediabetes was defined as a fasting plasma glucose between 100 mg/dL and 125 mg/dL, and/or HgbA1c between 5.7 and 6.4%. Individuals with fasting plasma glucose ≥126 mg/dL (7.0 mmol/L) on two separate tests, or HgbAlc ≥6.5%, or random glucose ≥200 mg/dL and who had not been diagnosed or treated for diabetes mellitus prior to study participation were assumed to have type 2 diabetes mellitus (T2DM). Participants with a previous diagnosis of diabetes mellitus and who reported use of oral diabetic medications with or without subsequent need for insulin, were also classified as having T2DM. Insulin-dependent diabetes was considered present among participants treated with insulin from the time of onset of diabetes and thereafter. Among female participants, gestational diabetes mellitus (GDM) was defined17 among those who reported diabetes during pregnancy, not present prior to gestation and whose symptoms resolved postpartum. Participants were classified as having drug-induced diabetes mellitus (DIDM) if they experienced treatment-induced hyperglycemia transiently requiring insulin therapy while undergoing treatment for ALL. All records were centrally reviewed by the study endocrinologist (WC).
Treatment exposures
Treatment exposures, including types and doses of chemotherapy, and sites and doses of radiation were abstracted from medical records. Chemotherapeutic agents considered in analysis were mercaptopurine, methotrexate, cytarabine, and cumulative doses of vincristine, anthracyclines,18 asparaginase (in pegaspargase-equivalents),19 dexamethasone and prednisone.20–22
Other co-variates
Demographic factors identified a priori to include sex, race, ethnicity, and educational attainment. Self-reported physical activity in metabolic equivalent minutes per week were calculated from participant answers to the National Health and Nutrition Examination Survey Physical Activity Questionnaire (NHANES-PAQ).23 Body weight was measured on an electronic scale in kilograms (kg) and height in meters (m) with a wall mounted stadiometer (Model 5002, Scale-Tronix, Inc., Wheaton, IL). Body Mass Index (BMI) (kg/m2) was calculated and categorized as underweight (<18.5 kg/m2), normal (18.5–24.9 kg/m2), overweight (25–29.9 kg/m2), or obese (≥30 kg/m2). Fat- and lean-mass were measured with dual x-ray absorptiometry (DXA, Hologic Model QDR 4500, software version 13.3:3, Bedford, MA) in the total body scanning mode.24–28 The scanner was calibrated weekly with known phantoms to minimize machine drift. Percent body fat and percent lean mass were calculated by dividing fat mass and fat free mass by total body mass and multiplying the result by 100.
Analysis
Descriptive statistics characterized demographic and treatment variables. Comparisons between participants and non-participant survivors and between survivors and controls utilized Chi-square, Fishers exact or t-tests as appropriate (Tables 1 and 2). Log binomial regression with a log link and Poisson distribution compared prevalence of prediabetes and T2DM between survivors and controls adjusting for age, race, sex, level of educational attainment, and BMI (Table 3). The cumulative incidence function was used to estimate T2DM risk with age as the time scale, taking into account the competing risk of death. Estimates were compared between survivors and community controls using Gray’s test. Logistic regression models were used to evaluate associations between host, lifestyle, DIDM and T2DM among all survivors (Table 4, Supplemental Tables 4 and 5) and GDM among female survivors who ever reported a pregnancy (Supplemental Table 1). Variables were selected a priori and included age, sex, race, highest level of educational attainment, BMI (with additional models replacing BMI with percent fat or percent lean mass in Supplemental Tables 2 and 3), level of physical activity, treatment exposures, and a diagnosis of DIDM while undergoing ALL therapy. Associations between host and treatment factors and DIDM were also evaluated with logistic regression (Table 5). Analyses were performed using SAS v9.4 (Cary, N.C.).
Table 1.
Demographic and treatment characteristics of ALL survivor participants (N=1044) and non-participants (N=316)
| Participants | Non-Participants | ||
|---|---|---|---|
| N (%) | N (%) | P-value | |
| Sex | |||
| Female | 513 (49.14%) | 165 (52.22%) | 0.67d |
| Male | 531 (50.86%) | 151 (47.78%) | |
| Race | |||
| Non-Hispanic White | 905 (86.69%) | 258 (81.65%) | 0.009d |
| Non-Hispanic Black | 95 (9.10%) | 33 (10.44%) | |
| Hispanic | 31 (2.97%) | 11 (3.48%) | |
| Asian | 3 (0.29%) | 2 (0.63%) | |
| Other | 10 (0.96%) | 12 (3.80%) | |
| Spinal radiation | |||
| No | 960 (91.95%) | 283 (89.56%) | 0.18d |
| Yes | 84 (8.05%) | 33 (10.44%) | |
| Cranial radiation dose (Gray)a | |||
| None | 468 (44.83%) | 139 (44.13%) | 0.65d |
| 1–23 | 251 (24.04%) | 70 (22.22%) | |
| ≥24 | 325 (31.13%) | 106 (33.75%) | |
| Bone marrow transplantb | |||
| No | 1021 (97.80%) | 309 (97.78%) | 0.99d |
| Yes | 23 (2.20%) | 7 (2.22%) | |
| 6-Mercaptopurine | |||
| No | 8 (0.77%) | 3 (0.95%) | 0.72d |
| Yes | 1036 (99.23%) | 313 (99.05%) | |
| Methotrexate | |||
| No | 1 (0.10%) | 0 (0.00%) | 1.00d |
| Yes | 1043 (99.90%) | 316 (100.0%) | |
| Age at diagnosis (years) | |||
| Mean (SD) | 6.59 (4.41) | 6.07 (4.39) | 0.06e |
| Median (IQR) | 5.07 (3.10, 9.26) | 4.31 (3.02, 8.04) | |
| Cumulative vincristine dose (mg/m2) | |||
| Mean (SD) | 37.43 (27.55) | 33.90 (27.95) | 0.05e |
| Median (IQR) | 43.83 (6.45, 58.14) | 35.27 (6.11, 56.88) | |
| Cumulative cytarabine dose (mg/m2) | |||
| Mean (SD) | 6473.72 (5031.27) | 6478.51 (5124.89) | 0.99e |
| Median (IQR) | 5224.02 (2069.59, 10653.37) | 5167.36 (2407.76, 10199.79) | |
| Cumulative prednisone dose (mg/m2) | |||
| Mean (SD) | 4967.36 (4257.59) | 4273.24 (4232.82) | 0.06e |
| Median (IQR) | 224.00 (1120.00–9520.00) | 2240.00 (1120.00–8880.00) | |
| Cumulative dexamethasone dose (mg/m2) | |||
| Mean (SD) | 1541.27 (572.98) | 1509.19 (574.33) | 0.74e |
| Median (IQR) | 1568.00 (1288.00–1620.00) | 1620.00 (1288.00–1620.00) | |
| Cumulative anthracycline dose (mg/m2) | |||
| Mean (SD) | 123.85 (88.62) | 117.06 (78.31) | 0.28e |
| Median (IQR) | 101.79 (58.82, 150.67) | 101.12 (51.44, 139.19) | |
| Cumulative asparaginase dose (units/m2)c | |||
| Mean (SD) | 5895.25 (6505.10) | 5592.54 (6667.92) | 0.50e |
| Median (IQR) | 3629.27 (2063.18, 6281.45) | 3000.00 (2008.93, 6130.14) | |
| Drug induced diabetes mellitus | |||
| No | 962 (92.15%) | 306 (96.84%) | 0.004d |
| Yes | 82 (7.85%) | 10 (3.16%) |
Abbreviations: SD=standard deviation; IQR=interquartile range;
one non-participant survivor missing cranial radiation dose;
survivors who received bone marrow transplant all received total body irradiation;
cumulative asparaginase dose reported as pegaspargase-equivalents;
Chi-square;
t-test
Table 2.
Comparison between 1044 survivors and 368 community control participants
| Characteristics | ALL Survivors (n=1044) | Community Controls (n=368) | P-value |
|---|---|---|---|
| Age at most recent interview | |||
| Mean (SD) | 33.97 (9.14) | 35.33 (10.21) | 0.02i |
| Median (IQR) | 33.19 (26.78, 40.53) | 34.48 (28.10, 41.72) | |
| N (%) | N (%) | ||
| Sex | |||
| Female | 513 (49.14) | 200 (54.35) | 0.09h |
| Male | 531 (50.86) | 168 (45.65) | |
| Race | |||
| Non-Hispanic White | 905 (86.69) | 310 (84.24) | <0.001h |
| Non-Hispanic Black | 95 (9.10) | 23 (6.25) | |
| Hispanic | 31 (2.97) | 14 (3.80) | |
| Asian | 3 (0.29) | 3 (0.82) | |
| Other | 10 (0.96) | 18 (4.89) | |
| Education | |||
| <College Graduate | 700 (67.05) | 171 (46.72) | <0.001h |
| College Graduate | 344 (32.95) | 195 (53.28) | |
| Body mass index | |||
| Underweight (<18.5 kg/m2) | 17 (1.63) | 12 (3.26) | 0.004h |
| Normal (18.5–24.9 kg/m2) | 266 (25.48) | 126 (34.24) | |
| Overweight (25.0–29.9 kg/m2) | 288 (27.59) | 102 (27.72) | |
| Obese (≥30 kg/m2) | 473 (45.31) | 128 (34.78) | |
| Prediabetesa | |||
| No | 767 (73.47) | 300 (81.52) | 0.002h |
| Yes | 277 (26.53) | 68 (18.48) | |
| Insulin-dependent diabetesb | |||
| No | 1041 (99.71) | 365 (99.18) | 0.19h |
| Yes | 3 (0.29) | 3 (0.82) | |
| Type 2 diabetesc | |||
| No | 966 (92.53) | 354 (96.20) | 0.01h |
| Yes | 78 (7.47) | 14 (3.80) | |
| Gestational diabetesd | |||
| No | 233 (93.57) | 115 (96.64) | 0.23h |
| Yes | 16 (6.43) | 4 (3.36) | |
| Drug induced diabetese | |||
| No | 962 (92.15) | NA | |
| Yes | 82 (7.85) | NA | |
| Chronic conditions CTCAE Grade 2–4f | |||
| Chronic kidney disease | |||
| Type 2 diabetes | 1 (1.28) | 0 (0.00) | 1.00j |
| Prediabetes | 0 (0.00) | 0 (0.00) | - |
| Cardiomyopathy | |||
| Type 2 diabetes | 4 (5.13) | 0 (0.00) | 1.00j |
| Prediabetes | 3 (1.08) | 1 (1.47) | 0.59j |
| Coronary artery disease | |||
| Type 2 diabetes | 1 (1.28) | 0 (0.00) | 1.00j |
| Prediabetes | 3 (1.08) | 0 (0.00) | 1.00j |
| Exercise Intolerance | |||
| Type 2 diabetes | 46 (58.97) | 4 (28.57) | 0.04j |
| Prediabetes | 105 (37.91) | 13 (19.12) | 0.003h |
Abbreviations: SD=standard deviation; IQR=interquartile range; kg=kilogram; m2=square meter
Fasting plasma glucose 100–125 milligrams per deciliter (mg/dL) and/or hemoglobin A1c 5.7–6.4%;
Insulin required when diabetes diagnosed;
Fasting plasma glucose ≥126 mg/dL on two occasions, random glucose ≥200 mg/dL, hemoglobin A1c ≥6.5% and/or initially on oral medication when diabetes diagnosed;
diabetes during gestation, not present prior to pregnancy that resolved postpartum;
required brief insulin therapy during treatment for leukemia;
Common Terminology Criteria for Adverse Events version 4.02;
<490 meters in the six minute walk test;
Chi-square;
t-test;
Fisher’s exact.
Table 3.
Prevalence of prediabetes and type 2 diabetes among 1044 ALL survivors and 368 controls
| Prediabetes N=277 | Type 2 Diabetes* N=78 | |||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Adjusted for BMI | Not adjusted for BMI | Adjusted for BMI | Not adjusted for BMI | |||||||||||
| % | PR | 95% CI | P-value | PR | 95% CI | P-value | % | PR | 95% CI | P-value | PR | 95% CI | P-value | |
| Survivors | 26.53 | 1.33 | 1.06–1.66 | 0.01 | 1.41 | 1.12–1.78 | 0.003 | 7.47 | 2.07 | 1.11–3.87 | 0.02 | 2.36 | 1.27–4.39 | 0.007 |
| Controls | 18.48 | Ref | . | . | . | . | . | 3.80 | Ref | . | . | . | . | . |
Abbreviations: PR=prevalence ratio, adjusted for age, sex, race and level of educational attainment; 95% CI=95% confidence interval; BMI=body mass index; Ref=referent group
Logistic regression was used to estimate odds ratio as an approximation to PR for rare outcomes (<10%), marked with ‘*’.
Table 4.
Risk factors for type 2 diabetes mellitus in adult survivors of childhood acute lymphoblastic leukemia
| Type 2 Diabetes Mellitus (N=78) | |||||||||
|---|---|---|---|---|---|---|---|---|---|
| Characteristics | Adjusted for BMI | Not Adjusted for BMI | |||||||
| N | Row % | OR | 95% CI | P-value | Row% | OR | 95% CI | P-value | |
| Age | |||||||||
| Age at assessment (years) | 1044 | 7.50 | 1.05 | 1.02–1.08 | <0.001 | 7.50 | 1.06 | 1.04–1.09 | <0.001 |
| Sex | |||||||||
| Male (reference) | 531 | 7.34 | 1.00 | - | - | 7.34 | 1.00 | - | - |
| Female | 513 | 7.60 | 1.01 | 0.61–1.67 | 0.97 | 7.60 | 1.02 | 0.62–1.67 | 0.94 |
| Race | |||||||||
| Non-Hispanic white (reference) | 905 | 6.96 | 1.00 | - | - | 6.96 | 1.00 | - | - |
| Other | 139 | 10.79 | 1.96 | 0.99–3.89 | 0.05 | 10.79 | 1.87 | 0.96–3.63 | 0.06 |
| Education | |||||||||
| < College graduate (reference) | 700 | 8.57 | 1.00 | - | - | 8.57 | 1.00 | - | - |
| College graduate | 344 | 5.23 | 0.57 | 0.32–1.02 | 0.06 | 5.23 | 0.51 | 0.29–0.90 | 0.02 |
| Body mass index | |||||||||
| Underweight/Normal (<25.0 kg/m2) | 283 | 1.41 | 1.00 | - | - | 1.41 | - | - | - |
| Overweight (25.0–29.9 kg/m2) | 288 | 5.21 | 3.04 | 0.98–9.47 | 0.05 | 5.21 | - | - | - |
| Obese (≥30 kg/m2) | 473 | 12.47 | 7.40 | 2.61–20.97 | <0.001 | 12.47 | - | - | - |
| Physical activity | |||||||||
| Per 100 MET minutes per week | 1044 | 7.51 | 1.00 | 0.99–1.01 | 0.81 | 7.51 | 1.00 | 0.99–1.01 | 0.64 |
| Drug induced diabetes mellitus | |||||||||
| Yes | 82 | 24.39 | 4.67 | 2.53–8.61 | <0.001 | 24.39 | 4.84 | 2.67–8.76 | <0.001 |
| No (reference) | 962 | 6.03 | 1.00 | 6.03 | 1.00 | ||||
Abbreviations: kg=kilograms; m2=square meter; MET=metabolic equivalent; OR=odds ratio; 95% CI=95% confidence interval
Table 5.
Risk factors for drug induced diabetes mellitus in 1044 survivors of childhood acute lymphoblastic leukemia, stratified by age at diagnosis
| Age at Diagnosis <15 years | Age at Diagnosis ≥15 years | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| N | Row% | OR | 95% CI | P-value | N | Row % | OR | 95% CI | P-value | |
| Sex | ||||||||||
| Male (reference) | 488 | 5.33 | 1.00 | - | - | 43 | 30.23 | 1.00 | - | - |
| Female | 490 | 7.76 | 1.47 | 0.87–2.50 | 0.15 | 23 | 21.74 | 0.55 | 0.13–2.28 | 0.41 |
| Race | ||||||||||
| Non-Hispanic white (reference) | 846 | 5.79 | 1.00 | - | - | 59 | 25.42 | 1.00 | - | - |
| Other | 132 | 11.36 | 1.83 | 0.96–3.48 | 0.07 | 7 | 42.86 | 1.20 | 0.17–8.48 | 0.86 |
| Chemotherapy | ||||||||||
| Asparaginase dose (per 1000 units/m2)a | 978 | 6.54 | 0.99 | 0.94–1.04 | 0.55 | 66 | 27.27 | 1.12 | 1.02–1.23 | 0.01 |
| Prednisone dose (per 1000 mg/m2) | 951 | 6.52 | 1.05 | 0.98–1.12 | 0.16 | 65 | 26.15 | 1.07 | 0.90–1.28 | 0.44 |
| Dexamethasone dose (per 1000 mg/m2) | 208 | 9.62 | 1.58 | 1.05–2.37 | 0.03 | 23 | 30.43 | 0.53 | 0.15–1.84 | 0.32 |
Abbreviations: OR=odds ratio; 95% CI=95% confidence interval; mg=milligrams; m2=square meter
asparaginase dose reported as pegaspargase-equivalents
Results
Participants
Among 1,360 eligible ALL survivors as of June 30, 2016, 199 declined participation, 68 completed questionnaires only, and 49 were lost to follow-up, leaving 1,044 (76.76%) with completed clinical and laboratory evaluations. Table 1 shows characteristics of participants and non-participants. Participants were more likely to be non-Hispanic white and received higher doses of glucocorticoids. Participants were 50.86% male and a median age of 5.07 years (interquartile range: 3.10, 9.26) at diagnosis.
Comparison of survivors and controls
Characteristics of the study population are shown in Table 2. Mean age at assessment was slightly higher in controls (35.33 years, SD 10.21) than in survivors (33.97 years, SD 9.14, p=0.02). Survivors were more likely than controls to be obese and less likely to be college graduates. Survivors were more likely than controls to have T2DM with a cumulative incidence by age 50 of 16% (vs. 9% in controls, p=0.01) (Supplemental Figure 1). The prevalence of insulin dependent diabetes was less than 1% in both survivors and controls. After adjusting for age, sex, race, and educational attainment, survivors were more likely to have prediabetes and T2DM than controls. Adjusting for BMI slightly attenuated the estimates (Table 3).
Risk factors for type 2 diabetes mellitus in survivors
Age at most recent interview, obesity (as BMI, percent fat, or percent lean mass), and DIDM while undergoing treatment were associated with an elevated risk of T2DM in survivors. Sex, race, physical activity, and level of educational attainment were not associated with increased T2DM risk (Table 4, Supplemental Tables 2 and 3). Stratifying by cranial radiation exposure did not modify these associations (Supplemental Tables 4 and 5).
Drug-induced diabetes mellitus
A total of 82 (7.85%) survivors developed DIDM while undergoing treatment for ALL, which was most prevalent in survivors ≥15 years at diagnosis. Among survivors ≥15 years at diagnosis, every 1000 units/m2 increase in asparaginase dose increased the odds of developing DIDM by 1.12 (95% CI, 1.02–1.23). Cranial radiation and chemotherapy exposures other than asparaginase were not associated with DIDM in survivors ≥15 years at diagnosis. In survivors <15 years at diagnosis, dexamethasone exposure (per 1000 mg/m2) was associated with a 1.58 (95% CI, 1.05–2.37) increased odds of DIDM (Table 5). Cranial radiation and chemotherapy exposures other than dexamethasone were not associated with DIDM in survivors <15 years at diagnosis.
Gestational diabetes
Of 531 female survivors, a total of 249 (48.54%) had been pregnant at least once with 16 (6.43%) developing GDM during pregnancy, six (37.50%) of whom had been previously diagnosed with DIDM (supplemental Table 1). Race, age at diagnosis, and specific chemotherapy exposures were not independently associated with an increased risk of GDM. GDM was not a risk factor for future development of T2DM among females in our cohort.
Discussion
This evaluation, in a well characterized cohort of cancer survivors, provides a robust estimate of the prevalence of T2DM among adult survivors of childhood ALL, and identifies that obesity, a known risk factor for T2DM in the general population,17 increases risk for this outcome. In addition, we identify that DIDM, prevalent among children ≥15 years at ALL diagnosis exposed to higher doses of asparaginase, and among children <15 years at ALL diagnosis exposed to higher doses of dexamethasone, is a risk factor for both T2DM and among females for GDM. These data both provide a target for early intervention (obesity) and indicate that survivors of childhood ALL who develop DIDM during therapy, or who are older teenagers when diagnosed with ALL should be screened for development of T2DM. Female survivors of childhood cancer with a history of DIDM should be closely monitored during pregnancy for development of GDM.
Approximately 8% of survivors in our cohort were diagnosed with T2DM compared to 4% of controls, where the prevalence of T2DM was comparable to that in the general population. The Centers for Disease Control and Prevention estimate that 4% of U.S. adults aged 18–44 years have diabetes mellitus.29 Additionally, nearly a third of survivors were prediabetic suggesting that this outcome will become more prevalent as survivors age. These data are consistent with previous reports from the CCSS and the Bone Marrow Transplant Survivor Study (BMTSS). Survivor participants in the CCSS treated for ALL during childhood were 1.8 times more likely (95% CI, 1.2–2.6) to be diagnosed with diabetes mellitus compared to their siblings.6 This increased risk was associated with total body irradiation, in both the CCSS (OR=12.6; 95% CI, 6.2–25.3) and BMTSS (OR=3.42; 95% CI, 1.55–7.52) cohorts.6,30 In our cohort, 23 survivors were treated with total body irradiation, two of whom developed T2DM. We speculate that this exposure, which includes radiation to both the cranium and abdomen, results in early injury to pancreatic beta cells and predisposes to obesity,31 which was associated with T2DM in our cohort.32
While previous studies have reported the prevalence of DIDM in children undergoing treatment for ALL to be between 10 and 20%,9–11 this, to our knowledge, is the first study that we know of to report an association between DIDM and later development of T2DM, and among females, for GDM. It is possible that the development of DIDM in some patients unmasks an underlying predisposition leading to an increased risk of T2DM and/or GDM. During remission induction therapy, patients receive high doses of glucocorticoids and asparaginase, both of which are known to induce hyperglycemia.9 It is possible that pancreatic beta cell injury caused by these medications results in predisposition to insulin resistance in these individuals and an increased risk for developing T2DM or GDM later in life.33,34
Prophylactic pharmacotherapy has not been evaluated in adult survivors of childhood ALL. However, survivors who develop DIDM during treatment may benefit from preventive pharmacotherapy (i.e. metformin) for T2DM prophylaxis. It is also possible that at risk children may benefit if metformin is given during ALL therapy. Several ongoing clinical trials are investigating the clinical and biological effects (antitumor) of metformin in the setting of high risk and refractory/relapsed ALL (; ).35,36 Early reports from have found that among ALL patients who express high levels of the ABCB1 gene, those who receive metformin in combination with conventional chemotherapy had higher survival compared to those who only received conventional chemotherapy (83.33% vs. 26.47%; p=0.025).35 These studies are relevant as they may provide safety data for the potential administration of metformin in children with ALL during therapy. Although speculative, an additional benefit of metformin in high risk groups, like the older children in our cohort exposed to higher doses of asparaginase, may be to prevent DIDM and perhaps future development of T2DM during adulthood.
Findings from this study should be considered within the context of study limitations. It is possible the results of our analysis were influenced by selection bias due to the less than 100% participation rate. Although demographic differences between participants and non-participants were not substantial,37 it is possible that survivors who completed on-campus evaluations differ in health status compared to non-participants. For example, in the general population, T2DM is more common among non-white than among white persons. Because non-participants were more likely to be non-white, our estimate of the prevalence of T2DM may be low. Conversely, because we found an association between dexamethasone exposure and DIDM, and because non-participants had on average a lower dose of this agent, our estimate of the prevalence of DIDM may be high. In addition, survivors were only recruited from one institution, which may limit the generalizability of our findings, particularly to survivors of non-white race who comprised only 13% of our cohort. However, the modified Berlin-Frankfurt-Münster backbone used in St. Jude Total Therapy is similar to regimens used throughout the world. The complex pathophysiology of diabetes mellitus and multiplicity of factors interfering with glucose homeostasis represent significant challenges to the classification of affected individuals.17 Participants with new onset insulin-dependent or other rare forms of diabetes mellitus may have been misclassified as having T2DM by our criteria. In our model for GDM, we were unable to include patient BMI at time of GDM diagnosis, as this information was not available in our medical records. Due to the relationship between high BMI and insulin resistance, it would be of interest to include patient BMI at GDM diagnosis in a future analysis. It is also possible that our sample size of survivors who developed GDM was too small (n=16) to detect any associations between exposures and GDM risk. The use of standardized follow-up medical assessments in SJLIFE has allowed for the continued collection of objective measures such as height, weight, and percent body fat, among others. However, some self-reported measures such as smoking, alcohol intake, and total met-minutes per week, established risk factors for diabetes, may have been over- or under- estimated in our population.
Conclusion
Adults diagnosed with diabetes are at an increased risk for numerous chronic health conditions, ranging from metabolic syndrome to cardiovascular disease.38 Among SJLIFE survivors of childhood ALL, there is a high prevalence of prediabetes (26.53%) and T2DM (7.47%), compared to controls (18.48% and 3.80%, respectively). After controlling for BMI, survivors had a 33% increased risk for prediabetes and a two-fold increased risk of developing T2DM compared to controls. Our findings show that age, obese BMI, and a diagnosis of DIDM while on treatment, are all predictors of T2DM in survivors of childhood ALL. As the survival rate for childhood ALL continues to improve, it is important to investigate late-effect health outcomes in this aging population. The high prevalence and increased risk of diabetes among SJLIFE ALL survivors, combined with the potentially modifiable risk of diabetes, underscores the importance of surveillance and early interventions among survivors.
Supplementary Material
Cumulative incidence of type 2 diabetes mellitus among childhood cancer survivors and controls by age
FUNDING SUPPORT:
National Cancer Institute, the Cancer Center Support (CORE) grant (P30 CA21765, C. Roberts, Principal Investigator); CA195547, M. Hudson, Principal Investigator; and the American Lebanese and Syrian Associated Charities (ALSAC).
Footnotes
CONFLICT OF INTEREST DISCLOSURES: HW has equity interest in and is employed by Color Genomics, Inc. All other authors declare no conflicts of interest.
References
- 1.Ward E, DeSantis C, Robbins A, Kohler B, Jemal A. Childhood and adolescent cancer statistics, 2014. CA Cancer J Clin. 2014;64:83–103. [DOI] [PubMed] [Google Scholar]
- 2.Pui CH, Yang JJ, Hunger SP, et al. Childhood Acute Lymphoblastic Leukemia: Progress Through Collaboration. J Clin Oncol. 2015;33:2938–2948. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Mody R, Li S, Dover DC, et al. Twenty-five-year follow-up among survivors of childhood acute lymphoblastic leukemia: a report from the Childhood Cancer Survivor Study. Blood. 2008;111:5515–5523. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Bhakta N, Liu Q, Ness KK, et al. The cumulative burden of surviving childhood cancer: an initial report from the St Jude Lifetime Cohort Study (SJLIFE). Lancet. 2017;390:2569–2582. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Chemaitilly W, Cohen LE, Mostoufi-Moab S, et al. Endocrine Late Effects in Childhood Cancer Survivors. J Clin Oncol. 2018;36:2153–2159. [DOI] [PubMed] [Google Scholar]
- 6.Meacham LR, Sklar CA, Li S, et al. Diabetes mellitus in long-term survivors of childhood cancer. Increased risk associated with radiation therapy: a report for the childhood cancer survivor study. Arch Intern Med. 2009;169:1381–1388. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Nottage KA, Ness KK, Li C, Srivastava D, Robison LL, Hudson MM. Metabolic syndrome and cardiovascular risk among long-term survivors of acute lymphoblastic leukaemia - From the St. Jude Lifetime Cohort. Br J Haematol. 2014;165:364–374. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Gurney JG, Ness KK, Sibley SD, et al. Metabolic syndrome and growth hormone deficiency in adult survivors of childhood acute lymphoblastic leukemia. Cancer. 2006;107:1303–1312. [DOI] [PubMed] [Google Scholar]
- 9.Lowas SR, Marks D, Malempati S. Prevalence of transient hyperglycemia during induction chemotherapy for pediatric acute lymphoblastic leukemia. Pediatr Blood Cancer. 2009;52:814–818. [DOI] [PubMed] [Google Scholar]
- 10.Roberson JR, Spraker HL, Shelso J, et al. Clinical consequences of hyperglycemia during remission induction therapy for pediatric acute lymphoblastic leukemia. Leukemia. 2009;23:245–250. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Pui C-H, Burghen GA, Bowman WP, Aur RJA. Risk factors for hyperglycemia in children with leukemia receiving l-asparaginase and prednisone. J Pediatr. 1981;99:46–50. [DOI] [PubMed] [Google Scholar]
- 12.Yeshayahu Y, Koltin D, Hamilton J, Nathan PC, Urbach S. Medication-induced diabetes during induction treatment for ALL, an early marker for future metabolic risk? Pediatric Diabetes. 2015;16:104–108. [DOI] [PubMed] [Google Scholar]
- 13.Ojha RP, Oancea SC, Ness KK, et al. Assessment of Potential Bias From Non-Participation in a Dynamic Clinical Cohort of Long-Term Childhood Cancer Survivors: Results From the St. Jude Lifetime Cohort Study. Pediatr Blood Cancer. 2013;60:856–864. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Hudson MM, Ness KK, Nolan VG, et al. Prospective Medical Assessment of Adults Surviving Childhood Cancer: Study Design, Cohort Characteristics, and Feasibility of the St. Jude Lifetime Cohort Study. Pediatr Blood Cancer. 2011;56:825–836. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Hudson MM, Ness KK, Gurney JG, et al. Clinical Ascertainment of Health Outcomes among Adults Treated for Childhood Cancer: A Report from the St. Jude Lifetime Cohort Study. JAMA. 2013;309:2371–2381. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Hudson MM, Ehrhardt MJ, Bhakta N, et al. Approach for Classification and Severity Grading of Long-term and Late-Onset Health Events among Childhood Cancer Survivors in the St. Jude Lifetime Cohort. Cancer Epidemiol Biomarkers Prev. 2017;26:666–674. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.American Diabetes Association. 2. Classification and Diagnosis of Diabetes: Standards of Medical Care in Diabetes-2018. Diabetes Care. 2018;41:S13–S27. [DOI] [PubMed] [Google Scholar]
- 18.Feijen EA, Leisenring WM, Stratton KL, et al. Equivalence Ratio for Daunorubicin to Doxorubicin in Relation to Late Heart Failure in Survivors of Childhood Cancer. J Clin Oncol. 2015;33:3774–3780. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Pui CH, Pei D, Sandlund JT, et al. Long-term results of St Jude Total Therapy Studies 11, 12, 13A, 13B, and 14 for childhood acute lymphoblastic leukemia. Leukemia. 2009;24:371–382. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Meikle AW, Tyler FH. Potency and duration of action of glucocorticoids. Am J Med. 1977;63:200–207. [DOI] [PubMed] [Google Scholar]
- 21.Meikle AW, Tyler FH. Potency and Duration of Action of Glucocorticoids - Effects of Hydrocortisone, Prednisone and Dexamethasone on Human Pituitary-Adrenal-Function. Am J Med. 1977;63:200–207. [DOI] [PubMed] [Google Scholar]
- 22.Pui CH, Pei D, Sandlund JT, et al. Long-term results of St Jude Total Therapy Studies 11, 12, 13A, 13B, and 14 for childhood acute lymphoblastic leukemia. Leukemia. 2010;24:371–382. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.U.S. Department of Health and Human Services, Centers for Disease Control and Prevention, National Center for Health Statistics. National Health and Nutrition Examination Survey Data. 2008; https://www.cdc.gov/nchs/nhanes/index.htm. Accessed May 1, 2018.
- 24.Thomas SR, Kalkwarf HJ, Buckley DD, Heubi JE. Effective dose of dual-energy X-ray absorptiometry scans in children as a function of age. J Clin Densitom. 2005;8:415–422. [DOI] [PubMed] [Google Scholar]
- 25.Jensen MD, Kanaley JA, Roust LR, et al. Assessment of body composition with use of dual-energy x-ray absorptiometry: evaluation and comparison with other methods. Mayo Clin Proc. 1993;68:867–873. [DOI] [PubMed] [Google Scholar]
- 26.Njeh CF, Samat SB, Nightingale A, McNeil EA, Boivin CM. Radiation dose and in vitro precision in paediatric bone mineral density measurement using dual X-ray absorptiometry. Br J Radiol. 1997;70:719–727. [DOI] [PubMed] [Google Scholar]
- 27.Kalender WA. Effective dose values in bone mineral measurements by photon absorptiometry and computed tomography. Osteoporos Int. 1992;2:82–87. [DOI] [PubMed] [Google Scholar]
- 28.Njeh CF, Fuerst T, Hans D, Blake GM, Genant HK. Radiation exposure in bone mineral density assessment. Appl Radiat Isot. 1999;50:215–236. [DOI] [PubMed] [Google Scholar]
- 29.U.S. Department of Health and Human Services, Centers for Disease Control and Prevention, National Center for Health Statistics. National Diabetes Statistics Report, 2017. 2017; https://www.cdc.gov/diabetes/pdfs/data/statistics/national-diabetes-statistics-report.pdf. Accessed May 1, 2018.
- 30.Scott Baker K, Ness KK, Steinberger J, et al. Diabetes, hypertension, and cardiovascular events in survivors of hematopoietic cell transplantation: a report from the bone marrow transplantation survivor study. Blood. 2007;109:1765. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Wilson CL, Liu W, Yang JJ, et al. Genetic and clinical factors associated with obesity among adult survivors of childhood cancer: A report from the St. Jude Lifetime Cohort. Cancer. 2015;121:2262–2270. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.de Vathaire F, El-Fayech C, Ben Ayed FF, et al. Radiation dose to the pancreas and risk of diabetes mellitus in childhood cancer survivors: a retrospective cohort study. Lancet Oncol. 2012;13:1002–1010. [DOI] [PubMed] [Google Scholar]
- 33.Tamez-Pérez HE, Quintanilla-Flores DL, Rodríguez-Gutiérrez R, González-González JG, Tamez-Peña AL. Steroid hyperglycemia: Prevalence, early detection and therapeutic recommendations: A narrative review. World J Diabetes. 2015;6:1073–1081. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Fine NHF, Doig CL, Elhassan YS, et al. Glucocorticoids Reprogram β-Cell Signaling To Preserve Insulin Secretion. Diabetes. 2018;67:278–290. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Ramos-Peñafiel C, Olarte-Carrillo I, Cerón-Maldonado R, et al. Effect of metformin on the survival of patients with ALL who express high levels of the ABCB1 drug resistance gene. J Transl Med. 2018;16:245. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Biondani G, Peyron J-F. Metformin, an Anti-diabetic Drug to Target Leukemia. Front Endocrinol (Lausanne). 2018;9:446. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Ojha RP, Oancea SC, Ness KK, et al. Assessment of potential bias from non-participation in a dynamic clinical cohort of long-term childhood cancer survivors: results from the St. Jude Lifetime Cohort Study. Pediatr Blood Cancer. 2013;60:856–864. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Menke A, Casagrande S, Geiss L, Cowie CC. Prevalence of and trends in diabetes among adults in the united states, 1988–2012. JAMA. 2015;314:1021–1029. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Cumulative incidence of type 2 diabetes mellitus among childhood cancer survivors and controls by age