Characteristics of Type 2 Diabetes in Female and Male Youth

We identified 716 youth <19 years of age with a clinical diagnosis of type 2 diabetes who were seen in our center between July 2016 and July 2019. Steroid-induced and cystic fibrosis–related diabetes were not included. We also excluded one child who tested positive for islet autoantibodies, resulting in a final sample of 715. The study protocol was approved by the Baylor College of Medicine Institutional Review Board (H-45325).

We collected demographic, clinical, and laboratory characteristics by electronic medical record review. Physical exam information included Tanner staging for pubertal development (I–V) within 3 months of diabetes diagnosis, performed by a pediatric endocrinologist using breast development staging for girls and testicular size for boys.

Laboratory data included random serum glucose and C-peptide levels measured simultaneously at the time of diabetes diagnosis, A1C within 1 month of diagnosis, history of clinician-diagnosed dyslipidemia and/or nonalcoholic fatty liver disease (NAFLD), lipid profile, and urine albumin-to-creatinine ratio obtained within 1 year of diagnosis.

Microalbuminuria was defined as two consecutive abnormal spot urine results with microalbumin-to-creatinine ratio >30 mg/g. Diabetic ketoacidosis (DKA) was defined using the standard criteria of the International Society for Pediatric and Adolescent Diabetes 2018 clinical practice consensus guidelines (2) in those with available biochemical data or by documentation of DKA in the medical record.

We first conducted a univariable comparison of males and females for each of the characteristics of interest, using χ², Fisher exact, and Wilcoxon rank sum tests as appropriate. To adjust for potential confounding, we conducted multivariable logistic regression analysis with the variables that were statistically significant in the univariable analysis. With Bonferroni correction for the number of independent comparisons, P <0.0045 was considered statistically significant in the univariable analyses, and P <0.05 was considered statistically significant for multivariable analyses. All analyses were conducted using STATA, v. 12 statistical software (StataCorp, College Station, TX).

We studied 715 children (448 female and 267 male) with type 2 diabetes who presented to the clinic within the 3-year study period. At diagnosis, their mean age was 13.7 years (SD 2.4 years), glucose was 13.9 mmol/L (SD 7.60 mmol/L) (250.5 mg/dL [SD 136.9 mg/dL]), and A1C 9.6% (SD 2.6%). Puberty was assessed at stage IV or V in 60.4% of the patients. The mean BMI z score at diagnosis was 2.25 (SD 0.48). Within 1 year of diagnosis, mean HDL cholesterol was 0.98 mmol/L (SD 0.26 mmol/L) (37.9 mg/dL [SD 10.1 mg/dL]), and the mean triglyceride level was 5.52 mmol/L (SD 6.35 mmol/L) (488.9 mg/dL [SD 562.4 mg/dL]) (Table 1).

In univariable analysis, female patients were younger (13.4 vs. 14.1 years of age), had lower A1C (9.3 vs. 10.1%), and had lower blood glucose (13.2 vs. 15.0 mmol/L [237.8 vs. 270.3 mg/dL]) at diagnosis. A larger percentage of female patients were at Tanner stages IV or V than male patients (68.4 vs. 44.9%). Additionally, 5.2% of females presented with DKA at diagnosis, compared with 12.6% of males. HDL cholesterol levels measured within 1 year of diagnosis were higher in female patients (1.02 vs. 0.92 mmol/L [39.4 vs. 35.6 mg/dL]) (Table 1).

Race/ethnicity, family history of type 2 diabetes, age- and sex-adjusted BMI z score, and C-peptide levels at diagnosis were not significantly different between females and males. Non-HDL cholesterol, LDL cholesterol, triglycerides, history of NAFLD, history of retinopathy, and microalbuminuria collected within 1 year of diagnosis were also not significantly different between the sexes.

In multivariable analyses, female sex was significantly associated with both younger age after adjusting for Tanner stage (P <0.0001) and more advanced Tanner stage (P <0.0001) after adjusting for age. Male sex was statistically associated with both DKA after adjusting for A1C (P = 0.035) and A1C after adjusting for DKA (P = 0.006) at diagnosis. The significant association between female sex and higher HDL levels persisted (P = 0.001) after adjusting for age and Tanner stage at diagnosis.

We studied sex differences in a large cohort of youth with type 2 diabetes and observed that females are diagnosed at a younger age and with more advanced pubertal development than males and present with milder clinical characteristics as indicated by lower A1C, lower frequency of DKA, and higher HDL cholesterol levels.

The earlier age at type 2 diabetes diagnosis in females is possibly the result of higher insulin resistance from earlier pubertal development than males. This hypothesis is supported by the higher percentage of females who presented at Tanner stages IV and V and previous studies demonstrating that diagnosis of type 2 diabetes is rare before the onset of puberty (14). Puberty is characterized by a growth spurt driven by a marked increase in growth hormone secretion that results in growth velocities averaging 9.5 cm/year in males and 8.25 cm/year in females (15). Growth hormone is known to induce insulin resistance, with a consequent rise in serum glucose levels, through complex mechanisms related in part to growth hormone’s facilitation of adipose tissue lipolysis (16) and hepatic gluconeogenesis (17). It is also possible that female sex hormones have a greater influence on type 2 diabetes risk through their influence on adiposity or other mechanisms, further contributing to the earlier age of onset in females. Girls are more insulin resistant than boys during early to midpuberty (16,18). This sex difference is partially explained by differences in adiposity (16,18), IGF-1 levels (16), as well as lower physical activity in females (16). On the other hand, obesity is associated with higher androgens, which relate to insulin resistance during puberty only in girls (19). These factors may explain some of the sex differences in age of onset of type 2 diabetes. An additional potential explanation for the milder presentation in females could be earlier medical attention prompted by health concerns related to the earlier pubertal development that is typically observed in girls compared with boys. However, longitudinal studies are needed to examine these relationships and account for additional confounders, including race/ethnicity and physical activity.

In addition to the implications for current knowledge of the physiopathology of type 2 diabetes, the observation that girls develop type 2 diabetes earlier has clinical significance. Screening guidelines for diabetes and prediabetes may need to be stratified by sex, and females with risk factors for type 2 diabetes should be monitored starting earlier in life than males.

We also observed that serum A1C was higher in males than in females. Given the well-known associations between A1C and diabetes outcomes, this finding suggests that males may be at risk for worse diabetes outcomes than females. The mechanisms underlying these sex differences are not clear.

HDL cholesterol levels were higher in females than in males. This finding may be related to known sex differences in HDL cholesterol levels. Pubertal development has been reported to improve the lipid profile, but, in the present data, the difference in HDL cholesterol was present after adjustment for pubertal development. However, we cannot exclude incomplete adjustment because we did not have Tanner stage information for all of the children in the study. Additionally, we could not account for the potential effect of glucose control on the lipid profile at the time of the assessment. The observed sex differences in HDL cholesterol are in contradistinction to findings from the TODAY (Treatment Options for Type 2 Diabetes in Adolescents and Youth) study, in which females had lower HDL cholesterol than males, whereas males were more likely to have elevated blood pressure early in the disease course (20). Interestingly, in the TODAY study, males were more likely than females to have been born large for gestational age and to have been exposed to maternal gestational diabetes (20). This finding supports the importance of more studies to gain a better understanding of the underpinnings of these observations.

Limited studies have compared clinical characteristics in males and females with type 2 diabetes. Strengths of our study include the large sample size in a single-center study and the availability of Tanner stage information evaluated by a pediatric endocrinologist for a large percentage of the sample. Our study has some limitations, however. Because of the cross-sectional design, we did not have follow-up information on the trajectory of glucose control or other diabetes outcomes. We were not able to assess birth history or other environmental, physical activity, or nutritional factors that may contribute to sex differences in diabetes risk. Additionally, we lacked detailed information on specific relatives and diabetes types (e.g., gestational diabetes or established type 2 diabetes in mothers of participants), which could be helpful in identifying genetic patterns in this multifactorial disease. Longitudinal studies with comprehensive assessment of in utero exposures, lifestyle factors, and the hormonal milieu in childhood will increase our understanding of pathophysiologic and prognostic differences between male and female youth with type 2 diabetes. Extended follow-up of participants will allow determination of whether the observed differences between males and females persist over time. In addition, it will be important to evaluate potential differences in diabetes outcomes such as risks for and types of chronic diabetes complications and comorbidities.

In conclusion, we observed that females are diagnosed with type 2 diabetes at a younger age, at a more advanced pubertal stage, and with milder disease characteristics than males. There are likely sex-specific factors affecting the development and progression of type 2 diabetes in children that are not yet well understood. Continued efforts to understand these processes could lead to more precise approaches to the diagnosis and treatment of type 2 diabetes in children.

This work was supported by a U.S. Department of Agriculture-Agricultural Research Service Current Research Information System award (F.B.) and National Institutes of Health grant R01 DK124395 (M.J.R.).

No potential conflicts of interest relevant to this article were reported.

B.C.R. contributed to the study design, data analysis, and interpretation; communicated results at local and international meetings; and wrote the first draft of the manuscript. M.A. collected the data, contributed to data analysis and interpretation, and edited the manuscript. M.T., S.M., F.B., and A.B. contributed to data interpretation and edited the manuscript. A.R. contributed to data collection and edited the manuscript. M.J.R. conceptualized the study; contributed to the study design, data analysis, and interpretation; and edited the manuscript. All the authors approved the final version of the manuscript. M.J.R. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Parts of the content of this article were presented in abstract form at the European Academy of Paediatric Societies’ virtual 8th Congress, 16–19 October 2020.