Low Birth Weight and Development of Type 2 Diabetes in a Japanese Population

According to epidemiological studies in the U. K. and other countries, individuals with low birth weights often develop insulin resistance–based disorders (1–5). Among Pima Indians and Taiwanese school children, a higher prevalence of type 2 diabetes was observed in individuals with both low and high birth weights (U-shaped relationship of birth weight and diabetes) (6,7). We conducted a study to examine the relation of low birth weight with type 2 diabetes and insulin resistance in Japanese type 2 diabetic patients.

To obtain data on birth weight, weight at age 20 years, maximum weight in the past, height, and other variables, we asked 2,471 employees (2,259 men and 212 women) of two companies in Tokyo and 815 patients (514 men and 301 women) with type 2 diabetes who were treated at Saiseikai Central Hospital, both ≥40 years of age, to complete a questionnaire in April 2001. Among them, we selected 1,960 male employees (occupational cohort) and 164 male diabetic patients (hospital cohort), both aged 40–59 years, who could provide their birth weights, either through maternal and child health notebook records (issued by each municipal office) or their mother’s memory (if the notebook was missing), and who had agreed to participate in this study with informed consent. Birth weights <2,500, 2,501–3,699, and >3,700 g were defined as low, normal, and high, respectively. In the occupational cohort, subjects with known type 2 diabetes and with HbA_1c >6.5% were defined as diabetic patients. In the hospital cohort, current use of antihypertensive agents was defined as hypertension. In both cohorts, subjects who had a diabetic parent, sibling, and/or offspring were considered to have a family history of diabetes. The significance of difference of the results in the different groups was tested by χ² analysis or Student’s t test.

The prevalence of low birth weight in the 301 diabetic subjects, including those in the occupational cohort and the 1,823 nondiabetic subjects in the occupational cohort, was 18.6% (56 of 301) and 9.8% (178 of 1,823), respectively (P < 0.001), whereas the prevalence of high birth weight in the same groups was 9.3 and 11.6%, respectively (NS). Of the 56 diabetic subjects with low birth weight, 32 had a family history of diabetes (57.1%), and of the 178 corresponding nondiabetic subjects, only 25 had a family history of diabetes (14.0%) (P < 0.0001). Mean BMI at the age of 20 was 21.8 ± 3.3 and 20.6 ± 2.2 kg/m² (P < 0.01), and mean maximum BMI was 26.1 ± 3.3 and 24.8 ± 3.0 kg/m² (P < 0.01) in the same groups. The prevalence of hypertension in diabetic subjects with low and normal birth weight in the hospital cohort was 46.2% (18 of 39) and 23.4% (22 of 94) (P < 0.01), respectively, and the prevalence in those with high birth weight was 29.0% (9 of 31) (NS compared with low and normal weight groups), while mean BMI did not differ significantly among the three groups (23.0 ± 2.3, 23.9 ± 3.2, and 25.4 ± 4.2 kg/m²).

Our results showed that low birth weight was also associated with the development of type 2 diabetes in Japanese subjects and that not only genetic influences but also higher BMI in adulthood seemed to be important in the development of type 2 diabetes in individuals with low birth weight. The significantly higher incidence of hypertension in diabetic subjects with low birth weight compared with diabetic subjects with normal birth weight suggests that insulin resistance might be stronger in the former. Further prospective studies with a large number of participants using more appropriate indices of insulin resistance are required to clarify this association between birth weight and insulin resistance in individuals with type 2 diabetes.