Abstract
BACKGROUND: Height growth is affected by longterm childhood morbidity. OBJECTIVES: To compare the growth curves of Indian children diagnosed with Type-1 diabetes mellitus (T1DM) and a control group of children without diabetes, and to see how parental height and disease severity affect the growth pattern. Subjects and Methods. The data came from: (i) the Sweetlings T1DM (STDM) study with 460 subjects aged 4-19 years, previously diagnosed with T1DM and followed for 2-6 (median 3) years, with repeat measurements of height and glycated hemoglobin (HbA1c), and (ii) the Pune School-Children Growth (PSCG) study with 1,470 subjects aged 4-19 years, and height measured annually for median 6 years. Height growth was modeled using SuperImposition by Translation and Rotation (SITAR), a mixed effects model which fits a cubic spline mean curve and summarizes individual growth in terms of differences in mean size, and pubertal timing and intensity. RESULTS: SITAR explained 99% of the variance in height, the mean curves by sex showing that compared to controls, the children with diabetes were shorter (by 4/5 cm for boys/girls), with a later (by 1/6 months) and less intense (-5%/-10%) pubertal growth spurt. Adjusted for mean height, timing and intensity, the diabetic and control mean curves were very similar in shape. SITAR modeling showed that mean HbA1c peaked at 10.5% at age 15 years, 1.0% higher than earlier in childhood. Individual growth patterns were highly significantly related to parental height, age at diabetes diagnosis, diabetes duration, and mean HbA1c. Mean height was 3.4 cm more per + 1 SD midparental height, and in girls, 2 cm less per + 1 SD HbA1c. CONCLUSION: The results show that the physiological response to T1DM is to grow more slowly, and to delay and extend the pubertal growth spurt. The effects are dose-related, with more severe disease associated with greater growth faltering.