Relationship Between Short Stature and Postchallenge Glycemia in Pregnancy

Short stature has been identified as a risk factor for gestational diabetes mellitus (GDM), although the pathophysiological basis for this association is not clear (1,–3). In this context, we sought to evaluate the relationship between height and parameters of glucose homeostasis, including insulin sensitivity and β-cell function, in 671 Caucasian women undergoing a 3 h 100-g oral glucose tolerance test (OGTT) in late pregnancy. As previously described (4), the OGTT stratified subjects into three groups: GDM (n = 126), gestational impaired glucose tolerance (GIGT) (n = 117), and normal glucose tolerance (NGT) (n = 428). Insulin sensitivity and β-cell function were assessed by the insulin sensitivity index for OGTT (IS_OGTT) and the Insulin Secretion-Sensitivity Index-2 (ISSI-2), respectively.

After adjustment for age, current weight, and family history of diabetes, mean height decreased from NGT (164.1 cm) to GIGT (163.9 cm) to GDM (162.5 cm) (P = 0.027). On univariate analysis, height was weakly associated with IS_OGTT (r = −0.09, P = 0.022) but not with ISSI-2 (r = 0.04, P = 0.363). Notably, height was inversely correlated with postchallenge glucose measurements at 1 h (r = −0.13, P = 0.0006), 2 h (r = −0.13, P = 0.0009), and 3 h (r = −0.18, P < 0.0001) on the OGTT, but not with fasting glucose (r = 0.04, P = 0.319). On multiple linear regression analyses adjusting for age, family history of diabetes, and current weight, height emerged as a significant negative predictor of each postchallenge glucose value (1 h: β = −5.62, P < 0.0001; 2 h: β = −4.59, P< 0.0001; 3 h: β = −5.12, P < 0.0001). Importantly, none of these relationships were attenuated by further adjustment for IS_OGTT, ISSI-2, or even GDM, respectively. Because these relationships were independent of GDM, we queried whether they existed within the normoglycemic range by repeating the analyses in only the 428 women with NGT. Indeed, as before, significant negative correlations were observed between height and the normal 1-h glucose (β = −0.76, P = 0.0006), 2-h glucose (β = −0.92, P = 0.0003), and 3-h glucose (β = −0.96, P < 0.0001) values of the women with NGT.

This study confirms the independent negative association between height and postchallenge glycemia previously observed in other ethnically-homogeneous populations (1,–3). Importantly, however, we extend this literature by showing that these relationships remain significant after adjustment for GDM and, in fact, are present in the normoglycemic range. This observation potentially supports the notion of an underlying biological basis and underscores the need to examine the relevant pathophysiology. In this context, however, our multiple linear regression analyses suggest that neither insulin sensitivity nor β-cell function account for the relationship between height and postchallenge glycemia. The possibility of an artifactual basis for this relationship (wherein the standardized 100-g glucose load represents a greater glycemic stimulus in shorter women than in their taller peers) has been debated in the literature (1,5). At present, we can only postulate that short stature may be a modest marker or epiphenomenon identifying women at risk of GDM, with the underlying biology still unclear.

Our study is limited by the absence of assessment of leg length, which Moses and Mackay (2) have previously shown to be the precise anthropometric measurement that relates height to glycemia. In addition, we did not assess diet and socio-economic status, which are potential confounding factors. Further study will be needed to address the basis for the inverse relationship between stature and postchallenge glycemia in pregnancy.