Diurnal variation of sex hormone binding globulin and insulin-like growth factor binding protein-1 in women with polycystic ovary syndrome

Abstract

Objectives: The aim of this study was to examine (1) the diurnal variation in SHBG and (2) the inter-relationships of insulin, IGF-I, SHBG and IGFBP-1 over 24 hours in 10 women with anovulatory PCOS and compare them with weight-matched ovulatory controls.

Patients and methods: The two groups comprised 10 anovulatory women with PCOS (as defined by clinical, ultrasound and biochemical criteria) and 10 weight matched controls. Serum samples were taken at two-hourly intervals for 24 hours and stored for measurement of SHBG, IGFBP-1, insulin and IGF-I. Differences between the groups were compared using the Wilcoxon ranked paired tests of the individual peak and trough concentrations in each group. The variation in insulin, IGFBP-1 and SHBG concentrations over 24 hours was tested using two-way analysis of variance with the factors time and subject. Spearman's correlation coefficient was calculated from the subjects' median value over 24 hours.

Results: The median (interquartile range) body mass index (BMI) was 25.2 (22.2-29.3) in the PCOS group and 24.3 (23.2-25.7) kg/m2 in the control group. Serum testosterone (T) and LH levels were significantly raised in the PCOS group compared to the control group; T 3.8 (2.9-5.6) vs 1.9 (1.9-2.5) nmol/l (P < 0.007) and LH 12 (10-15) vs 4.1 (3.6-4.5) IU/I (P < 0.005) respectively. There was no diurnal variation in SHBG. The median (interquartile ranges) of the peak SHBG concentrations was lower in the PCOS group: 29.4 (14.9-39.4) vs 52.1 (39.4-61) nmol/l in the control group (P < 0.01). The fasting levels of insulin at 0600 h (median (interquartile ranges)) were not significantly different between the groups; 6.6 (5.4-9.8) and 6.2 (1.9-7.6) mU/l, respectively, although the peak median concentrations were significantly different; PCOS 66.1 (50.9-129.2) vs 40 (36.1-74.2) mU/l (P < 0.05). Two-way analysis of variance showed a diurnal variation in insulin concentrations in the control group (P = 0.001) but not in the PCOS group (P = 0.1). The diurnal variation in IGFBP-1 was similar in the two groups but the peak median levels were lower in the women with PCOS 54.9 (22.3-79.2) vs 71.5 (60.5-99.3) micrograms/l (P < 0.03). The decline in IGFBP-1 concentrations correlated with the increase in insulin concentrations. The IGF-I concentrations were similar in the two groups. There was a significant negative correlation between SHBG and insulin (P < 0.05) and between insulin and IGFBP-1 (P < 0.01).

Conclusion: This study demonstrates that there is no diurnal variation in SHBG concentrations and confirms the finding of a marked diurnal variation in the concentration of IGFBP-1. Women with PCOS who are anovulatory have an abnormal pattern of insulin secretion with an absence of diurnal variation compared to weight matched controls. This provides further evidence of the relative insulin resistance which is independent of weight found in women with anovulatory PCOS. The inverse correlations of insulin concentrations with SHBG and IGFBP-1 support the role of insulin as a possible regulator of the circulating levels of these binding proteins although the difference in the time course of their response makes it unlikely that they are co-regulated.