Correlation Between Glucose Metabolism and Serum Steroid Hormones in Patients With Polycystic Ovary Syndrome

Xuelin Li; Tianyue Zhang; Shengxian Li; Yuying Deng; Lihua Wang; Tao Tao; Shujie Wang; Yanyun Gu; Weiqiong Gu; Jie Hong; Wei Liu; Weiqing Wang; Yifei Zhang


Clin Endocrinol. 2020;92(4):350-357. 

In This Article


Clinical and Metabolic Characteristics and Serum Steroid Hormone Levels of Patients With Different Glucose Metabolic Statuses

In the PCOS population of the two centres, the median age of patients was 26 years (IQR: 21–30), and the median BMI was 28.93 kg/m2 (IQR: 24.16–33.20). Of all the subjects, 59.4% (543/914) had normal glucose metabolism, and 40.6% (371/914) had abnormal glucose metabolism (including 29.3% [268/914] with prediabetes and 11.3% [103/914] with T2DM).

The women with abnormal glucose metabolism, which included prediabetes and T2DM, were older and had a higher BMI and WHR than those with normal glucose metabolism. Serum aldosterone (P < .001) and androstenedione (P = .011) levels and FAI (P < .001) were significantly increased in the PCOS patients with abnormal glucose metabolism, and SHBG (P < .001) was decreased, and these differences remained significant between the two groups after adjusting for age and BMI (Table 1). We further performed a subgroup analysis and found that after adjusting for age and BMI, the differences in aldosterone (P = .004), androstenedione (P = .004) and cortisone (P = .038) levels were significant between the normal glucose metabolism and prediabetes groups, and the differences in aldosterone (P = .031) and SHBG (P = .001) levels and FAI (P = .001) were significant between the normal glucose metabolism and T2DM groups, whereas the differences in corticosterone levels (P = .026) were significant between the prediabetes and T2DM groups (Supplemental table 2).

Relationship Between Serum Steroid Hormones and the Risk of Abnormal Glucose Metabolism

In the correlation analysis, FPG had a positive correlation with aldosterone (r = 0.152, P < .001), pregnenolone (r = 0.077, P = .022) and FAI (r = 0.238, P < .001); 2hPG had a positive correlation with aldosterone (r = 0.114, P = .001), androstenedione (r = 0.065, P = .048) and FAI (r = 0.230, P < .001) and a negative correlation with progesterone (r = −0.075, P = .023); glycated haemoglobin (HbA1c) was positively associated with aldosterone (r = 0.174, P = .007), oestrone (r = 0.150, P = .015), pregnenolone (r = 0.141, P = .023) and FAI (r = 0.350, P < .001), and negatively associated with progesterone (r = −0.152, P = .014; Table 2 and Figure 1).

Figure 1.

Heat map of the Spearman correlation analysis between serum steroid hormones and glucose metabolism parameters. Colours in the heat map indicate the pairwise Spearman correlation between the different data sets. The colour of each square represents the strength of the correlation, and the plus and minus indicate the direction of the correlation. FPG, fasting plasma glucose; 2hPG, 2-h postload plasma glucose; HbA1c, glycated haemoglobin; DHEAS, dehydroepiandrosterone sulphate; and FAI, free androgen index. *P < .05, **P < .01, ***P < .001

Table 3 shows the association between the parameters identified through the above correlation analysis and the risk of abnormal glucose metabolism in a logistic regression model using abnormal glucose metabolism as a dependent variable. In model 1, without modification for any confounding factors, aldosterone, oestrone and FAI (in quartiles) were positively correlated with the risk of abnormal glucose metabolism (P values were < .001, .020 and < .001, respectively), whereas progesterone was negatively correlated with the risk of abnormal glucose metabolism (P value was .013). After adjusting for age and BMI in model 2, only aldosterone, androstenedione and oestrone showed a positive correlation with the prevalence of abnormal glucose metabolism (adjusted P values for trend were .002, .021 and .003, respectively); further adjustment for fasting insulin levels in model 3 did not change the results of the correlation trend. As hypertension may influence aldosterone secretion, we further excluded patients with this condition (13.8% of patients with hypertension in normal glucose metabolism and 28.1% in prediabetes and T2DM) from the logistic analysis, and we found that aldosterone still showed a positive correlation with abnormal glucose metabolism. Refer to the regression model 3, the odds ratios (95% confidence intervals) of abnormal glucose metabolism were 1.49 (0.95–2.35), 1.59 (1.01–2.49) and 2.13 (1.35–3.36) for increasing quartiles of aldosterone level, 1.59 (1.03–2.45), 1.41 (0.91–2.18) and 1.84 (1.19–2.85) for increasing quartiles of androstenedione level, and 0.82 (0.54–1.25), 1.64 (1.08–2.49) and 1.14 (0.74–1.77) for increasing quartiles of oestrone level (Figure 2). Moreover, there was no association of pregnenolone, progesterone, testosterone and FAI with the risk of abnormal glucose metabolism.

Figure 2.

Associations of serum steroid hormones with the risk of abnormal glucose metabolism. Odds ratio for abnormal glucose metabolism events in quartiles of aldosterone (A), androstenedione (B) and oestrone (C) levels. The midpoint symbol indicates the median, and the line indicates the 95% CI for the estimate. Model 1 is unadjusted, Model 2 is adjusted for age and BMI, and Model 3 is further adjusted for age, BMI and fasting insulin levels. CI, confidence interval; Q1, quartile 1; Q2, quartile 2; Q3, quartile 3; Q4, quartile 4