Circulating Prolactin and Risk of Type 2 Diabetes

A Prospective Study

TiangeWang; Yu Xu; Min Xu; Guang Ning; Jieli Lu; Meng Dai; Baihui Xu; Jichao Sun; Wanwan Sun; Shenghan Lai; Yufang Bi; Weiqing Wang*


Am J Epidemiol. 2016;184(4):295-301. 

In This Article


During a mean follow-up period of 3.7 years, 189 cases of diabetes were documented (incidence rate: 3.4/100 person-years). Baseline characteristics of study participants by quartiles of prolactin are presented in Table 1 . In men and in postmenopausal women, compared with the participants in the lowest quartile, those in the highest quartile were older (P trend ≤ 0.041). In men, a higher prolactin level was significantly associated with a lower proportion of current smokers and a lower mean level of low density lipoprotein cholesterol (all P's ≤ 0.033). In postmenopausal women, FPG and 2-hour OGTT plasma glucose levels gradually decreased across quartiles of prolactin (both P's ≤ 0.008). The prolactin concentrations were higher in women than in men and were lower in participants who developed diabetes (men: mean = 8.92 (standard deviation (SD), 3.16) ng/mL; median, 8.57 (interquartile range, 6.71–11.12) ng/mL; women: mean = 9.40 (SD, 4.36) ng/mL; median, 8.38 (interquartile range, 6.36–11.19) ng/mL) than in those who were free of diabetes at follow-up (men: mean = 8.96 (SD, 3.22) ng/mL; median, 8.42 (interquartile range, 6.59–10.64) ng/mL; women: mean = 9.93 (SD, 4.14) ng/mL; median, 9.12 (interquartile range, 6.94–11.76) ng/mL), but no significant difference was observed (Web Table 1 available at

The associations between circulating prolactin and incident type 2 diabetes are shown in Table 2 . In men, the crude rates per 100 person-years of incident diabetes were 3.1%, 3.6%, 3.6%, and 4.0% across the lowest quartile to the highest quartile of prolactin, respectively. No significant association between prolactin and incident diabetes was found in men. In women, the crude rates per 100 person-years of incident diabetes decreased from 4.6% to 2.5% across serum prolactin quartiles. The age-adjusted hazard ratios for incident diabetes were 0.58 (95% confidence interval: 0.35, 0.97) for the second, 0.54 (95% confidence interval: 0.32, 0.90) for the third, and 0.45 (95% confidence interval: 0.26, 0.77) for the fourth quartiles of prolactin, compared with the first quartile. In the multivariate-adjusted model, the hazard ratio for incident diabetes in the highest compared with the lowest quartile of prolactin was 0.48 (95% confidence interval: 0.26, 0.90). There was a significant interaction between sex and prolactin levels for the association with incident diabetes (P interaction = 0.028).

Given that women in the lowest quartile of prolactin (n = 378) had the highest incidence of type 2 diabetes, we further divided these women into finer subdivisions (quartiles) to further estimate the characteristics of this group. As shown in Web Figure 1, in the 378 women, the crude rate per 100 person-years for type 2 diabetes was 6.2%, 3.7%, 2.7%, and 4.5% across the 4 subdivisions of the lowest quartile of prolactin, respectively. Women with circulating prolactin less than 5 ng/mL showed the highest incidence, suggesting that women with prolactin less than 5 ng/mL might be at particular risk for incident type 2 diabetes.