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



Study participants were enrolled from the Songnan Community, Baoshan District, Shanghai, People's Republic of China, in 2 phases as reported previously.[8] In phase 1 (June and July, 2008), we recruited 10,185 registered permanent residents aged 40 years or older to receive a screening examination. We tested fasting plasma glucose (FPG) and preliminarily categorized participants into 3 groups: normal glucose regulation, defined as a FPG of <100 mg/dL and with no history of diabetes; impaired glucose regulation, defined as a FPG from 100 to <126 mg/dL and with no history of diabetes; and diabetes, defined as a FPG of ≥126 mg/dL or with a history of diabetes. In phase 2 (June through August, 2009), we randomly selected participants from the 3 groups in a ratio of 1.0 (diabetes) to 1.2 (impaired glucose regulation) to 1.44 (normal glucose regulation), oversampling people with lower glucose levels because they might have a lower participation rate than those with higher glucose levels. A total of 4,012 participants were randomly selected and participated in phase 2 and had similar characteristics such as age, sex, and body mass index compared with those who did not participate (6,173 participants). Among the 3,455 study participants with blood and urine samples included in phase 2, we tested the concentrations of serum prolactin, and those who met the following criteria were excluded: 1) without the results of plasma glucose from an oral glucose tolerance test (OGTT) at 0 and 2 hours (n = 32); 2) without sufficient serum for prolactin measurement (n = 280); 3) with a history of pituitary disease or breast tumor (n = 226); 4) with hyperprolactinemia (serum prolactin higher than the laboratory reference: prolactin of >19.40 ng/mL for men and >26.53 ng/mL for women) (n = 122); and 5) premenopausal women (n = 418). Finally a total of 2,377 participants (including 1,034 men and 1,343 postmenopausal women) were included in the baseline analysis of prolactin and diabetes.[8] From March to May, 2013, the participants were invited to have a comprehensive follow-up examination. Of the 1,596 participants without baseline diabetes at phase 2, 11 died and 75 were lost during the follow-up. Finally, 1,510 participants were included in the present study. The flow chart of the study design is shown in Figure 1. The Committee on Human Research at Shanghai Jiao Tong University School of Medicine, Rui Jin Hospital, approved the study protocol, and all study participants provided written, informed consent. All procedures used in this study were in accordance with institutional guidelines.

Figure 1.

Flow chart of the study design, Shanghai, People's Republic of China, 2009–2013.


Information on sociodemographic characteristics, medical history, family history, and lifestyle factors was collected by using a standard questionnaire at baseline in 2009. Family history of diabetes was positive if any first-degree relative (mother, father, and siblings) had diabetes. Data on hormone therapy and parity (both women only) were also collected by asking the questions: "Have you ever had hormone replacement therapy?" and "How many liveborn children have you delivered?" Weight, height, and waist circumference were measured by experienced nurses according to a standard protocol. Body mass index was calculated as weight (kg)/height (m)2.

All participants underwent a 75-g OGTT (without diabetes present at baseline) after an overnight fast of more than 10 hours, and blood samples were collected at 0 and 2 hours. FPG, OGTT 2-hour plasma glucose, low density lipoprotein cholesterol, high density lipoprotein cholesterol, and triglycerides were measured by using the glucose oxidase method on an autoanalyzer (ADVIA-1650 Chemistry System; Bayer Corporation, Leverkusen, Germany). Serum insulin was measured by using an electrochemiluminescence assay (Roche Diagnostics, Basel, Switzerland), and hemoglobin A1c was determined by an automated high-performance liquid chromatography analyzer (Bio-Rad, Hercules, California). Serum prolactin was determined by using a chemiluminescent microparticle immunoassay by the Architect assay (Abbott Laboratories, Abbott Park, Illinois). The laboratory reference range of prolactin was 3.46–19.40 ng/mL for adult men and 5.18–26.53 ng/mL for adult women.[8] We also performed stability verification of the prolactin measurement. Prolactin concentrations were measured in the fresh serum samples randomly selected from 20 men and 20 women after serum extraction. After storage, the same samples were retested before measurement in the present study. The deviation of the results before and after storage was within the laboratory allowable range.[8]

Incident Type 2 Diabetes

Type 2 diabetes was determined according to the 1999 World Health Organization criteria supplemented by a definite diagnosis by physicians. Information included whether participants had been diagnosed with diabetes; with a positive response, further information was collected on the type of diabetes, date, and hospital where it was first diagnosed and whether by OGTT, and diabetes treatments. Participants who received antidiabetic therapies, those who responded positively to the OGTT as a FPG of ≥126 mg/dL, and/or the OGTT 2-hour plasma glucose of ≥200 mg/dL, and/or those who had a definite diagnosis of type 2 diabetes during the follow-up were diagnosed as having incident type 2 diabetes. For each participant, person-years of follow-up were calculated from the date of enrollment in the study to the date of reported physician diagnosis (data collected by questionnaire) or the date of the follow-up visit, whichever occurred first.

Statistical Analysis

Statistical analysis was performed by using SAS software, version 9.3 (SAS Institute, Inc., Cary, North Carolina). Continuous variables are summarized as means and standard deviations, and categorical variables are summarized as percentages. Sex-specific cutpoints were used to define quartiles of prolactin. For the differences of characteristics among participants across quartiles of prolactin, P values for trend were tested by using general linear regression for continuous data or logistic regression for categorical data. Cox proportional hazards analysis was used to evaluate the hazard ratios and 95% confidence intervals for each quartile of circulating prolactin, with the lowest quartile as the reference. Potential confounding factors were controlled for in multivariable analyses, including age, body mass index, waist circumference, family history of diabetes (yes or no), smoking status (never, former, or current), triglycerides, high density lipoprotein cholesterol, low density lipoprotein cholesterol, hormone use, and parity, which are known risk factors for type 2 diabetes. The interaction between the prolactin quartiles and sex on type 2 diabetes was tested by including a multiplicative interaction term in the models. To test for P values for trend across the prolactin quartiles, we modeled prolactin as an ordinal variable, coding with regard to its quartiles as 1, 2, 3, and 4. The reported P values were 2 sided, and P < 0.05 was considered statistically significant.