The prevalence of hyperprolactinaemia and association with markers of autoimmune thyroid disease in survivors of the Whickham Survey cohort

Abstract

Objective: Few data exist on the prevalence of hyperprolactinaemia in the community. This study was intended to determine the prevalence of hyperprolactinaemia in a sample closely matched to the current British population aged 38 years and over.

Design and patients: The 1877 survivors at the 20-year follow-up of the Whickham Survey were a cross-sectional sample of the community aged 38 years and over. Serum was frozen and stored at -30 degrees C from 90% of the survivors (751 men, 924 women, median age 58 years (range 38 to 93 years)) who participated in the follow-up survey.

Measurements: Two years after the follow-up survey, serum prolactin concentrations were measured by ELISA/1 step sandwich assay (reference range < or = 600 mU/l in men and women). A repeat prolactin measurement was made in those subjects who had prolactin levels within the top 2.5% of men and women in this sample.

Results: At screening, 0.7% of the men and 2.5% of the women had serum prolactin levels greater than 600 mU/l. For men, 2.5% were above 400 mU/l. The prevalence of hyperprolactinaemia, if defined as greater than 400 mU/l in men and greater than 600 mU/l in women on repeat testing, was 1.4% in the men and 1.2% in the women. The aetiology in men was prolactin-raising drugs (n = 3), renal failure (n = 1), microprolactinoma (n = 1), and unknown (n = 2), and in women it was prolactin-raising drugs (n = 7), microprolactinoma (n = 1), and unknown (n = 1). Logarithmic transformation of serum prolactin concentrations produced Gaussian distributions with 95% reference ranges of 60-430 mU/l in men and 40-560 mU/l in women. No significant relationship was found in either sex between hyperprolactinaemia and age or evidence of autoimmune thyroid disease at either survey. In women, there was no association with age, distance beyond the menopause or duration of reproductive years but prolactin levels were slightly higher in those on oestrogen therapy (geometric mean prolactin 226 mU/l compared to 178 mU/l; t-test on log prolactin t = 3.79; P < 0.0001).

Conclusions: This study has demonstrated that a gender-related reference range for serum prolactin is necessary. Pituitary pathology is not common and screening with measurement of serum prolactin is not warranted in middle-aged and elderly subjects. In asymptomatic subjects with modestly elevated serum prolactin levels (< 3 SD above the mean), extensive pituitary imaging and investigation is unwarranted. Autoimmune thyroid disease was not a significant cause of hyperprolactinaemia in this sample.