Ethnic differences in manifestations of PCOS in 1010 Danish women
Dorte Glintborg1, Pernille Ravn2, Hanne Mumm1 & Marianne Andersen1
Background: Clinical manifestations, metabolic risk factors and diabetes risk may differ in ethnical subgroups of women with PCOS.
Material and methods: A total of 1010 premenopausal women were referred with the diagnoses hirsutism or PCOS during 19972008. The patients underwent clinical evaluation, hormone analyses and transvaginal ultrasound (US). Oral glucose tolerance tests (OGTT) (n=500) and ACTH tests (n=434) were performed in a subgroup of patients. The patients were divided according to ethnicity: Caucasian (CA, n=792), Middle East (ME, n=190), Asian (n=14), various (n=14).
Results (CA versus ME women): The median BMI (median (2575% quartiles)) for the study population was 26 (2232) kg/m2. CA women were significantly older (32 (2537) vs 24 (1832) years), less hirsute (total FG-score 11 (615) vs 16 (1122)) and had increased waist circumference (90 (78103) vs 82 (7398) cm) than ME women. BMI levels were not significantly different (P=0.07). The Rotterdam criteria were fulfilled in 55% of both populations, but PCO was diagnosed during US in 47% (314/672) CA versus 29% (50/172) ME women and clinical or paraclinical hyperandrogenaemia in 86 vs 93%, both P<0.01.
CA women had significantly increased cholesterol, triglycerides, LDL and blood pressure (systolic and diastolic) and smoked at a higher frequency (39.4 vs 22.5%). Glucose levels (fasting and during OGTT) were comparable in the two groups, but insulin levels was decreased in CA versus ME women, P<0.05. Diabetes or IGT was diagnosed in 12% and increased to 21% in obese patients (NS CA versus ME). Cortisol and 17-hydroxyprogesterone responses during ACTH tests were similar in the two ethnical groups.
Conclusion: In the present study, CA women had a more adverse cardiovascular profile than ME women, whereas no significant differences were found in glucose tolerance. The prevalence of the individual Rotterdam criteria differed significantly in the two study populations.