Abstract Study question What is the role of androgen excess as a contributing factor to insulin resistance and abnormal glucose metabolism (AGM) in women? Summary answer There was a positive association between early adulthood hyperandrogenaemia with AGM. Serum SHBG levels could help identifying women at risk for disordered glucose metabolism. What is known already It is commonly recognised that insulin resistance induces compensatory hyperinsulinaemia which promotes ovarian androgen secretion. Studies in rodents have also suggested that testosterone causes prolonged activation of androgen receptor in pancreatic islet β-cells, inducing insulin hypersecretion and eventually secondary β-cell failure, thus predisposing to type 2 diabetes (T2D). However, the exact physiology behind the association between androgens, insulin resistance and T2D in women is not well understood. Many previously published studies are limited by cross-sectional study design, unrepresentative clinic populations, as well as variying steroid hormone measurement methods and definitions of androgen excess. Study design, size, duration A prospective longitudinal population-based cohort (n = 5,889) to investigate whether serum levels of testosterone (T, measured using LC-MS/MS) and free androgen index (FAI) at ages 31 and 46 associated with AGM at age 46 years. After exclusion of pregnant women, users of hormonal intrauterine device, contraceptive pills, hormone therapy, minipills and statins, there were 4,421 women at age 31 and 4,457 women at age 46. At age 46 a two-hour OGTT was performed in 2,780 women. Participants/materials, setting, methods Serum fasting glucose and insulin, insulin resistance (HOMA–IR) and pancreatic β-cell function (HOMA–B) assessments were performed at ages 31 and 46. Elevated T levels (age 31: >2.3nmol/l; age 46: >1.7nmol/l) were defined according to the 97.5% percentile. T2D diagnoses were gathered from postal questionnaire at age 46, and verified and completed from the hospital discharge and national medication reimbursement registers. Impaired fasting glucose, impaired glucose tolerance or T2D were categorised as AGM. Main results and the role of chance At age 31, hyperandrogenic (HA) women displayed increased insulin resistance estimated by HOMA-IR (1.35±0.96 vs. 1.03±0.44, P = 0.05), increased insulin secretion estimated by HOMA-B (115.05±34.67 vs. 99.25±25.47, P = 0.006), and higher fasting insulin level (10.48±7.54 mU/l vs. 7.93±3.42 mU/l, P = 0.034) compared with normoandrogenic (NA) women, after BMI adjustment. At age 46, HA women had comparable HOMA-B levels (98.04±60.03 vs. 96.27±65.89, P = 0.93) but their fasting glucose (5.57±1.06 mmol/l vs. 5.37±0.77 mmol/, P = 0.07) and glycated haemoglobin levels (37.47±7.83 mmol/mol vs. 36.18±4.99 mmol/mol, P = 0.07) tended to be higher. Women in the highest T quartile (Q4 odds ratio [OR]=1.77, 95%CI: [1.14–2.76]) and in the two highest FAI quartiles at age 31 (Q4: OR = 3.61 [2.16–6.03] and Q3: OR = 2.11 [1.24–3.59]) had increased risk for AGM at age 46, independently of BMI, when compared with women in the lowest quartile. Similarly, women in the two highest FAI quartiles at age 46 had increased risk for AGM (Q4: OR = 2.91 [1.82–4.64]) when compared with women in the lowest quartile, after BMI adjustment. The three highest sex hormone-binding globulin (SHBG) quartiles inversely associated with AGM, both at ages 31 and 46, independently of BMI (at age 31: Q4: OR = 0.38 [0.24–0.62], at age 46: Q4: OR = 0.27 [0.17–0.44]). Limitations, reasons for caution We used only serum T as a marker of HA, even though other androgens, such as androstenedione and adrenal androgens have a place in the evaluation of androgenicity in women. Further studies of other large populations are needed to confirm our results. Wider implications of the findings This is the first longitudinal, general population based study to confirm a positive association between early adulthood hyperandrogenaemia with AGM in middle adulthood independently of confounding factors. Our results further suggest that SHBG levels could help to identify women at risk for AGM. Trial registration number NA