Background: Many clinical and biochemical criteria are used to assess hyperandrogenism in subjects with polycystic ovary syndrome (PCOS). Standard indicators used to confirm hyperandrogenism are based predominantly on western data. Whilst the phenotype of PCOS has ethnic specific variation, specific cutoffs for hyperandrogenism in South Asians have not been defined.

Objectives: To evaluate the effectiveness of modified Ferriman-Gallwey score (FG Score), serum total testosterone and free androgen index (FAI) in the assessment of hyperandrogenism in PCOS.

Materials and methods: A case control study was conducted on 100 women aged 20-45 years (mean age= 30) attending a specialized endocrine clinic in Colombo, Sri Lanka from 1st January 2010 to 1st June 2013. Confirmed cases with PCOS (Rotterdam criteria 2003) were age matched for controls from healthy volunteers.

Recommended cut-offs for hyperandrogenism: FG score≥ 8, testosterone (T)> 3.5 nmol/L and FAI> 5 were applied and receiver operating characteristics (ROC) curves were drawn to compare the diagnostic power of each parameter.

Results: 50 cases with PCOS and 50 controls were studied. Cases versus controls had significantly greater FG score, testosterone and FAI: median FG= 10 vs 3, mean testosterone 2.762±1.78 vs 1.045±0.40 (p= 0.0001), mean FAI 7.31±7.55 vs 3.64±4.87 (p= 0.01); 76% cases and 4% controls had FG score≥ 8 (p= 0.0001), 30% cases had elevated (T) with none among controls (p= 0.00001), 43.3% cases and

14.7% controls had FAI≥ 5 (p= 0.002). The diagnostic power of serum testosterone was greater than that of FAI in…