Polycystic Ovarian Syndrome: Assessment of Approaches to Diagnosis and the Social Determinants of Risk

Presentation Number: LB SAT 75
Date of Presentation: April 1st, 2017

Adrian H Heald*1, Christopher J Duff2, Dave Holland3, Gabriela Y Cortes4, Mark Livingston5, Pensee Wu6 and Anthony A Fryer7
1University of Manchester, Greater Manchester, United Kingdom, 2University Hospital of the North Midlands, Stoke-on-Trent, United Kingdom, 3Keele University, Stoke-on-Trent, United Kingdom, 4The School of Medicine, IPN, Mexico City, Mexico, 5Walsall Manor Hospital, Walsall, United Kingdom, 6University of Keele, Stoke-on-Trent, United Kingdom, 7University Hospitals of North Midlands NHS Trust, Stoke-on-Trent, United Kingdom



Polycystic ovarian syndrome (PCOS) is one of the commonest endocrine disorders affecting women of reproductive age (1) and is a predisposing factor for type 2 diabetes (2).

We examined the specific tests that are done in primary care to diagnosis and monitor PCOS and investigated demographic factors influencing cardiometabolic profile in PCOS women.


In this primary care based study, 1797 women were identified from a pooled GP practice database. The search included all patients defined with PCOS or related terms. Records included demographic information, medical history (diagnoses & blood test results).


The diagnosis of PCOS was most commonly made in the age group 20-29 years with women at greater socioeconomic disadvantage (as measured by Townsend index) more likely to be diagnosed at a younger age (23.7 vs 25.2 years, p=0.008).

During the 24 months prior to the diagnosis of PCOS being made, 30.6% of women underwent a measurement of their serum total testosterone level while 29.7% had their serum SHBG measured. For serum oestradiol the corresponding statistics were 28.4% checked before diagnosis. For LH 45.6% were checked before diagnosis and for FSH 45.7%.

Fasting or random blood glucose and/or HbA1c was checked in 33.7% of women before diagnosis but in only 20.6% of women in the 24 months after diagnosis. Similarly, cholesterol level was measured in 10.1% of women before diagnosis and in only 9.2% of women in the 24 months after diagnosis.

There was a tendency for endocrine tests (oestradiol, LH, FSH, testosterone, SHBG) to be performed in the months prior to diagnosis in order to aid the diagnostic conclusion, with a peak in the weeks immediately before diagnosis. For plasma glucose / cholesterol, testing was performed more evenly over time.

A higher BMI (closest BMI to the point of diagnosis with PCOS) associated with a higher Townsend index (denoting greater socioeconomic disadvantage) (r2 =0.04) as did a lower SHBG (r2 =0.013). This relation held when adjustment was made for age, BMI, and systolic BP.

BMI was more likely to increase in women with a higher Townsend Index when analysed by quintiles of disadvantage. Specifically for the most disadvantaged women, BMI change over time was 18.6% compared with the most advantaged quintile at 13.7%.

Of all women diagnosed with PCOS 589/1797 (32.8%) were prescribed metformin, 66/1797 (3.7%) were prescribed antihypertensives, 39/1797 (2.2%) statins and 1141 / 1797 (63.5%) an oestrogen containing contraceptive pill or HRT at some point in the follow-up period.


Social disadvantage influences BMI and SHBG and also influences BMI change over time in women with PCOS. There is significant variation in testing approach in investigation of suspected PCOS.

Laboratory driven protocols for investigation of PCOS would provide the necessary information to enable the right therapeutic choice to be made, in management of this condition.


Nothing to Disclose: AHH, CJD, DH, GYC, ML, PW, AAF