Hyperandrogenism, postprandial hyperinsulinism and the risk of PCOS in a cross sectional study of women with epilepsy treated with valproate

Abstract

Among a sample of 43 women with epilepsy treated for at least 2 years with valproate (n=22) or other antiepileptic drugs (AEDs) (n=21), polycystic ovary syndrome (PCOS) was diagnosed in three women, two of them were treated with valproate. Although the rate of PCOS and of menstrual disturbances, weight body mass index (BMI) and waist to hip ratio as well as fasting blood glucose levels, fasting insulin, proinsulin and C-peptide values was similar in this small sample of women treated with valproate and other AEDs, valproate exposure was associated with higher androgen levels, higher postprandial (pp) insulin and proinsulin levels, as well as lower cholesterol and low density lipoprotein (LDL) cholesterol levels. The pronounced increase in pp insulin levels during VPA treatment may indicate an effect of the fatty acid derivate VPA on pancreatic islet cells.

Introduction

Reproductive endocrine disorders are common in women with epilepsy, and recent papers have evoked the possible increased frequency of polycystic ovary syndrome (PCOS) in epileptic women, and the role of AEDs, more particularly valproate, in the expression of this syndrome.

PCOS is the commonest endocrine disorder affecting women (Balen, 1999). It covers a series of morphological and functional changes in the ovaries and the hypothalamic–pituitary–ovarian axis, which combine with greater or lesser frequency with several clinical characteristics. This makes it difficult to establish definitive criteria for diagnosing PCOS and since the first description by Stein and Leventhal (1935), a number of definitions have been proposed. The most widely accepted clinical definition is the association of ovulatory dysfunction with clinical evidence of hyperandrogenism (hirsutism, acne, male pattern baldness) and/or hyperandrogenaemia, at the exclusion of related disorders such as hyperprolactinaemia, thyroid disorders and non classical adrenal hyperplasia (Zawadski and Dunaif, 1992).

PCOS must be distinguished from polycystic ovary (PCO) that is a morphological finding established on ultrasound. PCO is a largely asympomatic condition and a frequent and commonly incidental finding even in an unselected population.

Most authors have stressed that several interlinking factors affect the expression of PCOS including both genetic predisposition and environmental factors (Balen, 1999, Franks et al., 1997). One explanation is that those various causes ultimately result in androgen excess and/or hyperinsulinaemia with insulin resistance, and are thus responsible for the onset of PCOS.

Changes in weight seem to affect the expression of PCOS in women with PCO, with a gain in weight being associated with a worsening of the symptoms of PCOS (Balen, 1999).

Epilepsy per se may also play a significant part in the expression of PCOS (Balen, 1999) with a possible additional effect of AEDs (Mattson and Cramer, 1985). Isojärvi and colleagues have suggested that valproate may cause PCO and other reproductive endocrine abnormalities, as a result of valproate induced weight gain (Isojärvi et al., 1993, Isojärvi et al., 1996, Isojärvi et al., 1998), though they have more recently questioned the role of obesity and elevated serum insulin levels (Vainiopää et al., 1999; Isojärvi and Tapanainen, 2000). In 1997, Murialdo et al. found that 40% of women on polytherapy including valproate had PCO in contrast to 13% of women on polytherapy not including valproate. Furthermore in that study, hyperandrogenism was reported in 44% of women on valproate, and in 4% of women on other medications. In their 1998 paper, Murialdo et al. found that 63.6% of valproate treated women had luteal progesterone levels consistent with impaired ovulation. In a more recent report, Bauer et al., 2000 found no increased rate of PCOS in women treated with valproate and thus the controversy remains.

Valproate is a widely used antiepileptic drug which has a broad spectrum activity covering both generalised and partial epilepsies, and at present it is the drug of first choice in the treatment of idiopathic generalised epilepsies. Whether there is a possible risk of increased frequency of PCOS and related symptoms that could outweigh its benefits in a population of young women with epilepsy and then justify, as suggested by Isojärvi and colleagues, the selection of an alternative drug, is still uncertain and needs further investigation.

This cross-sectional study aimed at detecting whether valproate and others AEDs exert differing effects on various endocrine and metabolic parameters and ovarian morphology in epileptic women, and to investigate the possible association between valproate, obesity and PCO/PCOS.