Oestradiol, amenorrhoea and lipids in adolescent girls with eating disorders: do they affect long-term cardiovascular risk?

Abstract

Anorexia nervosa is a potentially life-threatening disorder that has its onset during adolescence. Mortality rates are 12 times higher than those for women in the general population 1. While the acute medical complications of this disorder are well described, not much is known about the long-term medical morbidities such as the impact of the disease on bone and cardiovascular health. In this issue of the journal, Swenne describes an association between plasma cholesterol and non-HDL cholesterol levels with menstrual status and weight loss in a large cohort of adolescent girls with eating disorders 2. Elevated total cholesterol levels were found in 38% of premenarcheal girls, 32% of those with secondary amenorrhoea and 17% of regularly menstruating girls. Total cholesterol concentrations were inversely correlated with oestradiol levels even after correcting for BMI Z-scores. More importantly, non-HDL cholesterol, a newly recognised measure of atherosclerosis risk, which can be assessed using nonfasting samples 3, showed a similar inverse relationship to oestradiol levels. Some of the participants had total cholesterol and non-HDL cholesterol concentrations high enough to warrant concern for future cardiovascular health. The author suggests that oestrogen may mediate the effect of starvation on cholesterol and concludes that because of concerns about future cardiovascular health, restoring menstruation is an important goal of treatment. Amenorrhoea in anorexia nervosa develops as a result of hypothalamic–pituitary–ovarian suppression secondary to an energy deficit, possibly mediated via leptin 4. Serum levels of luteinising hormone, follicle-stimulating hormone and oestradiol are low, and patients may present with delayed puberty, primary amenorrhoea as well as secondary amenorrhoea. Weight gain is associated with resumption of spontaneous menses and is recognised by many as an important objective measure of return to biologic health. Resumption of menses is accompanied by increases in gonadotropins and serum oestradiol levels, and an oestradiol level above 110 pmol/L (30 pg/mL) has been found to be predictive of resumption of menses within 6 months 5. Hypercholesterolaemia, first described in anorexia nervosa by Klinefelter in 1965 6, is unexpected, given the fat-restricted diets consumed by patients with this disorder. Elevations have been found in total cholesterol, triglycerides, LDL cholesterol, HDL cholesterol as well as apolipoproteins A, B, C2 and C3. The aetiology of hypercholesterolaemia in eating disorders remains unresolved. Possible mechanisms include increased synthesis of cholesterol, delayed metabolism of cholesterol as well as increased mobilisation of lipids from adipose stores to meet energy requirements. Most studies have shown that in anorexia nervosa, hypercholesterolaemia improves or resolves with weight restoration. Endogenous oestrogen has important metabolic effects on lipid metabolism, affecting enzymes of cholesterol and lipoprotein synthesis and degradation 7. In premenopausal women, endogenous oestrogen production is associated with low levels of LDL cholesterol and high levels of HDL cholesterol. Oestrogen also inhibits LDL receptor degradation via suppression of a recently recognised protein, PCSK9 7, 8. These effects explain both the lower cholesterol levels in females compared to males and the resulting relative reduction in atherosclerosis seen in premenopausal women. Endogenous oestrogens affect not only the lipid precursors but also the inflammatory mechanisms of atherosclerotic cardiovascular disease. The long-term cardiovascular implications in adolescent girls with eating disorders are less clear and will depend on the duration, and severity of the oestrogen deficit and the lipid abnormalities. The associations found in this study are interesting but, because of its cross-sectional design, causality cannot be established. Both amenorrhoea and hypercholesterolaemia may be secondary to malnutrition and not necessarily related to each other. This study was not designed to assess future cardiovascular health nor was it designed to assess the effect of resumption of menses on cardiovascular health. The conclusion of the author that ‘re-establishing menstruation is an important goal in the treatment of eating disorders in order to avoid dyslipidaemia and the risk of future cardiovascular disease’ may be true, but further prospective longitudinal studies will be required to validate this hypothesis. We do know that resumption of menses is important for bone health – subjects with anorexia nervosa who gain weight but have not resumed menses have lower bone mineral density than those who have both gained weight and resumed menses 9. We do not yet know the effect of resumption of menses on the risk of cardiovascular health but can speculate that restoration of oestrogen levels expected with restoration of normal menstruation may have beneficial effects. The findings of this study are sufficiently interesting and important to invite further investigation either to support or refute the hypothesis that cardiovascular risk is increased in eating disorders and ameliorated by resumption of menses.