Metabolic inflexibility identified as a feature of PCOS

Andrea E. Dunaif, MD

Metabolic flexibility is the capacity to adapt fuel oxidation to fuel availability.  It is characterized by the ability to switch from fat oxidation in the fasted state to carbohydrate oxidation with insulin stimulation in the fed state.  Metabolic flexibility can be assessed by measuring the difference between fasting and insulin-stimulated respiratory quotient (RQ).  Metabolic inflexibility is defined as a reduction in the change in RQ  and is a feature of insulin resistant states such as type 2 diabetes, impaired glucose tolerance and obesity.  In the majority of studies, differences in metabolic flexibility are accounted for by differences in glucose uptake, suggesting that defects in glucose transport rather than intrinsic decreases in glucose oxidation account for metabolic inflexibility.  Decreases in glucose transport in these common insulin resistant states are the result of impaired metabolic insulin receptor signal transduction.

Di Sarra and colleagues examined insulin action and metabolic flexibility using state-of-the-art techniques in a large cohort (almost 90) of women with polycystic ovary syndrome (PCOS) diagnosed by Rotterdam criteria.  They further stratified this cohort based on hyperandrogenemia.  The control subjects were a historical sample of nonobese women with regular menses, normal ovarian morphology and no clinical evidence of hyperandrogenism whose data were used to determine reference ranges for hormonal and metabolic parameters.  The criterion for insulin resistance given in the authors’ other publications was a glucose disposal rate below the 25^th percentile for these nonobese control women; criteria for reference intervals for other parameters were not included in the publication. 

The authors reported that insulin-mediated glucose disposal and metabolic flexibility were significantly reduced in the hyperandrogenemic (based on elevated free testosterone levels by dialysis) compared to the normoandrogenemic women with PCOS.  In multiple regression analyses, metabolic flexibility was independently associated with insulin sensitivity, basal substrate oxidation rates, free testosterone and insulin-mediated suppression of free fatty acid levels.  The authors conclude that metabolic inflexibility is a feature of PCOS and that both insulin resistance and hyperandrogenism may contribute to defects in insulin-mediated glucose oxidation.

The strengths of the study are the large sample of women with PCOS studied using rigorous metabolic techniques.  It is the first study to examine metabolic flexibility in PCOS and the suggestion that androgen excess independently contributes to metabolic inflexibility in affected women is novel.  However, there several concerns about the study design.  First, stratifying the subjects by hyperandrogenemia rather than PCOS phenotype could result in combining several phenotypes in the hyperandrogenemic and normoandrogenemic subgroups.  It is well accepted that insulin sensitivity and other metabolic abnormalities differ by PCOS phenotype with women with the classic phenotype diagnosed by NIH criteria having the most severe defects.  In contrast, ovulatory women with polycystic ovary morphology (PCO) and hyperandrogenemia as well as anovulatory women with PCO and normoandrogenemia have minimal to absent metabolic defects.  Further, approximately 10% of women with the classic PCOS phenotype have clinical but not biochemical hyperandrogenism.  The study would be strengthened by stratifying by PCOS phenotype rather than by hyperandrogenemia.

Second, the hyperandrogenemic group was significantly more obese than the normoandrogenemic group.  Accordingly, the independent impact of adiposity on insulin action and metabolic flexibility could have confounded the results.  The authors did adjust for obesity in the regression analyses, however, the relationship between obesity and metabolic endpoints is not always linear as my group has shown in PCOS (Dunaif et al. Diabetes 41:1257, 1992).  This limitation could be addressed by including simultaneously studied groups of control women of comparable weight to the hyperandrogenemic and normoandrogenemic groups, respectively, as has been done in many previous studies. 

There are also alternative interpretations of the study findings.  It is equally plausible that the hyperandrogenemic group had higher testosterone levels as a consequence rather than a cause of their insulin resistance, although there are studies to suggest that androgens play a modest role in sustaining insulin resistance in PCOS.  Further, the authors suggest that metabolic inflexibility may reflect mitochondrial dysfunction in PCOS.  However, Eriksen and colleagues (JCEM 96:E1298, 2011) found that mitochondrial function was normal in cultured myotubes from insulin resistant women with PCOS.  In summary, the study of Di Sarra and colleagues suggesting metabolic inflexibility is novel feature of PCOS is provocative.  Further studies are needed to determine whether metabolic inflexibility is independent of obesity and insulin resistance as well as the role androgens in the pathogenesis of this abnormality in PCOS.