Women with polycystic ovary syndrome (PCOS) could be at higher risk of developing bipolar disorder, although the risk may be mitigated by use of metformin, research suggests. 

The investigators analyzed data from a large health insurance database to compare 7175 women with PCOS to 28,697 age-matched women without PCOS, classifying them into three subgroups: those treated with metformin, those treated with hormone therapy, and those who received neither treatment. 

During an 8-year follow-up, patients with PCOS had a significantly increased risk of developing bipolar disorder compared with unaffected controls, after adjusting for variables such as age, comorbidities, and differing treatment options. Patients with PCOS taking metformin had a significantly lower risk for bipolar disorder compared with patients who were not taking metformin. 

"This study shows that patients with PCOS are associated with an increased risk of developing bipolar disorder and that the use of metformin for the treatment of PCOS is associated with a reduction in its risk," the authors write. 

They suggest that patients with PCOS should "be screened for bipolar disorder in case of emotional instability." 

The study was published online December 5 in the Journal of Affective Disorders. 

Role of Comorbidities, Pharmacotherapy 

Previous research has found much higher rates of bipolar disorder in patients with PCOS compared with the general population, but some of the studies did not take into account comorbidities or the influence of drugs being used to treat PCOS, the authors write. 

PCOS is commonly treated with hormonal therapies (clomiphene or cyproterone) and hypoglycemic agents (metformin), they state. As some PCOS treatments (for example, clomiphene and cyproterone) can be associated with psychiatric symptoms, the influence of these agents "may complicate epidemiological studies of the relationship between PCOS and bipolar disorder." 

To examine the association, investigators drew upon data from the National Health Insurance Database of Taiwan, a nationally representative database of medical claims data. They followed women with PCOS (n = 7175) to see if they went on to develop bipolar disorder at higher rates than their counterparts without PCOS (n = 28,607). 

Patients newly diagnosed with PCOS between 2000 and 2012 were each matched with four control participants by age, index year, and Charlson Comorbidity Index (CCI) score. 

Participants were followed for roughly 8 years until they received a new clinical diagnosis of bipolar disorder, or the date of withdrawal from the insurance program, or the end of 2013. 

Participants were divided into three subgroups to investigate the potential influence of PCOS treatment on future development of bipolar disorder: patients treated with metformin, patients treated with hormone therapy (clomiphene or cyproterone), and patients who received neither treatment. 

In both cohorts, roughly one quarter of participants (25.9%) were younger than 25 years, a little more than half (52.7%) were aged 25-35 years, and the remaining participants were older than 35 years. 

Most participants (92.4%) had a CCI score of 0, 6.9% had score of 1-2, and 0.7% had a score of ≥ 3. 

The PCOS group had a higher prevalence of several physical comorbidities, including asthma, migraine, irritable bowel syndrome, head injury, substance abuse, obesity, diabetes, hypertension, and dyslipidemia, compared with the control cohort (P < .01 for all). 

Eightfold Higher Risk 

The incidence rates of bipolar disorder were more than eight times higher in the PCOS cohort compared with the control cohort (1.05 vs 0.12 per 1000 person-years, respectively). 

Moreover, after adjusting for covariates (age, CCI score, comorbidities, and different treatment options), the hazard ratio (HR) for the development of bipolar disorder was significantly higher for the PCOS than the control cohort (HR, 8.29). 

There were no significant differences in HRs between the various age groups; however, compared with patients with a CCI score of 0, the risk for development of bipolar disorder was significantly higher in patients with a baseline CCI score of 1-2 and ≥ 3 (HR, 4.54 and 1.66, respectively) after adjustment for all other covariates. 

Moreover, compared with patients without the included comorbidities, the risk for bipolar disorder was significantly higher in patients with PCOS and those comorbidities, after adjustment for covariates (HR, 2.25). 

The Role of Inflammation 

Most participants, 95.7%, in the control group were not receiving either of the two medication protocols — hormone therapy or metformin — compared with 13.7% in the PCOS group. 

In particular, a considerably smaller percentage of the control cohort was receiving hormone therapy (0.6% vs 24.8%) and metformin (3.7% vs 61.3%) compared with the PCOS group. 

The researchers studied the incidence of the development of bipolar disorder in each of the medication subgroups in the PCOS cohort. 

Those taking metformin had a significantly reduced risk for bipolar disorder compared with non-users, after adjusting for other variables (HR, 0.36). 

Participants with PCOS receiving hormone therapy also had a lower incidence rate of bipolar disorder development compared with non-users, but the risk was not significant (HR, 0.68). 

The researchers suggest several possible mechanisms to account for the higher association between PCOS and the development of bipolar disorder. 

Inflammation may "play an important role in the link between hyperinsulinemia and hyperandrogenism of PCOS and metabolic dysregulation of PCOS, as well as psychiatric disorders," they speculate. 

Moreover, they hypothesize that, "from a psychological point of view, the chronic burden of PCOS symptoms...can affect the psychological well-being of an individual, which can exacerbate mood swings in patients with bipolar disorder subclinical vulnerability and trigger the onset of the disease." 

The protective mechanism of metformin against bipolar disorder development in patients with PCOS may "result from an improvement in hyperinsulinemia, thereby decreasing androgen synthesis and decreasing systemic inflammation associated with PCOS." 

One study limitation the researchers note is that they did not assess PCOS severity, "which reduced the chances of showing the effects of PCOS severity on bipolar disorder development." 

They conclude that further studies are necessary to "elucidate the mechanisms linking PCOS and bipolar disorder and to determine if there is a PCOS intervention other than metformin that could reduce this risk." 

Beyond the screening of patients with PCOS and emotional instability for bipolar disorder, they recommend "treatments in an integrated mind–body perspective" as potentially "beneficial" for these patients.