Women with BD are often treated with psychotropic agents, including mood stabilizers, atypical antipsychotics, and antidepressants, whose effects on reproductive function and IR are not fully understood. Medications may impact weight and endocrine abnormalities in a variety of ways that contribute to IR. For example, it has been theorized that valproate can influence the development of menstrual abnormalities by decreasing estrogen levels, increasing luteinizing hormone and increasing testosterone . The increase in testosterone can lead to an arrest in maturation of ovarian follicles, leading to the development of polycystic ovaries. Additionally, valproate may cause an increase in leptin resulting in increased body weight , or lead to glucose-stimulated insulin secretion by pancreatic cells . Weight gain itself, not specific to valproate use, may lead to menstrual abnormalities or IR. In our studies, we demonstrated a correlation between BMI and both testosterone and insulin levels in women with BD . Weight gain on psychotropic medications also predicted new onset menstrual abnormalities . Thus, weight gain (whether it is inherent to BD or a result of any weight-liable mood stabilizer) may contribute to IR, particularly in women.
Here, we review the evidence for the association between certain medications implicated in IR.
Valproate has been reported to have associations with PCOS in women with epilepsy and BD. Concerns over an association between valproate and PCOS first arose in val-proate-treated women with epilepsy who were found to have high rates of PCOS . As valproate is commonly used to treat women with BD, questions about its possible impact on endocrine and reproductive function in psychiatric populations emerged. Several cross-sectional studies have found increased rates of PCOS, menstrual abnormalities and/or hyperandrogenism in bipolar patients using valproate [22, 57, 58]. A study of 38 bipolar women  receiving either valproate or lithium monotherapy for at least 2 years demonstrated higher rates of menstrual abnormalities in patients receiving valproate (50%) than in patients receiving lithium (15%). Menstrual abnormalities were more prevalent in overweight/obese patients than in lean patients. Free testosterone and androstenedione levels were significantly higher than the reference range in valproate-treated patients, and luteinizing hormone was elevated in both groups. A similar study demonstrated that women receiving valproate had a significantly greater rate of menstrual abnormalities (47%) than women receiving non-val-proate therapy (13%) and control women (0%) . Forty-one percent of women with BD taking valproate had PCOS. A larger study of 300 women in the STEP-BD study found 10.6% (n = 9) of 86 patients treated with valproate had new onset menstrual cycle irregularities and hyperandrogenism within the first year of valproate use, most of of whom were also obese and had IR . A follow-up study demonstrated reversal of PCOS reproductive features in 3 of 4 women upon discontinuation of valproate, while symptoms continued in 3 women continuing valproate .
Our own findings on this topic have yielded mixed results about the association of valproate with PCOS features; however we have consistently shown high rates of menstrual abnormalities in women with BD, even preceding diagnosis and treatment for the disorder. A pilot study of 22 women with BD, receiving lithium monotherapy, valproate monotherapy, or lithium-valproate combination therapy did not reveal a significant association between PCOS and valproate or lithium therapy . All patients on lithium monotherapy or combination therapy and 60% of patients on valproate monotherapy reported menstrual disturbances. There were no significant differences in BMI or hirsutism, and hormone levels were within normal limits for all 3 groups. None of the subjects met NIH-defined criteria for PCOS. In a follow-up cross-sectional study , we examined 80 reproductive-aged women who were receiving valproate therapy or non-valproate therapy for BD. Fifty-two of the 80 women (65%) reported current menstrual abnormalities, 40 of which (50%) reported one or more menstrual abnormalities that preceded the diagnosis of BD. Fifteen women (38%) reported developing menstrual abnormalities since beginning treatment for BD, 14 of whom developed abnormalities since treatment with valproate. No significant differences were observed between valproate and non-valproate groups in mean levels of free or total serum testosterone. Three of the 50 women (6%) taking valproate met the criteria for PCOS, compared to 0% of the 22 taking other mood-stabilizer medications.
We conducted the first longitudinal study of this relationship, evaluating reproductive endocrine and metabolic markers in 25 women treated for BD over a 2-year time period . At baseline, 10 (40%) women were currently undergoing treatment with valproate. Consistent with other published reports described above, 41.7% of subjects reported current oligomenorrhea, while 40% reported oligomenorrhea before starting medication. Rates of oligomenorrhea and clinical hyperan-drogenism did not differ by type of medication use. Eighty percent of all subjects had a high homeostatic model assessment of IR (HOMA-IR) at baseline. Valproate use was associated with an increase over time in total testosterone. In contrast to other studies, we did not find co-occurrence of new-onset menstrual abnormalities with hyperandrogenism. We concluded that oligomenorrhea, hyperandrogenism, and IR are common in women with BD, and increases in androgen hormones may be related to valproate use .
Lithium could theoretically influence IR in women through weight gain or through effects of hypothyroidism. Lithium exposure increases the incidence of hypothyroidism. Clinical and subclinical hypothyroidism is more commonly seen in women than men with BD and in the general population, and the risk increases with age [61-63]. Hypothyroidism can cause alterations in lipid metabolism, weight gain, and menstrual cycle abnormalities, which in turn could impact IR.
Lithium treatment in patients with BD has been associated with weight gain in numerous studies [64-68] and has been found to range from 5 kg within 1-2 years to 4.5-15.6 kg over 2 years [69, 70]. A recent animal study found that lithium increased gastrointestinal weight of male and female rats but only increased total body weight in females . The mechanisms for lithium-induced weight gain remain unclear despite several conjectures. Lithium appears to exert insulin-like activity on carbohydrate metabolism, leading to increased glucose absorption in adipose tissue [72-75]. Lithium may also have direct effects on the hypothalamus to stimulate appetite and/or thirst and can also result in increased fluid retention . Although lithium treatment can result in significant weight gain and/or hypothyroidism, long-term lithium treatment does not appear to be associated with IR or increased risk of developing diabetes mellitus . In fact, one study suggested that lithium exerts anti-diabetic effects by lowering blood glucose levels in a manic-depressive female with adult onset diabetes mellitus .
Atypical antipsychotic (AAP) treatment is also associated with a greater risk of metabolic abnormalities, such as dyslipidemia, weight gain, and new-onset type 2 diabetes melli-tus . Sex-specific impacts of atypical antipsychotics on IR have not been elucidated. Recent data suggest that AAPs may be more weight-liable medications than valproate . In general, clozapine and olanzapine treatment are associated with the greatest risk of weight gain, while risperidone, quetiapine, ziprasidone, aripiprazole, amisulpride and zotepine have relatively lower levels of risk [80, 81]. In several studies examining IR via glucose tolerance testing, patients who received clozapine or olanzapine demonstrated higher degrees of resistance compared with those treated with risperidone [82-84]. Furthermore, in a recent study of 242 patients, patients treated with olanzapine or clozapine had significantly higher prevalence of dyslipidemia compared to unmedi-cated patients, independent of BMI . A large matched case-control study of patients treated with AAPs conducted by Olfson et al.  found that treatment with clozapine, risperidone, quetiapine, olanzapine, ziprasidone, but not aripiprazole, was associated with a significant risk for hyperlipidemia, as compared to no AAP medication.
There is clear correlation between obesity and IR, as adipocytes secrete several hormones (leptin, adipsin, adiponectin, tumor necrosis factor-a) that impair insulin secretion and/or insulin effects . With the exception of adiponectin, all the hormones secreted by the endocrine adipose tissue exert antagonistic effects on insulin secretion and action . As mentioned above, the use of AAPs is strongly associated with clinical weight gain and obesity; therefore, IR may be the result of increased adipose tissue and their hormone secretions. AAPs can also produce hyperinsuline-mia through their antagonistic effects on dopamine (D2) receptors. D2 receptors are expressed in the ^-cells and mediate glucose-stimulated secretion of insulin. Impaired dopamine receptor activation by via the effects of AAP may result in a compensatory increase in insulin secretion, which over time may develop into insulin insensitivity and resistance [89, 90]. Hormones important to appetite regulation and maintenance of adipose homeostasis such as leptin (anorectic) and ghrelin (orexigenic) may be disrupted by AAP treatment. It is hypothesized that AAP treatment disrupts appetite regulation by increasing ghrelin and/or decreasing leptin, thereby increasing Agouti-related peptide (AGRP), neuropeptide Y (NP-Y), and orexin, leading to increased food intake and obesity. Ghrelin levels in patients treated with AAP for 1 year were significantly higher than in placebo-treated controls, with no difference between the various atypical agents that were used (clozapine, olanzapine, risperidone, or quetiap-ine) . However, several other reports have indicated no difference in ghrelin levels or a reduction in ghrelin levels [92-94].
Carbamazepine (CBZ) may be associated with weight gain and metabolic abnormalities, but to a lesser extent than valproate or lithium . In a study of 105 epileptic women treated with valproate and CBZ monotherapy, rates of PCOs were similar (in both groups) to rates in the general population. However, BMI, triglycerides, and postprandial insulin were higher in valproate-treated patients than CBZ-treated patients. Of note, the enzyme-inducing properties of CBZ may reduce levels of free and total thyroxin in patients with thyroid dysfunction . To date, the literature has not supported an association between lamotrigine (LTG) and IR. In a study of 54 women with epilepsy, MetS was more frequently associated with VPA-treated patients (41.7%) than CBZ (5.3%), LTG (0%), or topiramate (TPM) groups (0%) . A post-hoc analysis was performed from a 12-week prospective open-label study of 1,175 patients with BD initiated on lamotrigine either as monotherapy or adjuvant treatment to valproate, lithium, antipsychotics, or antide-pressants. No weight or BMI changes were noted after lamotrigine monotherapy or adjuvant therapy . These findings are consistent with studies of lamotrigine in epilepsy populations .
While not approved for use as mood stabilizers, topiramate and zonisamide have potential roles in treating patients with BD. Both medications are considered weight neutral, and may possibly even contribute to weight loss. In patients with epilepsy and type 2 diabetes, patients showed better glycemic control when treated with topira-mate . Evidence has supported the utility of topiramate in treating binge-eating episodes in patients with obesity . Both topiramate and zonisamide may have a role in promoting weight loss in patients with BD .
Was this article helpful?