In this issue, Dias and collaborators (1) prospectively examine the course of bipolar illness in 191 women with self-identified premenstrual exacerbation and 102 women without it, finding a worse course of illness with a more rapid rate of relapse and more syndromal and subsyndromal episodes in the women with premenstrual exacerbation. Despite the increase in the number of episodes, women with self-identified premenstrual exacerbation were not more likely to qualify for a diagnosis of rapid cycling. The authors suggest that women with a sensitivity to times of hormonal change may suffer more mood instability in general. The authors conclude that self-reporting of premenstrual exacerbation may be a clinical marker predicting a more symptomatic and relapse-prone type of bipolar disorder in reproductive-age women.

Women with bipolar disorder commonly report premenstrual exacerbation. Similar to the prevalence rate in the report by Dias et al., estimates are generally in the range of 60%—70% (2, 3) with retrospective reporting. There have been a number of case reports of women with bipolar disorder who regularly experience hypomanic or manic symptoms premenstrually. Despite the high prevalence of self-reported premenstrual exacerbation, several prospective studies have not demonstrated consistent mood changes associated with menstrual cycle phase in the general population of women with bipolar disorder. For example, Shivakumar et al. (4) examined the charts of 41 women with bipolar disorder (including types I, II, not otherwise specified, and schizoaffective) who prospectively completed NIMH Life Chart Method daily ratings of mood over 3 consecutive months, and the authors did not find any association between changes in depression or mania with phases of the menstrual cycle. This echoed earlier work by Leibenluft et al. (5), who followed 25 women with rapid-cycling bipolar disorder with daily mood ratings for 3 months or longer and found no association between mood and menstrual cycle. Although the numbers of subjects were small, these studies indicate that there is no direct and consistent link between mood and menstrual cycle in the general population of women with bipolar disorder. However, these studies do not rule out the possibility that there is a subgroup of women with bipolar disorder who are sensitive to times of hormonal change and who experience premenstrual exacerbation.

Dias and colleagues followed women prospectively for a year, but they did not prospectively confirm premenstrual exacerbation. It therefore remains unclear if the women with premenstrual exacerbation in their study actually had premenstrual exacerbation or were instead misidentifying frequent relapse as premenstrual exacerbation. The premenstrual exacerbation group did endorse more classic physical and psychological premenstrual symptoms than the comparison group, which provides supporting evidence for a premenstrual mood syndrome.

Our group previously demonstrated that mood symptoms associated with the reproductive cycle (including premenstrual, postpartum, and perimenopausal periods) significantly correlated in women with major depressive disorder but not in women with bipolar disorder (3). While this finding needs to be replicated, the finding that one type of reproductive-cycle-associated mood symptom predicts others in a trait-like fashion in women with major depression supports the idea that there may be a subgroup of women with major depression who are consistently sensitive to times of hormonal fluctuation. Further support for the existence of a subgroup sensitive to hormonal change was given by the small but elegant studies by Schmidt et al. (6) and Bloch et al. (7), who demonstrated that women who have a history of mood syndromes associated with the reproductive cycle (premenstrual syndrome and postpartum depression) reexperience depressive symptoms in the setting of experimentally induced hormonal change.

Our finding of a lack of correlation between different types of reproductive-associated mood symptoms in women with bipolar disorder could indicate that the women in the study with bipolar disorder may have been more likely than women with major depression to take medications during reproductive life events, such as pregnancy, the postpartum period, and perimenopause. Medication use might have in turn obscured the correlation in women with bipolar disorder.

In support of this hypothesis, Karadag et al. (8) prospectively followed 34 women with bipolar disorder who were taking mood stabilizers and were affectively euthymic and 35 comparison subjects, and they found that the proportion of women who demonstrated a change in depressive symptoms premenstrually was higher among the comparison subjects than among the bipolar group. The authors concluded that mood-stabilizing treatment may have a prophylactic effect on premenstrual exacerbation. In light of the findings by Dias and colleagues, another interpretation is that women with bipolar disorder who respond well to mood stabilizers are unlikely to experience premenstrual exacerbation. Alternatively, women with premenstrual exacerbation may be less likely to fully respond to mood stabilizers. Further work in this area needs to be done.

As a psychiatrist who specializes in women's mental health, I see numerous cases of women seeking consultation for "PMS." Although I have not kept track of specific proportions, most women who are seeking treatment for premenstrual syndrome actually have an underlying mood disorder that is either untreated or inadequately treated. The first step in such cases is adequate treatment of the underlying mood disorder. The work presented by Dias et al. demonstrates that premenstrual exacerbation is a clinical marker of illness that is prone to relapse. Whether this is due to inadequate treatment or a less than optimal response to medications remains unclear. For clinicians, the message is clear: women who complain of premenstrual exacerbation should be followed closely and monitored for relapse.

In summary, it remains to be proven that there is a subgroup of women with bipolar disorder who are sensitive to times of hormonal change and that this subgroup is prone to more mood instability in general. This article offers us tantalizing but indirect support for this hypothesis, given the limitations of the study, including the retrospective reporting of premenstrual exacerbation. Despite this, one conclusion can be directly drawn from the data presented and is further supported by the literature: the biggest risk factor for future illness is not achieving full remission. Thus, the best defense against relapse is to get fully well and to not have premenstrual exacerbation.