{"paper_id":"4359fd9f-bfc9-4da0-a558-f3cdc269b437","body_text":"ORIGINAL ARTICLE\nArchives of Women's Mental Health (2025) 28:573–581\nhttps://doi.org/10.1007/s00737-024-01533-2\nIntroduction\nPrevious research has shown that key events in the female \nreproductive lifecycle (premenstrual, perinatal and peri -\nmenopausal period) often coincide with an exacerbation of \nmood symptoms in women with bipolar disorder (BD). One \nstudy of 158 women with BD found that up to 77% reported \n \r Lisa Jones\nlisa.jones@worc.ac.uk\n1 Psychological Medicine, University of Worcester, Worcester, \nUK\n2 Division of Psychological Medicine and Clinical \nNeurosciences, Cardiff University, Cardiff, UK\n3 Three Counties Medical School, University of Worcester, \nSevern Campus, Hylton Road, Worcester WR2 5JN, UK\nAbstract\nPurpose The premenstrual phase of the menstrual cycle, childbirth and perimenopause often coincide with a worsening of \nmood symptoms in women with bipolar disorder (BD). To date, findings from the limited number of studies investigating \nassociations between these events among women with BD have been inconsistent. This study aimed to investigate associa -\ntions between episodes in relation to the perimenopause and (i) premenstrual symptoms and (ii) postpartum mood episodes \nin a large sample of postmenopausal women with BD.\nMethods Among 567 postmenopausal women with BD, recruited as part of the UK Bipolar Disorder Research Network, \nrelationships between reproductive event-associated mood symptoms/episodes were examined. Multivariate binary analyses \nwere carried out to identify if history of premenstrual symptoms and/or postpartum episodes predicted the occurrence of \nmood episodes in relation to the perimenopause, controlling for potential confounders including number of mood episodes \nper illness year.\nResults History of premenstrual symptoms was associated with experiencing any type of mood episode, and depression \nspecifically, during the perimenopause (OR 6.189, p < 0.001 and OR 2.709, p = 0.019 respectively). History of postpar -\ntum depression within 6 weeks of delivery was associated with depressive episodes during the perimenopause (OR 2.635, \np = 0.027). Postpartum mania was not a significant predictor.\nConclusions Our findings suggest that women with BD with a history of premenstrual symptoms and postpartum depression \nare potentially at increased risk of experiencing episodes of depression in relation to the perimenopause. There are clinical \nand self-management implications in identifying a subgroup of women with BD who may be particularly vulnerable to epi-\nsodes of mood disturbance during reproductive events.\nArticle highlights\n ● Relationships between reproductive events investigated among 567 postmenopausal women with bipolar disorder.\n ● History of premenstrual symptoms and postpartum depression was associated with episodes of depression during the \nperimenopause.\n ● No significant relationships were identified with (hypo)mania in relation to the perimenopause.\n ● A subgroup of women with bipolar disorder may be particularly vulnerable to episodes of mood disturbance during \nreproductive events.\nKeywords Bipolar disorder · Women · Lifetime reproductive events · Perimenopause\nReceived: 29 July 2024 / Accepted: 4 November 2024 / Published online: 14 November 2024\n© The Author(s) 2024\nAssociations between lifetime reproductive events among \npostmenopausal women with bipolar disorder\nKatherine Gordon-Smith1  · Amy Perry1  · Arianna Di Florio2  · Nicholas Craddock2  · Ian Jones2  · \nLisa Jones3\n1 3\n\nK. Gordon-Smith et al.\nworsening of symptoms associated with at least one repro -\nductive event (Perich et al. 2017a). The association with \nthe perinatal period has been studied in depth, with a meta-\nanalysis of 37 studies investigating the risk of postpartum \nrecurrence in women with BD finding an overall risk of \n37% (Wesseloo et al. 2016). Much research in this field has \nfocused on the prevalence of postpartum psychosis, which \nis considerably higher in women with BD (23%) (Perry et \nal. 2021) than the general population (0.89 to 2.6 in 1000 \nwomen) (VanderKruik et al. 2017).\nThere has also been growing interest in the occurrence of \npremenstrual syndrome (PMS) and premenstrual dysphoric \ndisorder (PMDD) in BD. PMS is characterised by affective, \nbehavioural, cognitive and physical symptoms during the \nlate luteal phase of the menstrual cycle which end at, or \nwithin a few days after, menstruation onset. A meta-analysis \nreviewing epidemiological studies reported a general popu-\nlation pooled prevalence rate of 47.8% for PMS worldwide \n(Direkvand-Moghadam et al. 2014). Reported rates among \nwomen with BD vary, with the findings of a review high -\nlighting that retrospective studies indicate that 25–77% of \nwomen with BD report PMS (Teatero et al. 2014). PMDD \nis less common than PMS and characterised by more \nsevere affective, psychological and physical symptoms \nalso occurring in the luteal phase of the menstrual cycle. \nPMDD became an official diagnosis under the Depression \nand Depressive Disorders chapter in the fifth edition of the \nDSM-5 (American Psychiatric Association 2013). DSM-5 \ndiagnostic criteria for PMDD require the presence of at \nleast five symptoms, one being mood related, with symp -\ntoms resulting in functional impairment. In a recent meta-\nanalysis of 44 studies consisting of 50,659 participants, \nthe pooled general population prevalence was 3.2% for a \nconfirmed diagnosis of PMDD (requiring prospective moni-\ntoring of symptoms over two cycles) and 7.7% for a provi -\nsional diagnosis (Reilly et al. 2024). Among women with \nBD, a systematic review found the proportion who reported \nsymptoms consistent with a provisional diagnosis of PMDD \nranged from 27 to 76% (Sharma et al. 2022). This suggests \nPMDD is more common among women with BD than in the \ngeneral population. However, none of the studies included \nin the review confirmed the provisional diagnosis of PMDD \nusing prospective daily ratings.\nWhile the relationship between BD and the perinatal \nperiod in particular is well established, there is a lack of \nresearch examining the occurrence of mood episodes and \nsymptoms in women with BD in relation to the perimeno -\npause. A recent UK study on over 100,000 women has found \nthat the risk of a first occurrence of mania in the perimeno -\npause is over 3 times that of the late reproductive years (Shi-\ntomi-Jones et al. 2024). For women with pre-existing BD, \na systematic review identified that the perimenopause could \nbe a time of increased vulnerability (Perich et al. 2017b). \nThe review included 9 studies (3 longitudinal and 6 cross-\nsectional) and a total of 273 women with BD. The meno -\npause was found to be associated with increased symptoms \noverall and with depression in particular being the most \ncommon mood symptoms experienced. Specifically focus -\ning on mood episodes, Marsh et al. (2008) found among 47 \nperimenopausal aged women with BD, 68% experienced at \nleast one depressive episode over a longitudinal monitoring \nperiod of on average 17 months and that episode frequency \nwas significantly increased during the perimenopause com-\npared to during women’s reproductive years.\nThe findings of a recent review suggest that a subset \nof women with BD may be particularly vulnerable to the \nimpact of menstrual cycle events (Aragno et al. 2022) Fur-\nther research investigating associations between reproduc -\ntive events in women with BD support this suggestion. \nAmong 158 women with BD, those who reported they had \nexperienced postnatal episodes were significantly more \nlikely to report experiencing worse mood symptoms in the \npremenstrual and menopausal period (Perich et al. 2017a). \nThe authors indicated their findings suggested some women \nwith BD have a greater vulnerability to episodes of mood \ndisturbance in relation to reproductive events. Furthermore, \na case series of five women with BD type I (BDI), high -\nlighted a potential link between postpartum and menopausal \nmood episodes (Robertson Blackmore et al. 2008). All five \nwomen had experienced an episode of postpartum psycho -\nsis and an episode of mania around the time of the meno -\npause and the majority did not experience mania or other \nmood episodes outside of these periods. In contrast, among \n197 women with BDI, Payne et al. ( 2007) did not find any \nsignificant associations between reproductive cycle-asso -\nciated mood symptoms, with premenstrual and postpar -\ntum mood symptoms not predicting perimenopausal mood \nsymptoms. Marsh et al. ( 2015) also found that among 44 \nwomen with BD, self-reported history of mood exacerba -\ntion premenstrually and/or during the postpartum was not \nsignificantly associated with mood symptom severity during \nthe perimenopause.\nTo date the findings of the limited number of studies \ninvestigating associations between reproductive events \namong women with BD have been inconsistent. Studies \ninvolving women who have experienced or were experienc-\ning the perimenopause have included small sample sizes, \nlimiting the ability to stratify by episode types during this \nperiod. In a large sample of postmenopausal women with \nBD, this study aimed to investigate associations between \nboth (i) premenstrual mood symptoms and (ii) postpartum \nmood episodes and episodes in relation to the perimeno -\npause, stratifying by episode type in the postpartum and \nperimenopause periods.\n1 3\n574\n\nAssociations between lifetime reproductive events among postmenopausal women with bipolar disorder\nMethods\nThis study was part of an on-going research programme into \nthe genetic and non-genetic aetiology of BD (Bipolar Disor-\nder Research Network, bdrn.org). The research programme \nhas UK National Health Service (NHS) Health Research \nAuthority (HRA) approval – Research Ethics Committee \n(REC) reference (MREC/97/7/01) and local approvals in all \nparticipating NHS Trusts/Health Boards.\nRecruitment of participants\nParticipants were recruited into the research programme \nfrom across the UK using both systematic and non-system -\natic recruitment methods. Participants were recruited sys -\ntematically via National Health Service (NHS) psychiatric \nservices (community mental health teams and lithium clin -\nics). Non-systematic recruitment methods included adver -\ntising for volunteers using local and national media and \nthrough UK-based charities for individuals with BD, for \nexample Bipolar UK.\nThe research programme inclusion criteria were (i) \nDSM-IV/DSM-5 diagnosis of bipolar disorder (bipolar dis-\norder type I, bipolar disorder type II, schizoaffective disor -\nder bipolar type or bipolar disorder not otherwise specified) \n(American Psychiatric Association 2000, 2013), (ii) aged at \nleast 18 years, (iii) mood symptoms to have started before \nthe age of 65 years, (iv) able to provide written informed \nconsent, and (v) of European ancestry due to the focus of \nthe research programme on genetic risk factors. Potential \nparticipants were excluded if they: (i) had only experienced \naffective illness in relation to, or as a consequence of, alco-\nhol or substance abuse or dependence; (ii) had only experi -\nenced affective illness as a consequence of medical illness \nor medication; (iii) were biologically related to another \nstudy participant.\nThis study was carried out using a subset of women in the \nBDRN research programme who were postmenopausal and \nhad provided information about mood episodes in relation \nto the perimenopause, the postpartum period and premen -\nstrual symptoms (n = 567).\nPsychiatric assessment\nA modified version of the semi-structured Schedule for \nAssessment in Neuropsychiatry (SCAN) interview was \nused to collect detailed information on lifetime psycho -\npathology (Wing et al. 1990). In most cases, participants’ \npsychiatric and medical case notes were also accessed to \ncorroborate information given during the interview. Diag -\nnoses and lifetime-ever clinical ratings including age at \nillness onset and number of lifetime illness episodes were \nmade using all available clinical data. All interviews and \ndiagnostic and rating procedures were carried out by trained \npsychologists and psychiatrists. In cases where there was \ndoubt, ratings were made by at least two members of the \nresearch team blind to one another’s ratings and consensus \nwas reached by discussion where necessary. Inter-rater reli-\nability was formally assessed using 20 cases and was found \nto be high. The mean kappa score for DSM diagnosis was \n0.84 and those for categorical data ranged from 0.81 to 0.97. \nMean intra-class correlation coefficients for continuous data \nranged from 0.92 to 0.99.\nLifetime reproductive events\nData regarding the menopause were collected using a self-\nreport questionnaire designed by the authors, either at the \ntime of the initial clinical interview or subsequently as part \nof a questionnaire mail out sent to all BDRN participants \nin current contact with the research programme at that time \n(just under 4000 participants). Women who self-reported \nbeing postmenopausal were asked to report: (i) the age they \nwent through the menopause; (ii) if they had experienced \nmood episodes related to the menopause; and, (iii) written \ndetails about any episodes. A pragmatic decision to ask post-\nmenopausal women only was made due to the self-report \nmethod of data collection and women potentially being \nunsure if they were in the early stages of the perimenopause. \nThe details provided were used to classify the episodes \ndescribed creating a broad category of episode of any type \n(including unspecified mood episodes) and narrower defini-\ntions of episodes of depression and (hypo)mania. The nar -\nrower definitions of depression and (hypo)mania were rated \nwhere sufficient information was provided to be able to clas-\nsify the polarity of the episode(s). The ratings were made \nbased on the most significant episode reported during the \nperimenopause which was rated hierarchically according to \nthe written descriptions of the episodes and symptoms. In \ncases where a woman reported an episode of mania requir -\ning hospitalisation and/or psychosis and also described an \nepisode of depression, the episode of mania was rated hier -\narchically as the most impairing episode. These ratings were \nmade by at least two members of the research team blind to \none another’s ratings and consensus was reached by discus-\nsion where necessary. At the same time the Premenstrual \nSymptoms Screening Tool (PSST)(Steiner et al. 2003) was \nused to ask women to rate their past usual experience of 14 \npremenstrual syndrome symptoms and the resultant level of \noccupational and social impairment, during their lifetime, \nusing a four-point scale: not at all, mild, moderate or severe. \nThe responses were used to make a broad and narrow defi -\nnition of symptoms suggestive of moderate to severe pre -\nmenstrual syndrome (PMS) and premenstrual dysphoric \n1 3\n575\n\nK. Gordon-Smith et al.\nResults\nSample\nThe sample of 567 women is described in Table 1. Women \nhad a median age of 59 years, 23.1% had been recruited \nsystematically, 40.9% had completed higher education and \n91.2% had lived as married. All women were of European \nancestry due to the focus of the research programme on \ngenetic risk factors with 94% of UK White ethnicity. The \nmajority had a DSM diagnosis of bipolar I disorder (72.1%) \nor bipolar II disorder (22.4%). A further 2.6% had a diagno-\nsis of schizoaffective disorder bipolar type and 2.8% bipolar \ndisorder not otherwise specified. The median age of onset \nof bipolar disorder was 22.5 years. The median number of \nmood episodes (of any polarity), depressive episodes and \nhypo(manic) episodes per illness year respectively was \n0.51, 0.29 and 0.21. Median age at menopause was 50 years.\nRates of reproductive event-associated mood \nsymptoms/episodes\n359 out of 567 postmenopausal women were able to report \nthe definite absence or presence of a mood episode in rela -\ntion to the perimenopause, with almost two thirds of those \n(58.5%) experiencing an episode (Table 2). 21.4% and 8.9% \nof women reported an episode of depression and (hypo)\nmania in relation to the perimenopause respectively. 51.3% \nand 22.7% of women retrospectively reported symptoms \nsuggestive of PMS and possible PMDD. Among parous \nwomen, 49.5% experienced a mood episode of any type \nwithin 6 weeks postpartum, 26.4% and 19.6% had experi -\nenced an episode of depression and mania within 6 weeks of \ndelivery respectively.\ndisorder (PMDD) respectively according to the PSST tool’s \ncriteria. For possible PMDD this included at least one of \nfour core symptoms rated severe, at least four additional \nsymptoms “moderate to severe” and at least one of five \ndomains of functioning rated severe. At interview, parous \nwomen ( n = 432) provided information about the lifetime \noccurrence of mood episodes in each perinatal period. Case \nnote data were used to corroborate the information provided \nabout perinatal mood episodes. Interview and case note data \nwere combined to rate the time of onset in weeks follow -\ning delivery and polarity of the worst mood episode in each \nperinatal period. If this information could not be established \nat interview or from case note data a rating of unknown was \nmade.\nAnalysis\nSPSS version 29.0 was used to analyse the data. Associa -\ntions between the occurrence of menopausal episodes and \n(i) PMS/PMDD and (ii) postpartum mood episodes were \ninitially explored using chi squared tests.\nThree separate binary logistic regressions, using the \nenter method, were used to determine whether a history \nof postpartum episodes and premenstrual symptoms could \npredict the occurrence of mood episodes in relation to the \nperimenopause ((i) of any type, (ii) depression, (iii) (hypo)\nmania) after controlling for potential confounders: current \nage, number of mood episodes per illness year, parity (num-\nber of deliveries), age of onset of BD and reported age at \nmenopause.\nTable 1 Demographic and clinical characteristics of sample of postmenopausal women with bipolar disorder (n = 567^)\nDemographic\nCurrent age, years, median (IQR) 59 (10)\nSystematically recruited, % (n) 23.1 (127/550)\nHighest educational level (higher education), % (n) 40.9 (218/533)\nMarital history (has married), % (n) 91.2 (495/543)\nEthnicity (UK White), % (n) 94.4 (535/567)\nClinical\nBipolar disorder type I, % (n) 72.1 (409/567)\nBipolar disorder type II, % (n) 22.4 (127/567)\nSchizoaffective disorder bipolar type, % (n) 2.6 (15/567)\nBipolar disorder not otherwise specified, % (n) 2.8 (16/567)\nAge at illness onset, years, median (IQR) 22.5 (13)\nAverage number of mood episodes (of any polarity) per illness year, median (IQR) 0.51 (0.66)\nAverage number of depressive episodes per illness year, median (IQR) 0.29 (0.38)\nAverage number of hypo(manic) episodes per illness year, median (IQR) 0.21 (0.33)\nAge at menopause, years, median (IQR) 50 (7)\nIQR, interquartile range\n^ denominators vary due to missing/unknown data\n1 3\n576\n\nAssociations between lifetime reproductive events among postmenopausal women with bipolar disorder\nparous women who did not experience a postpartum mood \nepisode (26.3% vs. 15.6%, p = 0.036). This association was \nstronger when only including women who experienced post-\npartum depression (32.8% vs. 18.1%, p = 0.013) (Table 4). \nPostpartum mania was not associated with depression in \nrelation to the menopause. No significant associations were \nfound between the occurrence of postpartum mood episodes \nand the occurrence of (hypo)mania in relation to the meno -\npause (Table 4).\nPredictors of menopausal episodes after adjusting \nfor covariates\nBinary logistic analyses were carried out to identify whether \na history of reported (i) PMS, (ii) depression within 6 weeks \nof delivery and (iii) mania within 6 weeks of delivery \npredicted an episode in relation to the menopause (three \nseparate models: any episode, depression, mania). After \nadjusting for potential confounding variables (current age, \ndepressive episodes per illness year, hypo(manic) episodes \nper illness year, parity (number of deliveries), age of onset \nof BD and reported age at menopause) the following repro-\nductive events were significant predictors of episodes in \nrelation to the perimenopause:\nRelationships between lifetime reproductive events\nWomen with a possible history of PMS symptoms were sig-\nnificantly more likely to have experienced mood episodes \n(of any type) in relation to the menopause compared to those \nwithout PMS symptoms (77% vs. 42.2%, p < 0.001) and epi-\nsodes of depression in relation to the perimenopause (31.7% \nvs. 12.2%, p < 0.001) (Table 3). Similar significant associa-\ntions were found using the narrower definition of possible \nPMDD: 76.4% vs. 55.5%, p = 0.002 (episode in relation to \nmenopause of any type, possible PMDD vs. no PMDD) and \n37.5% vs. 17.8%, p < 0.001 (depression in relation to meno-\npause, possible PMDD vs. no PMDD) (Table 3). There were \nno significant associations found between history of possi -\nble PMS or PMDD symptoms and the occurrence of (hypo)\nmania in relation to the menopause (Table 3).\nThere were no significant differences in the overall rates \nof menopausal mood episodes (of any type) among women \nwith and without a history of postpartum mood episodes \n(Table 4). However, when explored further, depressive \nepisodes in relation to perimenopause were significantly \nmore common among women who experienced a postpar -\ntum mood episode within 6 weeks of delivery compared to \nTable 2 Rates of reproductive event-associated mood symptoms/episodes among sample of postmenopausal women with bipolar disorder \n(n = 567^)\nEpisodes in relation to the perimenopause a\nEpisode (of any type) in relation to menopause, % (n) 58.5 (210/359)\nDepression in relation to menopauseb, % (n) 21.4 (77/359)\n(Hypo)mania in relation to menopauseb, % (n) 8.9 (32/359)\nPremenstrual history\nSymptoms suggestive of PMS (broad definition) % (n) 51.3 (244/476)\nSymptoms suggestive of PMDD (narrow definition) % (n) 22.7 (108/476)\nPostpartum mood episodes (parous women only n = 432)\nEpisode (of any polarity) within 6 weeks of delivery, % (n) 49.5 (202/408)\nDepression within 6 weeks of delivery, % (n) 26.2 (107/408)\nMania within 6 weeks of delivery, % (n) 19.6 (80/408)\nPMDD, premenstrual dysphoric disorder; PMS, premenstrual syndrome\na excludes 208 cases where women were unsure if they had experienced episodes in relation to the perimenopause\nb rated where sufficient information was provided to be able to classify the polarity of the episode\n^ denominators vary due to missing/unknown data\nTable 3 Relationships between premenstrual symptoms and mood episodes in relation to the perimenopause\nPMS\n(n = 161)\nNo PMS\n(n = 147)\np value\nEpisode (of any type) in relation to perimenopause, % (n) 77.0 (124/161) 42.2 (62/147) < 0.001\nDepression in relation to perimenopause, % (n) 31.7 (51/161) 12.2 (18/147) < 0.001\n(Hypo)mania in relation to perimenopause, % (n) 9.9 (16/161) 9.5 (14/147) 0.903\nPMDD\n(n = 72)\nNo PMDD\n(n = 236)\np value\nEpisode (of any type) in relation to perimenopause, % (n) 76.4 (55/72) 55.5 (131/236) 0.002\nDepression in relation to perimenopause, % (n) 37.5 (27/72) 17.8 (42/236) < 0.001\n(Hypo)mania in relation to perimenopause, % (n) 4.2 (3/72) 11.4 (27/236) 0.068\n1 3\n577\n\nK. Gordon-Smith et al.\nperiods. Specifically, we found that a history of experienc -\ning PMS and severe premenstrual symptoms (suggestive of \nPMDD) was associated with experiencing episodes of any \ntype and depression during the perimenopause. Postpartum \ndepression within 6 weeks of delivery was further associ -\nated with depression during the perimenopause. In contrast, \nno significant relationships were identified with (hypo)\nmania in relation to the perimenopause.\nAlthough there is very limited research to date, our find-\nings support those of Perich et al. ( 2017a) who found that \nwomen with BD who experienced postnatal episodes were \nmore likely to experience worse mood symptoms meno -\npausally. In contrast to a case series of five women with BDI \nwhich highlighted a potential link between postpartum psy-\nchosis and menopausal mood episodes (Robertson Black -\nmore et al. 2008), postpartum mania was not associated \nwith the occurrence of any type of episode in relation to the \nmenopause in our sample. The findings in our sample spe -\ncifically suggest a relationship between postpartum depres-\nsion and depression in relation to the menopause.\nThe association between premenstrual symptoms and \nepisodes in relation to the menopause has not been found in \nthe small number of studies that have previously explored \nthis relationship in BD. For example, in a study by Payne et \nal. (2007) using a similar retrospective study design, among \n197 women with BDI neither premenstrual nor postpar -\ntum mood symptoms were associated with perimenopause \nsymptoms. The authors did not however stratify their analy-\nses by mood symptom type during the postpartum or meno-\npausal period. In our study we also did not find an overall \nassociation between menopausal mood episodes (of any \ntype) and postpartum mood episodes. Significant relation -\nships were specifically found between depression in relation \nto the menopause and postpartum depression. Furthermore \nthe authors only included women with BDI in their analy -\nsis of associations between reproductive cycle-associated \nPredictors of episode (of any type) in relation to \nperimenopause\nHistory of PMS was significantly associated with any type \nof mood episode during the perimenopause (OR 6.189, \n95% CI 2.890–13.250, p < 0.001). Postpartum depression \nand postpartum mania within 6 weeks of delivery were not \nsignificant.\nPredictors of episode of depression in relation to \nperimenopause\nHistory of PMS (OR, 2.709, 95% CI 1.175–6.248, p = 0.019) \nand postpartum depression within 6 weeks of delivery (OR \n2.635, 95% CI 1.119–6.208, p = 0.027) were significantly \nassociated with depressive episodes during the perimeno -\npause. Postpartum mania was not significant.\nPredictors of (hypo)mania in relation to perimenopause\nPMS, postpartum depression and postpartum mania within \n6 weeks of delivery were all not significant. The same pat -\ntern of results was found when using the narrower defini -\ntion of possible PMDD which predicted both an episode of \nany type (OR 3.014, 95% CI 1.229–7.390, p = 0.016) and \nepisode of depression (OR 3.033, 95% CI 1.286–7.156, \np = 0.011) in relation to the perimenopause.\nDiscussion\nTo date this is the largest study to investigate relationships \nbetween illness exacerbation in relation to lifetime reproduc-\ntive events among postmenopausal women with BD strati -\nfying by episode type in the postpartum and perimenopause \nTable 4 Relationships between postpartum and perimenopausal episodes\nPostpartum episode (of any polar-\nity) within 6 weeks (n = 133)\nNo postpartum episode within \n6 weeks (n = 122)\np \nvalue\nEpisode (of any type) in relation to perimenopause, % (n) 60.2 (80/133) 62.3 (76/122) 0.726\nDepression in relation to perimenopause, % (n) 26.3 (35/133) 15.6 (19/122) 0.036\n(Hypo)mania in relation to perimenopause, % (n) 6.8 (9/133) 13.1 (16/122) 0.089\nPostpartum depression within 6 \nweeks (n = 67)\nNo postpartum depression \nwithin 6 weeks (n = 177)\np \nvalue\nEpisode (of any type) in relation to perimenopause, % (n) 64.2 (43/67) 60.5 (107/177) 0.593\nDepression in relation to perimenopause, % (n) 32.8 (22/67) 18.1 (32/177) 0.013\n(Hypo)mania in relation to perimenopause, % (n) 4.5 (3/67) 11.3 (20/177) 0.104\nPostpartum mania within 6 weeks \n(n = 59)\nNo postpartum mania within \n6 weeks\n(n = 186)\np \nvalue\nEpisode (of any type) in relation to perimenopause, % (n) 61.0 (36/59) 61.8 (115/186) 0.911\nDepression in relation to perimenopause, % (n) 22.0 (13/59) 22.0 (41/186) 0.999\n(Hypo)mania in relation to perimenopause, % (n) 11.9 (7/59) 9.1 (17/186) 0.540\n1 3\n578\n\nAssociations between lifetime reproductive events among postmenopausal women with bipolar disorder\nrecruited as part of a large on-going genetic epidemiol -\nogy study of mood disorders within the UK it is likely our \nsample is not fully representative of the broader population \nof women with BD. Furthermore, the majority of women \nhad a diagnosis of BDI (72.1%) which may limit the gen -\neralisability of the results. A limitation of our study is that \nour measure of mood episodes during the perimenopause \nwas based on self-report, and we are unable to establish if \nthese would meet criteria for DSM mood episodes including \nmixed episodes. As women were asked to self-report if they \nwere postmenopausal we were unable to stratify our analy -\nsis further by menopausal stage for example those outlined \nin the Stages of Reproductive Aging Workshop (STRAW) \ncriteria (Harlow et al. 2012). A recent systematic review \nand meta-analysis of 17 non-clinical cohort studies found \nthat the perimenopausal stage (but not the post-menopausal \nstage) was associated with the risk of depressive symptoms \ncompared to the pre-menopause (Badawy et al. 2024). We \nwere however unable to separate out women who may have \nexperienced mood episodes in the early postmenopausal \nstage in our sample. Due to numbers we were also unable \nto stratify our analysis according to whether or not women \nexperienced the onset of BD in relation to the menopause. \nThis would be interesting to explore in larger samples given \nthe recent link found between the perimenopause and first \nonset BD (Shitomi-Jones et al. 2024). We also did not collect \ndata on other potential confounding factors including age at \nmenarche and medication including hormone replacement \ntherapy (HRT).\nDue to the retrospective reporting of episodes, some \nwomen with a history of premenstrual symptoms and/or \nepisodes of postpartum depression may be more likely to \nattribute episodes to other reproductive life events i.e. the \nmenopause. Among women who were able to establish the \nabsence/presence of an episode in relation to the perimeno-\npause, just under 60% reported experiencing an episode in \nrelation to the menopause. This is lower than the findings \nof a longitudinal study among 47 women with BD which \nfound 79% (37/47) experienced at least one mood episode \nduring the perimenopause with 68% of women experiencing \nat least one depressive episode, 23% (hypo)mania episodes, \nand 13% episodes with both depressive and elevated mood \n(Marsh et al. 2008). This could suggest that the rates in our \nstudy might be an underestimate due to the method of data \ncollection used but nevertheless our findings are similar in \nterms of depression being found to be the predominant mood \nepisode over the menopausal period among women with BD. \nThe findings of a more recent systematic review also high -\nlighted that, among women with BD, the menopause was \nreported to be associated with increased symptoms over -\nall and with depression in particular (Perich et al. 2017b). \nThe rate of mood episodes in relation to the menopause in \nmood symptoms whereas the current study included women \nwith bipolar spectrum disorder (predominantly BDI and \nBDII) meaning the results are not directly comparable. In \nthe same study by Payne et al. (2007) among a larger sample \nof 509 women with major depressive disorder, premenstrual \nand postpartum mood symptoms predicted perimenopausal \nsymptoms. Using a prospective design, Marsh et al. ( 2015) \ndid not find a significant association between self-reported \nhistory of mood exacerbation premenstrually and/or during \nthe postpartum and mood symptom severity during the peri-\nmenopause among 44 women with BD. However the differ-\nent study design, definitions and measurements may account \nfor the different findings with our study. For example Marsh \net al. ( 2015) asked for self-reported endorsement of a his -\ntory of premenstrual or postpartum mood exacerbation and \ndid not differentiate between types of mood exacerbation in \nthe postpartum.\nOur finding of an association between reproductive life \nevents, in particular premenstrual symptoms and postpar -\ntum depression and depression in relation to the menopause, \nsuggest some women with BD might have a greater vulner-\nability to episodes of mood disturbance during reproductive \nevents. If replicated the findings of this study have clini -\ncal implications in terms of potentially identifying women \nwith BD who may be particularly vulnerable to episodes of \ndepression during the perimenopause based on a previous \nhistory of severe premenstrual symptoms and postpartum \ndepression. Providing women with information about this \nrisk could also be incorporated into BD self-management \nprogrammes to increase awareness. Worsening of mood \nsymptoms during reproductive events may have shared \npsychosocial and biological susceptibility factors. The fact \nthat associations between reproductive events remained \nsignificant when controlling for previous number of mood \nepisodes suggests our findings cannot be explained simply \nby women having an increased susceptibility to depression \nmore generally across their lives. It has been suggested that \nthere may be a subgroup of women with BD that have sen -\nsitivity to hormonal changes that occur during reproduc -\ntive events (Teatero et al. 2014). In relation to psychiatric \nsymptoms related to reproductive life events more gener -\nally there is growing interest in the role of allopregnano -\nlone (ALLO), a neuroactive metabolite of progesterone and \nmodulator of the GABA(A) receptor complex (Martinez et \nal. 2016; Osborne et al. 2017; Slopien et al. 2018; Hantsoo \nand Epperson 2020).\nThe strengths of the current study are that our study \npopulation was recruited from across the UK, and provided \na large sample of postmenopausal women with BD. Diag -\nnostic and rating procedures were informed by rich clinical \ndescriptions and inter rater reliability was formally assessed \nand found to be high. However, as women in our study were \n1 3\n579\n\nK. Gordon-Smith et al.\nimplications as well as for women with BD themselves. \nIf further research confirms our findings, clinicians would \npotentially be able to predict the occurrence of a mood epi -\nsode in relation to the menopause for each of their female \nBD patients based on history of postpartum mood episodes \nand premenstrual symptoms. This knowledge would enable \nclinicians to put measures in place to help educate their \npatients and to support them through this period of increased \nrisk, for example, self-monitoring of mood symptoms, more \nfrequent medication reviews and access to tailored psycho -\neducation and self-management programmes.\nAcknowledgements We would like to thank all of the BDRN partici -\npants who have kindly given their time to participate in our research \nand all the mental health professionals who were involved in the \nrecruitment of participants. Thank you also to Dr Francesca Serra for \nher help with collating the menopause data.\nFunding This study was supported by grants from the Wellcome \nTrust [078901] and the Stanley Medical Research Institute [6045240-\n5500000100]. The funding sources had no role in the study design, in \nthe collection, analysis, and interpretation of data, in the writing of the \nreport and in the decision to submit the article for publication. ADF is \nfunded by MRC grant MR/W004658/1 and by the European Research \nCouncil grant 947763.\nDeclarations\nEthical approval  The study was performed in accordance with the \nethical standards as laid down in the 1964 Declaration of Helsinki and \nits later amendments or comparable ethical standards. The research \nprogramme has UK National Health Service (NHS) Health Research \nAuthority (HRA) approval – Research Ethics Committee (REC) ref -\nerence (MREC/97/7/01) and local approvals in all participating NHS \nTrusts/Health Boards.\nInformed consent All participants gave written consent to participate.\nConflict of interest The authors declare no competing interests.\nOpen Access   This article is licensed under a Creative Commons \nAttribution 4.0 International License, which permits use, sharing, \nadaptation, distribution and reproduction in any medium or format, \nas long as you give appropriate credit to the original author(s) and the \nsource, provide a link to the Creative Commons licence, and indicate \nif changes were made. The images or other third party material in this \narticle are included in the article’s Creative Commons licence, unless \nindicated otherwise in a credit line to the material. If material is not \nincluded in the article’s Creative Commons licence and your intended \nuse is not permitted by statutory regulation or exceeds the permitted \nuse, you will need to obtain permission directly from the copyright \nholder. To view a copy of this licence, visit  h t t  p : / /  c r e  a t i  v e c o m m o n s . o \nr g / l i c e n s e s / b y / 4 . 0 /     .  \nReferences\nAmerican Psychiatric Association (2000) Diagnostic and Statisti -\ncal Manual of Mental Disorders, 4th Edition, Text Revision \n(DSM-IV-TR)\nour BD sample is also similar to previous research among \nwomen with a history of major depressive disorder (MDD). \nIn a follow up study of a US community-based prospec -\ntive investigation of the perimenopause (Study of Women’s \nHealth Across the Nation (SWAN: Bromberger et al. 2011), \n59% of women with a history of MDD experienced an epi -\nsode of MDD during midlife compared to 28% of women \nwith no history of MDD (Bromberger et al. 2015).\nA further limitation is that because women in our sample \nwere postmenopausal, symptoms of possible PMS/PMDD \nwere self-rated retrospectively. It was not possible in our \nstudy to confirm these ratings using prospective daily rat -\nings which are required over at least two cycles for a diag -\nnosis of PMDD in DSM-5. Despite premenstrual symptoms \nbeing rated retrospectively among women who were now \npostmenopausal, the rates of possible PMS and PMDD \n(52% and 23% respectively) are very similar to those found \nin a recent cross-sectional study among 262 Han Chinese \nwomen with BD using the same measure of PMS (58% and \n21% respectively) (Liang et al. 2024). The women in the \nstudy by Liang et al. ( 2024) were all currently menstruat -\ning suggesting the rates found in our study may not have \nbeen significantly affected by the fact that women were now \npostmenopausal and reporting on past premenstrual symp -\ntoms. However, as highlighted in a recent systematic review \nand meta-analysis, prevalence of PMDD using a confirmed \ndiagnosis via prospective ratings is lower compared to a pro-\nvisional diagnosis based on retrospective symptom report \nwhich are likely to produce artificially high prevalence rates \n(Reilly et al. 2024). Despite this, the rate of possible PMDD \nin our sample (23%) is higher than the pooled prevalence \nof a provisional diagnosis of PMDD found in the general \npopulation (7.7%) (Reilly et al. 2024).\nConclusions\nOur study suggests that women with BD with a history of \npossible PMDD and postpartum depression are potentially \nat increased risk of experiencing episodes of depression in \nrelation to the perimenopause. Although challenging given \nthe time periods between reproductive life events, longitudi-\nnal research is needed using prospective daily ratings of pre-\nmenstrual symptoms and differentiating between different \nmenopausal stages (early/late) which was beyond the scope \nof our study. Findings may have implications in terms of \nidentifying a subgroup of women with BD who may be par-\nticularly vulnerable to episodes of mood disturbance dur -\ning reproductive events, which may help to develop a better \nunderstanding of the aetiology of mood disorders. Identify-\ning women who may be particularly at risk of recurrence of \nmood episodes in relation to the perimenopause has clinical \n1 3\n580\n\nAssociations between lifetime reproductive events among postmenopausal women with bipolar disorder\nPayne JL, Roy PS, Murphy-Eberenz K et al (2007) Reproductive \ncycle-associated mood symptoms in women with major depres -\nsion and bipolar disorder. J Affect Disord 99:221–229.  h t t  p s : /  / d o  i \n. o  r g / 1 0 . 1 0 1 6 / j . j a d . 2 0 0 6 . 0 8 . 0 1 3       \nPerich T, Roberts G, Frankland A et al (2017a) Clinical characteristics \nof women with reproductive cycle–associated bipolar disorder \nsymptoms. Australian New Z J Psychiatry 51:161–167.  h t t  p s : /  / d o  \ni . o  r g / 1 0 . 1 1 7 7 / 0 0 0 4 8 6 7 4 1 6 6 7 0 0 1 5       \nPerich T, Ussher J, Meade T (2017b) Menopause and illness course \nin bipolar disorder: a systematic review. Bipolar Disord 19:434–\n443. https:/ /doi.or g/10.11 11/bd i.12530\nPerry A, Gordon-Smith K, Di Florio A et al (2021) Mood episodes in \npregnancy and risk of postpartum recurrence in bipolar disorder: \nthe bipolar disorder Research Network pregnancy study. J Affect \nDisord 294:714–722. https:/ /doi.or g/10.10 16/j. jad.2021.07.067\nReilly TJ, Patel S, Unachukwu IC et al (2024) The prevalence of pre -\nmenstrual dysphoric disorder: systematic review and meta-analy-\nsis. J Affect Disord 349:534–540\nRobertson Blackmore E, Craddock N, Walters J, Jones I (2008) Is the \nperimenopause a time of increased risk of recurrence in women \nwith a history of bipolar affective postpartum psychosis? A case \nseries. Arch Womens Ment Health 11:75–78.  h t t  p s : /  / d o  i . o  r g / 1 0 . 1 \n0 0 7 / s 0 0 7 3 7 - 0 0 8 - 0 2 1 5 - 2       \nSharma V , Mazmanian D, Eccles H (2022) Relationship of Premen -\nstrual Dysphoric Disorder with bipolar disorder. J Clin Psychiatry \n83:22r14416. https:/ /doi.or g/10.40 88/JC P.22r14416\nShitomi-Jones L, Dolman C, Jones I et al (2024) First onset psychiatric \ndisorders in the perimenopause. Nature Mental Health In press\nSlopien R, Pluchino N, Warenik-Szymankiewicz A et al (2018) Cor -\nrelation between allopregnanolone levels and depressive symp -\ntoms during late menopausal transition and early postmenopause. \nGynecol Endocrinol 34:144–147.  h t t  p s : /  / d o  i . o  r g / 1 0 . 1 0 8 0 / 0 9 5 1 3 5 \n9 0 . 2 0 1 7 . 1 3 7 1 1 2 9       \nSteiner M, Macdougall M, Brown E (2003) The premenstrual symp -\ntoms screening tool (PSST) for clinicians. Arch Womens Ment \nHealth 6:203–209. https:/ /doi.or g/10.10 07/s0 0737-003-0018-4\nTeatero ML, Mazmanian D, Sharma V (2014) Effects of the menstrual \ncycle on bipolar disorder. Bipolar Disord 16:22–36.  h t t  p s : /  / d o  i . o  \nr g / 1 0 . 1 1 1 1 / b d i . 1 2 1 3 8       \nVanderKruik R, Barreix M, Chou D et al (2017) The global prevalence \nof postpartum psychosis: a systematic review. BMC Psychiatry \n17:272. https:/ /doi.or g/10.11 86/s1 2888-017-1427-7\nWesseloo R, Kamperman AM, Munk-Olsen T et al (2016) Risk of \nPostpartum Relapse in Bipolar Disorder and Postpartum psy -\nchosis: a systematic review and Meta-analysis. Am J Psychiatry \n173:117–127. https:   //d oi. or g/10 .117  6/a ppi. ajp.2015.15010124\nWing JK, Babor T, Brugha T et al (1990) SCAN. Schedules for \nClinical Assessment in Neuropsychiatry. Arch Gen Psychiatry \n47:589–593\nPublisher’s note Springer Nature remains neutral with regard to juris-\ndictional claims in published maps and institutional affiliations.\nAmerican Psychiatric Association (2013) Diagnostic and Statistical \nManual of Mental Disorders, (5th edn) (DSM-5)\nAragno E, Fagiolini A, Cuomo A et al (2022) Impact of menstrual \ncycle events on bipolar disorder course: a narrative review of cur-\nrent evidence. Arch Womens Ment Health 25:257–266.  h t t  p s : /  / d o  \ni . o  r g / 1 0 . 1 0 0 7 / s 0 0 7 3 7 - 0 2 2 - 0 1 2 1 7 - 9       \nBadawy Y , Spector A, Li Z, Desai R (2024) The risk of depression in \nthe menopausal stages: a systematic review and meta-analysis. \nJ Affect Disord 357:126–133.  h t t  p s : /  / d o  i . o  r g / 1 0 . 1 0 1 6 / j . j a d . 2 0 2 4 \n. 0 4 . 0 4 1       \nBromberger JT, Kravitz HM, Chang Y-F et al (2011) Major depression \nduring and after the menopausal transition: study of women’s \nHealth across the Nation (SWAN). Psychol Med 41:1879–1888. \nhttps:/ /doi.or g/10.10 17/S0 03329171100016X\nBromberger JT, Schott L, Kravitz HM, Joffe H (2015) Risk factors \nfor major depression during midlife among a community sample \nof women with and without prior major depression: are they the \nsame or different? Psychol Med 45:1653–1664.  h t t  p s : /  / d o  i . o  r g / 1 0 \n. 1 0 1 7 / S 0 0 3 3 2 9 1 7 1 4 0 0 2 7 7 3       \nDirekvand-Moghadam A, Sayehmiri K, Delpisheh A, Kaikhavani S \n(2014) Epidemiology of premenstrual syndrome, a systematic \nreview and meta-analysis study. J Clin Diagn Res 8:106–109. \nhttps:/ /doi.or g/10.78 60/JC DR/2014/8024.4021\nHantsoo L, Epperson CN (2020) Allopregnanolone in premenstrual \ndysphoric disorder (PMDD): evidence for dysregulated sensitiv -\nity to GABA-A receptor modulating neuroactive steroids across \nthe menstrual cycle. Neurobiol Stress 12:100213.  h t t  p s : /  / d o  i . o  r g / \n1 0 . 1 0 1 6 / j . y n s t r . 2 0 2 0 . 1 0 0 2 1 3       \nHarlow SD, Gass M, Hall JE et al (2012) Executive summary of the \nstages of Reproductive Aging workshop + 10: addressing the \nunfinished agenda of staging reproductive aging. Fertil Steril \n97:843–851. https:/ /doi.or g/10.10 16/j. fertnstert.2012.01.128\nLiang X, Yang T, Liao Y et al (2024) The impact of comorbid pre -\nmenstrual syndrome or premenstrual dysphoric disorder on the \nclinical characteristics of bipolar disorder among Han Chinese \nwomen. Arch Womens Ment Health 27:67–75.  h t t  p s : /  / d o  i . o  r g / 1 0 . \n1 0 0 7 / s 0 0 7 3 7 - 0 2 3 - 0 1 3 8 0 - 7       \nMarsh WK, Templeton A, Ketter TA, Rasgon NL (2008) Increased fre-\nquency of depressive episodes during the menopausal transition in \nwomen with bipolar disorder: preliminary report. J Psychiatr Res \n42:247–251. https:/ /doi.or g/10.10 16/j. jpsychires.2006.12.006\nMarsh WK, Gershenson B, Rothschild AJ (2015) Symptom severity of \nbipolar disorder during the menopausal transition. Int J Bipolar \nDisord 3:17. https:/ /doi.or g/10.11 86/s4 0345-015-0035-z\nMartinez PE, Rubinow DR, Nieman LK et al (2016) 5α-Reductase \ninhibition prevents the Luteal phase increase in plasma allo -\npregnanolone levels and mitigates symptoms in women with \nPremenstrual Dysphoric Disorder. Neuropsychopharmacology \n41:1093–1102. https:/ /doi.or g/10.10 38/np p.2015.246\nOsborne LM, Gispen F, Sanyal A et al (2017) Lower allopregnanolone \nduring pregnancy predicts postpartum depression: an exploratory \nstudy. Psychoneuroendocrinology 79:116–121.  h t t  p s : /  / d o  i . o  r g / 1 0 \n. 1 0 1 6 / j . p s y n e u e n . 2 0 1 7 . 0 2 . 0 1 2       \n1 3\n581","source_license":"CC0","license_restricted":false}