SEARCH

SEARCH BY CITATION

Keywords:

  • bipolar disorder;
  • childbirth;
  • miscarriage;
  • termination

Abstract

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Clinical implications
  7. Acknowledgements
  8. Disclosures
  9. References

Objectives

To compare rates of bipolar episodes following miscarriage and termination with those occurring in the postpartum period.

Methods

Information in relation to pregnancy and childbirth was gathered retrospectively for 1,283 women with broadly defined bipolar disorder by interview and case-notes review.

Results

Rates of mania or affective psychosis were significantly higher after full-term delivery than after termination (p < 0.001) or miscarriage (p < 0.001). Rates of non-psychotic major depression were similar following full-term deliveries, miscarriages (p = 0.362), and terminations (p = 0.301).

Conclusions

While women with bipolar disorder and their clinicians should be aware of the possible onset of depression in the weeks following miscarriage or termination, episodes of mania or affective psychosis are less common in comparison with the high rates observed in the postpartum period.

Women with bipolar disorder face very difficult decisions in relation to pregnancy, with consistent evidence of a high risk of severe postpartum episodes [1]. However, there is a paucity of studies that can inform women and their clinicians about the risk of a recurrence after a miscarriage or termination. Data from a Danish population-based cohort showed that, although admission rates were high in women who had a termination, there was no evidence for an increased risk of readmission with bipolar disorder in the months following the induced abortion [2]—rates were high before and following the procedure. There is less information regarding the impact of miscarriage on bipolar episodes, with data limited to small studies and case reports [3].

In a large sample of women with bipolar disorder, we here compare the rates of episodes following miscarriage and termination with those occurring in the postpartum period.

Methods

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Clinical implications
  7. Acknowledgements
  8. Disclosures
  9. References

Recruitment

The recruitment, assessment methods, and detailed demographic and clinical characteristics of our mood disorder sample have been described in detail elsewhere [1]. In summary, women were recruited into our clinical and genetic study of bipolar disorder using systematic and non-systematic methods. All participants were ≥18 years and provided written informed consent. Potential participants were excluded if they had (i) a lifetime diagnosis of intravenous drug dependence, (ii) experienced affective illness only as a result of alcohol or substance dependence, or (iii) experienced affective illness only secondary to medical illness or medication.

Women were included in the current analyses if: (i) they had a lifetime DSM-IV diagnosis of bipolar disorder or schizoaffective disorder–bipolar type [4], (ii) we had detailed information for at least one pregnancy, (iii) they reported an age at onset before 50 years of age, and (iv) they were recruited not specifically because of a history of illness in relation to childbirth. Data on postpartum episodes were collected throughout the course of the study and are therefore available for all women in the sample; however, we collected data about psychiatric episodes following miscarriage and termination for a more limited period and hence these data are not available for all women.

Assessment

Participants were interviewed using the Schedules for Clinical Assessment in Neuropsychiatry [5], and psychiatric case notes were reviewed. Information was obtained for each participant about obstetric history and about episodes of bipolar disorder in relation to pregnancy. Best-estimate lifetime diagnoses were made according to DSM-IV, and key clinical variables, such as age at onset and number of episodes, were rated. We defined:

  1. postpartum, post-miscarriage, or post-termination mania or affective psychosis as a new episode of DSM-IV mania or psychotic depression with onset within six weeks of the end of the pregnancy.
  2. postpartum, post-miscarriage, or post-termination non-psychotic major depression as a new episode of DSM-IV non-psychotic major depression with onset within six weeks of the end of the pregnancy.

Statistical analyses

Statistical analyses were performed using R, version 2.13.0 (6). Estimates were calculated with 95% confidence intervals (CIs). Different pregnancies from the same woman cannot be considered independent events; thus we fitted mixed-effects models using the lmer function (package lme4) and setting the individual woman as a random effect and obstetric outcome and parity order as explanatory variables.

Results

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Clinical implications
  7. Acknowledgements
  8. Disclosures
  9. References

Complete information was available for 2,774 pregnancies in 1,283 women with bipolar I disorder (n = 873), bipolar II disorder (n = 341), schizoaffective disorder–bipolar type (n = 42), or bipolar disorder not otherwise specified (n = 27). Sample characteristics are summarized in Table 1. As described above, we collected information on postpartum episodes for the whole of the sample and therefore we had data on many more live births than other pregnancy outcomes. The episode after pregnancy was the first lifetime episode for 107 (42.5%) women with a history of post-pregnancy psychosis and 93 women (31.6%) with a history of post-pregnancy non-psychotic depression. Data were collected for 2,566 live births, 108 miscarriages, 88 terminations, and 12 stillbirths. Mean age at pregnancy was 26.6 years (standard deviation = 5.52 years). Distribution of lifetime diagnoses (p = 0.834), age at first impairment (p = 0.815), rates of lifetime psychosis (p = 0.757), and alcohol misuse (p = 0.107) did not differ across pregnancy outcome groups, while miscarriage and termination groups had significantly higher rates of substance misuse (p = 0.011). Proportions of pregnancies followed by mania/affective psychosis and non-psychotic depression are shown in Figure 1.

Table 1. Demographic and clinical characteristics of the sample (N = 1,283)a
  1. NOS = not otherwise specified.

  2. a

    Values of n may vary due to missing/unclear data.

  3. b

    Including cannabis.

Age at interview, years, median [range]47 [19–85]
Age at onset, years, median [range]21 [4–49]
No. of pregnancies, median [range]2 [1–15]
No. of deliveries, median [range]2 [0–7]
Age at first pregnancy, years, median [range]25 [13–42]
DSM-IV diagnosis, n (%)
Bipolar I disorder873 (68.0)
Bipolar II disorder341 (26.6)
Bipolar disorder NOS27 (2.1)
Schizoaffective disorder–bipolar type42 (3.3)
Comorbidity with substance use disorders, n (%)b123 (14.5)
Comorbidity with alcohol use disorders, n (%)208 (25.8)
Psychosis, n (%)364 (34.4)
image

Figure 1. Proportions of pregnancies followed by mania/affective psychosis and non-psychotic depression.

Download figure to PowerPoint

Overall, 332 pregnancies were followed by the onset of mania/affective psychosis within six weeks. Rates of mania/psychosis after a live birth were consistent with those previously reported by our group [1] for women with bipolar I disorder (n = 282, 19.2%; 95% CI: 17.2–21.3) and for those with bipolar II disorder (n = 14, 2.3%; 95% CI: 1.3–4.0). In women with broadly defined bipolar disorder, rates of mania/affective psychosis were significantly higher after a live birth (n = 320, 12.5%; 95% CI: 11.2–13.8) than after a termination (n = 4, 4.3%; 95% CI: <0.1–8.5; p < 0.001) or miscarriage (n = 6, 5.5%; 95% CI: 1.2–9.8; p < 0.001). Rates of mania/affective psychosis following stillbirth (2/12 = 16.7%) were consistent with rates following live birth. However, we excluded stillbirths from the comparative analysis due to the small sample size.

Non-psychotic major depression occurred after 518 pregnancies, including five episodes after stillbirth (41.7%). Again, because of the small sample size, stillbirths were excluded from further analysis. Rates were similar across pregnancy outcomes, with no significant differences between live births (n = 484, 18.9%; 95% CI: 17.5–20.6), miscarriages (n = 16, 14.6%; 95% CI: 7.8–21.2; p = 0.362), and terminations (n = 13, 15.9%; 95% CI: 8.3–23.5; p = 0.301).

Discussion

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Clinical implications
  7. Acknowledgements
  8. Disclosures
  9. References

In this study, we examined the occurrence of episodes of bipolar disorder post-miscarriage or termination and compared them with those after live birth in a large sample of women with bipolar disorder.

Childbirth was a more potent trigger for the early onset of severe bipolar episodes than termination. This finding is consistent with studies of induced abortion in the Danish registries [2]. To our knowledge, however, this is the first study reporting rates of bipolar episodes after miscarriage. About 5% of miscarriages were followed by affective psychosis, again suggesting a less potent trigger for severe episodes than childbirth. We are not able to address in this study, however, how the lower rates of mania and psychosis following miscarriage and termination compare with expected rates over a similar time period for bipolar women who have not been pregnant.

Extending analysis to non-psychotic major depression, we found similar rates of illness following all pregnancy outcomes. Non-psychotic depression affected about 15% of women following miscarriage or termination compared to about 19% of women following childbirth. Previous studies on bipolar depression following miscarriage or termination are lacking and our findings need to be replicated. It is also difficult to compare our results with previous reports on post-termination depression in women with and without mental disorders, because the incidence rates significantly varied across studies, probably due to methodological differences.

It is well established that childbirth is a very strong trigger for episodes of postpartum psychosis [1]. Rates of postpartum mania/psychosis reported here are in agreement with previous clinical [1] and registry-based [7] studies. Consistent with our previous study on the incidence of perinatal mood disorders [1], we found that women with bipolar I disorder had one in five live birth deliveries affected by postpartum mania/psychosis. For non-psychotic depression, there has been more debate about whether the postpartum is a period of high risk. It is possible that episodes of postpartum psychosis are triggered by the considerable biological changes in hormonal and other systems in late pregnancy and the early puerperium. Pregnancies ending earlier in gestation through miscarriage or termination may not share to the same extent these strong biological triggers and hence not lead to the same high risk of triggering mania or psychosis. The similar risks for non-psychotic depression we have found, however, may reflect childbirth, miscarriage, and termination all acting as potentially important life events and stressors for some women.

Our results need to be interpreted in light of the following limitations.

  1. Information was collected retrospectively; it is possible that women recall episodes in the postpartum period better than those following miscarriage and termination. However, psychiatric case notes were reviewed for 78% of participants and the reporting of episodes at interview was in agreement with medical records.
  2. Rates of episodes following miscarriages may be overestimated. It is, in fact, possible that women were not aware of miscarriages in early stages of pregnancy or were more likely to recall miscarriages occurring in relation to psychiatric episodes. However, this represents a conservative bias that made it less likely to detect the differences we found between live births and miscarriages.
  3. The numbers of pregnancies ending in termination or miscarriage for which we had data were much fewer than for childbirth, resulting in large CIs around our estimates. However, we were still able to detect a significantly higher rate of mania/psychosis in the postpartum period. Considering the current sample size for each group, we had also enough statistical power (1 − β = 0.8) to detect a small to medium effect (h = 0.3) in the analyses on non-psychotic depression.
  4. We did not have detailed information on the drug management of the women and could not, therefore, establish the influence of medication on the risk of a post-pregnancy episode. In particular, we were unable to determine if there are differences in the proportion of women on medication following the various pregnancy outcomes studied.

The limitations of studies in this area are difficult to overcome. Large register-based studies can provide more accurate estimates for severe episodes following terminations; however, they are less likely to detect less severe but more prevalent episodes, especially of non-psychotic depression. Only longitudinal clinical studies following up women preconception and throughout pregnancy can lead to more accurate information on miscarriages and for less severe episodes of illness. However, such longitudinal studies will require considerable resources that limit their feasibility.

Clinical implications

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Clinical implications
  7. Acknowledgements
  8. Disclosures
  9. References

Research into termination and mental health not only suffers from these methodological limitations but is also often influenced by strong personal opinions [8]. Despite the paucity of data, there is a need for women and their clinicians to understand the level of risk associated with all pregnancy outcomes [9]. It is important to promptly recognize the risk of bipolar episodes, especially postpartum psychosis, following childbirth [10]. Women with bipolar disorder and their doctors can be reassured to some extent that the risk of mania or psychosis does not appear to be as high following miscarriage or termination. However, in the case of non-psychotic depression we found no significant difference in the rates occurring postpartum and post-pregnancy with episodes following approximately 15% of miscarriages or terminations.

For women with bipolar disorder, therefore, clinicians should be aware of the possible onset of a depressive episode in the weeks following miscarriage or termination, while post-pregnancy mania or affective psychosis is less common, especially in comparison with the high rates observed in the postpartum period.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Clinical implications
  7. Acknowledgements
  8. Disclosures
  9. References

The authors thank all of the women who gave their time to participate in the study. ADF receives funding from a Cardiff MRC Centre Clinical Academic Mentorship (CAM) Scheme Fellowship. This work was supported by grants from the Wellcome Trust, the Stanley Medical Research Institute and the Medical Research Council.

Disclosures

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Clinical implications
  7. Acknowledgements
  8. Disclosures
  9. References

The authors of this paper do not have any commercial associations that might pose a conflict of interest in connection with this manuscript.

References

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Clinical implications
  7. Acknowledgements
  8. Disclosures
  9. References
  • 1
    Di Florio A, Forty L, Gordon-Smith K et al. Perinatal episodes across the mood disorder spectrum. JAMA Psychiatry 2013; 70: 168175.
  • 2
    Munk-Olsen T, Laursen TM, Pedersen CB et al. First-time first-trimester induced abortion and risk of readmission to a psychiatric hospital in women with a history of treated mental disorder. Arch Gen Psychiatry 2012; 69: 159165.
  • 3
    Brockington IF, Meakin CJ. Clinical clues to the aetiology of puerperal psychosis. Prog Neuropsychopharmacol Biol Psychiatry 1994; 18: 417429.
  • 4
    American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders: DSM-IV-TR. Washington, DC: American Psychiatric Association, 2000.
  • 5
    Wing JK, Babor T, Brugha T et al. SCAN. Schedules for Clinical Assessment in Neuropsychiatry. Arch Gen Psychiatry 1990; 47: 589593.
  • 6
    R Core Team. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. 2013. Available from: http://www.R-project.org/ [accessed October 2013].
  • 7
    Munk-Olsen T, Laursen TM, Mendelson T et al. Risks and predictors of readmission for a mental disorder during the postpartum period. Arch Gen Psychiatry 2009; 66: 189195.
  • 8
    Academy of Medical Royal Colleges by National Collaborating Centre. Induced Abortion and Mental Health: A Systematic Review of the Evidence–Full Report and Consultation Table with Responses. London: Academy of Medical Royal Colleges, 2011.
  • 9
    Casey P, Oates M, Jones I, Cantwell R. Invited commentaries on … Abortion and mental health disorders. Br J Psychiatry 2008; 193: 452454.
  • 10
    Jones I, Craddock N. Bipolar disorder and childbirth: the importance of recognising risk. Br J Psychiatry 2005; 186: 453454.