Preterm Birth (PTB) < 37 weeks:
Mei-Dan et al. [23] found that the prevalence of PTB was higher in women with BD (11.4%) than in the comparison group (6.2%). Lee et al. [22] found that women with BD were more likely to have PTBs than pregnant women with no history of mental illness (14.2% vs 6.9%); the odds of PTB for women with BD were 2.08 higher than healthy women after maternal, paternal, and infant characteristics were considered. Boden et al. [21] also found that the risk of PTB significantly increased for women with BD whether treated (8.1%) or untreated (7.6%) compared to women without BD (4.8%). Figures (2) shows the risk ratio (RR) forest plot of preterm birth < 37 weeks outcome. The meta-analysis result between bipolar and control groups in the 3 studies [21–23] is (RR 1.81, 95% CI 1.64 to 2.01; participants = 1294752; studies = 3; I2 = 0%, where CI is the confidence interval and I2 refers to the heterogeneity).
Another study [29] also shows a higher prevalence of PTB with women having BD (12.4%) compared to WA mothers (7.5%), (p < 0.001 “clinically significant”). However, Nguyen et al. [26] reported no significant difference in PTB in infants of women with BD compared to infants of WA mothers (8.9% vs. 7%, P = 0.548). Wisner et al. [25] also reported that the proportions of late preterm births were similar across the 3 groups: treated BD (8.2%), untreated (5.3%), and comparison group (6.9%) and that the risk remained non-significantly different, adjusting the potential confounding effects of maternal age, race, employment status, use of illicit drugs, and pre-pregnancy BMI; approximately 10% of all participants’ infants were delivered preterm overall.
PTB < 32 weeks and < 28 weeks:
Mei-Dan et al. [23] found that PTB < 32 weeks increased in the BD group compared to the referent group (9.1% vs.1.1%, Adjusted Odds Ratio (AOR): 1.70; 95% CI 1.16– 2.48), while no significant increased risk in PTB < 28 weeks was reported between BD (0.9%) and the referent group (0.6%) (AOR 1.66; 95% CI 0.92–3.02).
Small for Gestational Age (SGA) < 10th percentile:
Mei-dan et al. found that women with BD have an increased risk of SGA < 10th percentile compared to women without mental illness: BD: 14.1%, reference group 12.8%. AOR = 1.17 (95%CI 1.03–1.34), adjusted for maternal age and parity only. Lee et al. also reported that women with BD were more likely to have SGA than healthy pregnant women (22.3% vs.15.7%) (AOR = 1.47 (95% CI 1.14–1.91) when adjusted for maternal, paternal, and infant characteristics. Figures (3) shows the RR forest plot of the 2 studies [22] [23]; the meta-analysis result is (RR 1.81, 95% CI 1.64 to 2.01; participants = 1294752; studies = 3; I2 = 0%).
Severe SGA < 3rd percentile:
Mei-Dan et al. found that the risk of severe SGA was not significantly elevated in the bipolar group (4.6%) vs. women with no mental difficulties (3.9%), AOR 1.15; 95% CI, 0.92–1.43) when adjusted for maternal age, income quintile, parity, infant sex, obesity, substance abuse hypertension, venous thromboembolic disease, and gestational diabetes mellitus (DM), gestational hypertension (HTN), preeclampsia/ eclampsia. Boden et al. also showed that there are no significant differences for severe SGA reported between women with BD (neither treated with mood stabilizers nor not) and the control group, as the birth weight for untreated = 3.4%, treated = 2.5%, and no BD = 2.3%, birth length for untreated = 3.8%, treated = 3.2%, and no BD = 2.3%, and head circumference (HC) for untreated = 3.9%, treated = 3.3%, and no BD = 2.3%. This study shows that untreated BD was associated with the most increased risk of SGA regarding weight, length, and HC. After adjusting for confounders, the increased risks were somewhat attenuated and no longer statistically significant except for microcephaly, as the infants of women with untreated BD were at an increased risk of microcephaly than others. Figures (4) shows the RR forest plot of the 2 studies [21] [23]; the meta-analysis result is (RR 1.22, 95% CI 1.02 to 1.46; participants = 766354; studies = 2; I2 = 5%)
Jablensky defined the SGA as less than 37 weeks. The authors found no significant difference in SGA occurrence between women with BD and those with no mental illness (6.2% Vs 7.6%), as the mean gestational age (weeks) was 39.1 in the BD group, whereas 39.0 in the comparison group.
Large for Gestational Age (LGA):
Mei-Dan et al. divided the LGA into LGA > 90th percentile and severe LGA > 97 percentile: In LGA > 90th percentile, the authors noticed no significant difference in occurrence between women with BD (9.1%) compared to women with no BD (8.2%), AOR 1.13 95% CI 0.96 − 0.32, adjusted for maternal age and parity only. Regarding severe LGA > 97th percentile, it was more common among women with BD than in the control group (3.8% vs. 2.7%, AOR = 1.29; 95% CI, 1.08–1.54). Boden et al. also measured severe LGA and found no significant results between women with treated and untreated BD (2.5% vs. 2.3%). Overall, women with bipolar disorder (treated or not) were at an increased risk of having an LGA infant for weight than the control group, although this was not statistically significant. Treated women were at a non-significantly increased risk of having an LGA infant for length. Untreated women were less likely to have LGA for length. It seems that the infants of treated women with BD will not develop either macrocephaly or microcephaly, whereas women with untreated BD are more likely to develop microcephaly, but not macrocephaly.
Frayne et al. defined LGA at > 37 weeks and > 4000 grams, and they found that there was no much-increased risk of LGA for infants of women with BD compared to the WA population (7.1% vs. 11%, respectively, p-value = 0.285 “statistically no difference”).
Congenital Malformations:
Jablensky et al. found no difference in congenital abnormalities, defects of the cardiovascular system, and cerebral palsy, a rare complication of the neonate, in infants of women with BD compared with those of women with no mental health difficulties. Mei-Dan et al. [23] found that BD caused a significant risk of congenital malformations in infants of the BD group (RR between BD and control is 1.4). Boden et al. [21] suggest that psychotic treatment during pregnancy causes a high risk of congenital malformations, as the prevalence of congenital abnormalities was higher for infants of women with BD who are treated with mood stabilizers than those born to women with untreated BD or without BD (3.4% vs. 1.9% vs. 2%). Figures (5) shows the RR forest plot of the 2 studies [21] [23]; the meta-analysis result is (RR 1.39, 95% CI 1.16 to 1.66; participants = 766354; studies = 2; I2 = 0%).
Head Circumference (HC):
Three studies [21] [25] [28] found that the head circumference was smaller in the newborns of mothers with BD than in the newborns of mothers with no mental health difficulties; two of them [21] [25] reported that the incidence rate of small HC was higher in unmedicated women with BD compared to women undergoing bipolar treatment.
Fetal Distress:
Jablensky et al. found that fetal distress occurred non-significantly more often in women with BD than in unaffected women (13.6% vs. 11.8%). There was no significant difference in OR before and after adjustment for maternal characteristics (1.18 vs. 1.17, respectively). However, Nguyen et al. found that suspected fetal distress was significantly more common in women with BD compared to WA mothers (21.4% vs. 12.6%; P < 0.0001).
Low Birth Weight of Infant < 2500 grams:
Only Lee et al. found women with BD were more likely to have low birth weight infants than pregnant women with no history of mental health difficulties (9.8% vs. 5.7%). However, two studies [26] [29] found no difference in the low birthweight incidence rate or the mean birthweight between infants of women with BD and WA mothers and their babies. Wisner et al. [25] also found that the mean birth weight did not significantly differ (p = 0.079) between BD-NP and the other two groups, but tended to be lower in the BD-NP group. Another study [28] showed no significant difference between bipolar and control groups having LBW infants (9.9% vs. 9.3%).
5-minute Apgar Score < 7 or < 8:
Three studies [21] [25] [28] reported no difference found in low 5-min Apgar scores between infants of women with BD and those with no mental illness. However, Nguyen et al. [26] found that babies born to women with BD attending the CAMI clinic were less likely to have an Apgar score of 8 at 1 minute (64.3% vs. 85%; P < 0.0001) and were also less likely to have an Apgar score of 8 at 5 minutes (89.3% vs. 97.5%; P < 0.0001).
Stillbirth:
Mei-Dan et al. and Jablensky et al. noticed that there is no significant difference found in the rate of stillbirths between women with BD and those with no mental health difficulties.
Neonatal readmission at < 28 days of neonate life:
Mei-Dan et al. found that the BD group had an increased risk of neonatal admission (2.0%) compared to the control group (0.9%) (AOR: 2.41 (95% CI 1.76–3.31, when adjusted only for maternal age, income quintile, hypertension, venous thromboembolic disease, and gestational DM, gestational HTN, pre-eclampsia/eclampsia). However, Wisner et al. found no significant differences in the neonatal intensive care unit (NICU) admission occurred across the three groups: BD with treatment = 4.1%, untreated BD = 7.9%, and comparison = 8.1%. The risk remained non-significantly different across the three groups (p = 0.729) in the multiple regression models adjusting for the potential confounding effects of maternal characteristics, but it was noticed that the women taking treatment for bipolar are less likely to get their babies readmitted to NICU compared to the non-treated and comparison group.
Neonatal Hypoglycemia:
Only one study [21] reported infants of women with untreated BD were at a higher risk of neonatal hypoglycemia (4.3%) than treated (3.4%) BD women and women without BD (2.5%), but the risk estimates in treated BD were imprecise.
Neonatal Morbidity:
Mei Dan et al. defined neonatal morbidity as respiratory distress syndrome (RDS), seizures, sepsis, intravenous hyperalimentation (IVH), persistent fetal circulation, and neonatal abstinence syndrome. The risk of having any of these neonatal morbidities was higher in women with BD (5.4%) compared with those without mental illness (1.9%) (AOR 2.99, 95% CI 2.44– 3.66). When analyzed separately, babies of women with BD had higher rates than the referent group for RDS: 1.5% vs.1.0%, seizures: 0.6% vs. 0.2%, sepsis: 1.3% vs. 0.7%, IVH: 0.7% vs. 0.3%, and neonatal abstinence syndrome: 1.9% vs. 0.0% (great difference), respectively. However, no difference was seen between the two groups in the incidence rate of persistent fetal circulation.
Infant Mortality:
Two studies [23] [28] reported that there was no difference in neonatal mortality (< 28 days) or postnatal death (< 1 year) in the BD group compared with women with no mental illness group; one of them [28] also showed no difference in childhood death between the 2 groups.
Labor or Delivery Obstetric Complications:
Jablensky et al. found that there was no difference in the incidence rate of cephalopelvic disproportion, atypical presentation, cord anomalies, postpartum hemorrhage, early rupture of the membranes, prolonged labor, and threatened abortion between women with BD and unaffected women. Women with pre-existing BD had more risk of obstetric complications than those who developed BD after the index birth.
Gestational Diabetes Mellitus (GDM):
Boden et al. noticed no increased risk for GDM between either pharmaceutically treated BD, untreated BD, and control groups, with no difference in odds ratio even after adjustment for smoking, and diagnosis of alcohol or substance misuse disorder. Lee et al. and Mei-Dan et al. also found no significant difference in rates of GDM between women with and without BD. Figures (6) shows the RR forest plot of the 3 studies [21] [22] [23]; the meta-analysis result is (RR 0.92, 95% CI 0.76 to 1.12; participants = 1294752; studies = 3; I2 = 0%).
However, Frayne et al. reported that women with BD are more likely to have GDM than WA mothers, (10.7% vs. 7.4% respectively; p < 0.001 “clinically significant”). Nguyen et al. also supported this finding; they found that women with BD attending the CAMI clinic were at an increased risk of developing GDM than WA mothers (12.5% vs. 4.4% respectively; P < 0.0001).
Gestational Hypertension (GHTN) (Preeclampsia):
Lee et al. found women with BD were more likely to have GHTN (1.5% vs. 0.5%) than pregnant women with no history of mental health difficulties (p < 0.02). The odds ratio (AOR) was 2.81 when adjusted for maternal, and infant characteristics (95% CI 2.53–3.10). However, Mei-Dan found no difference in GHTN incidence rate between BD and control groups. Figures (7) shows the RR forest plot of the 2 studies [22] [23]; the meta-analysis result is (RR 1.13, 95% CI 0.84 to 1.51; participants = 962615; studies = 2; I2 = 77%).
Another 2 studies showed a great impact of BD on GHTN high prevalence; Frayne et al. found a higher prevalence of GHTN in women with BD (15.7%), compared to WA mothers (2.1%), (p < 0.001 “clinically significant”) and Nguyen et al. found that women with BD had an increased risk of GHTN than WA mothers (10.7% vs. 2.7%, respectively; P < 0.0001).
Jablensky et al. also reported that pregnant women with BD were more likely to have preeclampsia than pregnant women with no history of mental illness, (12.5% vs. 10.9%, unadjusted OR 1.17, 95% Cl = 0.93—1.48), but the authors did not mention the adjusted OR, which may differ from the unadjusted one.
Antepartum Hemorrhage (APH):
Jablensky et al. showed that Australian women with BD were more likely to have APH than other women (4.2% vs. 2.6%). It was noticed that it is a specific and significant complication in women with pre-existing BD not who developed it after the index birth. Compared to WA mothers (2.8%), Frayne et al. also found a higher prevalence of APH in women with BD (9%, p < 0.001 “clinically significant”). However, Nguyen et al. found no clinically significant (P = 0.232) difference between APH incidence rates in women with BD and WA mothers.
Threatened Preterm Labor:
Nguyen et al. found that BD caused an increased risk of developing threatened preterm labor compared with no mental illness in the WA obstetric population (7.1% vs. 2.5% respectively; P = 0.006). Another study [29] also found that the risk of threatened preterm labor was elevated in women with bipolar compared to WA mothers, (10.1% vs. 2.2%, respectively, p < 0.001 “clinically significant”). However, one study [28] found no significant difference in incidence between women with BD and unaffected women (4.2% vs. 3.8%).
Placenta Previa:
Only one [28] of the included studies showed that women with BD were more likely to have placenta previa (1.6%) than pregnant women with no BD (0.8%, unadjusted OR 2.04 95% CI 1.11–3.73). Placenta previa was also noticed to be specific in the BD group.
Time to Spontaneous Respiration and Intubation:
No difference was found [28] in time to spontaneous respiration > 2 minutes and intubation in women with BD compared with those with no mental health difficulties.
Naloxone Administration:
Jablensky et al. found that infants born to mothers with BD are more likely (3.4%) to be administrated with a narcotic antagonist (naloxone) compared to those born to healthy mothers (2.4%). The unadjusted OR was (1.44. 95% Cl = 0.99—2.11), and after adjustment for maternal age, parity, plurality, marital status, aboriginality, and sex, the odds remained significant (AOR 1.38; 95% CI 0.94—2.02).
Cesarean Section (CS):
Boden et al. found that women without BD had a rate of cesarean birth of 16.8%, while both untreated and treated women with bipolar had increased risks of 23.5% vs. 25.6%, respectively. However, Frayne et al. found that there was no difference in emergency and elective CS rates between the women with BD and WA mothers. Nguyen et al. found that women with BD are less likely to choose elective CS than WA mothers (12.5% vs 17.9% respectively, P = 0.954), but a non-significant difference in emergency CS was found.
Instrumental Delivery (Assisted Delivery):
Two studies found that BD caused an increased risk of assisted delivery [21] [26]; one of them [21] reported that women without BD had a rate of instrumental birth of 24.7%, while both untreated and treated women with bipolar disorder had increased risks (33.0% vs. 34.1%, respectively). The other study [26] also found that infants of women with BD are more likely to be born through instrumental delivery than WA mothers (28.6% vs. 13.7%, respectively, P = 0.019 “clinically significant”). Frayne et al. found that assisted delivery rates between the women with BD and WA mothers were 20.2% vs. 15.1%, respectively, P = 0.085).
Non-spontaneous Start of Labor (Induction of Labor):
Boden et al. found that BD caused high rates of the non-spontaneous start of labor, as the rate in women with no mental illness was 20.7%, while both untreated and treated women with BD had increased risks of 30.9% vs. 37.5%, respectively.
Probability of incidence of the outcomes:
Calculating the conditional probability of incidence for each complication for bipolar and control cohorts, we found that the most significant differences in the probability of incidence between women with BD and women with no mental illness were in preterm birth (13.2% vs. 5.9%, respectively; in a total population = 1,358,249), gestational HTN (10.5% vs. 0.73%, respectively; total population = 596,325), and microcephaly (10.8% vs. 2.3%, respectively; total population = 336,567). Women with bipolar also showed a higher probability for the incidence of gestational diabetes (5.2% vs. 2.4%, respectively; total population = 394,760), threatened preterm labor (5% vs. 2.4%, respectively; total population = 67,927), and APH (4.6% vs. 3.08%, respectively; total population = 67,927) than unaffected women.
Neonates of women with BD show a higher probability of incidence of some complications, when compared to mentally healthy pregnant women’s neonates: congenital malformations (5.02% vs. 3%, respectively; total population = 770,784), low birth weight (9% vs. 5.4%, respectively; total population = 556,464), neonatal readmission (2.5% vs. 0.9%, respectively; total population = 434,991), and SGA < 10th percentile (16.4% vs. 14%, respectively; total population = 963,215).