Key words
pregnancy outcome - neuropsychiatric illness - abortion - fetal outcome - pregnancy
Introduction
The care of women expecting a child is the central focus of routine obstetric practice;
optimal antenatal care is the primary goal. In order to achieve this, however, in
addition to the pregnancy itself other potential somatic and psychiatric illnesses
must be taken into consideration. Neuropsychiatric illnesses in particular are on
the increase among women of reproductive age worldwide [1]. In 2013 the prevalence of depression among German women was estimated at 11.8%
between 18 – 29 years of age, 10.5% between 30 – 39 years and 9.9% from 40 to 49 years
of age. The lifetime prevalence of depression in German women is as high as 15.4%
[2]. According to the WHOʼs “Global Burden of Disease” study, by 2020 unipolar depression
will be the second most common cause of illness (after ischaemic heart disease) responsible
for reduced quality-of-life [3]. A study from 2014 found that 27.7% of all Germans had suffered from at least one
psychiatric problem in the previous 12 months (amongst others: anxiety disorders,
depression, psychosis, addiction), young women being most affected [4].
An increasing proportion of women of reproductive age suffer from substance dependence/addiction
[5]. According to a study by Gopman et al. in 2014, in New Mexico 10.7% of non-pregnant
women/girls and 5.9% of pregnant women/girls between 15 – 44 years of age consumed
illegal drugs [6].
Studies have proven the negative effects of drugs, alcohol and nicotine consumption
during pregnancy on the unborn child [7].
Swedish studies show that rates of caesarean section, premature birth and low Apgar
scores are higher among women with depression and use of antidepressants than in the
general population [8], [9].
One Canadian study found an increased risk of prematurity and intrauterine growth
restriction (small for gestational age, SGA) for children of mothers with schizophrenia
[10].
However, apart from one study by Grunwald et al. there are no valid data available
on pregnancy outcomes among mentally ill women in Germany [11]. The aim of this study was to describe the differences in pregnancy outcomes between
women with neuropsychiatric illness or substance addiction (with and without medication)
compared to mentally healthy women in a cohort from central Germany, and to define
possible consequences for routine clinical practice.
Patients and Methods
Patient collective, clinical parameters and data collection
Data on 5092 pregnancies managed at a university womenʼs hospital between 01.01.2010
and 31.03.2014 were analysed retrospectively. The study was authorised by the medical
faculty ethics committee (113/14).
In this single centre, retrospective study we analysed data captured routinely for
every birth using the local labour ward program (Nexus, PDM, Version 6.3.3.1), those
captured by the hospitalʼs internal information system (Medico, Version 23.00, Cerner
HS Deutschland) and additional data taken from patient files. The parameters were
collected in tabular form (Microsoft Excel 2010) and anonymised for the analysis.
Patients were divided into a case group (CG) and a control group (CoG) according to
ICD 10 codes for neuropsychiatric diagnoses that had been diagnosed by a specialist
and stated by patients on hospital admission. Thus 185 case pregnancies, subsequently
termed “cases”, and 4907 control pregnancies, subsequently termed “controls” were
identified for comparison. For the case group medical records, consent forms, U1 and
U2 neonatal screening examination results and discharge summaries were studied for
additional information. In the event of in-patient treatment information was expanded
on using level 1 perinatal centre data. For some variables the case group was divided
into specific subgroups according to neuropsychiatric diagnosis. [Fig. 1] provides an overview of the study population.
Fig. 1 Flow diagram of study participants and division into subgroups. AD = antidepressants,
AED = anti-epileptic drugs, subst. = substitution therapy.
The following maternal parameters were analysed: Number of abortions on obstetric
history, age during current pregnancy, number of antenatal visits during current pregnancy
and mode of delivery.
The following neonatal parameters were analysed: interval to estimated date of delivery
(EDD), birth weight, length and head circumference at birth and Apgar scores (after
1, 5 and 10 minutes).
Statistical analysis
Statistical analysis was performed using the statistics software SAS, Version 9.4
(SAS Institute Inc., Cary, NC, USA). The sample was described using simple counting
of events, cross tabulation including χ2 tests, calculation of means and standard
deviations, and medians and quartiles represented in box plots.
Statistical comparison of maternal and neonatal characteristics was conducted at three
levels. At the first level cases were only differentiated from controls. For the next
level cases were subdivided according to the categories depression, epilepsy, substance
abuse, other psychiatric illnesses and other neurological disorders. For more definite
interpretability of results patients in the last two groups were then excluded from
the analysis. At the third level the remaining three subgroups were further subdivided
for use or non-use of disease-specific medications.
For the neonatal criteria birth weight, length and head circumference at birth and
the three Apgar scores comparison between groups was made using covariant analysis
(in SAS-Procedure PROC GLM) with sex and “twin yes/no” as factors in addition to group
allocation, and “interval to EDD” as a covariable. At the second and third levels
described above the Dunnett test was used to compare differences between the various
case subgroups and the control group. Similar comparisons were performed for “interval
to EDD”, only that here the interval to EDD was no longer a covariable.
Maternal age between the groups was compared using a one-way analysis of variance
and the above-stated group allocation (PROC GLM). The number of previous abortions
and extrauterine pregnancies in the obstetric history, and the number of antenatal
visits in the current pregnancy were compared between groups using a generalised linear
model (PROC GLIMMIX) with age as a covariable. Here too in each case the Dunnett comparisons
of the other groups were compared to the control group.
All tests were performed with an assumed error probability of 5%. Adjustment for multiple
tests was performed as described using the Dunnett procedure for comparison of groups.
No further adjustments for analysis of the various parameters or grouping levels was
performed.
Results
Maternal collective
Case numbers
The case group consisted of 169 pregnant women with neuropsychiatric diagnoses including
185 pregnancies with 193 live births. For comparison 4907 pregnancies with 5102 live
born children served as controls. 3.6% of pregnancies were affected by maternal substance
abuse or neuropsychiatric illness.
Maternal neuropsychiatric illness
[Table 1] shows the incidence distribution of neuropsychiatric diagnoses in the case group.
Depression (including bipolar disorder and anxiety disorders), epilepsy, substance
abuse, other psychiatric illnesses and other neurological disorders constitute the
five largest groups numerically. The depression group, with an absolute incidence
of 77 (41.6% of cases; 1.51% of the total collective), was the largest group overall
followed by epilepsy with an absolute incidence of 39 (21.1% of cases; 0.77% of the
total collective) and other neurological disorders with 28 (15.1% of cases; 0.55%
of the total collective). An additional 23 cases (12.5% of cases; 0.45% of the total
collective) formed the group with substance abuse and 18 cases (9.7% of cases; 0.35%
of the total collective) had other psychiatric illnesses.
Table 1 Incidence distribution of maternal neuropsychiatric diagnoses in the case group according
to ICD 10 GM 2014.
|
Maternal neuropsychiatric diagnoses according to ICD 10 (GM 2014)
|
Absolute incidence (n = 185)
|
Relative incidence (%)
|
|
Main group
|
Subgroup
|
|
I Depression (F31 – F33)
|
|
77
|
41.6
|
|
II Epilepsy (G40)
|
|
39
|
21.1
|
|
III Substance abuse (F10, 11, 19)
|
|
23
|
12.4
|
|
IV Other psychiatric illness (F60, 99, 20, 79, 50)
|
|
18
|
9.7
|
|
Of these:
|
|
6
|
3.2
|
|
|
3
|
1.6
|
|
|
4
|
2.2
|
|
|
3
|
1.6
|
|
|
2
|
1.1
|
|
V Other neurological disorders (G35, 70, 50, 91, 81, 43, 25, 71, 25)
|
|
28
|
15.1
|
|
Of these:
|
|
12
|
6.5
|
|
|
1
|
0.5
|
|
|
1
|
0.5
|
|
|
3
|
1.6
|
|
|
3
|
1.6
|
|
|
4
|
2.2
|
|
|
2
|
1.1
|
|
|
1
|
0.5
|
|
|
1
|
0.5
|
Number of previous abortions
Pregnant women in the case group had significantly more previous abortions on obstetric
history than controls (p = 0.015).
Women with depression were the chief contributors to this significant difference (p = 0.001).
On more specific analysis a significantly higher number of previous abortions compared
to controls was found among women with depression who were not taking antidepressants.
In contrast, women with depression who were taking antidepressants did not have significantly
more previous abortions compared to controls.
[Table 2] gives a detailed overview of means, standard deviations and p values.
Table 2 Overview of maternal findings part 1.
|
|
Case groups
|
Control group
|
Significance (p value)*
|
|
MW
|
SD
|
MW
|
SD
|
|
M = mean, SD = standard deviation, AED = antiepileptic drug)
* Statistical significance testing with one-way analysis of variance for the criterion
age, and generalised linear models with age as an additional covariable for the other
criteria (for details see text section statistical analysis). The p values for the
multiple group comparisons are derived from a Dunnett comparison between the subgroup
and the control group in each case.
|
|
Number of previous abortions on obstetric history
|
|
0.33
|
0.69
|
0.24
|
0.60
|
0.015
|
|
|
0.47
|
0.75
|
0.001
|
|
|
0.38
|
0.91
|
0.054
|
|
|
0
|
0
|
0.967
|
|
|
0.51
|
0.83
|
0.002
|
|
|
0.40
|
0.62
|
0.139
|
|
|
0.33
|
0.49
|
0.313
|
|
|
0.41
|
1.05
|
0.100
|
|
|
0
|
0
|
0.971
|
|
|
0
|
0
|
0.986
|
|
Age
|
|
29.40
|
5.86
|
29.67
|
5.62
|
0.543
|
|
|
31.23
|
4.93
|
0.043
|
|
|
28.87
|
6.29
|
0.765
|
|
|
24.83
|
5.89
|
< 0.001
|
|
|
31.34
|
5.29
|
0.222
|
|
|
31.07
|
4.38
|
0.675
|
|
|
27.25
|
6.12
|
0.589
|
|
|
29.59
|
6.34
|
1.000
|
|
|
24.11
|
5.88
|
< 0.001
|
|
|
28.25
|
5.32
|
0.997
|
|
Number of antenatal visits
|
|
12.63
|
3.90
|
12.05
|
3.97
|
0.026
|
|
|
13.75
|
3.94
|
< 0.001
|
|
|
11.45
|
3.29
|
0.323
|
|
|
10.70
|
3.64
|
0.198
|
|
|
14.06
|
3.57
|
< 0.001
|
|
|
13.21
|
4.51
|
0.112
|
|
|
11.67
|
2.67
|
0.831
|
|
|
11.35
|
3.59
|
0.294
|
|
|
10.41
|
3.73
|
0.145
|
|
|
12.33
|
3.21
|
0.888
|
Age during current pregnancy
[Table 2]: There was no significant difference in average age of women between the case and
control groups overall (p = 0.543).
Women in the case group with depression were significantly older than women in the
control group (p = 0.043).
Women in the case group with substance abuse were significantly younger on average
than women in the control group (p < 0.001) and significance was particularly pronounced
for those not on substitution therapy (p < 0.001).
Number of antenatal visits during current pregnancy
[Table 2]: The number of antenatal visits during the current pregnancy was significantly higher
for the case group compared to controls (p = 0.026). This significance was achieved
by the groups of women with depression and those with antidepressant use in particular
(p < 0.001).
Mode of delivery
[Table 3]: Significantly more mothers in the case group had a primary classical caesarean
section (18.3%, p = 0.015) than in the control group (11.9%). Secondary caesareans
and other forms of caesarean section were also significantly more frequent in the
case group (CG) than in the control group (CoG) (secondary caesarean: CG 14.7%; CoG
12.6%; other caesarean sections: CG 11.0%, CoG 7.9%). The overall caesarean section
rate in the case group was 44% compared to 32.4% in the control group. The most common
mode of delivery in both groups was spontaneous vaginal delivery, with a higher frequency
in the control group (CG 53.4%; CoG 64,8%).
Table 3 Overview of maternal results part 2.
|
|
Case group (%)
|
Control group (%)
|
Significance (p value)*
|
|
* Statistical significance testing with χ2 test, significance level p < 0.05
|
|
Mode of delivery
|
|
53.4
|
64.8
|
0.015
|
|
|
18.3
|
11.9
|
|
|
14.7
|
12.6
|
|
|
11.0
|
7.9
|
|
|
2.6
|
2.9
|
Neonatal collective
Interval to estimated date of delivery (EDD)
[Fig. 2]: Children of mothers in the case group were born significantly earlier on average
than children of mothers in the control group (p = 0.003).
Fig. 2 Box plot of interval to estimated date of delivery for case group and control group
children. Y-axis: interval to EDD in days, 0 = EDD (40/0 SSW).
[Table 4] shows means, standard deviations and p values.
Table 4 Overview of results for neonatal parameters.
|
|
Case groups
|
Control groups
|
Significance (p values)*
|
|
M
|
SD
|
M
|
SD
|
|
M = mean, SD = Standard deviation, AED = antiepileptic drug
* Statistical significance of testing with multifactorial analysis of covariance
with additional factors sex and multiple birth (yes/no) and interval to EDD as covariables.
For comparison of multiple groups the Dunnett test results of the other groups compared
to the control group are shown. Significance level p < 0.05.
|
|
Interval to EDD in days
|
|
− 14.84
|
22.55
|
− 10.43
|
21.65
|
0.003
|
|
|
− 12.51
|
20.27
|
0.739
|
|
|
− 12.20
|
18.75
|
0.830
|
|
|
− 16.77
|
24.70
|
0.155
|
|
|
− 13.84
|
19,78
|
0.971
|
|
|
− 10.30
|
21.22
|
0.996
|
|
|
− 11.00
|
12.56
|
1.000
|
|
|
− 12.78
|
21.29
|
0.841
|
|
|
− 17.61
|
27.06
|
0.293
|
|
|
− 13.00
|
9.90
|
0.998
|
|
Birth weight in g
|
|
3013.77
|
768.37
|
3162.96
|
780.09
|
0.378
|
|
|
3120.75
|
680.52
|
0.993
|
|
|
3106.50
|
719.56
|
0.987
|
|
|
2738.48
|
857.73
|
0.022
|
|
|
3116.40
|
728.15
|
0.945
|
|
|
3128.00
|
604.67
|
0.977
|
|
|
2952.69
|
610.50
|
0.594
|
|
|
3180.56
|
766.23
|
0.987
|
|
|
2698.42
|
870.44
|
0.049
|
|
|
2928.75
|
889.22
|
0.984
|
|
Birth length in cm
|
|
49.26
|
4.58
|
50.85
|
2.95
|
0.013
|
|
|
49.93
|
3.71
|
0.997
|
|
|
50.00
|
3.93
|
0.940
|
|
|
47.70
|
4.98
|
< 0.001
|
|
|
50.06
|
3.87
|
0.950
|
|
|
49.71
|
3.45
|
0.635
|
|
|
50.25
|
2.56
|
0.997
|
|
|
49.89
|
4.45
|
1.000
|
|
|
47.42
|
5.25
|
< 0.001
|
|
|
49.00
|
3.65
|
0.979
|
|
Head circumference in cm
|
|
33.75
|
5.54
|
34.40
|
1.74
|
0.916
|
|
|
34.53
|
7.71
|
0.012
|
|
|
33.87
|
2.60
|
0.916
|
|
|
32.17
|
3.10
|
< 0.001
|
|
|
33.64
|
2.62
|
0.962
|
|
|
36.05
|
12.24
|
< 0.001
|
|
|
34.00
|
1.45
|
0.995
|
|
|
33.81
|
2.99
|
1.000
|
|
|
31.97
|
3.32
|
< 0.001
|
|
|
33.13
|
1.75
|
0.975
|
|
Apgar score at 1 min
|
|
8.63
|
1.38
|
8.79
|
1.37
|
0.765
|
|
|
8.76
|
1.33
|
0.987
|
|
|
8.30
|
1.77
|
0.076
|
|
|
8.70
|
1.11
|
0.996
|
|
|
8.88
|
1.17
|
0.820
|
|
|
8.57
|
1.57
|
0.895
|
|
|
8.77
|
0.83
|
1.000
|
|
|
8.07
|
2.06
|
0.035
|
|
|
8.63
|
1.17
|
1.000
|
|
|
9.00
|
0.82
|
0.998
|
|
Apgar score at 5 min
|
|
9.40
|
1.01
|
9.50
|
1.03
|
0.940
|
|
|
9.51
|
0.89
|
0.891
|
|
|
9.25
|
1.24
|
0.386
|
|
|
9.30
|
1.02
|
0.974
|
|
|
9.58
|
0.84
|
0.725
|
|
|
9.40
|
0.97
|
0.989
|
|
|
9.46
|
0.66
|
1.000
|
|
|
9.15
|
1.43
|
0.413
|
|
|
9.21
|
1.08
|
0.976
|
|
|
9.75
|
0.50
|
0.985
|
|
Apgar score at 10 min
|
|
9.68
|
0.69
|
9.75
|
0.78
|
0.796
|
|
|
9.70
|
0.70
|
0.999
|
|
|
9.70
|
0.65
|
0.999
|
|
|
9.61
|
0.84
|
0.996
|
|
|
9.74
|
0.72
|
0.997
|
|
|
9.63
|
0.67
|
0.949
|
|
|
9.85
|
0.38
|
0.995
|
|
|
9.63
|
0.74
|
0.997
|
|
|
9.53
|
0.91
|
0.988
|
|
|
10.00
|
0.00
|
0.945
|
Birth weight in g, length in cm and head circumference in cm
[Table 4]: On average, and taking gestational age into account, children of case group mothers
were not significantly lighter (SGA) or heavier (LGA) than children of control group
mothers, nor did they have significantly smaller head circumferences (p = 0.378, p = 0.916).
Case group children were however significantly smaller in terms of birth length than
control group children (p = 0.013).
On subgroup analysis significant differences between control group children and children
of mothers with substance abuse were found for each of the parameters birth weight,
length and head circumference. These differences were particularly significant for
children of women with substance abuse without substitution therapy.
Also, children of women with depression and children of women on antidepressants had
significantly larger head circumferences than control group children (p = 0.012, p < 0.001).
Apgar scores
There were no significant differences in Apgar scores at 1, 5 and 10 minutes between
children of case group mothers compared to controls. Only children whose mothers had
epilepsy were treated with antiepileptic drugs showed significantly lower 1 minute
Apgar scores (p = 0.035).
[Table 4] provides an overview of the characteristics of case group and control group children.
Discussion
The prevalence of major depression among American pregnant women has been estimated
at 12.4% [12]. The prevalence of general depressive symptoms among pregnant women is as high as
18.4% and in combination with anxiety disorders between 4.4 and 39% [13]. In industrialised countries rates of antenatal depression are between 7 – 15% [14]. In our study collective the prevalence of depression and anxiety disorders was
1.5%, which is relatively low compared to the above-mentioned studies.
The prevalence of epilepsy in the study collective (0.8%) was similar to that in a
German study from 2016 [11]. An American study also showed similar epilepsy prevalence (0.3 – 0.5%) [15].
According to data from the American “National Survey on Drug Use and Health (NSDUH)”
from 2012 the prevalence of illegal drug use during pregnancy was 5.9% among American
women [6]. Data from central Germany show a prevalence of substance abuse among pregnant women
of only 0.7% [11]. In our single-centre cohort the prevalence of substance abuse was comparable at
0.5%. These differences in prevalence between American and central German data presumably
reflect population differences as well as differences in antenatal care provision
between the two countries.
Limited data have been published on pregnancy and neonatal outcomes in relation to
maternal mental health in Germany. In this study we were able to analyse 5092 pregnancies
in terms of maternal disease and disease-specific medication. Parameter documentation
occurred during routine clinical care with data collection starting prospectively
at first presentation to the maternity unit and being completed at delivery. Cases
and controls were recruited from the same basis population (central Germany) and thus
represent an homogenous cohort.
This study shows an increased number of previous abortions on obstetric history among
women with neuropsychiatric illnesses. Women with depression were most affected. A
study from 2015 found correlation between recurrent abortions, increased prevalence
of depression and high levels of emotional stress [16]. Another study showed that the risk of abortion (maternal age 25 – 29 years) increased
from 8.9 – 9.3% in women without previous abortions on obstetric history, to 11.8 – 12.4%
in women with one previous abortion, and to 17.7 to 22.7% with two previous abortions.
Thus women with neuropsychiatric illness also had an increased risk of future abortions
(due to higher numbers of previous abortions) [17].
It is notable in this study that, when medication is taken into account, abortion
risk was significantly increased in women with depression who were not taking antidepressants.
In contrast risk of abortion was not significantly increased among women with depression
who were taking antidepressants. Increased risk of abortion therefore does not primarily
appear to be a toxic effect of medication; rather “depression” itself is a risk factor
for recurrent abortion. This conclusion has also been reached in relation to use of
duloxetine (SSNRI = selective serotonin noradrenaline reuptake inhibitor) during pregnancy.
In this study the authors found that the risk of fatal pregnancy outcomes (spontaneous
abortion, premature birth) was not increased in women taking duloxetine [18]. Studies of the use of SSRIs (fluvoxamine, paroxetine, sertraline, escitalopram)
in pregnancy have also not shown any increased risk of spontaneous abortion [19], [20].
One study from 2014 however showed controversial results contradictory to those above,
finding a one to three times increased relative risk of spontaneous abortion for the
use of duloxetine during pregnancy [21].
Analysis of our data confirms the focus on depression as the relevant risk factor.
Women with depression were significantly older than controls so that age may be a
possible cofactor. One study has shown that the risk of spontaneous abortion increases
from 8.7% at the age of 22 to 84.1% at the age of 48 years [17], [22], [23].
In our study spontaneous vaginal delivery was the most common mode of delivery in
both the case and control groups, however approximately 10% fewer women in the case
group delivered spontaneously (CG = 53.4%; CoG = 64.8%). The proportion of caesarean
sections overall was 44% in the case group and only 32.4% in the control group, i.e.
10% more women in the case group had caesarean sections than in the control group.
An increased caesarean section rate has been described in women with bipolar disorder
(23.5%) compared to women without bipolar disorder (16.8%) [24], [25], [26]. Interestingly, in the case group more planned, primary caesareans (18.3%) than
secondary caesarean sections (14.7%) were performed, which may have been due to existing
maternal illness. The opposite was true in the control group (primary caesarean = 11.9%,
secondary caesarean = 12.6%), the decision to perform caesarean section more commonly
being made after labour had commenced, due to unpredictable (in contrast to primary
caesarean), often fetal birth complications. The remaining numbers were made up by
operative vaginal deliveries (forceps, vacuum extraction).
Our data also show a significant difference in the interval between birth and estimated
date of delivery between children of case group and control group mothers. Children
of women with neuropsychiatric illness were born on average 14.84 days before their
EDD while control group children were born 10.43 days before EDD. A meta-analysis
from 2013 also showed a relationship between women with medically treated depression
and birth before EDD [27].
Common to both groups in this case-control study was that although births occurred
early, they did not reach prematurity (≥ 37/0 weeks of gestation). In contrast, a
study from 2014 found that children of schizophrenic women were at increased risk
of premature birth. Increased risk of prematurity has also been shown in studies of
women with post-traumatic stress disorder, anxiety disorders, depression and heroin
consumption [10], [28], [29].
A possible reason for children in our data analysis on average not being born prematurely
is that there were more women in the case group not on medical treatment than women
taking medication. Numerous studies have shown that, amongst other medications, antidepressants
may increase the risk of premature birth, and that the risk of prematurity increases
with increasing severity of depression [30], [31]. Since the majority of women in our case group were not on antidepressants we presume
that their depression was milder, and therefore the risk of prematurity lower.
In contrast to gestational age at birth there was no significant difference for average
birth weight or head circumference between cases and controls, taking gestational
age into account. There was also no significant difference between the groups on comparison
of Apgar scores at 1, 5 and 10 minutes. Although children of case group mothers were
an average of 82.2 g lighter than control children, producing a trend, this is more
likely attributable to shorter gestations than to the maternal illnesses themselves.
Only the children of women with substance abuse and especially those without substitution
therapy had significantly lower birth weight and smaller length and head circumference
measurements, a finding consistent with other studies [32], [33].
The meta-analysis by Ross et al. provides contrasting results: children whose mothers
had depression were significantly lighter and had lower Apgar scores than children
of mothers without medically treated neuropsychiatric illness [27]. Other studies have also shown that newborns of women with neuropsychiatric illness
are at increased risk of low birth weight [11], [34]. This difference can also be explained by the greater proportion of control group
women in this study without medication than with medication, since according to study
data neuropsychiatric medication in particular can result in low birth weight [35], [36].
Large meta-analyses of maternal epilepsy, depression and anxiety disorders have identified
the disorders themselves as risk factors for abortion, prematurity and intrauterine
growth retardation independent of disease-specific medication [37] – [39]. The influence of cofactors such as tobacco smoking, alcohol consumption and illicit
drug use is controversial [40], [41].
Data for this study were collected continuously over a period of 3.25 years by specialised
staff at a single institution irrespective of case or control group allocation. Nevertheless
on critical review “recall error” is possible among these mothers with addiction or
neuropsychiatric illness. In order to minimise this error maternal and neonatal data
were also collected from postnatal documentation by doctors and nurses as well as
medical and midwifery documentation of antenatal visits. For the parameter “mode of
delivery” cases in the subgroup “other caesarean sections” could not be allocated
to either of the subgroups “primary classical caesarean” or “secondary caesarean”.
Practice relevant findings may have resulted if allocation to these groups had been
possible, however incomplete and implausible data had to be excluded from the analysis.
Conclusion for Clinical Practice
Conclusion for Clinical Practice
This study provides evidence that women with neuropsychiatric illnesses are at increased
risk of having an abortion.
The data also suggest that antidepressants themselves are not solely responsible for
this increased risk.
The rate of caesarean section is increased by 10% in the context of maternal neuropsychiatric
illness and substance abuse.
Children of mothers with neuropsychiatric illnesses are born at younger gestational
ages. These newborns however do not have worse perinatal outcomes: birth is generally
not preterm, and parameters such as birth weight, head circumference and Apgar scores
are comparable with newborns of mentally healthy women.
This study underscores the significance of maternal neuropsychiatric health for pregnancy
outcomes and highlights the need for specialised antenatal care of affected women.
