Key words
pregnancy - obstetrics - birth - migration
Schlüsselwörter
Schwangerschaft - Geburtshilfe - Geburt - Migration
Introduction
In the USA, Great Britain, the Netherlands and Scandinavia, obstetrical outcomes have been found to
differ between immigrant and non-immigrant women (cf. [1], [2], [3], [4]). Unfavorable
socio-economic circumstances associated with immigration, a lack of antenatal and perinatal care,
difficulties in communication, cultural factors and even insufficiently understood biological factors
affecting immigrant women can lead to significantly higher rates of preterm births, increased perinatal
mortality rates and higher cesarean section rates. This raises the question whether similar differences
in perinatal health and healthcare can also be observed in Germany.
Around 15 million people out of the current population of 80.2 million persons living in the Federal
Republic of Germany are first, second, or third-generation immigrants; this corresponds to almost 19 %
of the total population; in Berlin this amounts to 23.9 % of the resident population [5]. The first articles on “babies born to foreign women” in Germany were
published at the end of the 1960s, and the potential negative impact of a lack of German language skills
and limited means of communication on antenatal care and obstetrical outcomes was discussed [6], [7].
Only a few studies have been published since then; one important study was a (retrospective) analysis of
maternal mortality in Bavaria (for the years 1983 to 2000) which showed a significantly higher mortality
rate for immigrant women compared to German women, although the rates began to converge after several
years [8]. In 2006 David et al. also published a retrospective study of
obstetrical outcomes for immigrant women of mainly Turkish origin, based on perinatal data from Berlin
for the years 1993 to 1999. The study showed that important perinatal quality parameters such as infant
and maternal mortality and preterm birth rates have largely converged; however, differences between
immigrant and non-immigrant women (rates for cesarean section, epidural anesthesia, and anemia) could
indicate persisting differences in the quality of care [9]. The comment in the
review article of 2009 by Gissler et al. [11] on the quality of the studies
applies both to the above-mentioned studies and to the register-based analysis of maternal mortality
among non-German women published in 1999 by Razum et al. [10]: information on
the socio-demographic and immigration background of immigrant women is very limited, as this information
is based on population and healthcare registers; this means that detailed information on (immigrant)
parents/mothers and their newborn infants is lacking, making it impossible to identify certain causal
medical and/or socio-demographic factors. In particular, it was not possible to verify whether
progressive acculturation and integration have resulted in changes and which determinants could have
played a role in these changes.
Immigrant women of Turkish origin are the largest group of immigrant women in Germany (besides ethnic
German “resettlers”) [5]. This group of immigrants already has a long history
of immigration to Germany, so that acculturation and integration may have already led to better
perinatal outcomes. However, immigrants of Turkish origin in Germany are regarded as underprivileged
socio-economically and disadvantaged with regard to health and healthcare [12]. This study aimed to examine antenatal and intrapartum process indicators and outcomes
for these immigrant groups and compare them to the figures for non-immigrant German women, based on the
following 6 questions on antenatal and intrapartum care:
-
Are there differences in attendance rates for antenatal examinations between immigrant and
non-immigrant women?
-
Are high-risk pregnancies more common in immigrant women?
-
Is antenatal and postnatal anemia more common in immigrant women?
-
Do immigrant women have similar rates for PDA/CSE administration during delivery compared to
non-immigrant women?
-
Are there differences in the mode of delivery?
-
Are perinatal outcomes less favourable for neonates born to immigrant women of Turkish
origin?
Methods
Data collection
Data collection was done as part of a study financed by the German Research Foundation (Fkz.: DA
1199/2-1) in 3 maternity hospitals in Berlin (Charité/Campus Virchow-Klinikum, Vivantes Klinikum
Neukölln, Vivantes Klinikum am Urban) and was based on standardized interviews and validated
questionnaires which were available in both German and Turkish. The primary data obtained in this
manner was supplemented by information obtained from the Mutterpass, an official maternal
record of antenatal and natal care created for every pregnant woman in Germany. This data was then
linked to the perinatal data recorded by the respective hospital during delivery, which must be
reported to the AQUA Institute as part of quality assurance. The questionnaire consisted of 3 parts:
socio-demographic questions (23 items), questions on aspects of care (9 items) and, for immigrant
women, questions regarding immigration and acculturation (8 + 15 items, respectively).
Determining the immigration status of the women was done in accordance with the recommendation of
Schenk et al. (2006) [13]. Using the information provided by the women
about the country of birth of their parents, the number of years the women had lived in Germany,
their native language, their own assessment of their knowledge and understanding of German, and
their residence permit status in Germany, the women of Turkish origin were grouped as follows: 1st
generation immigrant = woman born in Turkey; 2nd generation immigrant = both parents born in Turkey;
3rd generation immigrant = native language of the woman is Turkish of Kurdish, the woman herself and
both her parents were born in Germany; and women with a so-called binational background (one
immigrant-background parent from Turkey and one non-immigrant parent).
After a preliminary test run, data collection started in January 2011 and was continued for a period
of 1 year. Data collection was done by trained multi-lingual study nurses and student researchers
working on the project. Data were collected in the labor rooms and maternity wards of the 3
above-listed hospitals using a 2-shift system. The aim was to question the women a few hours before
they gave birth.
Inclusion criteria for the study were: all women admitted to one of the three participating maternity
hospitals who gave birth to a child (at GW 24/0 and above, with the child showing vital signs)
during the study period, who were aged at least 18 years when their child was born, and who had
permanent residence in Germany. Women who were underage, tourists who did not reside permanently in
Germany, and women who had a termination of pregnancy including miscarriage and stillbirth (death of
the child ascertained on admission to hospital and before the onset of labor) were excluded from the
study.
Statistical analysis and data protection
The data of 4598 women were included for analysis in this study. Due to their limited numbers, 3rd
generation immigrant women are not listed as a separate group but are included in the group of 2nd
generation immigrant women. For some of the analyses, 1st, 2nd and 3rd generation immigrant women on
the one hand, as well as the women with one immigrant parent (binational women) and non-immigrant
women on the other hand, were grouped together.
In addition to descriptive analysis, multivariate analysis was done to assess the impact of
immigration and acculturation processes on pregnancy and birth. Generalized linear models (Poisson
regression models) and binary and multinomial logistic regression analysis were used for
multivariate analysis. Persons for whom one or more datasets were missing for independent variables
were not included in the respective descriptive and multivariate analyses. This approach explains
the differences in the numbers of assessed Mutterpässe. The level of significance was taken
as p < 0.05. All analysis was done using the statistical software SAS 9.2.
The survey of the women, the compilation and merging of primary and secondary data, and the data
assessment complied with German laws on data protection. The Ethics Commission of the Charité
University Medicine Berlin approved the study.
Results
Study population
A total of 8157 births occurred at the 3 maternity hospitals during the one-year observation period.
9.2 % of women who were approached for the study declined to take part in the study or could not be
reached. After applying the inclusion and exclusion criteria, a total of 7100 women were included in
the study, which corresponds to a response rate of 89.6 %. For the analysis presented here, only
immigrant women of Turkish origin (n = 1277), non-immigrant women (n = 2991) and so-called
binational women were included (n = 330).
The interview with each woman took 15–25 min. Interviews with 6.4 % of 1st generation immigrant women
(of Turkish origin) and with 0.8 % of 2nd and 3rd generation immigrant women were held in Turkish.
[Table 1] compares the socio-demographic data for the 4 groups (total
number of women: n = 4598).
Table 1 Comparison of socio-demographic data for the four investigated
groups.
|
Sociodemographic variables
|
1st generation immigrant women of Turkish origin
|
2nd/3rd generation immigrant women of Turkish origin
|
Binational women
|
Non-immigrant women
|
|
1 t-test
2 χ2-test (only the group of
gravida II/multipara)
3 χ2-test (only the
University Entrance Diploma group)
4 χ2-test (only
the group with a very good knowledge of German)
* p < 0.05
**
p < 0.01
*** p < 0.001
n. s.: not
significant
# if German was not their first native
language; different n are the result of the different responses to individual
questions
|
|
Age
|
n = 697
|
n = 580
|
n = 330
|
n = 2 991
|
|
|
21.1
|
25.3
|
26.1
|
15.9
|
|
|
29.7
|
34.0
|
22.7
|
24.7
|
|
|
27.8
|
26.0
|
30.6
|
32.6
|
|
|
21.4
|
14.7
|
20.6
|
26.8
|
|
|
29.7***
|
28.6***
|
29.4***
|
30.9
|
|
|
29.0
|
28.0
|
30.0
|
31.0
|
|
Parity
|
n = 695
|
n = 580
|
n = 328
|
n = 2 988
|
|
|
24.7
|
43.5
|
57.0
|
56.3
|
|
|
34.0
|
32.9
|
25.3
|
30.2
|
|
|
41.3***
|
23.6***
|
17.7*
|
13.5
|
|
BMI (kg/m2)
|
n = 602
|
n = 518
|
n = 294
|
n = 2 647
|
|
|
3.0
|
3.3
|
5.1
|
4.5
|
|
|
46.5
|
53.5
|
61.2
|
63.7
|
|
|
36.2
|
24.5
|
19.1
|
20.3
|
|
|
14.3
|
18.7
|
14.6
|
11.6
|
|
|
25.6***
|
25.6***
|
24.6 (n. s.)
|
24.2
|
|
Qualifications
|
n = 689
|
n = 579
|
n = 328
|
n = 2 982
|
|
|
40.2
|
7.1
|
6.7
|
3.0
|
|
|
48.2
|
79.5
|
52.7
|
46.5
|
|
|
11.6***
|
13.5***
|
40.6***
|
50.6
|
|
Knowledge of German# according to own assessment
|
n = 658
|
n = 421
|
n = 37
|
–
|
|
|
6.7
|
0.0
|
0.0
|
|
|
|
30.2
|
1.0
|
5.4
|
|
|
|
36.0
|
9.0
|
5.4
|
|
|
|
17.3
|
36.1
|
27.0
|
|
|
|
9.7***
|
53.9*
|
62.2
|
|
Antenatal check-up
In the Mutterpässe available for analysis (n = 3703) the numbers of antenatal check-ups ranged
between 1 and 35. The mean figure was 11.2 check-ups (median: 11). Non-immigrant women attended an
average of 11.4 antenatal check-ups during pregnancy (median value: 11), binational women attended
11.3 (median: 11), 1st generation immigrant women of Turkish origin attended 10.6 (median: 10) and
2nd/3rd generation immigrant women attended 11.2 antenatal check-ups (median: 11). No statistically
significant differences was found between non-immigrant women (reference population) and immigrant
women in the expanded Poisson regression model (immigrant status adjusted for additional variables)
of the regression analysis for the parameter “number of antenatal examinations attended during
pregnancy”.
High-risk pregnancies/pregnancy risks
Based on the data obtained from the Mutterpass, 38.4 % of all women studied (n = 3996) had a
high-risk pregnancy. High-risk pregnancies were recorded for almost 41 % of non-immigrant women;
this figure was only 32 % for 1st generation immigrant women of Turkish origin, 37 % for 2nd/3rd
generation immigrant women, and 34 % for binational women. The difference between non-immigrant
women (reference population) and 1st generation immigrant women of Turkish origin was statistically
significant (p < 0.001); the difference compared to 2nd/3rd generation immigrant women (p = 0.08)
and to women with binational parents (p < 0.05) was also marked.
There were entries under the heading “pregnancy risks” in the Mutterpass of 76.2 % of all
women. Documented pregnancy risks were slightly higher for non-immigrant women (76.4 %) compared to
1st generation immigrant women of Turkish origin (75.9 %), but levels for non-immigrant women were
similar to those of 2nd/3rd generation immigrant women (76.9 %). The most common pregnancy risks
listed for the 3996 women included in the study were familial history (n = 915) and pregnancy above
the age of 35 years (n = 634). After adjusting for age, BMI, number of previous births and
school-leaving qualifications, 1st generation immigrant women of Turkish origin had a statistically
significantly lower risk of having recorded pregnancy risks compared to non-immigrant women in the
generalized linear model (Poisson regression analysis). We also investigated whether immigrants of
Turkish origin were less likely to have one or more pregnancy risks compared to non-immigrant women.
This was found only to apply to 1st generation immigrant women of Turkish origin (OR: 0.59; 95 % CI:
0.36–0.97).
Antenatal and postnatal anemia
Antenatal anemia (defined as a hemoglobin [Hb] value of < 10 g/dl) was recorded in the
Mutterpass of 7.2 % of all investigated women (n = 3765); this figure was found to be
higher for 1st, 2nd and 3rd generation immigrant women of Turkish origin compared to binational
women and non-immigrant women ([Table 2]).
Table 2 Antenatal anemia recorded in the Mutterpass of pregnant women
according to the womanʼs immigrant status (n = 3 765).
|
|
n
|
%
|
|
1st generation immigrant women of Turkish origin
|
64
|
11.4
|
|
2nd/3rd generation immigrant women of Turkish origin
|
58
|
11.9
|
|
Binational women
|
9
|
6.9
|
|
Non-immigrant women
|
130
|
5.3
|
|
Total
|
261
|
7.2
|
While the base model of logistic regression analysis (containing only the variable ‘immigrant
status’) showed a significantly higher chance of antenatal anemia for immigrant women of Turkish
origin, only a trend remained in the expanded model which took account of a number of additional
co-factors ([Table 3]). The chance of anemia decreases with higher age
and increasing BMI, but increases with higher numbers of previous births.
Table 3 Chance of antenatal anemia (based on Mutterpass records),
logistic regression (n = 3 765) (bold: p-value < 0.05).
|
n = 3 765
|
OR
|
95 % CI
|
p-value
|
|
Non-immigrant women
|
1.00
|
|
|
|
1st generation immigrant women
|
1.37
|
0.66–2.86
|
0.4037
|
|
2nd generation immigrant women
|
1.59
|
0.81–3.11
|
0.1797
|
|
Binational women
|
1.16
|
0.69–1.95
|
0.5881
|
|
Age
|
|
|
|
|
|
1.00
|
|
|
|
|
0.79
|
0.56–1.13
|
0.2014
|
|
|
0.68
|
0.46–0.99
|
0.0436
|
|
|
0.71
|
0.46–1.08
|
0.1058
|
|
Parity
|
|
|
|
|
|
1.00
|
|
|
|
|
1.04
|
0.79–1.38
|
0.7794
|
|
|
1.64
|
0.99–2.71
|
0.0540
|
|
BMI (kg/m2)
|
|
|
|
|
|
1.00
|
|
|
|
|
0.78
|
0.57–1.06
|
0.1120
|
|
|
0.58
|
0.38–0.90
|
0.0134
|
|
Qualifications
|
|
|
|
|
|
1.00
|
|
|
|
|
1.23
|
0.88–1.70
|
0.2228
|
|
|
1.32
|
0.80–2.19
|
0.2764
|
|
Native language
|
|
|
|
|
|
1.00
|
|
|
|
|
1.44
|
0.74–2.80
|
0.2873
|
|
|
1.47
|
0.64–3.39
|
0.3628
|
|
Pregnancies in rapid succession
|
|
|
|
|
|
1.00
|
|
|
|
|
0.78
|
0.40–1.52
|
0.4671
|
The average Hb value postpartum was 10.9 g/dl. The mean Hb value for 1st generation immigrant women
was 10.9 g/dl (median 10.9 g/dl); the mean value was 10.6 g/dl (median 10.8 g/dl) for 2nd/3rd
generation immigrant women, 10.8 g/dl (median 10.7) for binational women, and 11.0 g/dl (median
11.1) for non-immigrant women. Postpartum anemia with Hb values below 10 g/dl was recorded in 26.7 %
(p < 0.05) of 2nd/3rd generation immigrant women compared to 20 % of non-immigrant women and
21.9 % (p = 0.4292) of 1st generation immigrant women (respective reference: non-immigrant women).
An expanded logistic regression analysis model confirmed that 2nd/3rd generation immigrant women had
a significantly higher risk of postpartum anemia. The strongest association was found, as expected,
between low Hb value and increased postpartum bleeding of > 1000 ml. Even when the rare
incidences of significant blood loss (n = 43 in our studied population) were not included in the
analysis, immigrant women of Turkish origin, particularly 2nd/3rd generation immigrant women, had
significantly lower Hb values compared to non-immigrant women.
Mode of delivery
[Table 4] shows the incidence of surgical delivery for the different
patient groups (total patient population n = 4595) as well as the incidence of episiotomies and 3rd
and 4th grade perineal tears during delivery. There was a significantly higher proportion of higher
grade perineal tears in the group of 2nd/3rd generation immigrant women compared to the other study
groups.
Table 4 Incidence of surgical deliveries, episiotomies and higher grade
perineal tears according to the motherʼs immigration status (in %).
|
|
1st generation immigrant women of Turkish origin
|
2nd/3rd generation immigrant women of Turkish origin
|
Binational women
|
Non-immigrant women
|
|
n. s. = not significant
* p < 0.05
*** p < 0.001
|
|
Total cesarean section rate
|
22.0***
|
31.7***
|
38.8 (n. s.)
|
39.3
|
|
|
10.5
|
12.2
|
13.6
|
15.4
|
|
|
11.5
|
19.5
|
25.2
|
23.9
|
|
Vacuum extraction
|
8.9 (n. s.)
|
8.6 (n. s.)
|
10.3 (n. s.)
|
9.5
|
|
Forceps delivery
|
0.0
|
0.2
|
0.0
|
0.1
|
|
Episiotomy
|
8.1*
|
10.2 (n. s.)
|
9.8 (n. s.)
|
11.4
|
|
3rd/4th grade perineal tear
|
0.7 (n. s.)
|
2.4*
|
0.6 (n. s.)
|
1.1
|
Multifactorial logistic regression analysis showed that, after adjustment, there were no significant
differences between the four study groups with regard to surgical vaginal deliveries, while
C-section rates were significantly lower in the group of immigrant women (OR: 0.71;
p < 0.001).
Compared to non-immigrant women, immigrant women of Turkish origin had an overall lower risk of
secondary C-section (OR: 0.69; 95 % CI: 0.54–0.87). Moreover, immigrant women had significantly
lower elective C-section rates (OR: 0.73; 95 % CI: 0.55–0.96). A poorer understanding of German was
also associated with a lower rate of elective C-section (OR: 0.47; 95 % CI: 0.23–0.96). However,
after adjustment, difference in the risk of episiotomy and of higher grade perineal tears between
immigrant and non-immigrant women was not statistically significant (OR: 1.15; p = 0.33 and OR:
0.79; p = 0.46 respectively).
Incidence of PDA/CSE
The recorded rates for epidural anesthesia (PDA) and combined spinal-epidural analgesia (CSE) during
delivery differed according to the study group and the mode of delivery. In the group of vaginal
deliveries, 44.4 % of non-immigrant women and 28.2 % (p < 0.001) of immigrant women of Turkish
origin had a PDA/CSE; in the group who underwent secondary C-section 84.4 % of non-immigrant women
and 75.4 % of immigrant women had this form of analgesia (p < 0.01). Multivariate analysis found
no significant difference in the chance of PDA/CSE between immigrant women of Turkish origin and
non-immigrant women. However, a limited understanding of German by immigrant women was associated
with a significantly lower chance of PDA/CSE (OR: 0.67; 95 % CI: 0.48–0.95).
Perinatal outcome
[Table 5] shows the rates for normal and critical pH values for arterial
umbilical cord blood and for 5-minute Apgar scores. Additionally (not shown in [Table 5]), more than 93 % of neonates (in multiple births, only the first child was
included) had a 1-minute Apgar score of between 7 and 10 points. Only 0.4 % of all neonates had a
10-minute Apgar score of less than 7 after delivery.
Table 5 pH values for arterial umbilical cord blood and Apgar scores for
neonates at 5 minutes after delivery according to the motherʼs immigration status (in %)
(n = 4 552).
|
|
1st generation immigrant women of Turkish origin
|
2nd/3rd generation immigrant women of Turkish origin
|
Binational women
|
Non-immigrant women
|
|
1 χ2-test (only group with normal pH
values)
2 χ2-test (only group with 7–10
points)
|
|
pH value for arterial umbilical cord blood
|
|
|
98.4 (n. s.)
|
97.4 (n. s.)
|
97.3 (n. s.)
|
96.8
|
|
|
1.3
|
2.6
|
2.5
|
2.9
|
|
|
0.3
|
0.0
|
0.3
|
0.3
|
|
Apgar scores
5 min after delivery
|
|
|
0.0
|
0.0
|
0.0
|
0.3
|
|
|
1.5
|
1.9
|
2.1
|
2.3
|
|
|
98.5 (n. s.)
|
98.1 (n. s.)
|
97.9 (n. s.)
|
97.5
|
For logistic regression analysis, neonates who had a pH value for arterial umbilical cord blood of
≤ 7.10 (dependent variable) were grouped together. 1st generation immigrant women of Turkish origin
had a lower chance of pH values of arterial umbilical cord blood of 7.10 or less, compared to
non-immigrant women. However, this association did not reach statistical significance (p > 0.06).
As expected, there was a strong association between adverse arterial umbilical cord pH values and
low 5-minute Apgar scores ([Table 6]).
Table 6 Chance of pH values of arterial umbilical cord blood ≤ 7.10; logistic
regression (n = 4 552) (bold: p < 0.05).
|
|
OR
|
95 % CI
|
p-value
|
|
Non-immigrant women
|
1.00
|
|
|
|
1st generation immigrant women
|
0.51
|
0.25–1.04
|
0.0631
|
|
2nd/3rd generation immigrant women
|
0.88
|
0.49–1.56
|
0.6576
|
|
Binational women
|
0.88
|
0.44–1.77
|
0.7199
|
|
Age
|
|
|
|
|
|
1.00
|
|
|
|
|
1.71
|
0.95–3.11
|
0.0757
|
|
|
1.74
|
0.96–3.16
|
0.0674
|
|
|
0.94
|
0.48–1.88
|
0.8595
|
|
Apgar score
|
|
|
|
|
|
1.00
|
|
|
|
|
4.53
|
2.26–9.09
|
< 0.0001
|
|
Birth weight
|
|
|
|
|
|
1.00
|
|
|
|
|
1.50
|
0.68–3.30
|
0.3105
|
|
Birth
|
|
|
|
|
|
1.00
|
|
|
|
|
0.60
|
0.27–1.36
|
0.2207
|
|
Qualifications
|
|
|
|
|
|
1.00
|
|
|
|
|
0.80
|
0.54–1.10
|
0.2753
|
|
|
1.03
|
0.47–2.26
|
0.9481
|
|
Multiple births
|
|
|
|
|
|
1.00
|
|
|
|
|
0.60
|
0.18–2.05
|
0.4178
|
Discussion
Two recent review articles have examined the most important associations between perinatal outcomes and
immigration in European countries. After a detailed analysis of evidence-based studies for the period
1970–2004, the review by Bollini et al., published in 2009, came to conclusion that immigrant women are
at a disadvantage, at least with regard to perinatal outcomes, and that this disadvantage persists over
several generations. These negative impacts were apparently much reduced in countries with comprehensive
and wide-ranging integration policies [14]. The study by Gissler et al., also
published in 2009, looked at all available 42 studies for these years to see whether they could identify
an association between increased perinatal mortality and motherʼs immigrant status. They found that some
immigrant populations had equally good, sometimes even better, perinatal health compared to the
population of the host country. One explanation for this was the so-called ‘healthy migrant effect’ (a
self-selection of particularly healthy people who immigrated). Other immigrant groups had poorer overall
perinatal outcomes; the authors explained this as caused by maternal milieu, health behavior,
socio-economic conditions, disparities with regard to access to and utilization of healthcare
facilities, and social factors [11].
Investigations in the USA, Australia and Germany in the 1980s established relatively consistently that
immigrant women generally attended fewer antenatal examinations during pregnancy [15], [16], [17]. Based on an analysis
of all singleton pregnancies registered in Finland in the years 1999 to 2001, Malin and Gissler
established that there was no significant difference in the number of attended antenatal check-ups
between immigrant women and women born in Finland [4]. In 2012,
Martinez-Garcia et al. published a hospital-based retrospective analysis on this issue and reported that
women from Eastern Europe and from the Maghreb generally attended fewer antenatal check-ups compared to
non-immigrant Spanish women [18]. For the federal state of Baden-Württemberg
in Germany, Simoes et al. already noted in 2003 after an analysis of perinatal data that the tendency
among non-German women to attend fewer antenatal examinations was decreasing compared to the 1970s and
1980s [19]. For Berlin we were unable to find any current differences in the
number of antenatal examinations attended during pregnancy between immigrant and non-immigrant groups of
women.
An Austrian study by Oberaigner et al. showed that immigrant women of Turkish origin had significantly
fewer pregnancy risks compared to immigrant women from other countries and non-immigrant women born in
Austria [20]. We were able to confirm this finding for Berlin for 1st
generation immigrant women of Turkish origin. The picture was different for 2nd/3rd generation immigrant
women of Turkish origin. It would appear that the results for this group tend to converge with the
results for the non-immigrant population.
Endl and Tatra already reported higher anemia prevalences in pregnant women of Turkish origin in the
1970s [21]. Chan et al. also described this phenomenon in their study,
published in 1988, for groups of immigrants from different countries living in Australia, and Jans et
al. reported the same phenomenon in 2009 for non-North European immigrant women living in Amsterdam
[16], [22]. In our study we found that
postpartum anemia prevalences were significantly higher for 2nd and 3rd generation immigrant women of
Turkish origin compared to non-immigrant women or 1st generation immigrant women. Because of their rare
clinical manifestation, hemoglobinopathies and thalassemias cannot serve as an explanation for this
discrepancy [23], [24].
In contrast to Walsh et al., who reported in their hospital-based study in Dublin that Irish women were
more likely to have epidural analgesia during delivery compared to women from Eastern Europe [25], we were unable to detect differences between German women and women of
Turkish origin in our Berlin study population. However, we did find a difference compared to women with
very limited or no understanding of German. This mirrored the results of the study by Oberaigner et al.
for various Austrian maternity hospitals. In their study, published in 2013, they found that the
epidural analgesia rates increased when immigrant women had stayed longer in Austria and their German
language skills had improved [20]. Problems of language and communication
when providing information about the available options for anesthetization during delivery need to be
considered as a possible cause of the lower rates for immigrant women.
The significantly higher rate of planned/primary cesarean sections in the German sub-group may also point
to a problem in the care system. In 2004, Rizzo et al. used matched pairs analysis to compare Italian
women with non-EU immigrant women and also noted a higher rate of primary/elective cesarean sections
among Italian women. They ascribed this to the medical staff being less worried about legal claims when
dealing with immigrant women, meaning that staff were more likely to encourage vaginal births in
low-risk immigrant women. Moreover, differences in socio-cultural ideas about pregnancy and birth
between immigrant and non-immigrant women and other traditional notions about birth processes could also
play a role [26].
In contrast to our findings, several studies have reported a higher incidence of secondary cesarean
sections in immigrant women and ethnic minorities [27], [28], [29]. In 2000, Vangen et al. carried out a detailed analysis
of 553 491 live births recorded in the Norwegian register of births for the period 1986–1995 and found a
higher rate of cesarean sections for immigrant women from the Indian subcontinent, from Africa and from
Latin America; however the rates for immigrant women of Turkish origin were similar to those of
non-immigrant women [30].
In 2012, von Katterfeld et al. published an analysis of cesarean sections in Western Australia for the
period 1998–2006; they found that certain groups of African and Asian immigrant women had significantly
higher rates of secondary C-sections, while other groups had lower rates of primary/elective C-sections
compared to Australian-born women. In addition to medical and obstetric considerations, the authors were
of the opinion that socio-cultural aspects and attitudes to cesarean sections must also be considered as
causative factors. Either obstetricians are less likely to offer elective cesarean sections to immigrant
women or immigrant women consciously reject this mode of delivery [31]. In
2010 Rio et al. published an analysis of 215 000 singleton births in Spain and found that the general
risk for immigrant women to deliver their baby by cesarean section was lower than for native-born
Spanish women (RR: 0.83) [32]. In a 2013 perinatal study of several Austrian
maternity units, Oberaigner et al. reported that immigrant women of Turkish origin had significantly
higher vaginal delivery rates and lower elective C-section rates [20].
In our study, the condition of neonates, as determined immediately after birth based on arterial
umbilical cord blood pH and 5-minute Apgar scores, did not show poorer outcomes for the immigrant group;
in fact, there were no significant differences in neonatal outcomes between groups. In contrast, the
2011 study of Cacciani et al., which analyzed all births in the Italian region of Lazio, found that
Apgar scores were lower for infants born to immigrant women [33]. A recent
(2013) study was published by Margioula-Siarkou et al. and consisted of a retrospective analysis of
around 7000 births in a hospital in Greece; in this study, the children of the immigrant group had
better Apgar scores at 1 and 5 minutes after delivery [34].
Our study has a number of methodological strengths; this includes the large number of cases studied and
the very high response rate, particularly from the subgroup of immigrant women of Turkish origin. In
contrast to many exclusively register-based studies done in Germany (Razum et al., 2011 [35]), our study included a carefully differentiated examination of immigration
status, degree of acculturation and socio-economic status and analyzed the impact of these different
factors separately. Differences in how data was entered in the Mutterpass and in obtaining
perinatal data could be a methodological weakness. Moreover, the results from a metropolis with a large
population of immigrants may not necessarily be transferrable to smaller towns or rural regions with a
lower proportion of immigrants in the general population. Despite these critical points, our study
currently offers the most valid results.
Conclusion
For the large group of immigrant women of Turkish origin living in Berlin we found that the results for
most perinatal parameters were similar to those for non-immigrant women. In particular, our study showed
that rates for attending antenatal examinations and the level of medical obstetric care in maternity
hospitals in Berlin were similar for immigrant women and non-immigrant women; these results probably
also apply to other large cities and industrial urban agglomerations in Germany. As regards the
(limited) differences between immigrant and non-immigrant groups found in this study, the following
conclusions must be drawn with respect to the obstetrical care offered to immigrant women of Turkish
origin:
-
The overall lower rate of C-section deliveries, particularly the lower rates of elective cesarean
sections in the immigrant group, should prompt a search for possible causal and protective
factors. The difference may be due to a different approach to giving birth on the part of the
mother. The attitudes of healthcare staff and the advice offered by physicians could also play a
role.
-
The lower rates for PDA/CSE administration during delivery among immigrant women of Turkish
origin with a limited understanding of German must be further investigated as it could be taken
as an indication of a difference in the quality of care which has not yet been remedied.
-
The unfavorably high anemia rates postpartum in a subgroup of immigrant women of Turkish origin
should lead to increased prophylactic measures and should also prompt physicians to look
actively for anemia and its causes in this group of women who have given birth, not merely in
connection with obstetrical problems.
Funding
Partly funded by German Research Foundation (DFG), DA 1199/2-1.
