Changes in Maternal Body Mass Index, Weight Gain and Outcome of Singleton Pregnancies from 2000 to 2015: A Population-based Retrospective Cohort Study in Hesse/Germany.

Introduction Maternal obesity and excessive gestational weight gain (GWG) affect the outcomes of women and their offspring. Our aim was to evaluate population-based data from Germany. Material and Methods Data from 583 633/791 514 mother-child pairs obtained from the perinatal database in Hesse for the period from 2000 to 2015 were used after excluding incomplete or non-plausible datasets. Early-stage pregnancy maternal body mass index (BMI) and GWG were evaluated. Significant outcome changes were calculated using linear or logistic regression models. Results The mean maternal age increased from 29.9 to 31.28 years; GWG increased from 445.1 to 457.2 g/week (p < 0.01). Similarly, rates for both overweight and obesity rose from 31.5 to 37.5% (p < 0.001). Cesarean section rates rose from 22.8 to 33.2% (p < 0.001) and rates of postpartum hemorrhage increased from 0.6 to 1% (p < 0.001). There was no significant change in the rates for stillbirth or perinatal mortality (p = 0.92 and p = 0.53 respectively), but there was an increase in the rates of admissions to neonatal intensive care units from 7.8 to 9.5% (p < 0.0001). The percentage of newborns with an Apgar score of < 7 at 5 minutes increased from 1 to 1.1% (p < 0.01) and the rate of neonates with an umbilical artery pH of < 7.1 rose from 1.7 to 2.4% (p < 0.01). Conclusions In just 15 years, pre-pregnancy BMI and GWG rates of women with singleton pregnancies have increased, and this increase has been accompanied by a significant rise in the rate of cesarean sections and a significant worsening of short-term maternal and neonatal outcomes. It is time to discuss the risks and the short-term and more worrying long-term consequences for mothers and their offspring and the future impact on our healthcare system.

Introduction

There has been a global shift from under- to overnutrition 1 . The WHO reported that in 2016 more than 1.9 billion adults aged > 18 years met the criteria for being overweight (body mass index [BMI] of 25 kg/m 2 and above) or obese (BMI ≥ 30 kg/m 2 ) 2 . Similarly, women of childbearing age have also shown rising rates of overweight and obesity. Both overweight and obesity increase the risk of gestational diabetes (GDM) and hypertensive disorders of pregnancy (HDP) and, what is even worse, have irreversible long-term effects on mothers and their offspring 3 .

Already in 2014, a study published in this journal showed that obesity and excessive weight gain (EGWG) in pregnancy have a serious impact on maternal, fetal and neonatal outcomes and are associated with weight gain in women following childbirth and long-term cardiovascular and metabolic risk factors 4 . Similarly, the long-term risks for offspring are early cardiovascular disease, metabolic syndrome and decreased life expectancy as adults, even after correction by lifestyle changes later in life 5 . The pathophysiological mechanisms of fetal programming have also been discussed in this journal 6 . Since lifestyle changes to control maternal weight before, during and after pregnancy must be done by the pregnant women themselves, obstetricians should inform their patients about the consequences and help improve their health literacy 7 . Additionally, midwives could also address the issue by stressing the importance of close cooperation during the monitoring period 8 ,  9 . Sadly, this important public health problem has not generated the level of concern it warrants among healthcare specialists, policy makers or individual members in our society.

In Germany, there are still no clinical guidelines on how to deal with maternal obesity before, during or after pregnancy. Even if such German guidelines were to be completed and published, obstetric guidelines alone will not necessarily change the behavior of healthcare providers or their patients as noted elsewhere 7 . In the US, both an ACOG Committee Opinion 10 and a Practice Bulletin 11 underlined the importance of proper information about evidence-based criteria and BMI-adapted weight gain recommendations during pregnancy 12 . Nevertheless, obesity among US-women of childbearing age has increased to such an extent that now 48% of all US pregnant women gain more weight than recommended 13 .

In order to demonstrate the necessity for change to policy makers and parents-to-be, we therefore decided to analyze the course of overweight, obesity and weight gain during pregnancy and the associated immediate risks and outcomes in a large regional cohort, with the data obtained from a German Perinatal Database (HESSEN) for the period from 2000 to 2015.

Materials and Methods

Study group

The office for quality management in Hesse (a federal state in Germany with 6 176 172 inhabitants in 2015 14 ) approved our application for this large project and allowed us to access their perinatal database for the period from 2000 to 2015. This database records all deliveries in Hesse except for home births. Data relating to patients and centers are anonymized. Since there is no direct quality control of the data when it is entered in the database, an extensive plausibility control had to be performed. Non-plausible data were registered as “missing data”.

Criteria for inclusion in the final analysis were: singleton pregnancy, an initial clinical and ultrasound examination before 14 weeks of gestation, a gestational age at delivery of at least 24 weeks, a maternal height equal to or higher than 120 cm, and availability of all data relating to maternal weight at the start of pregnancy and at delivery ( Fig. 1 ). The final study cohort therefore consisted only of 583 633 of the original 791 514 mother-child pairs (73.7%) ( Table 1 ).

Fig. 1

 Flow chart showing criteria for inclusion in the study.

Table 1  Characteristics of the study group. n = 583 633/791 514 singleton pregnancies delivered between 2000 and 2015 in Hesse, Germany; 207 881 pregnancies were excluded after carrying out a plausibility control.

Characteristics	Mean (SD); n (%)
SD = standard deviation, n = absolute number
Maternal age (years)	30.62 (5.37)
Maternal weight at 1st examination (kg)	67.85 (14.51)
Maternal height (cm)	166.49 (6.37)
Maternal BMI at 1st examination (kg/m 2 )	24.46 (4.93)
Week of 1st examination	8.41 (1.99)
Gestational age at birth (weeks)	39.33 (1.85)
Mother without partner (missing values: 63 703)
Yes	59 482 (11.44%)
No	460 448 (88.55%)
Motherʼs origin
German	458 015 (78.48%)
Other nationalities	125 618 (21.52%)
Previous deliveries (missing values: 5646)
0	288 478 (49.91%)
1	200 652 (34.72%)
2	63 270 (10.95%)
≥ 3	25 587 (4.43%)
Work (missing values: 124 688)
House wife (at start of pregnancy)	172 996 (37.69%)
Still in education	14 563 (3.17%)
Worker	17 743 (3.87%)
Employee	169 593 (36.95%)
Academic position	84 050 (18.31%)

Outcome criteria

Maternal age at delivery was calculated as the difference between the motherʼs year of birth and the year of delivery. BMI was calculated as kg/m 2 and classified using the WHO criteria: BMI < 18.5 kg/m 2 : “underweight”; BMI 18.5 – 24.9 kg/m 2 : “normal weight”; BMI 25.0 – 29.9 kg/m 2 : “overweight”; BMI ≥ 30 kg/m 2 : “obesity” 15 . The (weekly) maternal gestational weight gain (GWG) was calculated as the difference between maternal weight at the first examination and weight at delivery, respectively, divided by the duration of this interval in weeks. Based on the IOM guidelines, women were classified as GWG below, GWG within and GWG above the optimal range.

The birth weight (BW) was categorized into groups of 500 g ranging from < 1500 g to ≥ 4000 g. Similarly, gestational age at birth was categorized from < 28 weeks up to ≥ 37 weeks of gestation.

The incidence of HDP was calculated indirectly from recorded clinical findings and defined using the criteria of the International Society for the Study of Hypertension in Pregnancy (ISSHP) 16 .

Statistical analysis

Mean values per year were calculated for all numerical outcomes, and significant differences were analyzed during the observation period using a linear regression model. Rates (%) for categorical outcomes were analyzed using logistic regression models. The programs R for Windows 3.5.1, R Studio (Version 1.1.456) and Excel 2013 were used for statistical evaluation.

Results

Characteristics of the study group

Between 2000 and 2015, a total of 820 174 deliveries were registered in Hessen, of which 791 514 were singleton pregnancies (96.5%). After applying the inclusion criteria ( Fig. 1 ) 583 633 cases remained for the final analysis; the characteristics of this final study group are shown in Table 1 . Around 50% of the women were primiparous, 21.5% did not originate from Germany, and around 37% of pregnant women did not work.

Changes in BMI, GWG, and outcomes between 2000 and 2015

The mean BMI of all women between 2000 and 2015 at their first examination was 24.46 kg/m 2 , meaning that it was within the normal upper range. Mean maternal age increased from 29.9 to 31.28 years ( Fig. 2 a ) (p < 0.001), with the age of primiparous women rising from 28.48 to 30.00 years (p < 0.001). There were no records for 63 703/583 633 women showing whether the women were living with a defined partner or alone; of the remaining 519 930 women, 59 482 (11.5%) were single ( Table 1 ).

Fig. 2

 Course of maternal outcomes between 2000 and 2015. Total = total singleton deliveries (n = 583 633), cases = absolute incidence of specific outcome, + = positive correlation, − = negative correlation. Statistical analysis was carried out using linear ( a, b ) or logistic ( c, d ) regression. a  Mean maternal age (years): p = 0.001 (+). b  Mean maternal weight gain/gestational week (kg/week): p < 0.0001 (+): c  Maternal BMI categories (%) at 1st examination (< 14 weeks of gestation): underweight: n = 22 632, p = 0.014 (−); normal weight: n = 358 101, p < 0.001 (−); overweight: n = 130 937, p < 0.001 (+); obese: n = 71 963, p < 0.001 (+). d  Hypertensive disorders of pregnancy (%): n = 15 596, p < 0.001 (−).

The mean maternal height (166.49 cm) did not change significantly during the observation period (p = 0.17), but the mean maternal weight at the first examination rose from 66.79 kg in 2000 to 68.72 kg in 2015 (p < 0.0001). Accordingly, there was also a significant increase in mean maternal BMI in early pregnancy from 24.10 kg/m 2 in 2000 to 24.80 kg/m 2 in 2015 (p < 0.0001). The categories for the BMI classification using the WHO definition are shown in Fig. 2 c , demonstrating the increasing number of women with overweight or obesity at the first examination (p < 0.001) and the decreasing rates of underweight (p = 0.014) and normal-weight women (p < 0.001).

The mean weekly GWG between 2000 and 2015 rose from 445.1 g/week to 457.2 g/week ( Fig. 2 b , p < 0.001). The rate of women gaining weight within the range recommended by the IOM decreased from 23.5% in 2000 to 20.23% in 2015, while the rate of women with a GWG which was higher than the IOM recommendations rose from 64.54 to 68.55% ( Fig. 2 b , p < 0.001).

However, there was no increase in the rate of HDP during the observation period ( Fig. 2 d ). Fig. 3 shows the increasing rates for total and secondary cesarean delivery, which rose from 22.79 to 33.2% and from 11.67% to 17.06%. The postpartum hemorrhage (PPH) rate rose from 0.61% to 0.98% (both p < 0.001). The rate of emergency cesarean deliveries decreased from 4.38% in 2000 to 3.56% in 2015 (p < 0.0001).

Fig. 3

 Changes in the rates of cesarean sections and postpartum hemorrhage between 2000 and 2015. Total = total singleton deliveries (n = 583 633), cases = absolute incidence of specific outcome, + = positive correlation, − = negative correlation. All statistical analyses were carried out using logistic regression. a  Total cesarean sections (%): n = 178 764, p < 0.001 (+). b  Secondary cesarean sections (%) n = 78 719, p < 0.001 (+), * data from 2007 was excluded due to missing values. c  Emergency cesarean sections (%): n = 6397, p < 0.0001 (−) (no data available in 2000). d  Postpartum hemorrhage > 1000 ml (%): n = 6067, p < 0.001 (+).

In contrast to the declining rates of late preterm deliveries between 34 + 0 and 36 + 6 weeks (p < 0.001), the percentage of early preterm births < 28 weeks (p = 0.26) and of deliveries between 28 + 0 – 33 + 6 weeks of gestation (p = 0.12) did not change significantly ( Fig. 4 a ).

Fig. 4

 Course of neonatal outcomes (I) between 2000 and 2015. total = total singleton deliveries (n = 583 633), cases = absolute incidence of specific outcome, + = positive correlation, – =negative correlation. All statistical analyses were carried out using logistic regression. a  Preterm deliveries (%) according to category: delivery < 28 weeks: n = 1732, p = 0.26; delivery between 28 + 0 and 33 + 6 weeks: n = 7990, p = 0.12; delivery between 34 + 0 and 36 + 6 weeks: n = 31 440, p < 0.001 (−). b  Birth weight (%) according to category: < 1500 g: n = 4825, p < 0.001 (+); 1500 – 1999 g: n = 5127, p = 0.07; 2000 – 2499 g: n = 18 710, p = 0.37; 2500 – 2999 g: n = 89 097, p = 0.07; 3000 – 3499 g: n = 223 092, p < 0.001 (+); 3500 – 3999 g: n = 180 765, p = 0.66; ≥ 4000 g: n = 60 964, p < 0.001 (−). c  Stillbirths (%): n = 1566, p = 0.92. d  Perinatal mortality (%): n = 1910, p = 0.53.

The mean birth weight decreased from 3388 g in 2000 to 3362 g in 2015 (p < 0.001). Accordingly, the course over time showed a higher incidence of newborns with a BW < 1500 g and a lower rate of macrosomic newborns (≥ 4000 g, Fig. 4 b ). There was no significant change in the rates of stillbirth (p = 0.92) or perinatal mortality (p = 0.53) ( Fig. 4 c and d ), but there was an increase in the rate of poor immediate neonatal outcome (defined as an Apgar score of < 7 at 5 minutes or an umbilical artery pH of < 7.1 or both) ( Fig. 5 b to d , p < 0.01). NICU admission rates increased from 7.77% in 2000 to 9.48% in 2015 ( Fig. 5 a , p < 0.0001).

Fig. 5

 Course of neonatal outcomes (II) between 2000 and 2015. total = total singleton deliveries (n = 583 633), cases = absolute incidence of specific outcome, + = positive correlation, − = negative correlation. All statistical analyses were done using logistic regression. a  NICU admissions (%): n = 49 633, p < 0.0001 (+). b  Newborns with an Apgar score of less than 7 at 5 minutes (%): n = 5342, p < 0.01 (+). c  Newborns with an umbilical artery pH below 7.1 (%): n = 10 768, p < 0.001 (+). d  Newborns with an Apgar score of less than 7 at 5 minutes and an umbilical artery pH of less than 7.1 (%): n = 877, p < 0.001 (+).

Discussion

In this cohort study, the mean maternal age increased from 29.9 to 31.28 years between 2000 and 2015, which amounts to a rise of 4.7%. Simultaneously, there was a 3% increase in mean maternal body weight and BMI and a 2.7% rise in maternal GWG. The rate of cesarean sections rose from 22.79% in 2000 to 33.20% in 2015. A parallel significant increase was observed for the rates of PPH, immediate poor neonatal outcome, and NICU admissions. The overall increase in HDP rates from 2.15% in 2000 to 2.55% in 2015 was not significant, because the rates between 2001 and 2003 were even higher.

Our results showing increasing maternal age, increased maternal BMI and increased cesarean section rates do not differ from global trends. However, the parallel increase in the rates of poor immediate maternal and neonatal outcomes has not been previously shown in a population-based study.

Increasing maternal age is observed in many Western countries 17 ,  18 . In the USA, the percentage of births in women aged ≥ 30 increased 2-fold; it increased 3-fold in women aged ≥ 35 and nearly 4-fold in women aged ≥ 40. All of these changes occurred between 1980 and 2004 19 . Older maternal age is a risk factor for prolonged and dysfunctional labor, PPH, malpresentation, and primary cesarean delivery 19 .

The increases in maternal overweight, BMI, and maternal GWG are consistent with the alarming global trend of overweight and obesity 2 ,  20 and had already been shown previously for other German cohorts 21 ,  22 ,  23 . A systematic review, published in 2017, stated that both low and high maternal pre-pregnancy BMI increases the risk of maternal morbidity and mortality 24 . Only one publication stated that severe maternal morbidity was not associated with pre-pregnancy obesity 25 . High maternal BMI is also associated with an increased risk of stillbirth, perinatal mortality, and infant death before the age of one year 26 . The risk of other immediate adverse maternal and infant outcomes such as preeclampsia, HDP, GDM, preterm birth, and small or large size for gestational age at birth also increases with a higher maternal BMI. Women with both a high BMI and EGWG are most at risk for these adverse outcomes 27 . Meanwhile, overweight and obesity are responsible for more global deaths than underweight and are a huge financial burden on our healthcare system 2 . Overweight and obesity are risk factors for reduced fertility, cardiovascular, renal, gastrointestinal, musculoskeletal and pulmonary diseases, and type II diabetes 28 ,  29 ,  30 . What was especially alarming was that in our cohort, 67% of women had EGWG above the recommendations of the IOM during pregnancy.

In 2010, cesarean delivery rates already varied considerably across Europe, ranging from 14.3% in Iceland to 49.9% in Cyprus 31 . In Central Europe, the rates were 16.1% in the Netherlands, 19.9% in France and 29.9% in Germany 31 . Based on data from 121 countries, a trend analysis showed that between 1990 and 2014, the global average rate of cesarean sections increased by 12.4%, rising from 6.7 to 19.1%, while the European rate increased by 3.8%, from 11.2 to 15% 32 . Compared with other European countries, Germany has one of the highest cesarean delivery rates in Europe 33 .

In our cohort, we observed an increase in overweight and obesity, as well as in the rate of EGWG. Although both are described as a risk factor for a cesarean delivery 34 , 35 , 36 , the differences in cesarean delivery rates across Europe are not due to different levels of obesity but rather to differences in the quality of obstetric training, the general attitude of obstetricians and their incentives, and, last but not least, the malpractice policies in individual countries 37 .

The increased rate of postpartum hemorrhage in our German cohort has also been observed in other Western countries 38 , while global mortality from PPH has been steadily decreasing 38 ,  39 . Antenatal risk factors for PPH include multiple pregnancy, anemia, macrosomia, placenta previa or increta, previous PPH, as well as obesity and advanced maternal age 39 . The significant increase in maternal age, overweight, and obesity may have contributed to the rising rate of PPH, but this does not preclude improving maternal surveillance post partum, particularly for “at-risk” patients, to prevent serious harm.

The increasing rate of NICU admissions observed in our cohort, which rose from 7.7 to 9.48%, was also detected in the USA during a 6-year study period 40 . Increasing rates of cesarean sections as well as poor immediate neonatal outcomes as outlined above may be the reasons for this trend.

With regard to long-term outcomes, a study from Shanghai found an association between cesarean section and a higher risk of becoming overweight or obese in primary school children 41 . Excessive GWG is associated with long-term adverse maternal outcomes such as postpartum weight retention 42 and thus a higher risk in subsequent pregnancies. Being overweight or obese during pregnancy also has a negative impact on offspring, increasing their likelihood of being overweight or obese 43 even in adulthood, increasing their risk of being admitted for cardiovascular disease, and even increasing their risk of dying earlier compared to children born to women with a normal BMI 5 .

A major strength of this study is the large cohort size, which permitted significant changes occurring in singleton pregnancies over a period of 15 years to be determined.

However, our study has also limitations. The database did not allow us to precisely determine the incidence of GDM, either because data were not provided or because it was unclear what the diagnosis was based on. The frequency of HDP could only be calculated indirectly from the clinical findings (blood pressure, proteinuria). Calculated to be only 2.7%, the mean rate of HDP in our cohort is therefore smaller than the suggested prevalence of 5.2 to 8.2% reported in other studies 44 .

Another limitation of our study is that we used the BMI documented at the first visit (before 14 weeks of gestation) rather than the pre-pregnancy BMI, which is rarely available to the obstetrician. It is debatable whether recalled weight prior to pregnancy would be more reliable than the weight measured objectively at an early stage in pregnancy. Since the GWG in the first trimester is only about 0.5 – 2 kg according to the IOM 12 , the difference is likely to be negligible.

As the study was based on anonymous unlinked data, it was not possible to identify mothers who had more than one pregnancy during the observation period. The problem that there was no plausibility control of the data when it was entered in the database was compensated by a one-year full-time retrospective plausibility control by our group, which also resulted in an increase in “missing values” when we found non-plausible data.

Conclusion

In the present paper, we have demonstrated that increasing rates of maternal overweight and obesity and simultaneously increasing rates of excessive gestational weight gain are also present in German cohorts and are accompanied by increasing rates of HDP and cesarean delivery, with no improvements in neonatal outcomes. The risk factors for pre-pregnancy overweight and obesity as well as for insufficient or excessive weight gain during pregnancy and an analysis of how these associations could be the cause of poor maternal and neonatal outcomes is a subject for future investigations. We investigated the impact of pre-pregnancy BMI, as defined by the criteria of the WHO, on maternal and neonatal outcomes using univariate linear regression for continuous variables and multivariate regression analysis for categorical and numerical variables with many predefined confounders.

Our present results already point to alarming consequences for women and their offspring due to irreversible epigenetic mechanisms which are insufficiently communicated to women at risk. Midwives and obstetricians are ideally positioned to initiate algorithms before and after pregnancy which could interrupt the vicious circle of transgenerational consequences. Widespread education of patients and healthcare providers is required to identify and treat the modifiable intergenerational risks of obesity and the associated metabolic, cardiovascular and mental diseases for the benefit of future generations.