Keywords
maternal obesity - adverse perinatal outcomes - superobesity
There is an increasing prevalence of obesity among women in the United States. Notably,
37% of reproductive-age women are obese as defined as a body mass index (BMI) greater
than or equal to 30 kg/m2, and 10% have morbid obesity with a BMI more than or equal to 40 kg/m2.[1] There is a significant geographic disparity in prepregnancy weight among women.
In 2015 for 48 states, New York City, and District of Columbia, the overall incidence
of prepregnancy normal weight (BMI 18.5 to < 25) was 45%. West Virginia reported the
second lowest prepregnancy normal weight of 40% and the second highest prepregnancy
obesity weight (BMI ≥ 30) of 31%.[2]
Obesity in pregnancy is an important public health problem with short-term and long-term
implications for maternal and child health. Maternal obesity, when compared with mothers
of normal prepregnancy weight, has been linked to adverse pregnancy outcomes including
preeclampsia, gestational diabetes, cesarean deliveries, and prolonged postpartum
hospital stay as well as perinatal problems including congenital anomalies, birth
asphyxia, neonatal hypoglycemia, and stillbirth.[3]
[4]
[5]
[6]
[7]
[8]
[9] A previous study demonstrated a “dose-response” relationship between the severity
of maternal obesity with adverse perinatal outcomes,[10] supported by two subsequent meta-analyses.[9]
[11]
The objective of this study was to investigate the relationship between adverse perinatal
outcomes and increasing maternal prepregnancy weight in our population. We hypothesized
that increasing prepregnancy weight would correlate with an increase in maternal and
neonatal complications.
Materials and Methods
This was a retrospective, observational, cohort study from a single tertiary care
perinatal center of women giving birth from 1/1/2015 to 12/31/2018. The study was
based at Cabell Huntington Hospital in Huntington, West Virginia, the perinatal teaching
hospital for the Marshall University School of Medicine. Antepartum and intrapartum
patient information were collected following delivery using the Cabell Huntington
Hospital Clinical Data Warehouse. As the Warehouse did not include a preconception
BMI, a first documented weight of more than or equal to 90 kg was chosen for the search
criteria. Individual charts and obstetrical records were then manually reviewed for
self-reported maternal prepregnancy body weight and measured height to calculate BMI.
Five hundred consecutive mothers with a prepregnancy BMI of 18.5 to less than 25,
and their offspring, and 500 additional mothers with a BMI more than or equal to 30,
and their offspring, were identified. For this study, prepregnancy BMI was categorized
as normal weight (18.5 to < 25), class I obesity (30.0 to < 35), class II obesity
(35.0 to < 40), class III obesity (40.0 to < 50), and superobesity (≥ 50). Mother/baby
dyads were excluded from the study for missing medical record data, neonatal abstinence
syndrome, maternal substance use disorder, genetic/congenital anomalies, or miscarriages/stillbirths.
The institutional review board of Marshall University School of Medicine approved
this human subject research prior to its initiation.
Statistical analysis was planned a priori. The primary outcomes considered were preeclampsia,
gestational diabetes mellitus, mode of delivery, 1- and 5-minute Apgar scores, neonatal
intensive care unit (NICU) admission, and neonatal death.
Maternal/newborn metrics were stratified by maternal preconception BMI and trend analysis
was performed using Jonckheere-Terpstra and Cochrane-Armitage tests for continuous
and binary data, respectively. We performed a subsequent simple univariable and multivariable
logistic regression analysis adjusted for advanced maternal age (> 35 years), preexisting
maternal diabetes, preexisting maternal hypertension, and infant gestational age.
Regression model covariates were selected based on their clinical relevance to the
selected outcomes. Regression model residuals were evaluated for any gross deviations
from test assumptions. Regression model results were reported as odds ratios with
95% confidence intervals and p-values for our effect measure. Wald test was used to assess for evidence of a linear
trend within the regression models. Statistical analysis was performed using STATA
(StataCorp. 2022. Stata Statistical Software: Release 17. College Station. TX: StataCorp
LP).
Results
Between January 1, 2015, and December 31, 2018, 500 consecutive mother/baby dyads
with maternal prepregnancy BMI of 18.5 to less than 25 and the first 500 consecutive
dyads with maternal prepregnancy BMI more than or equal to 30 were selected for this
study. Of these dyads, 142 did not meet inclusion criteria. The study population resulted
in 858 pairs (386 with prepregnancy BMI 18.5 to less than 25 and 472 with maternal
BMI more than or equal to 30 ([Fig. 1]).
Fig. 1 Flowchart of patient selection. BMI, body mass index.
Maternal and newborn metrics were stratified by maternal preconception BMI as well
as trend analysis. Simple univariable logistic regression models demonstrated higher
preconceptional BMI was significantly associated with progressively higher rates of
cesarean section, preeclampsia, gestational diabetes mellitus, and postdelivery length
of stay. Higher preconceptional BMI was significantly associated with adverse neonatal
metrics including decreased weeks of gestational age, preterm birth, lower 1- and
5-minute Apgar scores, and postdelivery length of stay ([Table 1]).
Table 1
Descriptive statistics of obstetric/neonatal variables and maternal preconception
BMI
|
Maternal preconception BMI
|
p-Value
|
|
Normal weight
|
Obesity class 1
|
Obesity class 2
|
Obesity class 3
|
Super obesity
|
|
(18.5 to <25)
|
(30 to <35)
|
(35 to <40)
|
(40 to <50)
|
(≥ 50)
|
|
n
|
|
386
|
58
|
158
|
215
|
41
|
|
Maternal metrics
|
|
|
|
|
|
|
|
|
Age at time of delivery, years
|
Median (IQR)
|
26 (22, 31)
|
27 (24, 34)
|
27 (23, 32)
|
27 (24, 31)
|
28 (25, 31)
|
0.023
|
|
Preexisting diabetes
|
% (n/N)
|
1% (3/386)
|
5% (3/56)
|
2% (3/154)
|
7% (15/208)
|
3% (1/40)
|
<0.001
|
|
Preexisting hypertension
|
% (n/N)
|
3% (11/383)
|
16% (9/56)
|
16% (25/154)
|
24% (51/211)
|
40% (16/40)
|
<0.001
|
|
Preeclampsia
|
% (n/N)
|
8% (31/385)
|
16% (9/57)
|
15% (23/156)
|
20% (42/207)
|
32% (13/41)
|
<0.001
|
|
Gestational diabetes
|
% (n/N)
|
2% (8/386)
|
5% (3/56)
|
11% (17/153)
|
15% (32/209)
|
18% (7/40)
|
<0.001
|
|
Preconception weight, kg
|
Median (IQR)
|
59 (55, 63.5)
|
105 (100, 105)
|
105 (101, 111)
|
118 (112, 125)
|
150 (140, 159)
|
<0.001
|
|
Weight at delivery, kg
|
Median (IQR)
|
74 (67, 80)
|
118 (109, 123)
|
116 (110, 124)
|
127 (118, 133)
|
149 (138, 165)
|
<0.001
|
|
Weight change, kg
|
Median (IQR)
|
15 (11, 19)
|
14 (7, 18)
|
10 (5, 17)
|
8 (2, 13)
|
3 (-4, 10)
|
<0.001
|
|
Preconception BMI
|
Median (IQR)
|
22 (21, 23)
|
34 (33, 35)
|
37 (37, 39)
|
43 (41, 45)
|
52 (51, 56)
|
<0.001
|
|
BMI at delivery
|
Median (IQR)
|
28 (26, 29)
|
37 (36, 40)
|
41 (39, 44)
|
46 (44, 49)
|
54 (51, 58)
|
<0.001
|
|
Postdelivery length of stay, days
|
Median (IQR)
|
2 (2, 2)
|
2 (2, 3)
|
2 (2, 3)
|
2 (2, 3)
|
2 (2, 3)
|
<0.001
|
|
ICU admission
|
% (n/N)
|
4% (14/386)
|
5% (3/56)
|
3% (4/155)
|
3% (6/207)
|
5% (2/41)
|
0.733
|
|
Neonatal metrics
|
|
|
|
|
|
|
|
|
Delivery type
|
|
|
|
|
|
|
|
|
Vaginal
|
% (n/N)
|
76% (292/386)
|
62% (36/58)
|
56% (89/158)
|
47% (102/215)
|
34% (14/41)
|
<0.001
|
|
Spontaneous
|
% (n/N)
|
33% (128/386)
|
22% (13/58)
|
28% (44/158)
|
19% (41/215)
|
20% (8/41)
|
<0.001
|
|
Induced
|
% (n/N)
|
43% (164/386)
|
40% (23/58)
|
29% (45/158)
|
28% (61/215)
|
15% (6/41)
|
<0.001
|
|
Cesarean
|
% (n/N)
|
24% (94/386)
|
38% (22/58)
|
44% (69/158)
|
53% (113/215)
|
66% (27/41)
|
<0.001
|
|
Primary
|
% (n/N)
|
13% (51/386)
|
21% (12/58)
|
28% (44/158)
|
33% (70/215)
|
32% (13/41)
|
<0.001
|
|
Repeat
|
% (n/N)
|
11% (43/386)
|
17% (10/58)
|
16% (25/158)
|
20% (43/215)
|
34% (14/41)
|
<0.001
|
|
Nonemergent
|
% (n/N)
|
16% (61/386)
|
35% (20/58)
|
39% (61/158)
|
46% (99/215)
|
56% (23/41)
|
<0.001
|
|
Emergent
|
% (n/N)
|
9% (33/386)
|
3% (2/58)
|
5% (8/158)
|
7% (14/215)
|
10% (4/41)
|
0.438
|
|
Gestational age, weeks
|
Median (IQR)
|
39.0 (38.3, 39.6)
|
39.0 (37.9, 39.0)
|
39.0 (37.3, 39.1)
|
39.0 (37.0, 39.0)
|
38.0 (35.0, 39.0)
|
<0.001
|
|
Preterm
|
% (n/N)
|
11% (42/385)
|
16% (9/58)
|
17% (27/158)
|
15% (32/215)
|
37% (15/41)
|
0.001
|
|
1-minute Apgar
|
Median (IQR)
|
9 (8, 9)
|
8 (8, 9)
|
8 (8, 9)
|
8 (8, 9)
|
8 (7, 9)
|
<0.001
|
|
5-minute Apgar
|
Median (IQR)
|
9 (9, 9)
|
9 (9, 9)
|
9 (9, 9)
|
9 (9, 9)
|
9 (8, 9)
|
<0.001
|
|
5-minute Apgar < 7
|
% (n/N)
|
0% (1/386)
|
0% (0/58)
|
2% (3/158)
|
1% (3/213)
|
0% (0/41)
|
0.141
|
|
Postdelivery length of stay, days
|
Median (IQR)
|
2 (2, 3)
|
2 (2, 3)
|
2 (2, 3)
|
2 (2, 3)
|
3 (2, 7)
|
<0.001
|
|
NICU admission
|
% (n/N)
|
15% (59/385)
|
18% (10/57)
|
21% (33/157)
|
20% (43/214)
|
42% (17/41)
|
0.002
|
|
NICU length of stay, days
|
Median (IQR)
|
9 (6, 20)
|
12 (2, 17)
|
8 (4, 15)
|
7 (4, 17)
|
9 (6, 15)
|
0.426
|
|
Death
|
% (n/N)
|
0% (0/385)
|
0% (0/58)
|
1% (2/157)
|
1% (2/213)
|
0% (0/41)
|
0.104
|
Abbreviations: BMI, body mass index; IQR, interquartile range; NICU, neonatal intensive
care unit.
Note.
p-Values were calculated using Jonckheere-Terpstra and Cochran-Armitage trend analysis
tests for continuous and binary data, respectively.
Associations remained significant in multivariable logistic regression models for
gestational diabetes, preeclampsia, cesarean delivery adjusting for advanced maternal
age, preexisting maternal hypertension, preexisting maternal diabetes, gestational
age, and prior c-section ([Table 2]).
Table 2
Descriptive statistics and logistic regression analysis of obstetric/neonatal variables
and maternal preconception BMI, n = 861
|
Maternal preconception BMI (kg/m2)
|
% (n/N)
|
Nonadjusted
|
Adjusted[a]
|
|
odds ratio
|
95% CI
|
p-Value
|
odds ratio
|
95% CI
|
p-Value
|
|
Gestational diabetes
|
18.5 to <25
|
2% (8/386)
|
|
|
|
|
|
|
|
30 to <35
|
5% (3/56)
|
2.67
|
[0.69, 10.40]
|
0.156
|
1.56
|
[0.32, 7.75]
|
0.584
|
|
35 to <40
|
11% (17/153)
|
5.91
|
[2.49, 14.00]
|
<0.001
|
5.72
|
[2.36, 13.89]
|
<0.001
|
|
40 to <50
|
15% (32/209)
|
8.54
|
[3.86, 18.92]
|
<0.001
|
9.21
|
[4.01, 21.14]
|
<0.001
|
|
≥50
|
18% (7/40)
|
10.02
|
[3.42, 29.37]
|
<0.001
|
9.82
|
[3.11, 30.95]
|
<0.001
|
|
Preeclampsia
|
18.5 to <25
|
8% (31/385)
|
|
|
|
|
|
|
|
30 to <35
|
16% (9/57)
|
2.14
|
[0.96, 4.77]
|
0.062
|
1.20
|
[0.45, 3.16]
|
0.714
|
|
35 to <40
|
15% (23/156)
|
1.97
|
[1.11, 3.51]
|
0.020
|
1.17
|
[0.60, 2.29]
|
0.642
|
|
40 to <50
|
20% (42/207)
|
2.91
|
[1.76, 4.79]
|
<0.001
|
2.14
|
[1.21, 3.78]
|
0.009
|
|
≥50
|
32% (13/41)
|
5.30
|
[2.50, 11.26]
|
<0.001
|
2.64
|
[1.10, 6.30]
|
0.029
|
|
Cesarean [b]
|
18.5 to <25
|
16% (61/386)
|
|
|
|
|
|
|
|
30 to <35
|
35% (20/58)
|
1.90
|
[1.06, 3.39]
|
0.030
|
1.52
|
[0.67, 3.42]
|
0.316
|
|
35 to <40
|
39% (61/158)
|
2.41
|
[1.63, 3.56]
|
<0.001
|
1.74
|
[0.99, 3.05]
|
0.053
|
|
40 to <50
|
46% (99/215)
|
3.44
|
[2.41, 4.91]
|
<0.001
|
2.67
|
[1.62, 4.41]
|
<0.001
|
|
≥50
|
56% (23/41)
|
5.99
|
[3.02, 11.90]
|
<0.001
|
4.75
|
[1.96, 11.46]
|
0.001
|
|
Preterm[c]
|
18.5 to <25
|
11% (42/385)
|
|
|
|
|
|
|
|
30 to <35
|
16% (9/58)
|
1.50
|
[0.69, 3.27]
|
0.308
|
0.99
|
[0.41, 2.39]
|
0.982
|
|
35 to <40
|
17% (27/158)
|
1.68
|
[1.00, 2.84]
|
0.051
|
1.43
|
[0.83, 2.47]
|
0.201
|
|
40 to <50
|
15% (32/215)
|
1.43
|
[0.87, 2.34]
|
0.157
|
1.07
|
[0.62, 1.82]
|
0.818
|
|
≥50
|
37% (15/41)
|
4.71
|
[2.31, 9.60]
|
<0.001
|
3.22
|
[1.49, 6.98]
|
0.003
|
|
NICU admission[c]
|
18.5 to <25
|
15% (59/385)
|
|
|
|
|
|
|
|
30 to <35
|
18% (10/57)
|
1.18
|
[0.56, 2.46]
|
0.667
|
0.94
|
[0.43, 2.07]
|
0.883
|
|
35 to <40
|
21% (33/157)
|
1.47
|
[0.92, 2.36]
|
0.111
|
1.29
|
[0.79, 2.11]
|
0.313
|
|
40 to <50
|
20% (43/214)
|
1.39
|
[0.90, 2.15]
|
0.138
|
1.01
|
[0.63, 1.62]
|
0.974
|
|
≥50
|
42% (17/41)
|
3.91
|
[1.98, 7.73]
|
<0.001
|
2.85
|
[1.37, 5.92]
|
0.005
|
Abbreviations: BMI, body mass index; CI, confidence interval; NICU, neonatal intensive
care unit.
Note. Descriptive statistics are visually stratified by maternal preconception BMI. Logistic
regression derived odds ratios and p-values are based on nonstratified maternal preconception BMI.
a Adjusted results were based on a multivariable logistic regression analysis adjusting
for advanced maternal age, preexisting maternal hypertension, preexisting maternal
diabetes, and gestational age.
b Cesarean multivariable logistic regression model additionally included previous cesarean
section as a covariate.
c Prematurity and NICU admission multivariable logistic regression models did not include
gestational age as a covariate.
All the maternal metrics demonstrated a significant trend except for age at time of
delivery and ICU admission. All the neonatal metrics demonstrated a significant trend
except emergent cesarean delivery, 5-minute Apgar score less than 7, NICU length of
stay, and death. Relative to pregnancies with normal BMI, pregnancies with a pre-conception
BMI more than or equal to 30 gained less weight during pregnancy and increasing obesity
classes had an incremental decrease in absolute weight gain at time of delivery. Simple
univariable logistic regression models demonstrated higher preconception BMI was significantly
associated with progressively higher linear rates of gestational diabetes, preeclampsia,
cesarean section, and NICU admission ([Table 2]). Compared with normal preconception BMI 18.5 to less than 25, cesarean was higher
in BMI more than or equal to 30, gestational diabetes was higher in BMI more than
or equal to 35, preeclampsia was higher in BMI more than or equal to 40, prematurity,
and NICU admission was higher in BMI more than or equal to 50. Associations remained
significant in multivariable logistic regression models although compared with normal
preconception BMI 18.5 to less than 25, cesarean was higher in BMI more than or equal
to 40 and preeclampsia was higher in BMI more than or equal to 40.
Discussion
In this retrospective, case–control study, obese women (preconception BMI ≥ 30) were
more likely to have maternal complications and neonatal morbidity when compared with
women of normal weight (preconception BMI 18.5 to < 25), and perinatal complications
increased incrementally with increasing obesity with superobese mothers having more
adverse perinatal outcomes when compared with other classes of obesity.
In this study, with increasing maternal prepregnancy BMI, we observed significant
increases in trend in maternal adverse metrics including preeclampsia, gestational
diabetes mellitus, cesarean delivery, and postdelivery length of hospitalization.
In addition, adverse neonatal metrics increased in trend for prematurity, lower 1-
and 5-minute Apgar scores, postdelivery length of stay, and NICU admission. Simple
univariable logistic regression models demonstrated higher preconception BMI was associated
with progressively higher linear rates of gestation diabetes mellitus, preeclampsia,
cesarean delivery, and NICU admission.
A previous retrospective study of infants born to 64,272 obese Missouri mothers from
2000 to 2006, utilizing data extracted from birth certificate records and hospital
discharge information, demonstrated a “dose-response” relationship between worsening
obesity and the incidence of cesarean delivery, macrosomia, neonatal hypoglycemia,
and preeclampsia. Three classes of obesity (BMI 30 to < 40, 40 to < 50, and ≥50) were
used for comparisons and no mothers of normal BMI were included in the study.[10]
There have been two meta-analyses examining graded relationships of maternal obesity
to pregnancy outcomes. The first combined 59 previous publications and concluded a
graded relationship of adverse perinatal outcomes existed with increased severity
of obesity including increased incidence of premature birth, macrosomia, preeclampsia,
gestational diabetes, cesarean section, maternal bleeding, low umbilical artery pH
and Apgar scores, and NICU admission.[9]
The second identified 13 studies with a low risk-of-bias describing 3,722,477 pregnancies.
It was concluded that most adverse pregnancy outcomes increased with increasing maternal
BMI, including maternal complications of gestational diabetes mellitus, hypertension
of pregnancy, cesarean delivery, and neonatal complications including hypoglycemia,
macrosomia, birth trauma, respiratory distress, NICU admission, and death.[11]
This study is the first single-center confirmation of a graded “dose-response” to
the level of obesity with regard to adverse perinatal outcomes. Unlike previous studies,
our data were extracted directly from maternal and neonatal medical records, eliminating
possible omissions on reports extracted from birth certificates and hospital discharge
data.
In this study, the likelihood of cesarean delivery increased from 24% in women of
normal prepregnancy BMI to 66% in superobese women. The rate of emergent cesarean
sections, however, remained statistically unchanged with increasing BMI as has been
previously reported.[5] Our findings are consistent with previous reports showing that the rate of cesarean
delivery and failed labor induction leading to cesarean section increases by maternal
prepregnancy BMI category.[5]
[12]
[13]
[14] In 2020, 31.8% of live births in the United States were to women who had a cesarean
delivery. The rate rose steadily from 25.1% of women of normal weight to 52.3% of
women in obesity class III.[15]
In this study, parity and failed induction were not analyzed to determine factors
involved in the increased rate of cesarean delivery.[16] As previously recommended, future studies are needed to examine what factors are
involved in the increased incidence of cesarean delivery related to increasing obesity.[10]
The American College of Obstetricians and Gynecologists has supported the Institute
of Medicine (IOM) recommendation for pregnancy weight gain of 5 to 9.1 kg for all
obese women without differentiating between classes of obesity.[16] In this study, we have stratified gestational weight gain (GWG) by obesity classes
I and II gaining more weight than the recommended GWG, while superobese mothers gained
less than the recommended GWG. Studies subsequent to the IOM recommendation have shown
benefits of recommended GWG during pregnancy[7] as well as benefits of GWG less than recommended.[7]
[17]
[18] While weight loss in obese pregnant women is associated with an increased risk for
low-birth-weight neonates, it has also been shown to decrease or maintain risk for
other adverse maternal and neonatal morbidities.[19]
Limitations and Strengths
Limitations and Strengths
Our study has several limitations. This is a single-center study based on limited
numbers and may not be applicable to a larger sample in other centers. This was a
retrospective, observational, cohort study with the potential for confounding factors
not considered in our analysis including maternal race, smoking status, level of education,
and adequacy of prenatal care, all of which can impact perinatal outcome. An additional
limitation was the use of self-reported preconception weight, which has been shown
to be underestimated by respondants.[20]
[21] Other studies, however, have shown that although women underestimated weight, 84%
of the women remained in the appropriate BMI categories of obesity and self-reporting
is a reliable measure of preconception weight.[22]
The findings of this study, as well as previous reports, imply that preconception
weight loss might reduce the incidence of adverse perinatal outcomes. A published
meta-analysis looked at outcomes following bariatric surgery in 8,364 women who were
compared with control subjects matched for presurgery BMI. There were reduced rates
of gestational diabetes mellitus, large for gestational age infants, gestational hypertension,
all hypertensive disorders, postpartum hemorrhage, and cesarean delivery rates. However,
the surgical patients showed an increase in small-for-gestational-age infants, intrauterine
growth restriction, and preterm deliveries. There were no differences in rates of
preeclampsia, NICU admissions, stillbirths, malformations, and neonatal death.[23] Future prospective studies are needed to examine the risks and benefits of preconception
weight loss, including weight reduction surgery, on perinatal outcomes.
Conclusion
We have demonstrated maternal and fetal complications increase with increasing obesity
with superobese mothers having more adverse perinatal outcomes when compared with
other classes of obesity. It is reasonable to counsel weight loss prior to conception
of women with preconception BMI more than or equal to 30 to potentially reduce maternal
and neonatal morbidity related to pregnancy. The impact of preconception weight reduction
on the incidence of adverse perinatal outcomes and economic implications needs further
study.
