Академический Документы
Профессиональный Документы
Культура Документы
org
OBSTETRICS
Early-pregnancy percent body fat in relation
to preeclampsia risk in obese women
Lindsay K. Sween, MD; Andrew D. Althouse, PhD; James M. Roberts, MD
OBJECTIVE: The purpose of this study was to identify differences of (measured at 10.2 3.0 weeks of gestation) as predictors of
early-pregnancy body fat percentage and body mass index (BMI) preeclampsia outcomes.
between obese women that experienced preeclampsia and those
RESULTS: Three hundred seventy-three women were included in the
who did not.
analysis: 30 women had preeclampsia by clinical definition (8.0%),
STUDY DESIGN: We performed an analysis of the Prenatal and 14 women had preeclampsia by the research definition (3.8%).
Exposures and Preeclampsia Prevention 3 longitudinal cohort There was no relationship between BMI and preeclampsia risk in
study of preeclampsia mechanisms in obese and overweight obese women; however, body fat percentage was associated signifi-
women. Women completed questionnaires regarding their health cantly with increased risk of both the clinical definition of preeclampsia
behaviors; had hematocrit level, weight and height, and waist and and the research definition. In 239 obese women, a 1% increase in
hip circumferences measured, and had resistance and reactance body fat was associated with approximately 12% increased odds of
measured by bioelectric impedance analysis machine during the clinical preeclampsia and 24% increased risk of preeclampsia by the
first, second, and third trimesters. Total body water, fat mass, and research definition.
percent body fat were calculated with the use of pregnancy-
CONCLUSION: Early-pregnancy body fat appears to be important in the
specific formulas. Preeclampsia was assessed with the clinical
pathophysiologic condition of preeclampsia in obese women.
definition and a research definition (clinical preeclampsia
plus hyperuricemia). Logistic regression models were constructed Key words: bioelectric impedance analysis, body fat percentage, body
to analyze early-pregnancy BMI and body fat percentage mass index, obesity, preeclampsia
Cite this article as: Sween LK, Althouse AD, Roberts JM. Early-pregnancy percent body fat in relation to preeclampsia risk in obese women. Am J Obstet Gynecol
2015;212:84.e1-7.
Statistical methods values < .05 were considered statistically preeclampsia risk by both the research
Baseline data were described with the signicant. and clinical denition. For each 1% in-
mean standard deviation for contin- crease in body fat, the risk of clinical
uous variables and percentages for cate- R ESULTS preeclampsia increased 12% (odds ratio
goric variables in the total population Study participants were aged 23.7 4.1 [OR], 1.124; 95% condence interval
and separately by preeclampsia status. years; 63% of them were black, and [CI], 1.018e1.240), and the risk of
Potential differences between women 20% of them were smokers (Table 2). preeclampsia by the research denition
with normal pregnancies or with pre- The average BMI was 33.1 7.8 kg/m2; increased 24% (OR, 1.239; 95% CI,
eclampsia were evaluated with the use of by study design, most participants were 1.054e1.455). These relationships were
t tests for continuous variables (equal overweight (22.0%) or obese (64.1%). strengthened slightly when we adjusted
variances unless otherwise called for; Gestational diabetes mellitus was the model for age, race, and smoking
unequal variances test used where more common in women in the both status (Table 4). We considered models
appropriate) and c2 tests for categoric clinical (6.9%) and research-denitions that included both BMI and percent
differences (Fisher exact test in cases of preeclampsia (14.3%) than in those body fat together; in most of these
where expected cell counts were <5). with no preeclampsia (3.5%). Mean models, percent body fat showed a
This study focused on early pregnancy gestational age at delivery was 39.5 stronger relationship than did BMI.
BMI and body fat percentage as pre- 1.3 weeks (39.0 1.3 weeks for the 30 However, with the small number
dictors of preeclampsia outcomes; women with clinical preeclampsia and of cases, including 2 highly correlated
therefore, rst-trimester measurements 38.1 1.7 weeks for the 14 women variables in a single model, resulted in
of BMI and body fat were used in all with research preeclampsia). There model instability that was caused by
primary analyses. Body fat percentage were no signicant differences in the variance ination. Similarly, we asked
was examined as a function of BMI; the distribution of obesity metrics (BMI, whether the relationships between body
Pearson correlation coefcient is pre- waist circumference, waist-hip ratio, fat/BMI and preeclampsia might be
sented to assess the linear relationship. or body fat percentage) in the 30 dependent on central obesity as indi-
Lacking sufcient sample size to test women with clinical or the 14 women cated by waist circumference or waist/
appropriately for interaction between with research preeclampsia compared hip ratio, but there was no interaction
BMI and percent body fat, we instead with the 343 normal pregnancies with either of these variables in the
assessed the relationship between body (Table 2). model (P > .05 for all).
fat and preeclampsia by testing for dif- There was a moderate linear correla- In models that were stratied by race,
ferences in percent body fat between tion between early-pregnancy body fat the ORs were generally similar for white
women with preeclampsia and healthy percentage and BMI (R2 0.66; Figure), and black women. For clinical pre-
control subjects within each of the World but it was not a perfect linear correlation, eclampsia, the OR was 1.15 (95% CI,
Health Organization BMI classications which suggests that body fat percentage 0.98e1.36) for white women, compared
using t tests (a test with unequal variance may offer independent information with of the OR of 1.02 (95% CI,
where appropriate). Logistic regression about preeclampsia risk. 0.95e1.11) for black women. By the
models were constructed to analyze BMI When we analyzed the mean body fat research denition, the OR for white
and body fat percentage as continuous percentage in women within World women was 1.15 (95% CI, 0.95e1.39)
variables and allow adjustment for a Health Organization BMI classications, compared with 1.29 (95% CI,
limited selection of potential con- only in the highest category of obese 1.04e1.61) for black women. We did not
founders. Because the Prenatal Expo- women (BMI >40 kg/m2) was the body have a sufciently large sample to
sures and Preeclampsia Prevention 3 fat percentage signicantly higher in formally test interaction by race.
study was designed to compare obese women who experienced preeclampsia When we examined these relation-
women who did or did not experience compared with those women who did ships in all women (including the lean
preeclampsia, we initially limited our not (Table 3). and overweight women), neither BMI
logistic regression models only to obese We compared BMI and percent nor body fat percentage was associated
women. We also performed a secondary body fat as predictors of preeclampsia signicantly with increased risk of pre-
analysis that included all participants in obese women (BMI 30 kg/m2; eclampsia by the clinical denition or
because of the surprisingly high rates Table 4). Higher BMI was associated the research denition (Table 5), which
of preeclampsia in the lean and over- signicantly with higher risk for the suggested that the relationship between
weight women (by the clinical deni- research denition of preeclampsia, but percent body fat and preeclampsia
tion, 7.69% of lean women and 6.31% of not the clinical denition. The effect was was present only in the obese women.
overweight women experienced pre- no longer signicant after adjustment These relationships did not vary signi-
eclampsia). All statistical analyses were for age, race, and smoking status. In cantly when we adjusted for gestational
performed with SAS software (version contrast, there was a signicant rela- diabetes mellitus, waist circumference,
9.4; SAS Institute, Cary, NC); probability tionship between percent body fat and or waist/hip ratio.
Classification, %
Lean 13.9 14.0 13.3 .89 14.3 .35
Overweight 22.0 21.9 23.3 28.6
Obese 1 29.8 30.3 23.3 14.3
Obese 2 15.8 15.5 20.0 7.1
Obese 3 18.5 18.4 20.0 35.7
Obstetrics
Waist circumference, mm 1003 168 1002 168 1011 175 .78 1028 206 .57
Waist-hip ratio 0.86 0.07 0.86 0.07 0.87 0.08 .68 0.88 0.07 .30
Body fat percentage 45.9 10.2 45.8 10.1 46.9 11.3 .57 48.6 12.5 .31
a
Derived from comparison of 30 participants with clinical preeclampsia vs 343 participants with no preeclampsia; b Derived from comparison of 14 participants with research preeclampsia vs 343 participants with no preeclampsia.
Research
Sween. Percent body fat and preeclampsia risk. Am J Obstet Gynecol 2015.
84.e4
Research Obstetrics ajog.org
TABLE 3
Distribution of body fat percentages
No preeclampsia Preeclampsia
Body mass index classification n Mean body fat, % SD n Mean body fat, % SD P value
Lean 48 31.5 8.1 4 34.1 10.1 .54
Overweight 75 42.1 7.6 7 38.2 7.3 .20
Obese class 1 104 47.2 6.5 7 46.3 7.2 .73
Obese class 2 53 49.2 7.9 6 52.0 5.4 .40
Obese class 3 63 56.0 5.7 6 61.2 2.6 .03
Sween. Percent body fat and preeclampsia risk. Am J Obstet Gynecol 2015.
TABLE 4
Relationships between body fat percentage and risk of preeclampsia: obese participants only
Odds ratio
Predictor Outcome Unadjusted 95% confidence interval Adjusteda 95% confidence interval
Body mass index Clinical preeclampsia (n 19) 1.017 0.947e1.092 1.011 0.939e1.088
Research preeclampsia (n 8) 1.099 1.008e1.198 1.082 0.989e1.184
Body fat Clinical preeclampsia (n 19) 1.124 1.018e1.240 1.127 1.009e1.257
Research preeclampsia (n 8) 1.239 1.054e1.455 1.294 1.060e1.581
a
Adjusted for age, race, and smoking status.
Sween. Percent body fat and preeclampsia risk. Am J Obstet Gynecol 2015.
at 35.2 0.9 weeks. Body fat percentage values were not signicant. Because of preeclampsia than do obese white
as calculated by pregnancy-specic BIA study design, our sample had low women.3,15 Furthermore, this study was
models also seems to agree with ndings numbers of lean and overweight women designed to recruit women with BMI
in studies that used hydrostatic weigh- (total 35.9%), which may not result in a >30 kg/m2, and most studies that have
ing. Previous studies that used this dispersal of data sufcient to demon- investigated the relationship between
technique have found body fat percent- strate a relationship. Alternatively, there prepregnancy BMI and preeclampsia
age to be 29.6 6.15% in healthy, lean may truly not be a relationship between have had small sample sizes of women
women at 30 weeks of gestation.8,14 The body fat percentage and preeclampsia with BMI >35 kg/m2. Bodnar et al4
average body fat percentage among the risk in women with a low amount of body found that the ORs of preeclampsia
lean women in our study at 35 weeks of fat. There was also a surprisingly high began trending downward after BMI 35
gestation (n 13) was similar (29.6 incidence of preeclampsia in lean and kg/m2, although they remained >1.0
7.3%). With all women in our study overweight women and a relatively low compared with BMI 21 kg/m2. Rela-
cohort included, the average body fat incidence in obese women in this study tionship between BMI and preeclampsia
percentage at 35 weeks of gestation was cohort (7.69% of lean women and 6.31% may become more ill-dened at very
45.3 8.9%, because of the large num- of overweight women, compared with high BMIs, given the varying percent
ber of overweight and obese women. 7.95% of obese women by the clinical body fat and body fat distributions (eg,
The relationship between body fat denition), which may blunt the re- central, abdominal, peripheral, or
percentage and preeclampsia risk was not lationships between BMI, body fat per- visceral) between obese individuals.
signicant when lean and overweight centage, and preeclampsia risk. Another Despite these possible explanations, both
women were included in the analysis, nor reason the expected relationship between the deviation from the well-established
was the usual relationship between prepregnancy BMI and the incidence of association between increasing BMI and
increasing BMI and increasing pre- preeclampsia was not seen could be elevated preeclampsia risk and the un-
eclampsia risk seen in this cohort. Both the predominance of obese black women usually high incidence of preeclampsia
increasing body fat percentage and BMI in the cohort. Several studies have among the lean and overweight women
trended towards increased preeclampsia determined that obese black women do raise questions about the representa-
risk with all women included, but the have a lower risk of the development of tive nature of this cohort.
TABLE 5
Relationships between obesity metrics and risk of preeclampsia: all patients
Odds ratio
Predictor Outcome Unadjusted 95% confidence interval Adjusteda 95% confidence interval
Body mass index Clinical preeclampsia (n 30) 1.006 0.960e1.055 1.008 0.961e1.057
Research preeclampsia (n 14) 1.028 0.964e1.096 1.025 0.963e1.091
Body fat Clinical preeclampsia (n 30) 1.011 0.973e1.050 1.009 0.971e1.048
Research preeclampsia (n 14) 1.030 0.973e1.092 1.034 0.977e1.094
a
Adjusted for age, race, and smoking status.
Sween. Percent body fat and preeclampsia risk. Am J Obstet Gynecol 2015.
Another challenge encountered in this body fat percentage values. Further, we 6. Van Loan MD, Kopp LE, King JC, Wong WW,
study was the unusually high rates of did not adjust for multiple comparisons Mayclin PL. Fluid changes during pregnancy:
use of bioimpedance spectroscopy. J Appl
preeclampsia in the lean and overweight in our regression modeling, which Physiol 1995;78:1037-42.
women. The overall rate of preeclampsia allowed the possibility of a spurious 7. Lukaski HC, Siders WA, Nielsen EJ, Hall CB.
has not changed dramatically in our result being interpreted as positive; Total body water in pregnancy: assessment by
Pittsburgh population, which suggests however, we believed that the application using bioelectrical impedance. Am J Clin Nutr
that the high incidence of preeclampsia of a particularly harsh adjustment of the 1994;59:578-85.
8. van Raaij JMA, Peek MEM, Vermaat-
may be an aberrant nding that is related signicance level would have made it
Miedema SH, Schonk CM, Hautvast JGAJ. New
to the intentionally small numbers of impossible to detect any effect in a equations for estimating body fat mass in preg-
lean and overweight women. Alterna- cohort with relatively few cases. We nancy from body density or total body water. Am
tively, the cohort is 63% black women, adjusted the logistic regression model for J Clin Nutr 1988;48:24-9.
for whom the literature is mixed with age, race, and smoking during pregnancy 9. NHBPEPW Group. Report of the National
regards to preeclampsia risk.3,16,17 In one but could not account for other potential High Blood Pressure Education Program
Working Group on High Blood Pressure in
study, lean black women were more covariates, such as prepregnancy and Pregnancy. Am J Obstet Gynecol
likely to experience preeclampsia than pregnancy diet and exercise habits. 2000;183(suppl):S1-22.
lean white women, but the trend This study supports the relationship 10. Hawkins TLA, Roberts J, Mangos GJ,
reversed at BMI >25 kg/m2.3 Thus, the of fat to preeclampsia. Subsequent lon- Davis GK, Roberts LM, Brown MA. Plasma uric
large number of black women in the lean gitudinal, multicenter trials with larger acid remains a marker of poor outcome in hy-
pertensive pregnancy: a retrospective cohort
cohort may explain the high levels of case numbers are needed to further
study. BJOG 2012;119:484-92.
preeclampsia in this group. In our study, assess the utility of BIA-determined body 11. Roberts JM, Bodnar LM, Lain KY, et al. Uric
when we dichotomized by race, there fat percentage in the prediction of pre- acid is as important as proteinuria in identifying
was no signicant relationship between eclampsia onset. - fetal risk in women with gestational hyperten-
body fat percentage in white obese sion. Hypertension 2005;46:1263-9.
women and preeclampsia risk, probably 12. Roberts JM, Hubel CA. The two stage
model of preeclampsia: variations on the theme.
because of a smaller number of cases in REFERENCES Placenta 2009;23(suppl):S32-7.
this subset. In obese black women, there 13. Burton GJ, Woods AW, Jauniaux E,
1. American College of Obstetricians and Gy-
was a signicant relationship between necologists. Diagnosis and management of Kingdom JCP. Rheological and physiological
increasing body fat percentage and risk preeclampsia and eclampsia. ACOG Committee consequences of conversion of the maternal
of preeclampsia by the research, but not on Practice Bulletins: Obstetrics. ACOG practice spiral arteries for uteroplacental blood ow dur-
ing human pregnancy. Placenta 2009;30:
by the clinical denition. This nding is bulletin no. 33. Obstet Gynecol 2002;99:
159-67. 473-82.
also probably because of a small sample 14. Jaque-Fortunato SV, Khodiguian N, Artal R,
2. Roberts JM, Gammill HS. Preeclampsia
size in the research denition subgroup recent insights. Hypertension 2005;46:1-7. Wiswell RA. Body composition in pregnancy.
and not a true racial interaction. It was 3. Bodnar LM, Catov JM, Klebanoff MA, Semin Perinatol 1996;20:340-2.
an unfortunate limitation of this study Ness RB, Roberts JM. Prepregnancy body 15. Marshall NE, Guild C, Cheng YW,
that the sample of women who experi- mass index and the occurrence of severe hy- Caughey AB, Halloran DR. Racial disparities in
pertensive disorders of pregnancy. Epidemi- pregnancy outcomes in obese women. J Matern
enced preeclampsia was not large
ology 2007;18:234-9. Fetal Neonatal Med 2014;27:122-6.
enough to assess adequately the impact 4. Bodnar LM, Ness RB, Markovic N, 16. Sibai BM, Gordon T, Thom E, et al. Risk
of body fat percentage by race. Roberts JM. The risk of preeclampsia rises with factors for preeclampsia in healthy nulliparous
There were additional limitations in increasing prepregnancy body mass index. Ann women: a prospective multicenter study. Am J
this study. This cohort had a small Epidemiol 2005;15:475-82. Obstet Gynecol 1995;172:642-8.
5. McCarthy EA, Strauss BJG, Walker SP, 17. Knuist M, Bonsel GJ, Zondervan HA,
number of cases (30 cases by the clinical
Permezel M. Determination of maternal body Treffers PE. Risk factors for preeclampsia in
denition, of which 14 cases also t the composition in pregnancy and its relevance to nulliparous women in distinct ethnic groups: a
research denition), which led to wide perinatal outcomes. Obstet Gynecol Surv prospective cohort study. Obstet Gynecol
condence intervals for some BMI and 2004;59:731-42. 1998;92:174-8.