Abstract
BACKGROUND
Maternal retinol-binding protein 4 (RBP4) and lipids may relate to preeclampsia and preterm birth risk but longitudinal data are lacking. This study examines these biomarkers longitudinally during pregnancy in relation to preeclampsia and preterm birth risk.
METHODS
Maternal serum samples from the Calcium for Preeclampsia Prevention (CPEP) trial were analyzed at baseline: average 15 gestational weeks; mid-pregnancy: average 27 weeks; and at >34 weeks. We measured RBP4, total cholesterol, high-density lipoprotein, low-density lipoprotein, triglycerides and lipoprotein (a) (Lp(a)). Cross-sectional logistic regression analyses estimated the odds ratio (OR) and 95% confidence intervals (CI) for preterm preeclampsia (n = 63), term preeclampsia (n = 104), and preterm delivery (n = 160) associated with RBP4 and lipids at baseline and mid-pregnancy compared with controls (n = 136). Longitudinal trajectories across pregnancy were assessed using mixed linear models with fixed effects. Adjusted models included clinical and demographic factors.
RESULTS
RBP4 concentrations at baseline and mid-pregnancy were associated with a 4- to 8-fold increase in preterm preeclampsia risk but were not associated with term preeclampsia. RBP4 measured mid-pregnancy was also associated with preterm birth (OR = 6.67, 95% CI: 1.65, 26.84). Higher triglyceride concentrations in mid-pregnancy were associated with a 2- to 4-fold increased risk for both preeclampsia and preterm birth. Longitudinal models demonstrate that both preterm preeclampsia and preterm birth cases had elevated RBP4 throughout gestation.
CONCLUSIONS
Elevated RBP4 is detectable early in pregnancy and its strong relation with preterm preeclampsia merits further investigation and confirmation to evaluate its potential use as a predictor, particularly among high-risk women.
Keywords: blood pressure, hypertension, lipids, preeclampsia, pregnancy, preterm birth, retinol-binding protein 4
Hypertensive disorders of pregnancy are a major cause of maternal morbidity and mortality in the developed world1 and their prevalence appears to be increasing in the United States.2 Preeclampsia is new-onset hypertension after 20 weeks of pregnancy, typically with accompanying proteinuria that indicates altered kidney function. It is a serious complication, often triggering an early delivery, and can progress to eclampsia with maternal seizures and critical blood pressure elevations. Current clinical practice relies on blood pressure and proteinuria measurements routinely conducted as part of prenatal care as the basis for diagnosis.
The pathophysiology of preeclampsia is multifactorial and likely includes both maternal and fetoplacental factors. In particular, impaired development of placental blood vessels which leads to placental hypoperfusion is thought to be fundamental to the development of preeclampsia.3 While some patient characteristics, notably nulliparity and obesity,1 are associated with increased preeclampsia risk, they are not strong predictors. However, the link with obesity suggests that adipocyte function or related metabolic factors may play a role.4
Retinol-binding protein 4 (RBP4) is an adipokine that delivers vitamin A from the liver to the peripheral tissues, but it also plays a role in obesity-related insulin-resistance and in inflammatory processes.5 Maternal adipose tissue and liver both serve as sources of RBP4 and higher serum levels have been associated with other relevant weight-related pathologies including hypertension and cardiovascular disease risk factors6 as well as polycystic ovary syndrome.7 While RBP4 has been investigated in relation to adverse pregnancy outcomes, the physiologic significance of altered RBP4 is not clear.8
Prior studies have examined RBP4 early in gestation in relation to preeclampsia risk and found either no association9 or increased risks for early-onset severe preeclampsia.10 When samples were collected at or after diagnosis, RBP4 has been reported to be elevated among preeclampsia cases in some studies11,12 but not different from controls in other studies13,14 and even lower than controls when samples were collected during the delivery hospital admission15 or at diagnosis.16 The cross-sectional, retrospective design, small sample size, assessment of RB4 once in pregnancy, and inadequate adjustment for confounders could explain the inconsistency in previous reports and these limitations hinder inference regarding pathophysiology and potential clinical utility.
Cholesterol and other maternal lipids are required to support the rapid growth and cell differentiation in the embryo and fetus. The potential role for altered maternal lipid levels as an early marker for adverse pregnancy outcomes has also been debated. Higher triglyceride levels measured in mid-pregnancy have been shown to predict preeclampsia17 but not when measured earlier or later in pregnancy.18 A rate of triglyceride change below the median from preconception over the course of pregnancy was associated with pregnancy loss and preterm birth19 suggesting that a steeper increase in early pregnancy was protective.
No prior studies have examined longitudinal measures in pregnant women at risk for preeclampsia. Our aim was to examine repeated measures of maternal lipids and RBP4 over the course of pregnancy to evaluate the relationship of these biomarkers with preeclampsia and preterm birth risk.
SUBJECTS AND METHODS
Data are from the Calcium for Preeclampsia Prevention (CPEP) trial, a clinical trial conducted from 1992–1995 to evaluate whether calcium supplementation would prevent preeclampsia among 4,589 nulliparous women with singleton pregnancies.20 We conducted a nested case–control study using stored serum samples. All cases of preterm preeclampsia (n = 63) and preterm birth (total n = 160; preterm without preeclampsia n = 107) were selected for analyses along with a random selection of preeclampsia cases at term (n = 104). The case definition for preeclampsia was previously established as hypertension (with diastolic blood pressure of at least 90 mm Hg on 2 occasions) and proteinuria (based on urine dipstick results on 2 occasions or 24-hour urine collection).20,21 The control group was comprised of randomly selected women (n = 136) who remained normotensive and normoglycemic throughout pregnancy without proteinuria, who also delivered a term infant without evidence of growth restriction. Controls and term preeclampsia cases were randomly selected based on the number of participants and cases in each CPEP study site in order to preserve the geographic distribution of the original trial. Sample size was based on providing statistical power for hypotheses related to the prediction of preeclampsia22 and other related adverse pregnancy outcomes.
Nonfasting blood samples were collected before randomization to calcium supplement or placebo (prior to 21 weeks, 6 days of gestation) as well as at least twice during follow-up. We analyzed the baseline sample (prerandomization), a mid-pregnancy sample, collected between gestational weeks 22 to 34 and a third sample at >34 weeks when available. RBP-4 was measured in plasma using an enzyme-linked immunosorbant assay (ELISA) from R&D Systems (Minneapolis, MN). The assay coefficient of variation (CV) was <10%. Lipid determinations were performed in serum on the Roche Modular P Chemistry Analyzer (Roche Diagnostics, Indianapolis, IN). Total cholesterol was measured using a cholesterol oxidase method (Roche Diagnostics) with a laboratory CV of 1.6%. High-density lipoprotein (HDL) cholesterol was measured directly using the Roche HDL cholesterol third generation homogenous enzymatic direct method (Roche Diagnostics) with a laboratory CV of 2.3%. Triglycerides were measured using the Roche Triglyceride Glycerol Blanked enzymatic colorimetric assay (Roche Diagnostics) with a laboratory CV of 2.1%. Methods for measurement of total cholesterol, HDL cholesterol, and triglyceride were monitored by the CDC/NHLBI Lipid standardization program. Low-density lipoprotein (LDL) cholesterol was measured directly using the homogenous enzymatic colorimetric Roche LDL-C second generation method (Roche Diagnostics) with a laboratory CV of 2.9%. Lipoprotein (a) (Lp(a)) was measured using an immunoturbidimetric assay by Randox Laboratories Limited, United Kingdom, adapted for use on the Roche Modular P Analyzer. The laboratory CV was 2.5%.
Participants were linked to stored biospecimens by a unique study identification number with no personally identifiable information. This study was exempted from Institutional Review Board review by the National Institutes of Health’s Office of Human Subjects Research because it uses existing de-identified samples.
STATISTICAL ANALYSES
Descriptive statistics were calculated for participants’ characteristics and biomarker levels at each of the three sample collection points. In cross-sectional analyses, we explored whether lipids and RBP4 levels measured at approximately week 15 of gestation predicted preeclampsia and preterm delivery. Crude and adjusted odds ratios (OR) and 95% confidence intervals (CI) were estimated using logistic regression models for the association of lipids and RBP4 levels at baseline with preterm preeclampsia, term preeclampsia, all preeclampsia (either term or preterm), and preterm delivery. Concentrations of all biomarkers had a skewed distribution, and, therefore, were transformed using natural logarithms. Non-HDL cholesterol was derived by subtracting HDL from total cholesterol, both measured as specified above, prior to log-transformation. Models were adjusted for age, race, body mass index, insurance status, current smoking, marital status, study center, and gestational age at the time of sampling (to account for the variation in time of blood draw during gestation). To examine the prediction of pregnancy complications, each biomarker was evaluated in relation to each outcome of interest separately. Next, models were run with only covariates, with covariates and all lipid biomarkers and finally, with covariates, lipid biomarkers, and RBP4. The area under the curve for the receiver operating characteristic curve for each model was calculated and a cross-validation was attempted. The mid-pregnancy sample was analyzed similarly. Since third samples were scarce for the preterm groups, they were only examined in the longitudinal analyses described below.
We also examined potential nonlinear associations between biomarkers levels and pregnancy outcomes. To this aim, we categorized the biomarkers using tertiles and explored the ORs and CIs for adverse pregnancy outcomes using the middle tertile as the reference group.
To describe the longitudinal trajectories of biomarkers throughout pregnancy in cases and controls, we used linear mixed models with fixed effects and an unstructured covariance matrix and reported the differences in repeated measures of biomarkers during gestation in women with preterm preeclampsia compared to controls. Adjusted means, derived from these longitudinal models, were plotted for each biomarker in each assessment period. These 3 periods of blood draw were originally designed in the study: <22.0 weeks of gestation, 22.0–34.0 weeks of gestation, and >34.0 weeks of gestation. For each time window, only one set of biomarker measurements per woman was included, regardless of whether she had additional blood draws. A similar approach was used in women with preterm delivery. Models were adjusted for age, race, body mass index, insurance status, current smoking, marital status, study center. Calcium treatment status was evaluated as a potential covariate but not included in the final model because it did not change the results. Statistical analyses were conducted using SAS 9.4 (SAS Institute, Cary, NC).
RESULTS
As the CPEP was designed to study high-risk women, women in the study were young (mean age 20–21), on average had a body mass index in the overweight category (>25) and about half of the sample were Black (Table 1). Most women (85–96%) had no private insurance and <30% were married. The gestational age at biospecimen collection was similar for all groups.
Table 1.
Characteristics of preeclampsia and preterm cases and controls from the Calcium for Preeclampsia Prevention Trial, 1992–1995
| Characteristic | Controls | Preterm Preeclampsia | Term Preeclampsia | All Preeclampsia | All Preterm |
|---|---|---|---|---|---|
| n | 136 | 63 | 104 | 167 | 160 |
| Mean (SD) | Mean (SD) | Mean (SD) | Mean (SD) | Mean (SD) | |
| Age, year | 20.4 (3.7) | 21.1 (4.6) | 20.7 (4.4) | 20.9 (4.5) | 20.5 (4.3) |
| Weight, kg | 66.6 (14.3) | 71.4 (19.6) | 73.4 (17.7) | 72.8 (18.4) | 68.8 (19.6) |
| BMI, kg/m2 | 25.5 (5.7) | 27.9 (7.3) | 28.0 (6.0) | 28.0 (6.5) | 26.4 (7.1) |
| Current smoker, n (%) | 10 (7.4) | 6 (9.5) | 16 (15.4) | 22 (13.2) | 17 (10.6) |
| Race, n (%) | |||||
| White non-Hispanic | 47 (34.6) | 18 (28.6) | 33 (31.7) | 51 (30.5) | 35 (21.9) |
| Black | 63 (46.3) | 32 (50.8) | 57 (54.8) | 89 (53.3) | 104 (65.0) |
| Hispanic or other race | 26 (19.1) | 13 (20.6) | 14 (13.5) | 27 (16.2) | 21 (13.1) |
| Less than high school education, n (%) | 60 (44.1) | 25 (40.3) | 46 (44.7) | 71 (43.0) | 66 (41.5) |
| No private insurance, n (%) | 118 (86.8) | 56 (88.9) | 100 (96.2) | 156 (93.4) | 144 (90.0) |
| Married, n (%) | 30 (22.1) | 18 (28.6) | 18 (17.3) | 36 (21.6) | 27 (16.9) |
| Male infant, n (%) | 63 (46.3) | 32 (50.8) | 53 (51.0) | 85 (50.9) | 75 (46.9) |
| Birthweight, g | 3,314 (445) | 2,152 (726) | 3,304 (495) | 2,870 (814) | 1,699 (731) |
| Gestational age at delivery, week | 39.3 (1.5) | 34.4 (2.9) | 39.2 (1.2) | 37.4 (3.1) | 31.3 (4.1) |
| Randomized to calcium, n (%) | 68 (50.0) | 34 (54.0) | 41 (39.4) | 75 (44.9) | 83 (51.9) |
| Baseline biomarkers | |||||
| HDL mg/dl | 61.5 (13.1) | 60.5 (13.2) | 58.9 (12.1) | 59.5 (12.5) | 61.5 (13.3) |
| LDL mg/dl | 118.4 (32.9) | 115.5 (37.7) | 119.2 (36.8) | 117.8 (34.3) | 112.6 (35.2) |
| non-HDL cholesterol mg/dl | 138.9 (33.1) | 139.1 (42.7) | 141.5 (35.9) | 140.6 (38.50) | 133.6 (39.4) |
| Triglyceride mg/dl | 126.2 (41.0) | 132.6 (60.1) | 132.7 (55.5) | 132.6 (57.1) | 122.0 (51.9) |
| Lipoprotein A mg/dl | 40.7 (44.3) | 43.4 (40.1) | 42.4 (36.8) | 42.8 (38.0) | 44.9 (38.4) |
| RBP4 | 22.2 (5.4) | 24.0 (5.8) | 21.4 (5.2) | 22.4 (5.5) | 22.5 (5.8) |
| Gestational age at baseline sample (<22.0 weeks) | 15.6 (2.6)range: 9.0–21.0 | 15.3 (2.8)range: 7.0–20.0 | 15.4 (3.2)range: 2.0–21.0 | 15.4 (3.1)range: 2.0–21.0 | 15.6 (2.9)range: 6.0–20.0 |
| Second sample biomarkers | |||||
| n = 117 | n = 53 | n = 92 | n = 145 | n = 62 | |
| HDL mg/dl | 63.4 (13.1) | 62.6 (16.7) | 60.8 (13.6) | 61.5 (14.8) | 63.5 (16.0) |
| LDL mg/dl | 151.7 (37.0) | 142.8 (50.1) | 146.9 (42.0) | 145.4 (45.2) | 146.4 (47.4) |
| non-HDL cholesterol mg/dl | 177.9 (39.4) | 180.3 (63.8) | 175.9 (44.9) | 177.5 (52.5) | 182.8 (58.9) |
| Triglyceride mg/dl | 176.5 (61.1) | 218.2 (137.3) | 194.4 (70.2) | 203.1 (100.2) | 212.7 (133.5) |
| Lipoprotein A mg/dl | 44.4 (47.7) | 50.1 (47.7) | 48.8 (44.1) | 49.3 (45.3) | 52.4 (48.8) |
| RBP4 | 22.8 (5.7) | 25.2 (7.1) | 21.9 (5.8) | 23.1 (6.5) | 25.4 (7.4) |
| Gestational age at second sample (22.0–34.0 weeks) | 27.3 (1.6)range: 24.0–33.0 | 27.0 (1.8)range: 22.0–33.0 | 27.1 (1.7)range: 23.0–31.0 | 27.1 (1.7)range: 22.0–33.0 | 27.0 (1.7)range: 22.0–33.0 |
| Third sample biomarkers | |||||
| n = 100 | n = 16 | n = 88 | n = 104 | n = 16 | |
| HDL mg/dl | 58.3 (13.5) | 60.2 (16.9) | 56.4 (14.4) | 5.7 (14.8) | 60.2 (9.1) |
| LDL mg/dl | 160.3 (45.3) | 158.6 (122.0) | 143.7 (44.8) | 146.0 (62.4) | 166.5 (113.6) |
| non-HDL cholesterol mg/dl | 197.4 (45.4) | 209.1 (127.9) | 184.8 (50.0) | 188.5 (67.6) | 221.0 (119.1) |
| Triglyceride mg/dl | 243.9 (84.3) | 270.0 (123.0) | 251.0 (92.8) | 253.9 (97.6) | 297.6 (116.3) |
| Lipoprotein A mg/dl | 43.7 (51.5) | 50.7 (57.5) | 46.2 (41.8) | 46.9 (44.3) | 42.6 (52.5) |
| RBP4 | 22.8 (6.0) | 23.7 (8.0) | 22.1 (6.7) | 22.3 (6.8) | 24.7 (9.4) |
| Gestational age at third sample (>34.0 weeks) | 36.2 (0.9)range: 34.0–40.0 | 35.4 (0.8)range: 34.0–36.0 | 36.2 (0.9)range: 34.0–40.0 | 36.1 (0.9)range: 34.0–40.0 | 35.3 (0.9)range: 34.0–36.0 |
Abbreviations: HDL, high-density lipoprotein; LDL, low-density lipoprotein; RBP4, retinol-binding protein 4.
Per log-unit increment in the RBP4 concentrations measured at baseline (15 weeks on average) was associated with statistically significant 4- to 8-fold increased risk for preterm preeclampsia (Table 2). Adjustment for known risk factor covariates and gestational age at specimen collection increased the risk estimates. In contrast to preterm preeclampsia, RBP4 was not associated with preeclampsia at term. Combining all preeclampsia cases together attenuates the RPB4 results (OR = 1.66, CI: 0.56–4.92). With regard to all preterm delivery (with or without preeclampsia), RBP4 levels were elevated at baseline but the associations were not significant. In analysis with tertiles of RBP4, we observed that compared to the middle tertile, the lowest tertile was associated with a lower odds of preterm preeclampsia (OR = 0.41, CI: 0.18, 0.95), but no significant association was observed with the highest tertile.
Table 2.
Early pregnancya retinol-binding protein 4 (RBP4) and lipids in relation to preeclampsia and preterm birth
| Outcome and biomarkers | Unadjusted | Adjusted |
|---|---|---|
| OR (95% CI) | OR (95% CI) | |
| Preterm preeclampsia | ||
| RBP4 | 4.30 (1.17, 15.37) | 8.24 (1.73, 39.32) |
| HDL | 0.68 (0.17, 2.78) | 0.80 (0.15, 4.27) |
| LDL | 0.63 (0.23, 1.96) | 0.49 (0.14, 1.68) |
| non-HDL | 0.85 (0.26, 2.81) | 0.61 (0.15, 2.52) |
| Triglycerides | 1.11 (0.49, 2.48) | 1.18 (0.43, 3.23) |
| Lipoprotein A | 1.00 (0.76, 1.33) | 0.99 (0.73, 1.35) |
| Term preeclampsia | ||
| RBP4 | 0.61 (0.21, 1.73) | 0.70 (0.21, 2.38) |
| HDL | 0.40 (0.11, 1.31) | 0.38 (0.08, 1.71) |
| LDL | 1.11 (0.44, 2.82) | 0.91 (0.32, 2.56) |
| non-HDL | 1.24 (0.44, 3.47) | 1.05 (0.32, 3.45) |
| Triglycerides | 1.20 (0.60, 2.39) | 1.19 (0.51, 2.79) |
| Lipoprotein A | 1.04 (0.82, 1.33) | 0.91 (0.69, 1.21) |
| All preeclampsia | ||
| RBP4 | 1.23 (0.48, 3.13) | 1.66 (0.56, 4.92) |
| HDL | 0.48 (0.16, 1.41) | 0.50 (0.14, 1.83) |
| LDL | 0.91 (0.40, 2.07) | 0.72 (0.29, 1.76) |
| non-HDL | 1.07 (0.44, 2.61) | 0.85 (0.31, 2.33) |
| Triglycerides | 1.15 (0.64, 2.08) | 1.16 (0.55, 2.43) |
| Lipoprotein A | 1.03 (0.83, 1.27) | 0.95 (0.75, 1.20) |
| All preterm | ||
| RBP4 | 1.23 (0.49, 3.08) | 2.67 (0.90, 7.95) |
| HDL | 0.96 (0.34, 2.71) | 0.55 (0.16, 1.90) |
| LDL | 0.51 (0.23, 1.10) | 0.49 (0.21, 1.16) |
| non-HDL | 0.48 (0.20, 1.12) | 0.51 (0.20, 1.40) |
| Triglycerides | 0.65 (0.35, 1.21) | 0.95 (0.44, 2.06) |
| Lipoprotein A | 1.12 (0.90, 1.40) | 0.97 (0.76, 1.23) |
Models were adjusted for age, race, body mass index, insurance status, current smoking, marital status, study center, and gestational age at the time of blood sampling. Abbreviations: CI, confidence interval; HDL, high-density lipoprotein; LDL, low-density lipoprotein; OR, odds ratio; RBP4, retinol-binding protein 4.
The effect sizes are per log-unit increase in the biomarker concentrations. Biomarkers were log-transformed for normality.
In general, cholesterol concentrations at baseline (HDL, LDL, and non-HDL) were inversely related to risk for preterm preeclampsia, term preeclampsia, and preterm birth, suggesting a reduced risk with higher lipids but these were nonsignificant. Analysis using tertiles of cholesterol measures showed no threshold effect. The results of the Lp(a) and triglycerides measured at baseline were also null for all outcomes; however, we observed a nonlinear relation associated with the baseline triglycerides, where both low (<101) and high triglyceride levels (>139) were associated with preterm preeclampsia [adjusted OR = 2.50 (1.06, 5.81) for low; adjusted OR = 2.75 (1.11, 6.80) for high] when compared to the middle category.
When all biomarkers were included in the same model, baseline RBP4 remained a significant predictor of preterm preeclampsia with an adjusted OR = 8.29 (1.65, 41.66). Adding RBP4 to a prediction model increased the area under the curve to 0.7040, an improvement over the model with covariates only (area under the curve = 0.6775) and covariates plus lipids (area under the curve = 0.6800) but an attempt at cross-validation (leave-one-out) did not replicate the initial finding.
Triglycerides measured in the second, mid-pregnancy sample remained associated with term preeclampsia with a 2- to 3-fold increased risk (adjusted OR = 3.31, CI: 1.21–9.07). No other lipids or RBP4 measured at approximately week 27 were significantly related to term preeclampsia.
We also examined risk for preterm delivery that was not related to preeclampsia, but no biomarkers were significantly related to risk and point estimates were attenuated at baseline in this subsample (data not shown). Second samples were only available for 18 cases of preterm delivery without preeclampsia, which precluded multivariate analyses.
Longitudinal models revealed that women with preterm preeclampsia and also women with preterm delivery (with or without preeclampsia) had consistently higher levels of RBP4, compared to controls throughout pregnancy (Figures 1 and 2). Although the results were nonsignificant, women with preterm preeclampsia had consistently higher levels of Lp(a) throughout pregnancy compared to controls. Furthermore, the pattern of triglycerides throughout pregnancy indicated that women with preterm preeclampsia had a steeper slope for triglyceride increase from time 1 to time 2 compared to controls (P for slope difference = 0.07). A similar distinct pattern of triglyceride increase was observed for preterm delivery (Figure 2).
Figure 1.
Adjusted mean levels* of retinol-binding protein 4 and lipids among women with preterm preeclampsia (PE) and controls at 3 time points during pregnancy: <22 weeks, 22–34 weeks, and >34 weeks, Calcium for Preeclampsia Prevention Trial, 1992–1995. *Means were adjusted for age, race, body mass index, insurance status, current smoking, marital status, and study center. Abbreviations: RBP4, retinol-binding protein 4; HDL, high-density lipoprotein; LDL, low-density lipoprotein; Lp(a), Lipoprotein (a); TG, Triglycerides.
Figure 2.
Adjusted mean levels* of retinol-binding protein 4 and lipids among women with preterm delivery and controls at 3 time points during pregnancy: <22 weeks, 22–34 weeks, and >34 weeks, Calcium for Preeclampsia Prevention Trial, 1992–1995 *Means were adjusted for age, race, body mass index, insurance status, current smoking, marital status, and study center. Abbreviations: RBP4, retinol-binding protein 4; HDL, high-density lipoprotein; LDL, low-density lipoprotein; Lp(a), Lipoprotein (a); TG, Triglycerides.
DISCUSSION
We found that RBP4 levels were strongly and positively associated with early-onset preeclampsia with a 4- to 8-fold increase in risk, but no associations were observed for preeclampsia at term. Triglyceride levels at mid-pregnancy were associated with a 2- to 3-fold increase in risk for both early and term preeclampsia. We also observed consistent increases in risk for preterm delivery associated with both RBP4 and triglyceride levels at mid-pregnancy, but did not have a sufficient sample size to assess this risk only among preterm births without preeclampsia. In contrast to Laughon et al.,19 we observed that a steeper increase in triglycerides in early pregnancy was associated with an increased preterm birth risk rather than a protective effect.
Our findings for triglycerides are supported by a small study23 that measured lipids in 26 normotensive and 27 preeclamptic pregnancies. Triglycerides were elevated in patients with preeclampsia but no differences were observed for total, HDL, or LDL cholesterol. Given our longitudinal sampling and larger study, we were able to observe that both low and high triglyceride levels early in pregnancy were associated with a doubling of preeclampsia risk and this relation became linear, with higher risk associated with higher triglyceride levels, in the second trimester for both term and preterm preeclampsia.
Preeclampsia cases appeared to have higher total cholesterol, triglycerides, and LDL in a prospective study of 251 pregnant women with early maternal blood samples (4–12 weeks) in which 16 mild and 10 severe preeclampsia cases developed.24 Our baseline samples were generally later in pregnancy, on average 15 weeks (range 2–21 weeks), and nonfasting which might account for these differences, but we also have a larger sample of cases and adjustment for gestational age at sampling did not have a substantive impact on the risk estimates. A single first trimester screening sample analyzed for RBP4 among 14 preeclampsia cases and 11 control pregnancies found no difference related to subsequent preeclampsia risk14 but this might be due to a mixing of preterm and term cases since the gestational age at diagnosis was not described.
Using a 1-time blood sample collected at the time of diagnosis, Vaisbuch et al. reported an association between RBP4 and preeclampsia, with a stronger relationship observed for early-onset disease among 39 women; however, the association may have been the result of preeclampsia rather than antedating it.11 On the other hand, RBP4 did not appear to be related to clinical measures or lipids among women with blood samples collected after diagnosis in a sample of 16 preeclampsia patients compared with 20 controls.13 In contrast to our findings, a small study in Japan found RBP4 measured at diagnosis was not different between 18 healthy pregnant women and 18 early-onset preeclampsia cases, but did see higher RBP4 increased in later onset cases (n = 41 in each group).25
Our study benefits from detailed clinical data and longitudinal serum samples from the CPEP Trial.20 With more than 60 cases of preterm preeclampsia and over 100 cases of preeclampsia at term, this is the largest studies to relate preeclampsia to RBP4, one of the few with prospective measurements and the only study with longitudinal measures across gestation. However, since women had a high risk for preeclampsia by design (nulliparous, majority Black and overweight, etc.), our findings might not be generalizable to a lower-risk obstetric cohort. We are also limited in our assessment of lipids by having only nonfasting samples available, although we note that fasting status has been found to remain predictive of cardiovascular events as any differences in fasting time should be random with respect to the development of later disease.26 In consideration of these issues, we measured LDL directly, rather than calculating it based on total cholesterol and HDL levels. We note that our samples were stored for up to 20 years which could impact the absolute values of the lipids we studied. Reassuringly, long-term storage of lipids, including triglycerides, yielded stable results after 8–10 years in similar conditions.27 Since our main findings are based on the relative concentrations rather than absolute values, we expect any potential storage-related variation is independent of diagnosis and thus would not substantively impact the associations we observed. Finally, we had a relatively small number of samples available among preterm deliveries without preeclampsia which precluded a complete independent analysis of preterm birth. It is somewhat reassuring that the majority of those cases (83%) had a spontaneous onset of labor, so they were less likely due to medical indications.
In conclusion, most prior studies have been limited by single blood samples, often collected after diagnosis, among relatively small numbers of preeclampsia cases and control pregnant women. In our larger study based on a clinical trial of women at risk for preeclampsia, we find that RBP4 is significantly and consistently related to preterm preeclampsia but not preeclampsia that developed at term. RBP4 is also associated with an increased preterm birth risk when measured at mid-pregnancy. Triglyceride levels were elevated in mid-pregnancy among women who developed preeclampsia and among those who delivered preterm. RBP4 merits further investigation in relation to preeclampsia. If our results are confirmed, RBP4 may be helpful in identifying women early in pregnancy who are at high risk for preterm preeclampsia.
DISCLOSURE
The authors declared no conflict of interest.
ACKNOWLEDGMENTS
This research was supported by the Intramural Research Program of Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD). The manuscript has been cleared by NICHD for publication but the funding source had no role in the study design; in the collection, analysis, and interpretation of data; or in the writing of the report. The Calcium for Preeclampsia Prevention trial was supported by contracts (N01-HD-1-3121, -3122, -3123, -3124, -3125, and -3126; N01-HD-3154; and N01-HD-5-3246) with the NICHD, with co-funding from the National Heart, Lung, and Blood Institute. Recent biomarker assays, including retinol-binding protein 4, was supported by contract (HHSN275201100002I-HHSN27500003) with NICHD.
REFERENCES
- 1. Hutcheon JA, Lisonkova S, Joseph KS. Epidemiology of pre-eclampsia and the other hypertensive disorders of pregnancy. Best Pract Res Clin Obstet Gynaecol 2011; 25:391–403. [DOI] [PubMed] [Google Scholar]
- 2. Berg CJ, Mackay AP, Qin C, Callaghan WM. Overview of maternal morbidity during hospitalization for labor and delivery in the United States: 1993-1997 and 2001-2005. Obstet Gynecol 2009; 113:1075–1081. [DOI] [PubMed] [Google Scholar]
- 3. Kovo M, Schreiber L, Bar J. Placental vascular pathology as a mechanism of disease in pregnancy complications. Thromb Res 2013; 131(suppl 1):S18–S21. [DOI] [PubMed] [Google Scholar]
- 4. Boghossian NS, Yeung E, Mendola P, Hinkle SN, Laughon SK, Zhang C, Albert PS. Risk factors differ between recurrent and incident preeclampsia: a hospital-based cohort study. Ann Epidemiol 2014; 24:871–7e3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Yao-Borengasser A, Varma V, Bodles AM, Rasouli N, Phanavanh B, Lee MJ, Starks T, Kern LM, Spencer HJ, 3rd, Rashidi AA, McGehee RE, Jr, Fried SK, Kern PA. Retinol binding protein 4 expression in humans: relationship to insulin resistance, inflammation, and response to pioglitazone. J Clin Endocrinol Metab 2007; 92:2590–2597. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Solini A, Santini E, Madec S, Rossi C, Muscelli E. Retinol-binding protein-4 in women with untreated essential hypertension. Am J Hypertens 2009; 22:1001–1006. [DOI] [PubMed] [Google Scholar]
- 7. Olszanecka-Glinianowicz M, Madej P, Zdun D, Bożentowicz-Wikarek M, Sikora J, Chudek J, Skałba P. Are plasma levels of visfatin and retinol-binding protein 4 (RBP4) associated with body mass, metabolic and hormonal disturbances in women with polycystic ovary syndrome? Eur J Obstet Gynecol Reprod Biol 2012; 162:55–61. [DOI] [PubMed] [Google Scholar]
- 8. Miehle K, Stepan H, Fasshauer M. Leptin, adiponectin and other adipokines in gestational diabetes mellitus and pre-eclampsia. Clin Endocrinol (Oxf) 2012; 76:2–11. [DOI] [PubMed] [Google Scholar]
- 9. Nanda S, Nikoletakis G, Markova D, Poon LC, Nicolaides KH. Maternal serum retinol-binding protein-4 at 11-13 weeks’ gestation in normal and pathological pregnancies. Metabolism 2013; 62:814–819. [DOI] [PubMed] [Google Scholar]
- 10. Yliniemi A, Nurkkala MM, Kopman S, Korpimaki T, Kouru H, Ryynanen M, Marttala J. First trimester placental retinol-binding protein 4 (RBP4) and pregnancy-associated placental protein A (PAPP-A) in the prediction of early-onset severe pre-eclampsia. Metabolism 2015; 64:521–526. [DOI] [PubMed] [Google Scholar]
- 11. Vaisbuch E, Romero R, Mazaki-Tovi S, Erez O, Kim SK, Chaiworapongsa T, Gotsch F, Than NG, Dong Z, Pacora P, Lamont R, Yeo L, Hassan SS, Kusanovic JP. Retinol binding protein 4–a novel association with early-onset preeclampsia. J Perinat Med 2010; 38:129–139. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Shangguan X, Liu F, Wang H, He J, Dong M. Alterations in serum adipocyte fatty acid binding protein and retinol binding protein-4 in normal pregnancy and preeclampsia. Clin Chim Acta 2009; 407:58–61. [DOI] [PubMed] [Google Scholar]
- 13. Stepan H, Ebert T, Schrey S, Reisenbüchler C, Blüher M, Stumvoll M, Kratzsch J, Tönnessen P, Faber R, Fasshauer M. Preliminary report: serum levels of retinol-binding protein 4 in preeclampsia. Metabolism 2009; 58:275–277. [DOI] [PubMed] [Google Scholar]
- 14. Fruscalzo A, Biasioli A, Londero AP, Ceraudo M, Stel G, Bertozzi S, Marchesoni D, Driul L, Curcio F. Retinol binding protein as early marker of fetal growth restriction in first trimester maternal serum. Gynecol Endocrinol 2013; 29:323–326. [DOI] [PubMed] [Google Scholar]
- 15. Seol HJ, Kim JW, Kim HJ. Retinol-binding protein-4 is decreased in patients with preeclampsia in comparison with normal pregnant women. J Perinat Med 2011; 39:287–289. [DOI] [PubMed] [Google Scholar]
- 16. Park J, Cha DH, Lee SJ, Kim YN, Kim YH, Kim KP. Discovery of the serum biomarker proteins in severe preeclampsia by proteomic analysis. Exp Mol Med 2011; 43:427–435. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Zhou J, Zhao X, Wang Z, Hu Y. Combination of lipids and uric acid in mid-second trimester can be used to predict adverse pregnancy outcomes. J Matern Fetal Neonatal Med 2012; 25:2633–2638. [DOI] [PubMed] [Google Scholar]
- 18. Emet T, Ustüner I, Güven SG, Balık G, Ural UM, Tekin YB, Sentürk S, Sahin FK, Avşar AF. Plasma lipids and lipoproteins during pregnancy and related pregnancy outcomes. Arch Gynecol Obstet 2013; 288:49–55. [DOI] [PubMed] [Google Scholar]
- 19. Laughon SK, McLain AC, Sundaram R, Catov JM, Buck Louis GM. Maternal lipid change in relation to length of gestation: a prospective cohort study with preconception enrollment of women. Gynecol Obstet Invest 2014; 77:6–13. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Levine RJ, Hauth JC, Curet LB, Sibai BM, Catalano PM, Morris CD, DerSimonian R, Esterlitz JR, Raymond EG, Bild DE, Clemens JD, Cutler JA. Trial of calcium to prevent preeclampsia. N Engl J Med 1997; 337:69–76. [DOI] [PubMed] [Google Scholar]
- 21. Levine RJ, Lam C, Qian C, Yu KF, Maynard SE, Sachs BP, Sibai BM, Epstein FH, Romero R, Thadhani R, Karumanchi SA; CPEP Study Group. Soluble endoglin and other circulating antiangiogenic factors in preeclampsia. N Engl J Med 2006; 355:992–1005. [DOI] [PubMed] [Google Scholar]
- 22. Yeung EH, Liu A, Mills JL, Zhang C, Männistö T, Lu Z, Tsai MY, Mendola P. Increased levels of copeptin before clinical diagnosis of preelcampsia. Hypertension 2014; 64:1362–1367. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Kozan A, Yildirmak ST, Mihmanli V, Ayabakan H, Cicek YG, Kalaslioglu V, Doean S, Cebeci HC. Serum oxidized low density lipoprotein levels in preeclamptic and normotensive pregnants. Clin Exp Obstet Gynecol 2015; 42:746–748. [PubMed] [Google Scholar]
- 24. El Khouly NI, Sanad ZF, Saleh SA, Shabana AA, Elhalaby AF, Badr EE. Value of first-trimester serum lipid profile in early prediction of preeclampsia and its severity: A prospective cohort study. Hypertens Pregnancy 2016; 35:73–81. [DOI] [PubMed] [Google Scholar]
- 25. Masuyama H, Inoue S, Hiramatsu Y. Retinol-binding protein 4 and insulin resistance in preeclampsia. Endocr J 2011; 58:47–53. [DOI] [PubMed] [Google Scholar]
- 26. Langsted A, Freiberg JJ, Nordestgaard BG. Fasting and nonfasting lipid levels: influence of normal food intake on lipids, lipoproteins, apolipoproteins, and cardiovascular risk prediction. Circulation 2008; 118:2047–2056. [DOI] [PubMed] [Google Scholar]
- 27. Matthan NR, Ip B, Resteghini N, Ausman LM, Lichtenstein AH. Long-term fatty acid stability in human serum cholesteryl ester, triglyceride, and phospholipid fractions. J Lipid Res 2010; 51:2826–2832. [DOI] [PMC free article] [PubMed] [Google Scholar]