Objective
Lipoproteins are associated with atherogenic and inflammatory processes, and these processes may be related to adverse pregnancy outcomes. We therefore examined whether variations in lipoprotein particle size and concentration are associated with preterm birth (PTB) <35 weeks’ gestation.
Study Design
This is a case-control ancillary study to a randomized trial of omega-3 fatty acid supplementation to prevent recurrent PTB. We measured standard lipids and used nuclear magnetic resonance (NMR) spectroscopy to characterize 17 lipoprotein particles from plasma collected at the baseline randomization visit (16-21 weeks’ gestation) in 128 cases (PTB <35 weeks’ gestation) and 132 term controls. Logistic regression models controlled for study center, race/ethnicity, number of prior PTB, smoking, and treatment group, as well as total low-density lipoprotein (LDL), high-density lipoprotein, and triglyceride concentrations when examining LDL NMR , high-density lipoprotein NMR , and very LDL (VLDL) NMR , respectively.
Results
Only 1 of the 17 NMR lipoproteins was associated with recurrent PTB. We observed an increased odds of recurrent PTB of 1.04 (95% confidence interval, 1.01–1.08; P = .02) per nanometer increase in VLDL NMR particle size and an odds ratio of 3.00 (confidence interval, 1.40–6.43; P = .005) for the third tertile of VLDL NMR particle size compared with the first tertile.
Conclusion
In women with prior PTB, variations in midpregnancy lipoproteins were not associated with recurrent PTB overall, however the association observed with VLDL NMR particle size is suggestive that PTB may be amenable to lifestyle, nutritional, or pharmacologic interventions.
The adaptation of a mother to pregnancy culminates in the mobilization of fatty acids from maternal fat stores in response to increases in insulin resistance that peaks and plateaus at midgestation. Pregnancy-associated insulin resistance is induced by pregnancy hormones and is associated with higher fasting plasma triglycerides (TG) and lower high-density lipoprotein (HDL) concentrations. Perturbations in this response are more pronounced in obese women, are associated with increases in inflammatory markers, and are associated with adverse outcomes in pregnancy, such as gestational diabetes and preeclampsia. Moreover, these insulin resistance–induced metabolic and inflammatory changes may affect enzyme expression in maternal adipose tissue and be associated with preterm birth (PTB).
Lipoprotein perturbations are causal factors in an array of atherogenic and inflammatory diseases. Given that inflammation and vascular compromise are hypothesized as causal paths culminating in PTB, and that pregnancy profoundly alters lipid metabolism, it is reasonable to assume that lipoprotein changes might be associated with an increased likelihood of PTB. We therefore conducted this investigation to determine whether lipoprotein particle size and number (concentration) are associated with recurrent PTB using an advanced lipoprotein measure, nuclear magnetic resonance (NMR) spectroscopy, which allows one to measure lipoprotein particle size and concentration.
Materials and Methods
The data for this report are from the Eunice Kennedy Shriver National Institute of Child Health and Human Development Maternal–Fetal Medicine Units Network randomized clinical trial of omega-3 long-chain polyunsaturated fatty acid supplementation to prevent recurrent PTB. The trial was conducted at 13 network centers from January 2005 through October 2006 and recruited women who had a history of at least 1 spontaneous singleton PTB. A total of 434 women were randomized to receive daily supplementation of 1200 mg eicosapentaenoic acid (20:5n-3) and 800 mg of docosahexaenoic acid (22:6n-3); while 418 were assigned to matching placebos, beginning at 16 0 to 21 6 weeks’ gestation and continuing until 36 6 weeks’ gestation or delivery, whichever occurred first. As part of the trial, all enrolled women also received weekly injections of 17 alpha-hydroxyprogesterone caproate (250 mg). Women currently taking fish oil or omega-3 polyunsaturated fatty acid supplements were ineligible for the trial; detailed inclusion and exclusion criteria are reported elsewhere. Gestational age at delivery was available for all 852 participants. The study ( NCT00125902 at www.clinicaltrials.gov ) was approved by the institutional review boards of all participating centers and this secondary analysis was determined to be exempt from institutional review board review by the Office of Human Subjects Research, University of North Carolina at Chapel Hill, Chapel Hill, NC. All enrolled women gave written informed consent for participation in the primary study. Eligibility for this secondary analysis was restricted to participants consenting to the use of their blood for future research on prematurity and other pregnancy complications.
The current analysis is a nested case-control study in which patients who delivered ≥37 weeks’ gestation were selected as controls and matched on race/ethnicity and study center in an approximate 1:1 ratio to cases, defined as delivery <35 weeks’ gestation. Gestational age at birth was determined from the sonographically confirmed gestational age at randomization and the elapsed time from randomization to delivery.
Blood was collected at the baseline randomization visit (16 0 to 21 6 weeks’ gestation), before dispensing study drug. Subjects were not instructed to fast. Standard lipids were measured by NMR Lipoprofile autoanalyzer (Liposcience Inc, Raleigh, NC) and included total cholesterol, low-density lipoprotein (LDL) cholesterol, HDL cholesterol, and TG. We relied on a commercially available laboratory process that uses NMR technology to assess each individuals’ lipoprotein particle concentration and size, including very LDL (VLDL), LDL, and HDL classes and subclasses (NMR Lipoprofile test). NMR allows investigators to forego the high expense and labor-intensive approach of ultracentrifugation and has been used in >1000 clinical trials and cohort studies. In this study, NMR spectroscopy was used to characterize particle size (in nanometers) and number (concentration in particle nmol/L or μmol/L) of 17 lipoprotein particles from plasma. Particle size categorization was done using parameters previously delineated using this technology.
Prior to conducting the multivariable analysis, to determine whether the association between each lipid biomarker and recurrent PTB <35 weeks’ gestation was linear, we assessed the lipid biomarkers as continuous variables using a locally weighted scatterplot smoothing technique. The patients were ranked by lipid value to create 10 groups. For each group, the median lipid value was calculated along with the corresponding log (odds) of recurrent PTB. These points were plotted and fitted with 2 nonparametric smoothers with 2 separate bandwidths (0.5 and 1.0) and the linearity in the log (odds) was assessed. Where there was evidence of nonlinearity, and when assessing a lipid biomarker as a continuous variable in a logistic model, we included both linear and quadratic terms for the lipoprotein.
The association between the lipid biomarkers and recurrent PTB <35 weeks’ gestation was assessed using logistic regression, conditional on race/ethnicity (Hispanic, non-Hispanic black, non-Hispanic white) and study center and adjusting for treatment group. The following clinically relevant variables were assessed for confounding: age, number of prior preterm deliveries (1 or ≥2), smoking status, and prepregnancy body mass index (kg/m 2 ). To assess whether the association between NMR lipoproteins and recurrent PTB was independent of standard lipids, models were also adjusted for LDL, HDL, and TG when examining LDL NMR , HDL NMR , and VLDL NMR , respectively; the colinearity between these variables was first assessed. The lipid biomarkers were assessed as continuous variables, including a quadratic term when relevant, and divided into tertiles based on the distribution of the controls as an alternate approach to present associations. We also assessed whether the association between the NMR lipoproteins and recurrent PTB differed between the treatment groups by including NMR lipoprotein by treatment group interaction terms in the multivariable logistic models. The Hosmer-Lemeshow test was used to check for model fit. All analyses were 2-sided and a P value < .01 was considered statistically significant.
Results
The characteristics of the study population are shown in Table 1 . The women with recurrent PTB were more likely to have smoked during pregnancy and to have had ≥2 previous PTBs. Table 1 also provides the medians and interquartile ranges for the midpregnancy lipid biomarkers. Multivariable models controlled for race/ethnicity, center, treatment group, number of prior preterm deliveries, and smoking status; LDL NMR were also adjusted for LDL cholesterol; HDL NMR were also adjusted for HDL cholesterol; VLDL NMR were also adjusted for TG. Each multivariable logistic model examining the association between a NMR lipoprotein and recurrent PTB had good model fit as indicted by the Hosmer-Lemeshow test. None of the interactions between NMR lipoproteins and treatment group was significant.
| Characteristic | Cases (recurrent PTB <35 wk) | Controls (term birth ≥37 wk) | P value a |
|---|---|---|---|
| n | 128 | 132 | |
| Age (y), mean ± SD | 26.9 ± 5.5 | 27.1 ± 5.5 | .80 |
| Race/ethnicity, n (%) | .82 | ||
| Non-Hispanic black | 52 (40.6) | 50 (37.9) | |
| Hispanic | 15 (11.7) | 14 (10.6) | |
| Non-Hispanic white | 61 (47.7) | 68 (51.5) | |
| Prepregnancy body mass index (kg/m 2 ), mean ± SD | 26.9 ± 7.2 | 26.4 ± 6.2 | .82 |
| Smokers, n (%) | 32 (25.0) | 13 (9.9) | .002 |
| Preeclampsia or gestational hypertension, n (%) | 5 (3.9) | 4 (3.1) | .75 |
| Assigned to omega-3 group, n (%) | 61 (47.7) | 71 (53.8) | .39 |
| Prior preterm deliveries, n (%) | .001 | ||
| 1 | 78 (60.9) | 106 (80.3) | |
| ≥2 | 50 (39.1) | 26 (19.7) | |
| Total cholesterol, mg/d, median (interquartile range) | 203 (181–229) | 212 (186–238) | .12 |
| LDL measures | |||
| Standard LDL cholesterol, direct, mg/dL, median (interquartile range) | 97 (82–110) | 103 (81–120) | .21 |
| LDL NMR particle concentration, nmol/L, median (interquartile range) | |||
| Total | 1114.5 (897.0–1401.0) | 1154.0 (856.5–1365.5) | .89 |
| Large | 636.5 (497.5–811.5) | 684.0 (506.0–841.0) | .26 |
| Medium small | 70.0 (6.5–148.0) | 47.0 (9.0–124.0) | .57 |
| Small | 376.5 (26.0–767.5) | 285.0 (57.0–624.5) | .55 |
| Very small | 309.5 (23.5–613.5) | 212.0 (29.5–513.5) | .53 |
| IDL | 46.5 (15.0–81.0) | 46.5 (9.0–90.5) | .94 |
| LDL NMR average particle size, nm, median (interquartile range) | 21.9 (21.2–22.6) | 22.1 (21.5–22.6) | .34 |
| HDL measures | |||
| Standard HDL, mg/dL, median (interquartile range) | 58 (51–64) | 60 (52–68) | .11 |
| HDL NMR particle concentration, μmol/L, median (interquartile range) | |||
| Total | 32.9 (29.9–36.8) | 32.9 (29.9–36.6) | .88 |
| Large | 11.8 (9.5–13.9) | 11.6 (10.0–13.3) | .93 |
| Medium | 0.1 (0.0–2.1) | 0.0 (0.0–1.6) | .16 |
| Small | 19.4 (17.4–22.3) | 20.2 (17.0–23.0) | .73 |
| HDL NMR average particle size, nm, median (interquartile range) | 9.8 (9.5–10.0) | 9.8 (9.6–10.0) | .23 |
| VLDL measures | |||
| Standard triglycerides, mg/dL median (interquartile range) | 157 (125–214) | 148 (111–196) | .17 |
| VLDL NMR particle concentration, nmol/L median (interquartile range) | |||
| Total VLDL/chylomicrons | 70.1 (53.2–93.4) | 74.4 (47.1–101.4) | .79 |
| Large VLDL/chylomicrons | 2.3 (0.6–4.8) | 1.4 (0.5–3.4) | .04 |
| Medium | 26.6 (16.2–38.4) | 27.8 (15.1–39.8) | .83 |
| Small | 41.2 (28.4–54.2) | 42.4 (26.8–61.4) | .39 |
| VLDL NMR average particle size, nm median (interquartile range) | 50.7 (45.9–56.9) | 47.6 (43.7–53.1) | .003 |
a Continuous variables were compared using Wilcoxon test; categorical variables were compared using χ 2 or Fisher exact test.
LDL NMR
Neither particle size nor any of the LDL NMR concentrations were associated with recurrent PTB ( Table 2 ).
| Variable | Adjusted a odds ratio (95% CI) | Adjusted b odds ratio (95% CI) | |||
|---|---|---|---|---|---|
| Lipid biomarker as categorical variable | Lipid biomarker as categorical variable | Lipid biomarker as continuous variable | |||
| Tertile 2 (vs tertile 1) | Tertile 3 (vs tertile 1) | Tertile 2 (vs tertile 1) | Tertile 3 (vs tertile 1) | Per unit increase c in linear model | |
| Total cholesterol, mg/dL | 0.89 (0.48–1.67) | 0.72 (0.39–1.35) | |||
| LDL measures | |||||
| Standard LDL cholesterol, direct, mg/dL | 1.43 (0.78–2.61) | 0.68 (0.34–1.33) | |||
| LDL NMR particle concentration, nmol/L | |||||
| Total | 1.01 (0.53–1.92) | 0.77 (0.41–1.45) | 1.08 (0.51–2.30) | 0.86 (0.38–1.94) | 1.00 (1.00–1.00) |
| Large | 0.69 (0.35–1.36) | 0.75 (0.40–1.41) | 0.71 (0.35–1.43) | 0.82 (0.39–1.75) | 1.00 (1.00–1.00) |
| Medium small | 0.73 (0.37–1.43) | 1.05 (0.56–1.98) | 0.75 (0.38–1.47) | 1.10 (0.57–2.12) | 1.00 (1.00–1.00) |
| Small | 0.72 (0.37–1.39) | 1.00 (0.54–1.86) | 0.72 (0.37–1.40) | 1.05 (0.56–2.00) | 1.00 (1.00–1.00) |
| Very small | 0.70 (0.36–1.35) | 0.95 (0.51–1.77) | 0.70 (0.36–1.36) | 1.00 (0.53–1.90) | 1.00 (1.00–1.00) |
| IDL | 1.09 (0.58–2.02) | 0.84 (0.43–1.65) | 1.12 (0.60–2.12) | 0.91 (0.44–1.88) | 1.00 (0.99–1.01) |
| LDL NMR average particle size, nm | 0.70 (0.38–1.30) | 1.01 (0.52–1.97) | 0.68 (0.36–1.26) | 0.98 (0.50–1.92) | 0.94 (0.67–1.32) |
| HDL measures | |||||
| Standard HDL, mg/dL | 1.20 (0.66–2.20) | 0.79 (0.41–1.54) | |||
| HDL NMR particle concentration, μmol/L | |||||
| Total | 0.99 (0.52–1.88) | 1.31 (0.68–2.54) | 1.01 (0.53–1.91) | 1.37 (0.70–2.69) | d |
| Large | 0.91 (0.46–1.80) | 1.31 (0.70–2.45) | 1.18 (0.55–2.52) | 2.11 (0.88–5.06) | 1.10 (0.96–1.27) |
| Medium | 2.90 (1.18–7.12) | 1.65 (0.88–3.10) | 2.88 (1.17–7.11) | 1.63 (0.85–3.12) | 1.00 (0.90–1.10) |
| Small | 1.16 (0.61–2.19) | 0.92 (0.47–1.77) | 1.14 (0.60–2.16) | 0.90 (0.46–1.75) | 1.00 (0.94–1.06) |
| HDL NMR average particle size, nm | 0.81 (0.45–1.44) | 0.82 (0.41–1.64) | 0.83 (0.44–1.57) | 0.87 (0.36–2.07) | 0.86 (0.34–2.18) |
| VLDL measures | |||||
| Standard triglycerides, mg/dL | 1.44 (0.74–2.80) | 1.65 (0.83–3.31) | |||
| VLDL NMR particle concentration, nmol/L | |||||
| Total VLDL/chylomicrons | 1.92 (1.01–3.66) | 1.47 (0.72–2.99) | 1.76 (0.91–3.43) | 1.19 (0.53–2.70) | 0.99 (0.98–1.00) |
| Large VLDL/chylomicrons | 0.85 (0.43–1.68) | 1.54 (0.80–2.94) | 0.82 (0.40–1.69) | 1.41 (0.59–3.35) | 0.99 (0.87–1.12) |
| Medium | 1.25 (0.64–2.42) | 1.07 (0.53–2.16) | 0.96 (0.46–2.02) | 0.64 (0.24–1.67) | 0.98 (0.96–1.00) |
| Small | 1.21 (0.64–2.29) | 1.09 (0.56–2.13) | 1.19 (0.63–2.26) | 1.03 (0.52–2.02) | 1.00 (0.98–1.01) |
| VLDL NMR average particle size, nm | 1.75 (0.83–3.66) | 2.99 (1.48–6.03) | 1.75 (0.83–3.68) | 3.00 (1.40–6.43) | 1.04 (1.01–1.08) |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
