Abstract
Context: Preterm birth is associated with maternal cardiovascular risk, but mechanisms are unknown.
Objective: We considered that dyslipidemia may predispose women to both conditions and that prepregnancy lipids may be related to preterm birth risk. We hypothesized that low or high prepregnancy plasma lipids would be associated with preterm birth.
Design, Setting, and Participants: A total of 1010 women (49% black) enrolled in the multicenter, prospective Coronary Artery Risk Development in Young Adults study with at least one singleton birth during 20 yr of follow-up were evaluated.
Main Outcome Measure: Postbaseline preterm births less than 34 wk or 34 to less than 37 wk vs. greater than 37 wk gestation.
Results: We detected a U-shaped relationship between prepregnancy cholesterol concentrations and preterm birth risk. Women with prepregnancy cholesterol in the lowest quartile compared with the second quartile (<156 vs. 156–171 mg/dl) had an increased risk for preterm birth 34 to less than 37 wk (odds ratio 1.86; 95% confidence interval 1.10, 3.15) and less than 34 wk (odds ratio 3.04; 1.35, 6.81) independent of race, age, parity, body mass index, hypertension during pregnancy, physical activity, and years from measurement to birth. Prepregnancy cholesterol in the highest quartile (>195 mg/dl) was also associated with preterm birth less than 34 wk among women with normotensive pregnancies (odds ratio 3.80; 95% confidence interval 1.07, 7.57). There were no associations between prepregnancy triglycerides, low-density lipoprotein cholesterol, or high-density lipoprotein cholesterol and preterm birth.
Conclusions: Both low and high prepregnancy cholesterol were related to preterm birth risk. These may represent distinct pathways to the heterogeneous outcome of preterm birth. Additional studies are needed to elucidate mechanisms that link low or high cholesterol to preterm birth and later-life sequelae.
Both low and high pre-pregnancy cholesterol were related to preterm birth risk, perhaps marking two distinct pathways to the heterogeneous outcome of preterm birth.
Preterm birth (PTB) is the most common perinatal complication, affecting more than 12% of births in the United States. PTB is the leading cause of perinatal morbidity and mortality, and rates have increased for 2 decades (1). There is intriguing epidemiological evidence that women who deliver preterm infants are at increased risk later in life for cardiovascular disease (CVD). Large registry-based observational studies have reported that women who delivered a preterm infant after pregnancies with or without preeclampsia had a 2- to 11-fold higher risk for cardiovascular death compared with those who delivered at term (2,3,4).
Mechanisms that may link preterm delivery with excess maternal cardiovascular risk have not been studied, but two recent reports relating either very low or very high lipid concentrations in pregnancy to PTB risk raise the possibility that dyslipidemia may play a role in both conditions (5,6). Like CVD, preterm birth aggregates in families and women with a prior preterm birth are at high risk for recurrence (7). Thus, some risk factors likely precede pregnancy. The relationship between preconception metabolic factors such as lipids to risk of preterm birth is not well understood. The racially diverse prospective Coronary Artery Risk Development in Young Adults (CARDIA) cohort provides a unique opportunity to relate prepregnancy factors to risk of adverse pregnancy outcomes such as preterm birth.
Our objective was to relate prepregnancy plasma lipid concentrations to risk of preterm birth. Specifically we compared prepregnancy total cholesterol, high-density lipoprotein (HDL), low-density lipoprotein (LDL), and triglyceride concentrations among women with PTBs less than 34 wk and 34 to less than 37 wk with those with term births (≥37 wk). We hypothesized that women with either low or high lipids before pregnancy would be at increased risk of PTB. We also evaluated whether the relation between lipids and PTB was modified by parity, overweight status, or race and was independent of pregnancy hypertensive status.
Materials and Methods
The CARDIA study is a multicenter, longitudinal, observational study designed to describe the development of risk factors for coronary heart disease in young black and white men and women. Participants were recruited from four geographic areas: Birmingham, AL; Chicago, IL; Minneapolis, MN; and Oakland, CA. From 1985 to 1986, 5,115 subjects (2787 women; 52% black) aged 18–30 yr were enrolled at baseline and provided written informed consent. A variety of self-reported measures were obtained as well as blood specimens, and the study design, methodology and cohort characteristics have been previously described (8,9). Retention rates were 92, 86, 81, 79, 74, and 72% of the surviving cohort at yr 2, 5, 7, 10, 15, and 20 after baseline.
Of the 2787 women enrolled in CARDIA, we excluded women with diabetes mellitus or hypertension at baseline based on self-report or medication use (n = 280), those who were pregnant within 3 months of the baseline assessment (n = 19), those with hysterectomy at baseline (n = 24), those without baseline measurement of the lipids or covariates (n = 118), or those with no births between baseline and yr 20 of follow-up (n = 1200). We further excluded births with no gestational age or birth weight reported (n = 223), and 31 twin births were excluded. The final study population was comprised of 1010 women and 1626 births after baseline. Reproductive events were assessed at baseline and at each follow-up examination. Women were categorized into those who were nulliparous vs. parous at baseline. For each birth women reported the gestational age at delivery (weeks) and birth weight. Preterm birth was characterized as 34 to less than 37 wk and less than 34 wk to evaluate severity. For women with more than one birth, we characterized each birth as preterm or term. For women with all term births, we selected the first birth for analysis. For women with at least one preterm birth, we analyzed the first preterm birth because it is unknown whether first or subsequent preterm births are related to excess maternal CVD risk. In additional analysis we then limited the preterm group to those who reported that their first birth was less than 37 wk; 34 women who reported a first term birth but subsequent preterm birth were excluded from this analysis.
We conducted a validation study of maternal report of gestational age at delivery among a subset of 211 CARDIA women using medical record abstractions. All seven preterm births that occurred less than 34 wk according to hospital records were appropriately classified as such by maternal report; two births that were reported as less than 34 wk preterm were either 34 to less than 37 wk or term according the hospital record (sensitivity for early preterm birth was 100%; specificity was 99%). A total of eight of the 12 births that were delivered 34 to less than 37 wk according to the hospital record were appropriately classified via maternal report; 21 of the births reported to have occurred at 34 to less than 37 wk were delivered at term according to the medical record (sensitivity was 67% and specificity was 89% for moderate preterm births). The overall sensitivity for maternal report of ever delivering preterm (<37 wk) was 84% (16 of 19), and the specificity was 89% (170 of 192).
At CARDIA examinations, women were asked for each pregnancy whether they had developed hypertension without toxemia or toxemia. Hypertensive disorders of pregnancy were classified as either gestational hypertension (i.e. hypertension without toxemia) or preeclampsia (i.e. toxemia) based on self-report for each pregnancy. When compared with the medical record, self-reported hypertensive disorders of pregnancy was overreported (sensitivity was 40%), but the negative predictive value of no self-report of preeclampsia or gestational hypertension was 90% (10).
Study participants were asked to fast for 12 h before their clinic examination. Fasting blood samples were sent to the Northwest Lipid Research Laboratories, University of Washington (Seattle, WA) for lipid determination. The laboratory participates in the Centers for Disease Control and Prevention lipids standardization program, and the samples were analyzed continuously. Total cholesterol, triglycerides, and HDL-cholesterol were measured enzymatically within 6 wk of collection. LDL-cholesterol was calculated using the Friedewald equation (11). LDL was not calculated for participants with triglyceride levels 400 mg/dl or greater (n = 25).
Information on demographic characteristics (age, sex, and race) was obtained at yr 0 and verified at yr 2, whereas educational achievement was self-reported on standardized questionnaires during each examination. During the examination, height and weight were measured. Body mass index (BMI) was calculated as weight in kilograms divided by height in meters squared. Waist circumference was measured as the abdominal girth midway between the iliac crest and the bottom of the ribcage. Smoking was measured as self-reported current smoking status (nonsmoker, ex-smoker, and current smoker). Habitual physical activity was measured by use of the CARDIA Physical Activity History, a simplified version of the Minnesota Leisure Time Physical Activity Questionnaire (12). Three resting seated blood pressure measurements were obtained with a random-zero sphygmomanometer; the mean of the second and third readings was used for this report.
Lipids were evaluated as continuous variables and as quartiles; women with lipids in the second quartile were the referent to evaluate the possibility of a U-shaped relation between baseline lipids and preterm birth risk. Concentrations were compared according to preterm birth status using ANOVA and then adjusted for relevant covariates. Risk of preterm birth less than 34 wk and 34 to less than 37 wk was evaluated using polycotomous logistic regression, an extension of traditional logistic regression, which models more than two levels of outcome so that odds ratios (ORs) and adjusted ORs are estimated in the same model. Covariates known to be associated with either lipids or preterm birth were included: age at selected birth, parity, race, BMI, physical activity, hypertension during any reported pregnancy, and calendar year of the birth. These results were then replicated after excluding women with any pregnancies reported to be complicated by hypertension (n = 183, 18%). Results were also stratified by parity, race, and overweight status, given the relevance of each of these factors to preterm birth risk and lipid concentrations. Race-specific quartiles of lipids were used in models stratified by race.
Results
Lipids were measured in women who were, on average, 24 yr of age and 6 yr before the first analyzed birth. A total of 218 women (21.6%) reported at least one pregnancy resulting in a preterm birth. Of these, 184 were first births to women in the cohort, and 34 were subsequent births. Women with preterm births were more likely to be of black race and to have less than a high school education at baseline compared with women with term births (Table 1). They also were more likely to report pregnancies complicated by hypertension, had modestly lower prepregnancy concentrations of triglycerides, and reported on average less physical activity compared with women with term births. There were no other differences in mean lipid concentrations among women with and without later PTBs.
Table 1.
Maternal characteristics [mean (sd) or n (percent)] at baseline according to preterm birth status (n = 1010)
| Term births (n = 792) | PTB 34 to <37 wk (n = 146) | PTB <34 wk (n = 72) | P value | |
|---|---|---|---|---|
| Age (yr) | 24.2 (3.7) | 23.7 (3.8) | 24.0 (3.2) | 0.31 |
| Race (% black) | 349 (44.1) | 94 (64.4) | 54 (75.0) | <0.01 |
| Education (% high school or less) | 251 (31.7) | 67 (45.9) | 31 (43.1) | <0.01 |
| Baseline parity (% nulliparous) | 556 (70.2) | 100 (68.5) | 48 (66.7) | 0.78 |
| Oral contraceptive use (% current) | 182 (23.0) | 36 (24.7) | 13 (18.1) | 0.84 |
| Smoker (% current) | 204 (25.8) | 33 (22.6) | 24 (33.3) | 0.45 |
| GDM (ever)a | 91 (11.5) | 20 (13.7) | 11 (15.3) | 0.52 |
| Family history of diabetes (% ever)a | 220 (27.8) | 39 (26.7) | 21 (29.2) | 0.93 |
| Hypertension during pregnancy (% ever)a | 119 (15.0) | 43 (29.5) | 21 (29.2) | <0.01 |
| Age at first birth (yr)a | 30.1 (4.8) | 29.0 (5.1) | 30.0 (5.0) | 0.05 |
| Time interval from baseline to selected birth (yr)a | 5.9 (4.1) | 5.3 (4.1) | 6.0 (4.1) | 0.28 |
| BMI (kg/m2) | 23.4 (4.8) | 23.6 (4.9) | 24.0 (5.3) | 0.55 |
| Waist circumference (cm) | 71.8 (9.4) | 71.8 (10.0) | 72.0 (10.9) | 0.99 |
| Systolic BP (mm Hg) | 104.9 (8.8) | 106.4 (8.5) | 106.0 (8.6) | 0.11 |
| Diastolic BP (mm Hg) | 65.4 (8.3) | 65.7 (8.8) | 65.1 (8.4) | 0.84 |
| Fasting plasma | ||||
| Total cholesterol (mg/dl) | 176.9 (31.4) | 177.0 (32.9) | 176.1 (36.6) | 0.98 |
| HDL-cholesterol (mg/dl) | 55.7 (12.4) | 57.7 (13.1) | 56.6 (13.6) | 0.19 |
| Triglycerides (mg/dl) | 65.9 (31.2) | 62.6 (35.5) | 57.8 (22.8) | 0.07 |
| LDL-cholesterol (mg/dl) | 108.0 (29.6) | 106.7 (29.0) | 107.9 (34.3) | 0.89 |
| Glucose (mg/dl) | 79.3 (7.5) | 78.8 (7.5) | 79.9 (9.0) | 0.59 |
| Physical activity (exercise units)b | 304.0 (326.0) | 228.5 (271.0) | 264.0 (274.0) | <0.01 |
GDM, Gestational diabetes mellitus; BP, blood pressure.
At baseline and during follow-up.
CARDIA did not query duration of bouts of physical activity; therefore, the unit of physical activity is called exercise units, which increases with frequency and intensity of performance. Physical activity reported as median (interquartile range).
We detected a U-shaped relationship between prepregnancy cholesterol concentrations and PTB risk after adjustment for covariates. Women with prepregnancy total cholesterol concentrations in the lowest quartile (94–155 mg/dl) had an increased risk of PTB 34 to less than 37 wk [OR 1.86; 95% confidence interval (CI) 1.10, 3.15] and less than 34 wk (OR 3.04, 95% CI 1.35, 6.81) compared with women with cholesterol in the second quartile (Table 2). This association was attenuated for PTBs that occurred at 34 to less than 37 wk but strengthened for preterm birth less than 34 wk after excluding women with a history of hypertensive disorders of pregnancy (n = 183). In addition, prepregnancy total cholesterol in the highest quartile (>195 mg/dl) was associated with PTB less than 34 wk among women with normotensive pregnancies (OR 3.80, 95% CI 1.21, 11.96). There was no association between prepregnancy LDL cholesterol, triglycerides, or HDL cholesterol and preterm birth risk. Results were largely consistent after excluding 34 women who reported that their first birth was term, but a subsequent birth was preterm. Among this group, women with normotensive pregnancies who had either prepregnancy low or high total cholesterol had excess risk for PTB less than 34 wk (OR 3.16, 95% CI 1.11, 9.02; OR 2.69, 95% CI 0.91, 7.91, respectively).
Table 2.
ORs for quartiles of plasma lipids according to PTB status; referent are women with term births (n = 792)
| Crude analysis
|
Adjusted analysisa
|
Excluding women with hypertension in pregnancya
|
||||
|---|---|---|---|---|---|---|
| Prepregnancy lipid concentration | 34 to <37 wk (n = 146) | <34 wk (n = 72) | 34 to <37 wk (n = 146) | <34 wk (n = 72) | 34 to <37 wk (n = 103) | <34 wk (n = 51) |
| Total cholesterol (mg/dl) | ||||||
| Quartile 4 (196–318) | 1.48 (0.88, 2.51) | 2.43 (1.08, 5.46) | 1.38 (0.80, 2.36) | 2.21 (0.97, 5.05) | 1.55 (0.82, 2.93) | 3.80 (1.21, 11.96) |
| Quartile 3 (173–195) | 1.36 (0.80, 2.32) | 2.24 (0.99, 5.06) | 1.30 (0.75, 2.23) | 2.11 (0.92, 4.85) | 1.57 (0.84, 2.94) | 3.53 (1.12, 11.18) |
| Quartile 2 (156–172) | Referent | Referent | Referent | Referent | Referent | Referent |
| Quartile 1 (94–155) | 1.83 (1.09, 3.06) | 3.10 (1.41, 6.85) | 1.86 (1.10, 3.15) | 3.04 (1.35, 6.81) | 1.66 (0.89, 3.09) | 5.02 (1.66, 15.16) |
| LDL cholesterol (mg/dl) | ||||||
| Quartile 4 (124–233) | 1.21 (0.72, 2.02) | 0.95 (0.47, 1.94) | 1.07 (0.63, 1.83) | 0.86 (0.41, 1.78) | 1.30 (0.70, 2.41) | 0.96 (0.39, 2.36) |
| Quartile 3 (105–123) | 1.02 (0.60, 1.73) | 1.13 (0.57, 2.21) | 0.96 (0.56, 1.66) | 1.03 (0.52, 2.07) | 1.05 (0.56, 1.98) | 1.34 (0.59, 3.02) |
| Quartile 2 (89–104) | Referent | Referent | Referent | Referent | Referent | Referent |
| Quartile 1 (26–88) | 1.50 (0.91, 2.48) | 1.16 (0.58, 2.30) | 1.60 (0.96, 2.68) | 1.17 (0.58, 2.37) | 1.40 (0.77, 2.54) | 1.16 (0.50, 2.67) |
| Triglycerides (mg/dl) | ||||||
| Quartile 4 (78–318) | 0.88 (0.53, 1.46) | 0.81 (0.39, 1.69) | 0.99 (0.58, 1.67) | 1.03 (0.48, 2.20) | 1.02 (0.53, 1.94) | 0.68 (0.27, 1.71) |
| Quartile 3 (58–77) | 0.99 (0.60, 1.62) | 1.10 (0.56, 2.18) | 1.05 (0.63, 1.75) | 1.22 (0.61, 2.47) | 1.41 (0.77, 2.58) | 1.09 (0.49, 2.42) |
| Quartile 2 (44–57) | Referent | Referent | Referent | Referent | Referent | Referent |
| Quartile 1 (16–43) | 1.16 (0.71, 1.89) | 1.42 (0.73, 2.76) | 1.13 (0.68, 1.88) | 1.30 (0.66, 2.58) | 1.44 (0.78, 2.67) | 1.19 (0.54, 2.60) |
| HDL cholesterol (mg/dl) | ||||||
| Quartile 4 (64–118) | 1.39 (0.86, 2.24) | 1.27 (0.66, 2.45) | 1.52 (0.92, 2.49) | 1.36 (0.69, 2.70) | 1.32 (0.76, 2.30) | 1.29 (0.58, 2.89) |
| Quartile 3 (55–63) | 0.90 (0.54, 1.52) | 0.93 (0.46, 1.88) | 0.95 (0.56, 1.62) | 0.99 (0.48, 2.04) | 0.72 (0.39, 1.36) | 1.00 (0.42, 2.36) |
| Quartile 2 (47–54 ) | Referent | Referent | Referent | Referent | Referent | Referent |
| Quartile 1 (25–46) | 1.06 (0.64, 1.77) | 1.09 (0.55, 2.19) | 1.05 (0.62, 1.79) | 1.31 (0.64, 2.69) | 0.98 (0.53, 1.82) | 1.62 (0.70, 3.75) |
Adjusted for race, parity, BMI, physical activity at baseline, age at selected birth, ever gestational hypertension or preeclampsia during follow-up, time interval from baseline measurement to selected birth.
When stratified by parity status at baseline, nulliparous women with prepregnancy total cholesterol concentrations in the lowest quartile had increased risk for PTB 34 to less than 37 wk and less than 34 wk compared with women with concentrations in the second quartile (Table 3). Low prepregnancy concentrations among parous women were associated with particularly high risk of PTB less than 34 wk (OR 5.62, 95% CI 1.13, 27.96). High prepregnancy cholesterol among parous women was associated with increased risk of PTB 34 to less than 37 wk (OR 2.92, 95% CI 1.04, 8.19). Parous women with HDL cholesterol in the highest quartile had excess risk of PTB 34 to less than 37 wk. The precision of race-specific results was compromised; however, the trend of low or high prepregnancy cholesterol associated with PTB did not appear to be limited to one ethnic group (data not shown). In addition, the relation between prepregnancy lipids and PTB did not differ by overweight status (results not shown).
Table 3.
ORs for quartiles of plasma lipids related to PTB, stratified by baseline parity statusa
| Nulliparous at baseline (n = 704)b
|
Parous at baseline (n = 306)b
|
|||
|---|---|---|---|---|
| 34 to <37 wk (n = 100) | <34 wk (n = 48) | 34 to <37 wk (n = 46) | <34 wk (n = 24) | |
| Total cholesterol | ||||
| 4 | 0.93 (0.48, 1.81) | 2.05 (0.78, 5.34) | 2.92 (1.04, 8.19) | 2.64 (0.49, 14.19) |
| 3 | 1.11 (0.59, 2.09) | 1.76 (0.66, 4.65) | 2.07 (0.69, 6.22) | 2.92 (0.53, 16.06) |
| 2 | Referent | Referent | Referent | Referent |
| 1 | 1.90 (1.04, 3.48) | 2.35 (0.90, 6.16) | 1.92 (0.64, 5.74) | 5.62 (1.13, 27.96) |
| LDL cholesterol | ||||
| 4 | 0.87 (0.44, 1.73) | 0.79 (0.33, 1.85) | 1.44 (0.59, 3.50) | 0.92 (0.21, 4.09) |
| 3 | 0.97 (0.50, 1.88) | 0.78 (0.33, 1.81) | 1.03 (0.39,2.74) | 1.92 (0.51, 7.29) |
| 2 | Referent | Referent | Referent | Referent |
| 1 | 1.95 (1.06, 3.58) | 0.81 (0.34, 1.94) | 0.89 (0.31, 2.55) | 2.85 (0.75, 10.86) |
| Triglycerides | ||||
| 4 | 1.11 (0.58, 2.13) | 0.81 (0.33, 2.00) | 0.67 (0.27, 1.68) | 1.91 (0.35, 10.62) |
| 3 | 1.22 (0.65, 2.28) | 0.98 (0.43, 2.23) | 0.68 (0.28, 1.69) | 2.51 (0.48, 13.03) |
| 2 | Referent | Referent | Referent | Referent |
| 1 | 1.34 (0.74, 2.44) | 0.79 (0.35, 1.79) | 0.79 (0.28, 2.21) | 4.66 (0.92, 23.64) |
| HDL cholesterol | ||||
| 4 | 1.04 (0.58, 1.85) | 1.17 (0.48, 2.83) | 4.41 (1.56, 12.48) | 1.85 (0.60, 5.73) |
| 3 | 0.76 (0.41, 1.39) | 1.23 (0.52, 2.88) | 1.76 (0.55, 5.60) | 0.50 (0.10, 2.64) |
| 2 | Referent | Referent | Referent | Referent |
| 1 | 0.64 (0.32, 1.26) | 1.37 (0.55, 3.43) | 2.53 (0.93, 6.90) | 1.03 (0.31, 3.41) |
Referents are women with term births (n = 556 for nulliparas; n = 236 for multiparas).
Adjusted for race, BMI, physical activity at baseline, age at selected birth, ever gestational hypertension or preeclampsia during follow-up, time interval from baseline measurement to selected birth.
Discussion
Our results suggest there is a relation between both low and high prepregnancy cholesterol and increased risk of preterm birth. This relationship was detected, regardless of parity, race, and BMI. These results are consistent with reports of low or high cholesterol assessed during pregnancy among women destined to deliver preterm (6). Our data provide intriguing evidence that subtle but likely meaningful differences in prepregnancy metabolic factors such as lipids are related to pregnancy outcomes. If these aberrations persist postpartum, dyslipidemia could be a marker for an underlying metabolic disturbance that links PTB to later-life maternal cardiovascular disease risk.
Low prepregnancy total cholesterol and PTB risk is consistent with the report by Edison et al. (6) that low cholesterol in midgestation is a risk factor for PTB. Circulating concentrations of cholesterol among nonpregnant adults appear to be dominantly determined by diet (14,15), although the profound increases in lipids during pregnancy appear less affected by diet (16). Low gestational LDL cholesterol concentrations across pregnancy among women with growth restriction have been documented (17,18). How prepregnancy total cholesterol concentrations may affect adaptations required for normal pregnancy or perhaps be a marker that is affected by other processes warrants further study. We previously reported that the expected gestational triglyceride increase appears blunted in the first half of pregnancy among women destined to deliver preterm (5), but how these adaptations might further vary according to prepregnancy concentrations to our knowledge has not been directly studied.
Women in our study with high prepregnancy cholesterol were also more likely to deliver preterm. This is consistent with reports that higher cholesterol and free fatty acids assessed in early or late gestation were related to excess PTBs (5,6,19). Although there are no guidelines for normal cholesterol concentrations among young adult women, the women in our study with prepregnancy cholesterol in the highest quartile approached concentrations deemed clinically relevant among older nonpregnant adults. Thus, a proatherogenic phenotype may be a risk factor for preterm birth among normotensive women. Other atherogenic factors measured in pregnancy such as inflammation (20,21), thrombin activation, (22,23,24), and glucose (25) have also been related to PTB risk, and it is possible that these factors may also be different before pregnancy among women with subsequent PTBs. This possibility warrants further study. Our data support the notion that pregnancy might be a cardiometabolic stress test such that adverse reproductive events during young adulthood could unmask a subclinical predisposition to later life cardiovascular risk (26).
To our knowledge there are no other studies relating prepregnancy factors to PTB. Two studies related prepregnancy blood pressure and lipids to preeclampsia and fetal growth, both from a Norwegian cohort followed up for 10 yr. These reports indicated that prepregnancy serum concentrations of triglycerides and total, LDL, and non-HDL cholesterol were associated with increased risk for preeclampsia (27). In addition, maternal prepregnancy blood pressure was negatively associated with infant birth weight, whereas high prepregnancy lipids were positively associated with infant birth weight (28). Neither of these studies reported on PTB as an outcome, and therefore, our study provides intriguing and novel data suggesting that prepregnancy metabolic factors may be importantly related to this common perinatal complication. There is evidence that maternal cholesterol metabolizing polymorphisms are associated with PTB (29), providing support for the possibility that the prepregnancy phenotype may be important in the pathogenesis of PTB. In addition, our results indicate that prepregnancy dyslipidemia was associated with increased risk for PTBs less than 34 wk, a group with particularly high risk of infant morbidity and mortality.
Our study was not designed to examine biological mechanisms that may link either high or low prepregnancy lipids to PTB. We can therefore only speculate on these possibilities. PTB is known to be a complex, heterogeneous condition, and low and high prepregnancy cholesterol likely represents two distinct pathways leading to early parturition. Prepregnancy lipids have been related to infant birth weight (28) and likely provide important substrate for fetal growth. Deficient prepregnancy cholesterol may trigger inadequate fetal growth (17) that initiates a cascade resulting in PTB among some women. Alternatively, low prepregnancy cholesterol may be a marker of poor nutritional status that has been shown to be related to PTB (30,31,32). In addition, poor nutrition may enhance susceptibility to infection that is a known contributor to the pathogenesis of PTB (33). The evidence that treatment for bacterial vaginosis is effective only at reducing PTB risk among underweight women supports this possibility (34). It is known that cholesterol is reduced in an acute-phase immune response (35), perhaps mediated by circulating cytokines (36). Jacobs et al. (37) reported that plasma total cholesterol was also lower among CARDIA participants that reported minor illness in the 24 h before blood draw. It is possible that women in our study with low cholesterol were more likely to have chronic infections or low-grade inflammation that increased their risk for subsequent pregnancy complications. The only marker of immune response at baseline in CARDIA was white blood cell count (WBC). Although WBC counts did not differ according to postbaseline PTB risk, women and men in our study with low cholesterol did have modestly higher WBC counts at baseline (data not shown).
In contrast, excess prepregnancy cholesterol may be a marker of an atherogenic profile that predisposes women to inadequate placental perfusion that results in PTB, even in the absence of hypertension. The evidence that a third of spontaneous PTBs are associated with placental vascular insufficiency lesions supports this possibility (38,39). Future studies that relate gestational concentrations of lipids to evidence of vascular or infectious placental pathology may help further elucidate these pathways. The fact that low or high prepregnancy cholesterol, but not LDL-cholesterol or triglycerides, was related to subsequent PTBs is puzzling. Future studies that evaluate how lipid metabolism adaptations during gestation might be related to PTB and postpartum concentrations are needed to better understand these findings.
Our results should be considered in light of the study limitations. Pregnancy data were collected retrospectively in CARDIA via maternal recall and therefore misclassification of PTB is possible. Others have reported that maternal recall of delivery of preterm infants is accurate, especially for recent births (13,40). Births were assessed in CARDIA every 2–5 yr, and therefore, pregnancy characteristics were recalled proximal to the event. Our validation study indicated that the reporting of preterm births less than 34 wk was remarkably precise, providing assurance that our results for these births are robust. Preterm births 34 to less than 37 wk were more likely to be misclassified, although the overall sensitivity and specificity for this group was adequate. Our data regarding hypertension in pregnancy were less robust. The high specificity of this exposure reporting, however, ensures that our results among normotensive women are unlikely to include cases of preeclampsia or gestational hypertension. We were unable to distinguish clinical presentation of preterm births (medically indicated, preterm labor, or preterm premature rupture of membranes), and this will be important in future studies. Women in CARDIA were, on average, older at their first birth compared with women in the United States, and this may compromise the generalizability of our results. Strengths of our study include the availability of fasting prepregnancy lipid results and the excellent retention of this racially diverse cohort (72% at yr 20).
Our results indicate that both low and high prepregnancy cholesterol are related to preterm birth risk, including early preterm births delivered less than 34 wk gestation. These may represent distinct pathways to the heterogeneous outcome of preterm birth. Additional studies are needed to elucidate mechanisms that link low or high cholesterol to preterm birth as well as later life maternal and infant sequelae.
Footnotes
This work was supported by contracts N01-HC-48047, N01-HC-48048, N01-HC-48049, N01-HC-48050, and N01-HC-95095 from the National Heart, Lung, and Blood Institute. J.M.C. is supported by Grant K12-HD043441.
Disclosure Summary: J.M.C., R.B.N., M.F.W., D.R.J., J.M.R., and E.P.G. have nothing to declare.
First Published Online May 25, 2010
Abbreviations: BMI, Body mass index; CARDIA, Coronary Artery Risk Development in Young Adults; CI, confidence interval; CVD, cardiovascular disease; HDL, high-density lipoprotein; LDL, low-density lipoprotein; OR, odds ratio; PTB, preterm birth; WBC, white blood cell count.
References
- Institute of Medicine 2006 Preterm birth: causes, consequences, and prevention. Washington, DC: National Academies Press [Google Scholar]
- Irgens HU, Reisaeter L, Irgens LM, Lie RT 2001 Long term mortality of mothers and fathers after pre-eclampsia. BMJ 323:1213–1217 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Smith GC, Pell JP, Walsh D 2001 Pregnancy complications and maternal risk of ischaemic heart disease: a retrospective cohort study of 129,290 births. Lancet 357:2002–2006 [DOI] [PubMed] [Google Scholar]
- Smith GD, Sterne J, Tynelius P, Lawlor DA, Rasmussen F 2005 Birth weight of offspring and subsequent cardiovascular mortality of the parents. Epidemiology 16:563–569 [DOI] [PubMed] [Google Scholar]
- Catov JM, Bodnar L, Kip KE, Hubel CA, Ness RB, Harger G, Roberts JM 2007 Early pregnancy lipid concentrations and spontaneous preterm birth. Am J Obstet Gynecol 197:610.e1–610.e7 [DOI] [PubMed] [Google Scholar]
- Edison RJ, Berg K, Remaley A, Kelley R, Rotimi C, Stevenson RE, Muenke M 2007 Adverse birth outcome among mothers with low serum cholesterol. Pediatrics 120:723–733 [DOI] [PubMed] [Google Scholar]
- Wilcox AJ, Skjaerven R, Lie RT 2008 Familial patterns of preterm delivery: maternal and fetal contributions. Am J Epidemiol 167:474–479 [DOI] [PubMed] [Google Scholar]
- Cutter GR, Burke GL, Dyer AR, Friedman GD, Hilner JE, Hughes GH, Hulley SB, Jacobs Jr DR, Lie K, Manolio TA 1991 Cardiovascular risk factors in young adults. The CARDIA baseline monograph. Control Clin Trials 12:1S–77S [DOI] [PubMed] [Google Scholar]
- Friedman GD, Cutter GR, Donahue RP, Hughes GH, Hulley SB, Jacobs Jr DR, Liu K, Savage PJ 1988 CARDIA: study design, recruitment, and some characteristics of the examined subjects. J Clin Epidemiol 41:1105–1116 [DOI] [PubMed] [Google Scholar]
- Gunderson EP, Chiang V, Lewis CE, Catov J, Quesenberry Jr CP, Sidney S, Wei GS, Ness R 2008 Long-term blood pressure changes measured from before to after pregnancy relative to nonparous women. Obstet Gynecol 112:1294–1302 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Friedewald WT, Levy RI, Fredrickson DS 1972 Estimation of the concentration of low-density lipoprotein cholesterol in plasma, without use of the preparative ultracentrifuge. Clin Chem 18 499–502 [PubMed] [Google Scholar]
- Jacobs D, Hahn L, Haskell W, Pirie P, Sidney S 1989 Validity and reliability of short physical activity history: CARDIA and the Minnesota Heart Health Program. J Cardiopulm Rehabil 9:448–459 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Tomeo CA, Rich-Edwards JW, Michels KB, Berkey CS, Hunter DJ, Frazier AL, Willett WC, Buka SL 1999 Reproducibility and validity of maternal recall of pregnancy-related events. Epidemiology 10:774–777 [PubMed] [Google Scholar]
- Connor WE, Cerqueira MT, Connor RW, Wallace RB, Malinow MR, Casdorph HR 1978 The plasma lipids, lipoproteins, and diet of the Tarahumara Indians of Mexico. Am J Clin Nutr 31:1131–1142 [DOI] [PubMed] [Google Scholar]
- McMurry MP, Connor WE, Goplerud CP 1981 The effects of dietary cholesterol upon the hypercholesterolemia of pregnancy. Metabolism 30:869–879 [DOI] [PubMed] [Google Scholar]
- Hansen AE, Wiese HF, Adam DJ, Boelsche AN, Haggard ME, Davis H, Newsom WT, Pesut L 1964 Influence of diet on blood serum lipids in pregnant women and newborn infants. Am J Clin Nutr 15:11–19 [DOI] [PubMed] [Google Scholar]
- Sattar N, Greer IA, Galloway PJ, Packard CJ, Shepherd J, Kelly T, Mathers A 1999 Lipid and lipoprotein concentrations in pregnancies complicated by intrauterine growth restriction. J Clin Endocrinol Metab 84:128–130 [DOI] [PubMed] [Google Scholar]
- Catov JM, Patrick TE, Powers RW, Ness RB, Harger G, Roberts JM 2007 Maternal leptin across pregnancy in women with small-for-gestational-age infants. Am J Obstet Gynecol 196:558.e1–558.e8 [DOI] [PubMed] [Google Scholar]
- Chen X, Scholl TO 2008 Association of elevated free fatty acids during late pregnancy with preterm delivery. Obstet Gynecol 112:297–303 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Catov JM, Bodnar LM, Ness RB, Barron SJ, Roberts JM 2007 Inflammation and dyslipidemia related to risk of spontaneous preterm birth. Am J Epidemiol 166:1312–1319 [DOI] [PubMed] [Google Scholar]
- Pitiphat W, Gillman MW, Joshipura KJ, Williams PL, Douglass CW, Rich-Edwards JW 2005 Plasma C-reactive protein in early pregnancy and preterm delivery. Am J Epidemiol 162:1108–1113 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Elovitz MA, Baron J, Phillippe M, Elovitz MA, Baron J, Phillippe M 2001 The role of thrombin in preterm parturition. Am J Obstet Gynecol 185:1059–1063 [DOI] [PubMed] [Google Scholar]
- Rosen T, Kuczynski E, O'Neill LM, Funai EF, Lockwood CJ, Rosen T, Kuczynski E, O'Neill LM, Funai EF, Lockwood CJ 2001 Plasma levels of thrombin-antithrombin complexes predict preterm premature rupture of the fetal membranes. J Matern Fetal Med 10:297–300 [DOI] [PubMed] [Google Scholar]
- Catov JM, Bodnar LM, Hackney D, Roberts JM, Simhan HN 2008 Activation of the fibrinolytic cascade early in pregnancy among women with spontaneous preterm birth. 112:1116–1122 [DOI] [PMC free article] [PubMed] [Google Scholar]
- HAPO Study Cooperative Research Group, Metzger BE, Lowe LP, Dyer AR, Trimble ER, Chaovarindr U, Coustan DR, Hadden DR, McCance DR, Hod M, McIntyre HD, Oats JJ, Persson B, Rogers MS, Sacks D2008 Hyperglycemia and adverse pregnancy outcomes. 358:1991–2002 [Google Scholar]
- Sattar N, Greer IA 2002 Pregnancy complications and maternal cardiovascular risk: opportunities for intervention and screening? BMJ 325:157–160 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Magnussen EB, Vatten LJ, Lund-Nilsen TI, Salvesen KA, Davey Smith G, Romundstad PR 2007 Prepregnancy cardiovascular risk factors as predictors of pre-eclampsia: population based cohort study. BMJ 335:978 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Romundstad PR, Davey Smith G, Nilsen TI, Vatten LJ 2007 Associations of prepregnancy cardiovascular risk factors with the offspring’s birth weight. Am J Epidemiol 166:1359–1364 [DOI] [PubMed] [Google Scholar]
- Steffen KM, Cooper ME, Shi M, Caprau D, Simhan HN, Dagle JM, Marazita ML, Murray JC 2007 Maternal and fetal variation in genes of cholesterol metabolism is associated with preterm delivery. J Perinatol 27:672–680 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Catov JM, Bodnar LM, Ness RB, Marcovic N, Roberts JM 2007 Periconceptional multivitamin use is associated with reduced risk of preterm or small for gestational age births. Am J Epidemiol 166:296–303 [DOI] [PubMed] [Google Scholar]
- Scholl TO, Hediger ML, Schall JI, Fischer RL, Khoo CS 1993 Low zinc intake during pregnancy: its association with preterm and very preterm delivery. Am J Epidemiol 137:1115–1124 [DOI] [PubMed] [Google Scholar]
- Siega-Riz AM, Promislow JH, Savitz DA, Thorp Jr JM, McDonald T 2003 Vitamin C intake and the risk of preterm delivery. Am J Obstet Gynecol 189:519–525 [DOI] [PubMed] [Google Scholar]
- Goldenberg RL, Hauth JC, Andrews WW 2000 Mechanisms of disease: intrauterine infection and preterm delivery. N Engl J Med 342:1500–1507 [DOI] [PubMed] [Google Scholar]
- Hauth JC, Goldenberg RL, Andrews WW, DuBard MB, Copper RL 1995 Reduced incidence of preterm delivery with metronidazole and erythromycin in women with bacterial vaginosis. N Engl J Med 333:1732–1736 [DOI] [PubMed] [Google Scholar]
- Feingold KR, Funk JL, Moser AH, Shigenaga JK, Rapp JH, Grunfeld C 1995 Role for circulating lipoproteins in protection from endotoxin toxicity. Infect Immun 63:2041–2046 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Harris H, Grunfeld C, Feingold K, Rapp J 1990 Human very low density lipoproteins and chylomicrons can protect against endotoxin-induced death in mice. J Clin Invest 86:696–702 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Jacobs Jr DR, Hebert B, Schreiner PJ, Sidney S, Iribarren C, Hulley S 1997 Reduced cholesterol is associated with recent minor illness: the CARDIA Study. Am J Epidemiol 146:558–564 [DOI] [PubMed] [Google Scholar]
- Arias F, Rodriquez L, Rayne SC, Kraus FT 1993 Maternal Placental vasculopathy and infection: two distinct subgroups among patients with preterm labor and preterm ruptured membranes. Am J Obstet Gynecol 168:585–591 [DOI] [PubMed] [Google Scholar]
- Germain AM, Carvajal J, Sanchez M, Valenzuela GJ, Tsunekawa H, Chuaqui B 1999 Preterm labor: placental pathology and clinical correlation. Obstet Gynecol 94:284–289 [DOI] [PubMed] [Google Scholar]
- Seidman DS, Slater PE, Ever-Hadani P, Gale R 1987 Accuracy of mothers’ recall of birthweight and gestational age. Br J Obstet Gynaecol 94:731–735 [DOI] [PubMed] [Google Scholar]