Objective
We previously reported that elevated antiinflammatory cervical cytokines in early pregnancy were associated with spontaneous preterm birth. Our objective was to explore the relation between serum folate vitamers and the lower genital tract inflammatory milieu.
Study Design
Pregnant women (n = 417) at <16 weeks’ gestation had serum samples that were analyzed for folate species 5-methyltetrahydrofolate, 5-formyltetrahydrofolate, and cervical fluid that was assayed for cytokine concentrations. Patterns in proinflammatory cytokines (interleukin [IL]-1β, -6, -8, and -10; monocyte chemotactic protein-1) and antiinflammatory cytokines (IL-4, IL-10, IL-13) were identified with factor analysis.
Results
After confounder adjustment, maternal serum 5-methyltetrahydrofolate concentrations had a strong negative association with elevated antiinflammatory scores; serum 5-formyltetrahydrofolate concentrations were associated positively with elevated antiinflammatory scores (both P < .05). Maternal folate was not associated with proinflammatory scores.
Conclusion
Maternal serum folate vitamers are associated with cervical cytokine concentrations, which suggests a possible mechanistic link between folate and preterm birth risk.
A relation between folate deficiency and prematurity was suggested originally in 1944 by Callender, who reported an increased incidence of prematurity in women with megaloblastic anemia in pregnancy. Since that time, a number of studies have explored whether maternal folate status is linked with preterm birth. More recent investigations of dietary, supplemental, or biomarker folate in relation to preterm birth have produced conflicting results that were likely due to varying ranges and timing of folate assessment, assessment methods of gestational age, and population characteristics. The 3 primary folate species that comprise total folate in serum are 5-methyltetrahydrofolate (5MeTHF), 5-formyltetrahydrofolate (5FoTHF), and folic acid. These different folate species contribute to the range of biologic effects of folate.
We recently demonstrated that the relative concentrations of folate species in maternal circulation in early pregnancy might be more critical than total folate in the determination of the risk of preterm birth. We found a significant interaction between serum 5MeTHF and 5FoTHF concentrations and risk of preterm birth ( P = .01). When serum 5MeTHF concentrations are low, there is a positive linear relation between 5FoTHF and the risk of preterm birth. When 5MeTHF concentrations are high, there is a strong negative relation between 5FoTHF and preterm birth.
In 1970, Baumslag et al wrote that “The mechanism whereby folic acid reduces the incidence of prematurity is unknown.” In the last 40 years, few explorations of the mechanisms underlying this association have been undertaken. We present our study that explored the relation between serum folate and the lower genital tract inflammatory milieu as a possible link between folate and spontaneous preterm birth.
The inflammatory milieu of the cervix has been hypothesized to play a role in contributing to the risk of preterm birth. Both pro- and antiinflammatory components of the cervical immune system may contribute to the risk of subsequent intrauterine infection and preterm birth. In the current work, we hypothesized that the lower genital tract inflammatory milieu in early pregnancy is a biologic link between folate and preterm birth risk.
Materials and Methods
The Study of Nutrition and Pregnancy is an ongoing prospective cohort study of pregnant women who seek care at Magee-Womens Hospital resident antepartum clinics in Pittsburgh, PA. The antepartum clinics serve a predominantly publically insured, low-income population that is approximately 55% black and 44% white. Eligible women were <16 weeks’ gestation, were non-Hispanic white or non-Hispanic black (based on self-report), and had singleton pregnancies with no known preexisting conditions, vaginal bleeding, fetal anomalies, or current or planned cervical cerclage. All women provided informed, written consent. The study was approved by the University of Pittsburgh Institutional Review Board.
At enrollment, which occurred at a mean of 9.5 weeks gestation, women underwent a standard pelvic examination and provided a nonfasting blood sample that was banked for later analysis. Women also completed an interviewer-administered questionnaire to collect data on sociodemographic characteristics; on medical, reproductive, and sexual history; and on maternal behaviors.
Quantification of folate vitamers
Nonfasting serum samples were allowed to coagulate at room temperature for 30 minutes. Subsequently, tubes were kept on ice and covered with foil to protect from light. Within 2 hours of the blood draw, tubes were centrifuged, and serum was aliquoted into amber vials. Vials were stored at –80°F until they were transported on dry ice in 2009 to the local laboratory of Dr Raman Venkataramanan. No samples were thawed before assay for folate vitamers. The 3 primary folate vitamers (5MeTHF, 5FoTHF, and folic acid) were quantified with high-performance liquid chromatography-tandem mass spectrometry on the basis of a published method, the details of which were described previously. The mean percent deviations were <15% for each concentration level of all calibration curves including the lower limit of quantification, 1.8 ng/mL. The intra- and interassay coefficient of variation was <10%. The method was also accurate within the 15% acceptable limits at all levels that were tested.
Serum total folate concentrations were calculated by summing the concentrations of 5MeTHF, 5FoTHF, and free folic acid. Folate vitamers were categorized into tertiles for analysis.
Lower genital tract inflammation
During a speculum examination at enrollment, a Dacron swab was placed in the cervix and left there for 10 seconds to achieve saturation. The swab was then placed in a plastic tube that contained 400 μL of phosphate-buffered saline solution (final dilution of 1:5), immediately transported to the laboratory, and stored at –20°C. The buffer did not contain any other additives, including protease inhibitors. For analysis, the specimens were thawed at room temperature. The swab and the remaining diluent were centrifuged in a spin-X centrifuge filter unit (Costar, Cambridge, MA) at 12,000 rpm for 20 minutes. Cytokine concentrations of interleukin (IL)-1β, -6, -8, -4, -10, and -13 and monocyte chemotactic protein (MCP)-1 were measured with a multiplex system (Luminex LabMAP; Luminex Corporation, Austin, TX) and a BeadLyte bead kit (Upstate Biotechnology, Lake Placid, NY). Each assay was run with an intra- and interassay variation of <10%. For all analytes, the lower limit of sensitivity of the assay was 3.5 pg/mL. The percentages of women with cytokine concentrations below the detection limit were 12% (IL-1β), 7% (IL-4), 2% (IL-6), 0% (IL-8), 14% (IL-10), 18% (IL-13), and 0.2% (MCP-1). Women with IL-4, -10, and -13 below the detection limit (n = 5) or IL-1β and -6 below the detection limit (n = 2) were not included in the analysis because we believed that the inability to detect the concentration of multiple cytokines reflected an unusable sample. For the remaining observations with a value below the detection limit, we imputed the limit of detection divided by the square root of 2.
Cervical cytokine factor analysis
A factor analysis was performed on cytokines with a principal factor method, as described in detail previously. Factor analysis was used to discover the underlying structure of cytokines in the lower genital tract immunologic milieu. Briefly, the number of factors that were extracted was based on the eigenvalues, examination of the scree plot, and the interpretability of the solution. Two factors were identified: factor 1, which is highly loaded on proinflammatory/immunomodulatory cytokines (IL-1β, -6, -8, and -10 and MCP-1); and factor 2, which is highly loaded on antiinflammatory cytokines (IL-4, -10, and -13). Each subject has a score for factors 1 and 2, which are continuous variables. Factor scores were categorized into thirds on the basis of tertiles of the distribution and are dichotomized as elevated (highest tertile) or not elevated (lower 2 tertiles).
Covariates
Women self-reported their race/ethnicity as non-Hispanic white or non-Hispanic black. Data on maternal age, smoking status in the 3 months before pregnancy, maternal education, marital status, parity, household yearly income, and employment status were ascertained from the baseline interviewer-administered questionnaire. Prepregnancy body mass index (kilograms per square meter) was based on maternal self-report of pregravid weight and measured height at enrollment. Gestational age was based on best obstetric estimate (a reliable, self-reported estimate of last menstrual period or an ultrasound scan early in pregnancy).
Women were also classified as having a sexually transmitted infection (STI; Trichomonas vaginalis, Chlamydia trachomatis or Neisseria gonorrhoeae ) based on specimens that were obtained from the enrollment pelvic examination. Also at the enrollment visit, vaginal swabs were collected for culture and identification of vaginal flora. Bacterial vaginosis was diagnosed by vaginal pH ≥ 4.7 and a score of 7-10 from a Gram-stained vaginal smear that was interpreted with the method of Nugent et al. Intermediate flora was defined as having a Nugent score of 4-6; normal flora was defined as having a Nugent score of 0-3.
Statistical analysis
Pearson chi-squared tests were used to test for differences in the distribution of maternal characteristics and tertiles of folate vitamers. We used Spearman rank correlation coefficients to test the association between folate vitamer concentrations and factor 1 and 2 scores. To assess the independent association between maternal folate status and elevated factor 1 and 2 scores, we used multivariable Poisson regression to estimate prevalence ratios (using the “log” link). Prevalence ratio was selected instead of an odds ratio because this was a cross-sectional analysis, and elevated factor scores were common outcomes. Potential confounders were race/ethnicity, smoking status before pregnancy, maternal age, marital status, education, income, employment status, pregravid body mass index, gestational age at enrollment, presence of STI, and vaginal flora. We removed confounders if their inclusion did not satisfy our a priori change-in-estimate criterion (>8% change in a prevalence ratio), yet none of these factors met our definition of confounding. We maintained maternal race/ethnicity and STI in the model out of convention. We tested for interaction between the 5MeTHF and 5FoTHF using a likelihood ratio test (α = .10).
Results
From 2003-2007, 548 eligible women enrolled in the study (75% response rate). Of these, we excluded 31 women who did not have cervical fluid available for cytokine assay and 100 women who did not have an adequate volume of banked serum to perform the folate assays. A total of 417 women were included in the final analysis. There were no meaningful differences in maternal race, age, parity, smoking, education, or the prevalence of preterm birth between women who were included and excluded from the analysis (data not shown). Most women in the cohort were 20-29 years old, multiparous, high-school educated, unemployed, and low-income ( Table 1 ). Non-Hispanic black women made up 56% of the cohort. More than one-half of women smoked. Bacterial vaginosis was common (45%), and 11% of women had an STI.
