Folic acid supplementation in early second trimester and the risk of preeclampsia

Results 

A total of 4024 women were approached to participate in the study; 3134 (78%) agreed and were recruited. Among them, 70 women were excluded because of twin births and 113 women were excluded because of missing information such as gestational age at delivery, birthweight, or sex (59) and lost to follow-up because of the participant’s relocation outside the study center (54), leaving 2951 subjects for final analysis.

Ninety-two percent of the study subjects were taking folic acid supplementation in the early second trimester, most by taking multivitamins containing folic acid at a dose of 1.0 mg or higher (Table 1). More than half of the women initiated supplementation before conception and approximately 20% discontinued supplementation in the third trimester partly or completely (Table 1).

TABLE 1. Folic acid supplementation in pregnant women of the OaK Birth Cohort Study, October 2002-December 2005a

		Number	Percent
	Category of supplementation
	Multivitamins containing folic acid	2016	74.30
	Folic acid alone	250	9.22
	Two or more types of vitamins	447	16.48
	Dose of folic acid (mg)
	0.1–0.9	114	4.20
	1.0	2152	79.32
	1.1–1.9	178	6.56
	≥2.0	269	9.92
	Initiated before conception	1682	62.00
	Initiated after conception	1031	38.00
	Discontinued in third trimester (partly or completely)	544	20.10

Wen. Folic acid supplementation and the risk of preeclampsia. Am J Obstet Gynecol 2008.

a Doses of folic acid were derived from vitamin brands; doses of other vitamins in multivitamins were not presented in the table; the ranges (medians) of doses for these vitamins were: vitamin A, 1000-4000 IU (1800 IU); vitamin C, 60-150 mg (95.83 mg); vitamin B1, 1.4 to 3.0 mg (2.23 mg); vitamin B2, 1.4 to 3.75 mg (2.78 mg); vitamin B3, 10-40 mg (25 mg); vitamin B5, 0.03 to 10.0 mg (5.02 mg); vitamin B6, 1.9 to 10 mg (4.98 mg); vitamin B12, 2.6 to 14 μg (9.77 μg); vitamin D, 200-400 IU (360 IU); vitamin E, 25-30 IU (25.58 IU); beta-carotene,1050-3000 IU (2270 IU); biotin, 30-45 μg (36 μg); niacin, 15-20 mg (18.25 mg); pantothenic acid, 6-10 mg (9 mg). Minerals were: calcium carbonate, 175-450 mg (250 mg); copper, 1-2 mg (1.6 mg); chromium, 25-30 μg (27 μg); iodine, 0.15 to 0.22 mg (0.16 mg); iron, 10-30 mg (22 mg); lutein, 250-300 μg (275 μg); magnesium, 50-100 mg (60 mg); manganese, 2-50 mg (16 mg); molybdenum, 25-50 μg (38 μg); phosphorous, 125 mg (125 mg); selenium, 25-30 μg (28 μg); zinc, 7.5-25 mg (16.88 mg).

The majority of the study participants were white with high socioeconomic status. Women with no supplementation were more likely to be younger, multiparous, non-whites; to have lower education level and household income; and to smoke cigarettes during pregnancy than women with supplementation (Table 2).

TABLE 2. Demographic and clinical characteristic of the pregnant women who participated in the OaK Birth Cohort Study, October 2002-December 2005a

	Variables	Overall	No supplementation	Supplementation	P value
	Number of subjects	2951	238	2713
	Maternal age (y)
	<25	6.40	13.45	5.79	< .0001
	25-29	21.82	23.11	21.71
	30-34	40.09	36.55	40.40
	≥35	31.68	26.89	32.10
	Maternal background
	Aboriginal	0.75	0.42	0.77	< .0001
	White	83.46	74.38	84.26
	Middle Eastern	5.29	7.56	5.09
	African	3.42	10.08	2.84
	Asian	7.08	7.56	7.04
	Prepregnancy body mass index (kg/m2)
	<18.5 (underweight)	6.03	5.46	6.08	.1492
	18.5-24 (normal)	60.01	59.24	60.08
	25-29 (overweight)	20.40	17.23	20.68
	≥30 (obesity)	13.55	18.07	13.16
	Education level
	High school and below	13.15	27.97	11.85	< .0001
	College/university not completed	9.34	10.17	9.27
	College/university completed	77.51	61.86	78.88
	Nulliparous	34.46	23.53	35.42	.0002
	Household income ($,CAD)
	<25,000	6.76	18.50	5.75	< .0001
	25,000-49,999	14.07	22.47	13.35
	50,000-79,999	27.76	27.31	27.80
	≥80,000	51.41	31.72	53.10
	Smoking during pregnancy	7.73	13.03	7.27	.0014
	Chronic hypertension	1.58	2.14	1.53	.4761
	Type 1 diabetes	0.82	1.29	0.78	.4151
	Type 2 diabetes	0.86	0.86	0.86	.9990
	Previous preeclampsia	3.96	4.62	3.91	.5879
	Gestational age at recruitment (wks)
	≤12	33.82	27.31	34.39	.2079
	13-15	45.21	49.58	44.82
	16-20	20.98	23.11	20.79

Wen. Folic acid supplementation and the risk of preeclampsia. Am J Obstet Gynecol 2008.

a Because of missing information in some variables: smoking (2), chronic hypertension (39), type 1 diabetes (40), type 2 diabetes (40), numbers in this table did not add up.

Folic acid supplementation was associated with increased serum folate and decreased plasma homocysteine (Table 3). The association between supplementation and serum folate and plasma homocysteine was stronger in blood samples taken at later gestation and (especially) in participants with MTHFR thermolabile variant genes (Table 3).

TABLE 3. Serum folate and plasma homocysteine levels in subjects with folic acid supplementation vs those without, by gestational age at recruitment and MTHFR genotype, OaK Birth Cohort Study, October 2002-December 2005

	Variables	No supplementation	Supplementation	P value
	Folate level (mean ± SD, μmol/L)a	25.79±8.60	36.30±7.99	<.0001
	Folate level stratified by gestational weeks at recruitment (mean ± SD, μmol/L)
	≤12	25.93±8.78	37.00±7.71	<.0001
	13-15	26.42±8.69	36.63±7.82	<.0001
	16-20	24.27±8.15	34.35±8.52	<.0001
	Folate level stratified by MTHFR genotype (mean±SD, μmol/L)b
	677CC	26.93±8.47	36.24±8.16	<.0001
	677CT	24.77±7.96	36.25±7.79	<.0001
	677TT	24.29±10.43	36.63±8.14	<.0001
	Homocysteine level (mean ± SD, μmol/L)c	4.69±1.24	4.32±1.01	<.0001
	Homocysteine level stratified by gestational weeks at recruitment (mean ± SD, μmol/L)
	≤12	4.61±1.02	4.43±1.02	.1868
	13-15	4.85±1.36	4.36±1.02	.0002
	16-20	4.43±1.13	4.04±0.94	.0043
	Homocysteine level stratified by MTHFR genotype (mean ± SD, μmol/L)b
	677CC	4.58±1.07	4.28±0.99	.0019
	677CT	4.59±1.15	4.31±1.04	.0199
	677TT	5.43±1.83	4.48±0.96	.0112

Wen. Folic acid supplementation and the risk of preeclampsia. Am J Obstet Gynecol 2008.

a Thirty-five subjects missing information on serum folate level. b Forty-four subjects missing information on genotype c 36 subjects missing information on serum homocysteine

The rate of preeclampsia was lower in the supplementation group than in the no supplementation group, with an aOR of 0.37 and 95% CIs 0.18-0.75 (Table 4). Women with supplementation of folic acid alone also had a lower rate of preeclampsia than those women who had no supplementation, although the difference was not statistically significant (Table 4). No significant associations between low serum folate level, hyperhomocystinemia, and MTHFR thermolabile variant genes with risk of preeclampsia were found (Table 4). Whether women initiated supplementation before or after conception or whether women discontinued supplementation in the third trimester, the rate of preeclampsia was similarly lower than those women who did not have any supplementation (Table 4).

TABLE 4. Occurrences of pregnancy complications according to folic acid supplementation status, serum folate level, plasma homocysteine level, and MTHFR genotype, OaK Birth Cohort Study, October 2002-December 2005

	Variables	Number of subjects	Preeclampsia	ORs (95% CIa)
	Folic acid supplementation
	No	238	12(5.04)	Reference
	Yes	2713	59(2.17)	0.37(0.18-0.75)
	Supplementation of folic acid alone
	No supplementation	238	12(5.04)	Reference
	Yes	421	12(2.85)	0.46(0.16-1.31)
	Initiation time of supplementation
	Before conception	1031	24(2.33)	Reference
	After conception	1682	35(2.08)	1.04(0.59-1.82)
	Discontinuation of supplementation in late gestation
	Yes	544	14(2.57)	Reference
	No	2169	45(2.07)	0.77(0.41-1.46)
	Serum folate levelb
	Normal (≥10th percentile for each gestational week)	2649	64(2.42)	Reference
	Low (<10th percentile for each gestational week)	302	7(2.32)	1.22(0.52-2.90)
	Serum homocysteine level (μmol/L)c
	Normal (<90th percentile for each gestational week)	2529	56(2.21)	Reference
	High (≥90th percentile for each gestational week)	422	15(3.55)	1.25(0.66-2.36)
	Genotyped
	677CC	1282	29(2.26)	Reference
	677CT	1286	35(2.72)	1.27(0.75-2.15)
	677TT	339	6(1.77)	0.70(0.28-1.77)

OR, odds ratio. Wen. Folic acid supplementation and the risk of preeclampsia. Am J Obstet Gynecol 2008.

a Adjusted for maternal age, ethnic background, education level, parity, history of preeclampsia, chronic hypertension, diabetes, prepregnancy body mass index, household income, gestational age at recruitment, and cigarette smoking. b Thirty-five subjects missing information on serum folate level. c Thirty-six subjects missing information on serum homocysteine. d Forty-four subjects missing information on genotype.

Comment 

Our prospective cohort study in a cohort of Canadian women found that 92% had supplementation with folic acid or multivitamins containing folic acid in the early second trimester, and among them, most (>95%) had a supplementation of 1.0 mg or higher, twice the recommended level for the prevention of neural tube defects by Health Canada.12

Supplementation of multivitamins containing folic acid was associated with increased serum folate, lowered plasma homocysteine, and reduced risk of preeclampsia (by 63%). In our study, women with no supplementation were of lower socioeconomic status, which may lead to increased risk of preeclampsia.1 On the other hand, women with no supplementation were more likely to be younger and multiparous and to smoke cigarettes during pregnancy, which may lead to a decreased risk of preeclampsia.1, 13 Thus, the potential confounding effects of maternal age, parity, socioeconomic status, and cigarette smoking may have canceled each other and may have limited impact on the observed association between supplementation and preeclampsia. Our study did not observe an association between MTHFR gene mutation and preeclampsia, which is consistent with the study by Powers et al.14 We speculate that high-level supplementation of folic acid in our study population may have suppressed the genetic effect on preeclampsia.

A previous study reported an association between supplementation of multivitamins containing folic acid and reduced risk of preeclampsia.5 Bodnar et al5 conducted a prospective cohort study in 1835 women in Pittsburgh, PA, between 1997-2001 and found that regular use of multivitamins containing folic acid at less than 16 weeks’ gestation was associated with a 45% reduction in preeclampsia risk, compared with nonusers (OR, 0.55; 95% CI, 0.32-0.95). This finding was consistent with ours. However, a previous study on this issue has several weaknesses. The Pittsburgh study did not conduct tests for genetics or blood folate and homocysteine levels.5 Our study has the largest sample size with detailed information on supplementation patterns such as type, duration, and dose of supplementation. We managed to measure several aspects of folate metabolism, including genetics, supplementation, and blood folate and homocysteine levels simultaneously. We used a prospective cohort study design and blindly adjudicated the outcome, which helped to minimize bias.

There are compelling biologic rationale to believe that folic acid may reduce the risk of developing preeclampsia. Preeclampsia is likely a 2-stage disorder: at stage I (most likely at late first trimester or early second trimester), a decreased placental perfusion, secondary to abnormal placental developments, develops; and at stage II (most likely at the early third trimester), the maternal syndrome of preeclampsia, secondary to systemic endothelial dysfunction, develops.4

Supplementation of large doses of folic acid in early gestation may work at both stages of preeclampsia development. Folic acid, or folate, is one of the B vitamins. It is a coenzyme in the production of nucleic acids and therefore is required by all cells for growth. The placenta develops from a single cell to a complex entity with a weight of about 500 g during pregnancy. An adequate cellular folate supply may play an important role in the implantation and development of the placenta. Folate may also reduce the risk of developing preeclampsia by improving endothelial function at both placental and systemic levels, directly or indirectly by its effect on lowering blood homocysteine level.7, 15, 16, 17

Endothelial dysfunction is demonstrable within the myometrial arteries of women with preeclampsia, and the incubation of healthy vessels within plasma obtained from women with preeclampsia induces similar endothelial changes.6 Young women with folate deficiency or hyperhomocystinemia may be prone to not only systemic endothelial dysfunction but also placental endovasculature.9 In our study, whether women initiated folic acid supplementation before or after conception or whether the women discontinued supplementation in the third trimester, the rate of preeclampsia was similarly lower than in those women who did not have supplementation, suggesting that for preeclampsia prevention, folic acid supplementation in the late first trimester or early second trimester, the most critical time window for preeclampsia development, may be the most important.

We did not observe an association between serum folate or plasma homocysteine level with preeclampsia. In large-scale epidemiologic studies, it is difficult to tightly control for factors that may affect the measured values of folate or homocysteine. Variations in gestational age and timing at which blood samples were taken, folate consumption by the growing fetus, and other physical and pathological processes that may have an impact on the metabolism of folate/homocysteine during pregnancy may introduce variability in the measurement of folate/homocysteine.18

It is rather difficult to distinguish cause and effect for biomarkers that are the intermediate steps of a health problem occurring in a rather short period of time, such as preeclampsia. Most previous studies that observed an association between serum folate and plasma homocysteine levels with preeclampsia collected blood sample in late gestation or at delivery.19, 20 For those studies that collected blood samples in the early second trimester, the prediction of hyperhomocystinemia on preeclampsia was poor.19, 2, 22

These findings suggest that variation by gestational period is a major obstacle in the study of the cause-and-effect relationship between a particular metabolite and a particular pregnancy outcome. Increasing study samples to precisely measure gestation period specific values in future studies may help determine the association between metabolites and pregnancy outcomes. Using biomarkers with longer half-life biomarkers (eg, red blood cell folate) may be helpful as well. These markers measure not only the current status but also the previous status. Because the development of specific health problem is often a process rather than a point effect, biomarkers of accumulation measures may be better measures of the cause-and-effect relationship than biomarkers of point measures.

Most women in our study had supplementation of multivitamins containing folic acid. There are several reasons for us to believe that folic acid may have played a more important role in preeclampsia than other vitamins. First, there is strong biologic rationale to believe that folic acid may reduce the risk of developing preeclampsia. No similar biologic mechanisms could be found for other vitamins. Second, recent randomized controlled trials found that supplementation with vitamins C and E (without folic acid) during pregnancy at doses many times higher than the ones in our study had no protective effect on preeclampsia.23, 24 Third, in the subgroup of women who had supplementation of folic acid alone (n = 421), we did observe a statistically nonsignificant trend toward similar protective effect on preeclampsia. However, because the effect of supplementation by folic acid alone (odds ratio, 0.46) was smaller than supplementation dominated by multivitamins containing folic acid (odds ratio, 0.37), other vitamins (eg, vitamin B6) may also play some role in the prevention of preeclampsia.

Our study has several weaknesses. The sample sizes for women with no supplementation or with supplementation at the doses of <1.0 mg or >1.0 mg folic acid or with folic acid alone were small. As a result, we could not find statistically significant association of supplementation with folic acid alone or perform a dose-response analysis. Although we adjusted for several obstetric, socioeconomic, and lifestyle factors in our analysis, residual confounding may exist. We did not collect data on folic acid intake from food. Women with folic acid supplementation were of a higher socioeconomic status and therefore may have higher folic acid from food intake. Ray and Mamdani25 studied hospital discharge data in the Canadian province of Ontario before (1990-1997) and after (1998-2000) mandatory folic acid food fortification and found no change in the rate of preeclampsia, suggesting that folic acid from food intake was too low to make any impact on the risk of preeclampsia.

The findings of our study and others give hope of a new prevention strategy for preeclampsia, which needs further evaluation. Randomized, controlled trials could provide definitive evidence regarding the relationship between folic acid supplementation and the risk of preeclampsia. Findings from observational studies such as the current one can help the design of future randomized trials (eg, when the supplementation should be initiated and what dose should be used, etc) and to establish the equipoise for future randomized trials.21