Maternal Periconceptional Folic Acid Supplementation and the Risk of Non-Syndromic Oral Clefts in Offspring: A Population-Based Case-Control Study in Chengdu, China

Maternal periconceptional folic acid supplementation (FAS) has been documented to be associated with decreased risk of nonsyndromic oral clefts (NsOC). However, the results remain inconclusive. In this population-based case-control study of 807 singletons affected by NsOC and 8070 healthy neonates who were born between October 2010 and September 2015 in Chengdu, China, we examined the association of maternal FAS with the risk of nonsyndromic cleft lip with or without cleft palate (NsCL/P), and cleft palate (NsCP). Unconditional logistic regression analysis was used to estimate the crude and adjusted odds ratios (ORs) and 95% condential intervals (CI). Signicant associations were found between maternal periconceptional FAS and decreased risk of NsCL/P (aOR = 0.41, 95% CI: 0.33–0.51). This protective effect was also detected for NsCL (aOR = 0.42, 95% CI: 0.30–0.58) and NsCLP (aOR = 0.41, 95% CI: 0.31–0.54). Both maternal FAS started before and after the last menstrual period (LMP) were negatively associated with NsCL/P (before LMP, aOR = 0.43, 95% CI: 0.33–0.56; after LMP, aOR = 0.41, 95% CI: 0.33–0.51). The association between NsCP and maternal FAS initiating before LMP was signicant (aOR = 0.52, 95% CI: 0.30–0.90), but the statistical power seemed weak due to limited number of NsCP cases. The ndings suggest that maternal periconceptional FAS can reduce the risk of each subtype of NsCL/P in offspring, while the potential effect on NsCP needs further investigations.


Introduction
Oral clefts (OC), resulting from the incomplete fusion of primary palate and/or secondary palate during the 7th to 10th week of pregnancy, usually include syndromic or nonsyndromic cleft lip with or without cleft palate (CL/P) and cleft palate (CP). 1 Depending on geographic regions and races, oral clefts affect 1 in 500 to 2500 births. 1 The nonsyndromic clefts (NsOC) are one of most common birth defects in China. 2 The reported prevalence of NsOC can be as high as 1.8/1000 in some provinces of China, causing huge disease burden on family and society. [2][3][4] Maternal folic acid supplementation (FAS) before and during early pregnancy is a well-established prevention approach for neural tube defects (NTDs). As facial mesenchyme is derived from neural crest, the protective effect of maternal FAS on oral clefts has been studied in various populations, but the ndings are inconclusive. [5][6][7][8] Most observational studies conducted in Europe, [9][10][11][12][13][14] America, [15][16][17] and Asia [18][19][20][21] , have shown that maternal FAS during early pregnancy is associated with decreased risk of nonsyndromic CL/P (NsCL/P), while other investigations in these areas did not. [22][23][24][25][26][27] Some of them revealed a preventive effect for nonsyndromic cleft palate (NsCP), 9,13−15 , whereas a few studies suggested maternal FAS or multivitamin intake as risk factor of NsCL/P and NsCP. 28 Though two recent meta-analyses support the potential preventive effect of maternal FAS against NsCL/P and NsCP, 7,8 the currently available evidence is insu cient and con icting, particularly for some cleft subtypes. [29][30][31] In fact, population-based evidence from China is relatively scarce.
With the exception of one cohort in three provinces, 18 a population-based case-control study in Shenyang city, 21 and a propensity-matched study in Shaanxi province, 26 almost all the observational studies in China have been hospital-based. [32][33][34][35][36] Chengdu, the capital of Sichuan province, is a megacity located in the southwest of China where the prevalence of OCs is higher than the national average. 3 Following the Guide of National Folic Acid Supplementation Program, 2 Chengdu has provided free folic acid (400 µg/day) for women of childbearing age who reside in the area with a plan to be pregnancy since 2010. The information regarding maternal folic use, prenatal exposures, perinatal health care and birth outcomes are prospectively collected by the Chengdu Maternal and Infant Health Surveillance system (CMIHS). These data allowed us to examine additional effect of maternal FAS on NsOC subtypes by performing a population-based case-control study.

Results
General characteristics of the study subjects During October 2010 to September 2015, a total of 807 singletons with nonsyndromic clefts from CMIHS were available for the current analysis, including 247 NsCLs, 369 NsCLPs, and 191 NsCPs. Table 1 shows the maternal and infantile characteristics of the cases and 8070 controls. Compared with the control mothers, more NsCL/P mothers were under the age of 25 years or older than 35 years of age at the time of delivery. NsCL/P mothers were less educated (≤ 9 years), or more overweighted (BMI ≥ 24.0) than control mothers. Much more NsCL/P mothers exposed to environmental risks, and had a family history than control mothers. Male predominance in NsCL/P and female excess in NsCP were identi ed. No signi cant difference was found between cases and controls regarding the distribution of parity, and maternal medical conditions in the rst trimester.  Signi cant differences were found between controls, the overall and each subtype of nonsyndromic oral clefts.

Association of maternal FAS and NsOC
As shown in Table 3  Note: cOR = crude odds ratio; aOR = adjusted odds ratio, adjusted for maternal age, maternal education, urban-rural classi cation, residence, parity, medical condition and environmental exposure in the rst trimester, maternal BMI, and infant sex.

Discussion
In this population-based case-control study, we demonstrated that maternal periconceptional FAS was associated with a reduced risk of overall NsCL/P, and the reduced risk varied by cleft subtype and supplementation initiation timing. We did not identify the preventive effect of periconceptional FAS for NsCP, whereas we observed a signi cant association between maternal FAS started before LMP and NsCP, suggesting that earlier or longer supplementation may be protective.
Consistent with previous studies in some western [9][10][11][12][13][14][15][16][17]  reported a preventive effect of optimal FAS for overall birth defects (OR = 0.71, 95% CI: 0.57-0.89), but not for oral clefts. 26 The authors thought that non-randomized maternal FA use, small sample size and supplementation initiation timing could explain differences studies. 18,26 A recent meta-analysis revealed the preventive effect of maternal FAS in early pregnancy against NsCL/P and NsCP, and identi ed publication bias in previously published researches. 7 However, ndings from randomised or quasi-randomised trials seemed not to support the preventive effect of daily FAS on oral clefts. 31 It can be seen that inconsistences among various studies are obvious. In fact, there is still a lack of reliable evidence on the preventive effect of folic acid on oral clefts, especially on the dose and initiation timing of supplementation.
Nonsyndromic cleft palate has been regarded as a distinct condition from NsCL/P on embryologic origin and etiology. As noted in previous investigation, 6,11,18,21,32 24 On the contrary, a case-control study conducted in Northern Netherlands 27 and another cohort study 28 in Japan identi ed maternal FAS/multivitamin supplement use during the rst trimester as a risk factor of NsCP.
Whether the preventive effect of maternal FAS against NsCP can be observed in a certain population depends on various factors, including genetic background, maternal dietary folic acid intake, serum folate level and the compliance with FAS, etc. Considering the small number of NsCP cases in our analysis, the ndings about maternal FAS and NsCP need to be further studied. These controversial results may be due to the heterogeneities in research design, population, sample size, exposure assessment, and other potential confounders. 7,18,21,32−36 Overall, more prospective studies are needed to elucidate the relationship between maternal FAS and NsCP.
Several strengths and limitations should be considered when interpreting our study results. The population-based nature and large sample size of Chinese cleft cases of the study ensure robust OR estimates for NsCL/P subtypes. We got the statistical power of 0.82 for NsCL and 0.98 for NsCLP. In addition, the prospectively collected exposure data could minimize recall bias. Though information of maternal dietary and multivitamin intake was not available for analysis, the associations are less likely to be distorted because the OR estimates were based on randomly selected control data with adjustment for known risk factors such as maternal illness, social factors, environmental exposures, and maternal illness.
In conclusion, our study provides additional evidence that maternal FAS during periconceptional period can reduce the risk of NsCL/P. Larger-sample-size studies are warranted to elucidate the association with NsCP and to determine whether women can bene t more from supplementation starting before LMP. It is of paramount importance for women of childbearing age to become aware of that maternal FAS can not only reduce the risk of NTD, but also reduce the risk of NsCL/P.

Data source and study subjects
Data for this study were abstracted from the CMIHS. According to the "Maternal and Infant Health Care Protocol of Sichuan Province", 37 every pregnant woman was required to have at least ve prenatal medical examinations during her pregnancy and three postnatal clinical visits (on the 7th, 28th and 42nd day after delivery). The results of examinations or visits, risk factors (maternal diseases, family history, other medical conditions), and pregnancy outcomes (spontaneous abortion, stillbirth, live birth, birth defects, etc.), were recorded in the CMIHS system. The medical examination, data collecting, checking and auditing were performed by well-trained obstetricians and nurses. CMIHS adopted the diagnosis criterions, case ascertainment, quality assurance, data collection and encoding of birth defects proposed by Chinese Birth Defects Monitoring Network. 38 Between October 2010 and September 2015, 807 singletons with NsOCs were included in as the cases, for each case ten controls were randomly selected from the healthy singletons born in the same period.

Folic Acid Intake And Covariates
Brie y, women who joined in the prevention program were followed once a month by local community healthcare workers, and their registration dates, last menstrual period (LMP), dates of starting and ending use, and the information of folic acid supplementation were recorded. In the CMIHS system, a woman's FAS information was linked to her medical records once she had her rst prenatal exam in any of the local hospitals.
In this study, pregnant women who regularly took folic acid during the periconceptional period for at least one month were de ned as "periconception users". Of them, those who started intake before their LMP were termed as "preconception users", while those who started on or after their LMP were named as "post-conception users". On the contrary, pregnant women who did not have folic acid intake, or took it continuously less than one month were considered as "non-users".
Other variables included maternal age, nationality, education, residence, parity, medical condition and environmental exposures in the rst trimester. Speci cally, the maternal medical condition referred to any of such conditions as positive result of syphilis, human immunode ciency virus and hepatitis B virus testing, anemia, chronic kidney, liver and heart diseases, diabetes (type or ), primary hypertension disorders. The environmental exposure was an indicator (yes or no) to show whether the mothers had smoking, alcohol drinking, drug abuse, or exposure to radiation and hazardous substances.
The variables of infants included gestational age, date of birth, birth weight, sex, birth defects and infant outcomes. Data used in this study were anonymous and included no individually identi able information. This research was approved by the Medical Ethics Committee of the West China Second University Hospital, Sichuan University. All methods were performed in accordance with relevant regulations and the individual informed consent was waived because the study used non-identi ed data.

Statistical Analyses
Differences in maternal and infant characteristics between cases and controls were examined with independent sample ttests for continuous variables and with Pearson chi-square tests for categorical variables. Non-conditional logistic regression analysis was used to calculate the crude and adjusted odds ratios (ORs) and 95% con dence intervals (CI).
Statistical analyses were performed with R 3.5.3 (R Development Core Team 2019) and packages "rms". Declarations