Mother-fetus pairs were recruited from 2019 to 2020 at the first hospital of Jilin University, Changchun, China. This study adhered to the guidelines laid down in the Declaration of Helsinki, and all procedures involving human subjects were approved by the Chinese Clinical Trial Registry (ChiCTR2000034179). Written informed consent was obtained from all subjects. Maternal and infantile demographics and physiology characteristics were obtained through questionnaire and hospital medical records. In brief, 103 dyads of healthy pregnancy women were recruited at the last time visit to obstetric clinic before delivery and classified according to pre-BMI, namely, normal weight (NW, BMI = 18.5–23.99 kg/m², n = 60), overweight/obese (OW, BMI ≥ 24 kg/m², n = 26), and underweight (UW, BMI < 18.5 kg/m², n = 17). The inclusion criteria were maternal age ≥20 years and ≤40 years, singleton pregnancy, gestation duration ≥37 weeks, and newborn health (Apgar>8). Women were excluded if they were suffering from metabolic diseases (e.g., prepregnancy and gestational diabetes mellitus), HIV-infected disease, pulmonary tuberculosis, other acute infectious diseases, severe heart disease and renal disease, and were taking drugs that affect nutrient metabolism, and infants with congenital and hereditary diseases were excluded. The information on maternal age, pre-BMI, antenatal BMI, gestational weight gain, gestation duration, infant anthropometry, and sex was investigated by questionnaire and medical record.

A valid food frequency questionnaire (FFQ) was carried out by the investigators who were trained uniformly. We gave each participant a face-to-face interview and a semistructured FFQ. Based on 7 different food groups (e.g., meat, eggs, poultry, fish and seafood, fruits, vegetables, milk, etc.,) and 54 different food categories combined with food pictures and standard food mold for food weight evaluation, the questionnaire was used to assess the dietary intake of enrolled subjects during the last trimester. It included specific questions about the cooking styles, cooking oil types, and dosages sources containing docosahexaenoic acid (DHA, C22:6n-3), such as freshwater fish, seafood, and canned tuna. The intake and frequency of food categories per day/week/month were recorded. To better understand the data of DHA intake, in the interview, we also asked participants about the DHA supplement's brand and daily doses. The investigators checked the content of DHA in supplements and calculated the daily doses of DHA. The questionnaire was improved and used by our group to assess the dietary intake, especially fatty acids (23). The intake of energy and five kinds of PUFA dietary intakes were calculated according to the food composition table (24). Finally, 79 questionnaires were collected, 24 subjects did not complete the questionnaire well (OW = 5, NW = 12, UW = 7).

A total of 5 ml maternal blood samples after an overnight fast was collected either on the morning of admission for surgery in case of primary cesarean sections or at the last visit to the obstetric clinic, no longer than 3 days before delivery. Then, 5 ml cord blood samples were collected at delivery by a maternity nurse. All blood samples were collected in EDTA tubes and centrifuged at 3,500 rpm for 15 min to separate plasma, and stored at −80°C until fatty acids analysis.

A direct methylation procedure was performed on 100 μl of plasma. Then, 100 μl plasma, 100 μl C17:0 internal standard solution (5 mg/ml), and 600 μl methanol were mixed and vortexed for 30 s, and centrifuged for 5 min at 900 × g. The methanol phase was taken to another glass centrifuge tube, mixed with 25 μl sodium methoxide solution, and the solution was mixed for 3 min at room temperature. Also, 75 μl methanol hydrochloride solution was added to terminate reaction. Then, 300 μl n-hexane was added, and the mingled solution was mixed for 30 s to extract fatty acid methyl esters (FAMEs), then the upper n-hexane phase was transferred to a new glass centrifuge tube. The extraction was blown to dry by nitrogen, and 50 μl n-hexane containing 2 g/L butylated hydroxytoluene was added to dissolve the residue for gas chromatography (GC) analysis. The concentrations of 36 plasma fatty acids (μg/ml) were determined in relation to the peak area of internal standard. Each plasma fatty acid was expressed as a percentage of the total 36 fatty acid concentrations measured.

Plasma FAMEs were detected by using GC-2010Plus gas chromatography (Shimadzu Corp., Kyoto, Japan). The GC was equipped with a SP-2560 capillary column (100 m × 0.25 mm × 0.20 μm; Supelco, Bellefonte, PA). The chromatographic conditions were, namely, high-purity nitrogen was the carrier gas (linear velocity: 1 ml/min), split ratio was 1:50, and the injection volume was 1 μl; the initial temperature of the column box was set at 140°C, held for 5 min, and then the temperature rose to 260°C at the rate of 4°C /min, held for 20 min; the temperature of flame ionization detector was set at 280°C; the flow rate of hydrogen was 40 ml/min, and airflow rate was 500 ml/min. Shimadzu lab solutions chromatography workstation software was used to record the chromatogram, retention time, and peak area. The fatty acid concentrations were calculated by comparing the peak area of internal standard. The levels of plasma fatty acids are expressed as the percentage of total fatty acid (%).

Linoleic acid (LA; C18:2n-6) is the precursor of n-6 PUFA, whereas α-linolenic acid (ALA; C18: 3n-3) is the precursor of n-3 PUFA (25). The process by which PUFA are synthesized from LA and ALA in the human body involves Δ6 desaturase (D6D, LA to γ-linolenic acid, ALA to stearidonic acid) and Δ5 desaturase [D5D, dihomo-γ-linolenic acid (DGLA) to arachidonic acid (AA), eicosatetraene acid to eicosapentaenoic acid (EPA)]. Product-to-precursor ratios have been used to represent enzyme indices (26–28). In this study, the D6D index was estimated by DGLA (20:3n-6)/LA (18:2n-6), the D5D index was estimated by AA (20:4n-6)/DGLA (20:3n-6), and the ratio for estimation of the elongase activity was C18:1/C16:1 (29).

The mean and standard error of mean (SEM) were used for normal variables, and quartile was used to describe continuous variables of non-normally distribution. The minimum sample size was calculated according to relevant literature (30). One-way ANOVA and Kruskal-Wallis H test were used for the analysis of anthropometric parameters and fatty acids. Chi-square test was used for nonquantitative variables. Pearson and Spearman were used to analyze the association between maternal parameters and infant parameters. Multivariable linear regression was used to analyze the associations between maternal parameters and infant DHA. In linear regression model 1, all significant maternal factors were included, and maternal age and gestation weight gains were adjusted to assess the relative importance without the influence of dietary fatty acids. In linear regression model 2, dietary fatty acids were adjusted. Statistical significance was set at p < 0.05, data were analyzed using SPSS 24.0 software package (SPSS Inc., Chicago, IL, USA), and figures were drawn using R soft.

A total of 103 women enrolled and completed the study, and maternal-fetal anthropometric and demographic characteristics are presented in Table 1. As expected, maternal prepregnancy and antenatal BMIs were significantly higher in the OW group than the other two groups (p < 0.001). The gestational weight gains were significantly lower in the OW group than the other two groups (p < 0.001). Infants born to OW and NW mothers presented significantly higher birth weight than infants born to UW mothers (p = 0.01, p = 0.03). Infants born to OW mothers presented significantly higher birth BMI z-score than NW and UW mothers (p = 0.034, p = 0.012). Infants born to NW mothers presented higher birth length than infants born to UW mothers (p = 0.011).

Maternal plasma fatty acids are listed in Table 2, and two kinds of fatty acids (C11:0, C23:0) were below the limit of quantification; the proportions of C4:0, C6:0, C8:0, C10:0, C12:0, C13:0, C20:0, C21:0, C22:0, C24:0, C14:1, C15:1, C17:1, C18:1(trans), C24:1 were lower than 0.10%, thus they are not presented in Table 2.

OW mothers had significantly lower level of DHA compared with NW and UW mothers (p = 0.003, p = 0.042). Maternal plasma EFA and n-6 PUFA were significantly higher in the OW group than the NW group (p = 0.038, p = 0.048). The percentage of n-3 PUFA was lower in the OW group than the NW group (p = 0.011). No significant differences were observed in maternal plasma D5D, D6D, and elongase activity indices among the three groups. The indexes of D6D and AA/LA tended to a marginal decrease in OW mothers (p = 0.054, p = 0.084). The ratio of n-6/n-3 was higher in the OW group than the NW and UW groups (p = 0.001, p = 0.021).

Two kinds of fatty acids in cord plasma (C11:0, C23:0) were below the limit of quantification, and the proportions of C4:0, C6:0, C8:0, C10:0, C12:0, C13:0, C20:0, C21:0, C24:0, C14:1, C15:1, C17:1, C18:1(trans), C24:1, C18:2n-6(trans), C18:3n-3, C20:3n-3 were lower than 0.10%; therefore, they are not presented in Table 3.

The proportions of C14:0 and C20:2n-6 in cord plasma were significantly higher in the OW group than that in the NW group (p = 0.034, p = 0.002). The proportions of C15:0, total saturated fatty acid (SFA), EPA, DHA, and n-3 PUFA were significantly lower in the OW group than that in the NW group (p = 0.041, p = 0.029, p = 0.038, p = 0.027, p = 0.017). MUFA was higher in the OW group than that in the UW group (p = 0.011). The ratio of n-6/n-3 was higher in the OW group than the NW and UW groups (p < 0.001, p = 0.001).

Associations between maternal parameters and the infantile parameters are shown in Figure 1. The pre-BMI was positively associated with maternal n-6 PUFA (r = 0.203, p = 0.039), n-6/n-3 PUFA (r = 0.258, p = 0.009), but the prepregnancy BMI was negatively associated with maternal DHA (r = −0.266, p = 0.007) and n-3 PUFA (r = −0.255, p = 0.009). Maternal pre-BMI was positively associated with infant birth BMI z-score (r = 0.274, p = 0.006) and cord plasma n-6/n-3 PUFA (r = 0.325, p = 0.001). Maternal pre-BMI was negatively associated with cord plasma DHA (r = −0.303, p = 0.002) and n-3 PUFA (r = −0.298, p = 0.002). The gestational weight gain was positively associated with cord plasma SFA (r = 0.261, p = 0.009). Maternal plasma EPA, DHA, and n-3 PUFA were positively correlated with cord plasma EPA, DHA, and n-3 PUFA (p < 0.05). Maternal plasma LA was positively associated with cord plasma LA, n-6 PUFA, and n-6/n-3, but negatively associated with cord plasma DHA and n-3 PUFA (p < 0.05).

The negative correlation was found between pre-BMI and cord plasma DHA in all subjects (Figure 2A). The subgroup analysis was carried out to explore the association of prepregnancy OW and cord plasma DHA (Figure 2B). The pre-BMI was negatively correlated with cord plasma DHA (r = −0.561, p = 0.004) in the OW group, but there were no correlations between pre-BMI and cord plasma DHA in NW (r = −0.012, p = 0.925) and UW subjects (r = −0.414, p = 0.098).

The positive correlation was found between pre-BMI and cord plasma n-6/n-3 PUFA in all subjects (Figure 3A). Further subgroup analysis was carried out to explore the association of prepregnancy OW and cord plasma n-6/n-3 PUFA (Figure 3B). The pre-BMI was negatively correlated with cord plasma n-6/n-3 ratio (r = 0.558, p = 0.004) in the OW group, but there were no associations between pre-BMI and cord plasma n-6/n-3 in the NW (r = 0.041, p = 0.756) and UW groups (r = 0.439, p = 0.078).

A number of confounder factors could influence the association between maternal parameters and fetal parameters. Multiple linear regression analysis was conducted to explore the direct correlation between maternal parameters and cord plasma DHA (Table 4). The correlations can be found between maternal pre-BMI, LA, EPA, DHA, EFA, n-6 PUFA, n-3 PUFA, and n-6/n-3 with cord plasma DHA in all subjects (Figure 1). Owing to collinearity, pre-BMI, maternal plasma LA, EPA, and DHA were included in the regression model. Multiple linear regression analysis demonstrated the associations between maternal pre-BMI and DHA with cord plasma DHA in all subjects (p < 0.05). In the subgroup analysis, the correlation between maternal pre-BMI and DHA with cord plasma DHA can be observed only in OW mothers (p < 0.05), but not in NW subjects.

Multiple linear regression analysis was conducted to explore the direct correlation between maternal parameters and cord plasma n-6/n-3 PUFA (Table 5). The correlations can be found between maternal pre-BMI, LA, AA, EPA, DHA, EFA, n-6 PUFA, n-3 PUFA, n-6/n-3, and cord plasma n-6/n-3 in all subjects (Figure 1). Owing to collinearity, finally pre-BMI, maternal AA, EPA, DHA, and n-6 PUFA were included in the regression model. Multiple linear regression analysis demonstrated the associations between pre-BMI and maternal DHA with cord plasma n-6/n-3 in all subjects (p < 0.05). In the subgroup analysis, maternal pre-BMI was associated with cord plasma n-6/n-3 in OW subjects (p < 0.05), but not in NW subjects (Table 5).