Beyond protein intake: does dietary fat intake in the year preceding pregnancy and during pregnancy have an impact on gestational diabetes mellitus?

We used data from the NPGSC study, which is a prospective cohort study conducted among pregnant women and their children from 27 study centers (12 urban hospitals and 15 rural hospitals) of Sichuan, Yunnan, and Guizhou province in China, initiated in January 2014, aiming to investigate the relevance of maternal nutrition before and during pregnancy on health outcomes of mother and child, as described elsewhere [7].

The pregnant women were invited to the study center for interviews during their first visit for routine ultrasound examination at gestational weeks 9–11. To facilitate follow-up, participants who were cooperative and voluntary, who had lived in their current residence for at least 2 years, who signed an informed consent form were included in this cohort study. Generally, each visit included questionnaire, anthropometric measurements, and clinical examinations. The study was approved by the Ethics Committee of Sichuan University.

All participants were followed up for three times before women giving birth (the first routine ultrasound examination, Q1; gestational weeks 20–22, Q2; gestational weeks 33–35, Q3) and 8 times across offspring’s infancy and young childhood (Fig. 1). At Q1, participants were asked to complete a self-administered questionnaire approached by local nurses and trained interviewers, which collected information about their birth characteristics, demographic characteristics, educational and employment status, lifestyle characteristics (e.g., smoking behavior, alcohol consumption, tea/coffee consumption), medical history, annual family income, family history of chronic diseases, and so forth. In addition, each visit included face-to-face interviews by trained investigators about their diet [one food frequency questionnaire (FFQ) assessing usual dietary intake during the previous year before pregnancy and one 24-h dietary recall covering the consumption over the past 24 h] and physical activity (one questionnaire addressing physical activity and sedentary behavior over the past 12 months before pregnancy and from the start of the pregnancy, respectively). At Q2 and Q3, dietary data and physical activity were collected by trained investigators in face-to-face interviews with one FFQ assessing usual dietary intake over the past 12 weeks and one 24-h dietary recall each addressing intake over the past 24 h, as well as one physical activity questionnaire, respectively.

The data used in this study were identified from the baseline survey of the prospective cohort study between 2014 and 2017. Participants in the present study lived in the Sichuan Provence and Guizhou Province (18 study centers: 8 urban hospitals and 10 rural hospitals). Initially, 6886 pregnant women were recruited. Of these, 587 were excluded: 73 diagnosed with the preexisting diabetes mellitus before pregnancy, 127 had stillbirth, multiple birth, or missing the first general questionnaire, 184 had incomplete information on three FFQs for the dietary intakes one year before pregnancy, at 1st trimester, or at 2nd trimester, 106 had implausible energy intake, and 97 had incomplete information on potential confounders. Therefore, the present analysis was based on a final sample of 6299 women (Fig. 2).

At baseline, dietary intake 1 year before pregnancy was assessed using FFQ, that is, the present analysis was based on the dietary data collected by a 128-item FFQ. This FFQ was based on a validated questionnaire and additionally modified to the respective food groups [7]. At each assessment, participants were inquired to how often (daily, weekly, monthly, annually, never) on average they had consumed standard serving sizes of a specific type of food and or a common unit of weight in China (i.e., 1 liang = 50 g or natural units). Standardized tableware, including the bowls, plates, and glasses was provided for study respondents to improve the accuracy of the estimated serving sizes. Dietary intake data of each food and beverage reported by the FFQ were converted into grams per day and total energy and nutrients were calculated according to the continuously updated in-house nutrient database [25], which reflected the China Food Composition [26]. This nutrient database includes any food item ever recorded in previous studies of our institute, and is based on the information from standard nutrient tables, product labels, or recipe simulation of missing nutrient data for foods (e.g., new commercial food products) based on the labeled ingredients and national food tables.

Self-reported pregravid weight was recorded on the day of registration. Trained investigators in each study center obtained body weight measurements according to standard procedures at enrollment and at regular intervals (in 4-week intervals from enrollment to week 25, every 2-week until week 33 weekly thereafter) to birth (Fig. 1). Body weight and maternal height (at enrollment) were measured with an ultrasonic instrument (Dingheng, Zhengzhou Province, China) to the nearest 0.1 kg and 0.1 cm, respectively. The information was recorded in the Medical Birth Registry.

The first ultrasound scan (Eub 5500, Hitachi; Eub 7500, Hitachi; Logiq E9, GE) measured in a standard manner by trained ultrasonographers on the day of registration and self-reported data on the last menstrual period were used for estimating the gestational age (GA), namely, if both measures were available and there was agreement (± 14 days) self-report data was used, otherwise, ultrasound data were used.

All participants underwent a 2-h 75-g oral glucose tolerance test (OGTT) in the morning after at least 8 h of fasting between 24 and 28 weeks of gestation. Venous plasma glucose at fasting, 1 and 2 h after the glucose load were measured by trained nurses. According to the IADPSG`s criteria, a diagnosis of GDM was made if one or more plasma glucose values of the following cutoff values were met: fasting plasma glucose ≥ 5.1 mmol/L or 1 h plasma ≥ 10.0 mmol/L or 2-h plasma glucose ≥ 8.5 mmol/L.

SAS procedures (version 9.3, SAS Inc, Cary, NC) were used for all data analyses. All analyses were performed with a significance level at P < 0.05, except for interaction tests, in which P < 0.1 was considered significant in multivariable analysis. Preliminary analyses indicated no interactions between age and the relation of dietary fat intake with the GDM risk. Thus, the data from different ages were pooled for all analysis.

In this analysis, we conducted separate analyses using dietary data collected in the year preceding pregnancy and during the 1st trimester and 2nd trimester of pregnancy, aiming to differentiate the critical time window in which dietary fat intake impacts GDM onset. Dietary fat intakes in these three different periods were expressed each as residuals from their regression on energy intake. Multiple models of logistic regression were used to estimate the odds ratios (ORs) and 95% CIs of GDM risk in relation to energy-adjusted residuals of dietary fat intakes (into tertiles). Comparisons among tertiles of dietary fat intake were performed using an analysis of variance (ANOVA) for continuous variables and the chi-square test for categorical variables.

In the basic models, dietary fat intakes were the independent predictors. The following variables potentially affecting these associations were considered: maternal age, parity, residence (urban/rural), family history of diabetes, maternal education level (12 or more years of schooling; yes/no), maternal occupation (no, yes: part-time worker or full-time worker), monthly personal income (< 3000 CNY, 3000–6000 CNY, > 6000 CNY), pregravid BMI, gestational weight gain, physical activity, smoking/passive smoking before, or during pregnancy (never, past, current: 1–15, 16–24, or > 24 cigarettes/d), alcohol consumption before or during pregnancy (0, 0.1–9.9, 10.0–19.9, 20.0–29.9, ≥ 30 g/d), and intakes of carbohydrate and protein at the same time point of dietary fat. Each potential confounder was initially considered separately and included if it substantially modified the association of dietary fat with GDM or significantly predicted the outcome variable. Thus, maternal age, pregravid BMI and parity were retained in model 2. In the third step, we additionally adjusted for parental history of diabetes (yes or no), current smoking (combination of passive smoking and active smoking, yes or no), family income, total energy intake, gestational weight gain, and physical activity. In a further step, we controlled for confounding and/or mediation by carbohydrate intake (model IV a) and animal protein intake (model IV b), as they have been proposed to be relevant for GDM risk [6, 7, 14, 27]. Animal fat and vegetable fat were mutually adjusted for one another.

General characteristics of the sample in this study stratified by dietary fat intake tertiles are shown in Table 1. In the present analysis, participants were on average 26.5 years old. No significant difference was found for age (P = 0.76), maternal educational level (P = 0.47), monthly personal income (P = 0.65) and family history of diabetes (P = 0.54) across all dietary fat intake groups. When compared with those consuming a diet with lower energy intake from dietary fat, participants in the highest energy intake from fat had a higher urbanization level (P < 0.001) and higher pregravid BMI (P = 0.04). In addition, higher energy intake from protein and lower energy intake from carbohydrate was observed in the highest dietary fat intake in the year preceding pregnancy and during the 1st trimester and 2nd trimester of pregnancy (P < 0.05).

The results of logistic regression analyses for the association between total and source of dietary fat intakes and GDM risk (Table 2) showed that intakes of animal fat and vegetable fat in the year preceding pregnancy and during pregnancy were not associated with risks for GDM (P > 0.05).

Higher intake of total fat during 12–22 weeks of gestation resulted in a significant higher GDM risk in initial model (P = 0.04). Even after adjustment for maternal age, pregravid BMI, parity, parental history of diabetes, current smoking, family income, total energy intake, gestational weight gain, physical activity, and carbohydrate intake (model IV a), women with highest total fat intake had an approximately 121% higher odds to those in the lowest tertile (95% CI: 1.19, 4.20, P = 0.02). However, this association was considerably attenuated once animal protein intakes (model IV b) were adjusted for (95% CI: 0.93, 3.64, P = 0.11), but did improve the fit of the final model. Moreover, total fat intake neither in the year preceding pregnancy nor during the early pregnancy was associated with GDM risk.

The results of logistic regression analyses of specific dietary fat and cholesterol intakes and the risk of GDM are shown in Table 3. We did not observe the significant association between intakes of total dietary cholesterol (P = 0.13), saturated fatty acid (SFA) (P = 0.23), monounsaturated fatty acid (MUFA) (P = 0.16), and polyunsaturated fatty acid (PUFA) (P = 0.92) one year before pregnancy and during the first and second trimesters and GDM risk in fully adjusted models including both dietary and nondietary covariates.

When we used total, specific, and source of fats estimated by FFQ to examine their association between total, specific, and source of fats and triglycerides (TG) and total cholesterol (TC), relation estimates remained similar, showing that dietary fat one year before pregnancy, and during pregnancy derived in our study were relatively accurate (data not shown).