Association between plant-based or animal-based dietary pattern and plasma glucose during oral glucose tolerance test among Chinese women with gestational diabetes mellitus: a prospective cohort study

Abstract

Objective

This study explored whether plant-based and animal-based dietary patterns are associated with plasma glucose levels during oral glucose tolerance test in women with gestational diabetes mellitus (GDM).

Design

A prospective cohort study was conducted using a Food Frequency Questionnaire to collect dietary data. Dietary patterns were derived using principal component analysis. Multivariate logistic regression analysis was performed to explore the association between dietary patterns and plasma glucose levels. Stratified analyses were conducted according to maternal age, prepregnancy body mass index (BMI) and other confounders.

Setting and participants

The study, conducted in January 2022 in two hospitals in Fujian, China, involved 424 women diagnosed with GDM using a 75 g 2-hour oral glucose tolerance test at 24–28 gestational weeks.

Results

Six maternal dietary patterns (plant based and animal based) were identified. Participants with plant-based pattern had lower 2-hour plasma glucose levels than those with animal-based pattern (β=−0.314; 95% CI (−0.596 to –0.032)). After adjusting the regression model covariates, this significant association remained (β=−0.288; 95% CI (−0.568 to –0.008)) and appeared more pronounced in women aged 30 years or above and those with prepregnancy BMI<24 kg/m².

Conclusions

Plant-based pattern is associated with lower plasma glucose levels in women with GDM, which is valuable information for dietary counselling and intervention.

XQC and QZ are joint first authors.

STRENGTHS AND LIMITATIONS OF THIS STUDY

• By focusing on dietary patterns rather than individual nutrients or dietary components, this study offers practical information and guidance on healthy eating in pregnant women with gestational diabetes mellitus.

• The effects of potential confounders were mitigated by adjusting for several covariates.

• As the dietary intake information in the cohort study was self-reported, it is susceptible to measurement errors and recall bias.

• Data-driven methods for determining dietary patterns hinder replicating the results of other studies.

• An observational study design may limit the power of causal relationships between dietary patterns and plasma glucose levels. Further randomised controlled trials are required to confirm these findings.

Introduction

Gestational diabetes mellitus (GDM) refers to any degree of glucose intolerance with onset or first detection during pregnancy.¹ It is a common disease during pregnancy and a major public health issue in China and worldwide. According to data released by the International Diabetes Federation in 2021,² approximately 21.1 million (16.7%) pregnant women with live births worldwide had elevated blood glucose levels during pregnancy, and GDM accounted for approximately 80.3% (about 17 million). With the implementation of China’s three-children policy, the number of obese and elderly women of childbearing age increased, and the incidence rate of GDM skyrocketed. In 2021, approximately 8.737 million pregnant women with live births in China were affected by hyperglycaemia, with an overall GDM prevalence rate of 8.6%.³ GDM endangers both the short-term and long-term health of mothers and their children. In the short term, GDM can cause serious complications in mothers and their offspring, including pre-eclampsia, ketoacidosis, macrosomia, respiratory distress syndrome, asphyxia and hypoglycaemia.^{4 5} In the long-term, it can increase the risk of type 2 diabetes, obesity and cardiovascular diseases.⁵ The harm posed by GDM to mothers and children has greatly increased social and medical burdens. Consequently, increasing emphasis has been placed on its prevention and treatment.

An important factor in the occurrence and development of GDM is the dietary intake during pregnancy.^{6 7} High intakes of cholesterol, saturated fatty acids, transfatty acids, haeme iron, red meat and processed meat are associated with an increased risk of GDM,⁸ whereas higher consumptions of legumes, fish, nuts, fruits and vegetables are associated with a reduced risk of GDM.^{7 9} Considering the interaction between various foods and nutrients, overall dietary patterns can be used to evaluate comprehensively the role of diet in diseases.¹⁰ In addition to individual nutrients and foods, some studies have confirmed that dietary patterns are also closely related to GDM.^{8 11 12} Higher scores of ‘Mediterranean’, ‘prudent’ and ‘vegetable’ dietary patterns have been associated with a lower risk of GDM.⁸ Recently, plant-based diets have grown in popularity, emphasising the importance of consuming higher (lower) proportions of plant (animal)-based foods.¹³ The 2015–2020 Dietary Guidelines for Americans recommend plant-based diets among others for the prevention of chronic diseases,¹⁴ especially diabetes. Studies link plant-based dietary patterns to the prevention of GDM,¹⁵ hypertension during pregnancy,¹⁶ type 2 diabetes,¹⁷ cardiovascular disease¹⁸ and polycystic ovary syndrome.¹⁹ However, several studies have reported conflicting results.^{20 21} In a large birth cohort, a vegetable-based diet was linked to a lower risk of GDM,²⁰ but another study found no association between plant-based dietary patterns and GDM risk.²¹ Furthermore, in China, the effect of plant-based dietary patterns on plasma glucose levels during the oral glucose tolerance test (OGTT) in women with GDM is unclear. Therefore, this study investigated the association between plant-based dietary patterns and plasma glucose levels during OGTT in women with GDM.

Maternal age and body mass index (BMI) are the most important drivers of GDM. Strong evidence using prepregnancy BMI as a proxy for health status indicates that the risk of GDM among overweight/obese women is 4–8 times higher than that among normal-weight women,²² and overweight/obesity results from unhealthy dietary habits.²³ Age is also highly associated with the risk of developing GDM²⁴: the older the woman with GDM, the higher the plasma glucose level in the OGTT.²⁵ It is generally accepted that pregnant women aged over 35 years carry an increased risk of various obstetric problems, having become an immutable factor.²⁶ Moreover, the major force for childbirth is concentrated over the age of 30 in China. Hence, it would be also beneficial if a given diet could improve the plasma glucose levels of such women. In addition, well-known risk factors such as living conditions during pregnancy, parity, history of GDM and family history of diabetes have been associated with GDM.²⁷ Therefore, we further stratified by age (<30 years, ≥30 years), prepregnancy BMI (<24 kg/m², ≥24 kg/m²) and other potential confounders among women with GDM.

Methods

Study design and participants

The ongoing, population-based prospective cohort study in Fujian Province is supported by the Joint Funds for the Innovation of Science (No. 2020Y9133) and designed for investigating the association between maternal diets and the health outcomes of mothers and offspring. The study, conducted in January 2022, involved pregnant women who had their antenatal check-up in two specialised maternal and child hospitals in Fujian, China (Fujian Maternity and Child Health Hospital and Fujian Obstetrics and Gynecology Hospital). In this study, we wanted to identify an association between maternal diets and plasma glucose levels during the OGTT among participants who had a singleton pregnancy and were diagnosed as GDM after the OGTT within 24–28 gestational weeks. Women who (1) were previously diagnosed with type 2 diabetes, (2) had a severe chronic or infectious disease, (3) had undergone infertility treatment and (4) had severe mental disease were excluded from the study.

Sample size estimation

This study estimated the sample size for a cross-sectional survey based on total caloric food intake of women with GDM, using the formula N= (Z1−α/2×σ)δ2 (σ is the overall SD and δ is the allowable error). Referring to the study by Schoenaker et al,²⁸ which showed that the total caloric food intake of women with GDM was 1581.6 kcal (SD=530.4), σ was calculated to be 530.4 and δ was 53.04 (10% of σ). The estimation was set for a two-sided test at α=0.05; hence, ${\displaystyle {\mathrm{Z}}_{1-\alpha /2}}$ was 1.96. Using the aforementioned data and formula, a sample size of 385 participants was needed. Furthermore, considering a 10% turnover rate, 424 participants were finally included in the study.

Patient and public involvement statement

Patients or the public were not involved in the design, or conduct, or reporting, or dissemination plans of our research.

Ascertainment of GDM

During 24–28 weeks of gestation, the participants were routinely scheduled for an OGTT. All participants fasted overnight for at least 8 hours before the OGTT. Blood samples were collected at fasting, 1 hour and 2 hours after participants received the 75 g oral glucose. GDM was diagnosed when at least one of the following values was met, using the International Association of Diabetes and Pregnancy Study Groups diagnostic criteria²⁹: fasting plasma glucose ≥5.1 mmol/L, 1-hour plasma glucose ≥10.0 mmol/L, or 2-hour plasma glucose ≥8.5 mmol/L.

Dietary assessment

Dietary intake during the last 3 months was reported by participants at 24–28 gestational weeks using a Food Frequency Questionnaire (FFQ) via face-to-face interviews conducted by investigators. Initially, all investigators were trained and evaluated by nutrition experts on how to use the FFQ, to ensure that the investigation ran smoothly. The interviewer-administered semiquantitative FFQ has proven to be a reliable and valid tool for assessing the food intake of pregnant women in China.³⁰ The FFQ comprises 61 food items, covering more than 200 food types. In this study, drinks, oils and condiments, which were uncommon or deemed difficult to estimate, were excluded. The participants were asked to recall the intake frequency and average consumption per serving of each food item in the past 3 months. To assist participants in quantifying their food intake, we provided them with a visual aid booklet with coloured photographs of various food portion sizes. For the intake frequency, we standardised and adjusted the original categories of frequency into six grades, following Li et al³¹ (0: ‘never’, 1: ‘1–3 times per 4 weeks’, 2: ‘1–3 times per week’, 3: ‘4–6 times per week’, 4: ‘1–2 times per day’ and 5: ‘more than two times per day’). The daily intake of each food in grams was calculated by multiplying the newly assigned frequency by the consumption per serving and then dividing it by 28 in standard portions. For the analysis of dietary patterns, food items were assembled into sixteen food groups with similar nutrient profiles or culinary uses, including rice and wheat products, whole grains, beans and bean products, leafy and cruciferous vegetables, root vegetables, melons and solanaceous vegetables, mushrooms and algae, fruits, dairy products, red meat, poultry, freshwater fish, animal organs and blood, seafood, eggs, and nuts.

Covariates

During the enrolment interview, trained investigators used a structured questionnaire to collect the participants’ self-reported demographic characteristics, anthropometric parameters and lifestyle information. Any questionable participant data were further checked through the medical information system, using outpatient or inpatient service numbers. The questionnaire included information on maternal age; ethnicity; educational level; average monthly personal income; work status, living conditions and exercise frequency during pregnancy; prepregnancy weight and height; parity; GDM history; and family history of diabetes. Maternal age (years) was categorised into ages <25, 25–29, 30–34 and ≥35 years. Ethnicity was categorised as Han Chinese or a minority. Educational levels were categorised as middle school or below, high school, junior college degree, bachelor’s degree, or higher. Average monthly personal income (CNY) was divided into the lowest (<CNY3000), lower (CNY3000–CNY5999), medium (CNY6000–CNY8999), higher (CNY9000–CNY11 999) and highest (≥CNY12 000) income. Living conditions during pregnancy were categorised as living with a partner and living with a partner and parents. Exercise frequency during pregnancy was divided into none, 1–2times/week, 3–5times/week and 6–7times/week. Prepregnancy BMI (kg/m²), which was calculated as prepregnancy weight (kg) divided by height (m²), was classified into <18.5, 18.5–23.9, 24.0–27.9 and ≥28.0, according to Chinese standards.³² Parity was classified as primiparous or multiparous. Other covariates, including working during pregnancy, maternal history of GDM and family history of diabetes, were divided into yes or no answers.

Statistical analysis

Dietary patterns were derived using principal component analysis (PCA) with varimax rotation for greater interpretability. PCA is a data-driven technique that uses a few comprehensive factors to replace the original indicators for analysis and preserves the bulk of the information, thereby reducing the computing workload when numerous indicators exist. In this study, the common factors of food consumptions were determined by reducing the dimensions of the data and grouping the correlated variables using PCA. The number of factors retained was based on the eigenvalues, breakpoints of the scree plot, cumulative variance and factor interpretability.³³ Thereinto, the total variance of the original variables (food group intake) was decomposed into the sum of the variances of several independent common factors (dietary patterns). The greater the cumulative variance among the factors, the better their ability to combine with the original variables (food group intake). The weights obtained by PCA are objective and reasonable, thereby addressing the flaws of certain evaluation systems in establishing weights. The factor loads reflected the correlation between the original 16 non-overlapping food groups and newly extracted factors (dietary patterns). In this study, an eigenvalue above 1 based on the scree plot and factor loadings of 0.40 or higher on a factor represented a high correlation between the food group and dietary pattern.³³ Each dietary pattern score for each participant was obtained by summing the mean standardised intake of the food groups, weighted by their factor loadings. The higher the dietary pattern score, the greater the adherence to the derived dietary pattern.³¹ Participants were deemed to follow the highest-scoring dietary pattern. All dietary patterns were divided into two types that varied according to their primary food source: ‘plant based’, which was rich in grains, legumes, fruits and vegetables, and ‘animal based’, which comprised meats, eggs and cheeses.³⁴

The distributions of continuous variables are described as means and SD, and the distributions of categorical variables are expressed as frequencies and percentages. Comparisons between groups were performed using χ² tests or Fisher’s exact test for categorical variables (participants’ demographic characteristics, anthropometric parameters and lifestyle). Multivariate linear regression models were used to analyse the relationships between dietary patterns (animal based and plant based) and plasma glucose levels during the OGTT. In the adjusted model, maternal age; ethnicity; educational level; average monthly personal income; work status, living conditions, and exercise frequency during pregnancy; prepregnancy BMI; parity; history of GDM; and family history of diabetes were adjusted for (the assignment values of maternal characteristics for multivariate linear regression analysis are shown in online supplemental figure 1). For real-world observational data, conducting a sensitivity analysis is essential to examine the causal effects between exposure and outcome.³⁵ Hence, we added an interaction term (age and prepregnancy BMI) to the adjusted multivariate linear regression model in the sensitivity analysis. Stratified analyses were conducted by maternal age (<30, ≥30), living condition during pregnancy (living with a partner or living with a partner and parents), prepregnancy BMI (<24, ≥24), parity (primiparous/multiparous), history of GDM (yes/no), family history of diabetes (yes/no). All statistical analyses were performed using SPSS software V.25, and GraphPad Prism V. 9.0 was used to prepare graphs. The Kaiser-Meyer-Olkin test (KMO=0.7) indicated the adequacy of sampling. Bartlett’s test of sphericity was significant (p<0 001), indicating that the factor analysis was suitable for the data. A two-sided α of less than 0.05 was considered statistically significant.

Results

Dietary patterns

In total, 450 pregnant women with GDM constituted the cohort. After excluding those who did not complete the FFQ, 437 were included in the analysis. The KMO value was 0.710 (p<0 001 for Bartlett’s test of sphericity). From the scree plot, six dietary patterns were identified with eigenvalues above 1 (online supplemental figure 1), which accounted for 55.87% of the total variation; the rotated factor loadings and overall total variance for each pattern are presented in (table 1). The first pattern included a higher intake of seafood, freshwater fish, fruits, and leafy and cruciferous vegetables (SFFL). The second included a higher intake of poultry, beans, bean products and red meat (PBR). The third included a higher intake of whole grains, melons and solanaceous vegetables (WM). The fourth included a higher intake of mushrooms, algae and root vegetables (MR). The fifth included a high intake of rice, wheat products and nuts (RN). The sixth consisted of a higher intake of eggs and dairy products and a lower intake of animal organs and blood (EDA), with the emphasis remaining mostly on animal-based foods. The WM, MR and RN patterns, characterised by a high intake of plant-based foods, were classified as plant-based dietary patterns. The SFFL, PBR and EDA patterns were classified as animal-based dietary patterns. The loading coefficients of animal-based foods were greater than those of plant-based foods in the SFFL diet; therefore, SFFL was classified as an animal-based dietary pattern.

Table 1

Factor loading matrix for dietary patterns by principal component analysis among women with GDM* (n=437)

Food group	Dietary patterns
	SFFL	PBR	WM	MR	RN	EDA
Seafood	0.699	0.083	0.088	0.008	0.100	0.199
Freshwater fishes	0.648	0.137	0.022	0.158	0.018	0.026
Fruits	0.514	0.127	0.505	0.228	0.263	0.215
Leafy and cruciferous vegetables	0.491	0.259	0.038	0.325	0.370	0.137
Poultry	0.069	0.750	0.189	0.057	0.016	0.018
Beans and bean products	0.024	0.648	0.052	0.231	0.084	0.149
Red meat	0.477	0.558	0.077	0.076	0.115	0.132
Whole grains	0.099	0.090	0.764	0.118	0.053	0.019
Melon and solanaceous vegetables	0.221	0.088	0.608	0.108	0.082	0.310
Mushrooms and algae	0.007	0.006	0.021	0.768	0.120	0.059
Root vegetables	0.139	0.080	0.328	0.576	0.008	0.083
Rice and wheat products	0.069	0.103	0.063	0.053	0.824	0.097
Nuts	0.089	0.086	0.054	0.147	0.625	0.454
Eggs	0.061	0.016	0.272	0.156	0.010	0.600
Dairy products	0.340	0.223	0.005	0.188	0.037	0.557
Animal organ and blood	0.285	0.183	0.103	0.305	0.063	0.413

Total variance for SFFL, PBR, WM, MR, RN, EDA were 18.01%, 8.35%, 8.27%, 7.76%, 6.98%, 6.49%, respectively.

*Values are factor loadings between each food variable and the dietary pattern derived from principal component analysis. Food groups are sorted by size of loading coefficients. Total variation accounted for 55.87%.

EDA, eggs, dairy products, low animal organ and blood; GDM, gestational diabetes mellitus; MR, mushrooms and algae, root vegetables; PBR, poultry, beans and bean products, red meat; RN, rice and wheat products, nuts; SFFL, seafood, freshwater fishes, fruits, leafy and cruciferous vegetables; WM, whole grains, melon and solanaceous vegetables.

Characteristics of participants

Participants’ characteristics are presented in table 2. Overall, their mean age was 31.59 years (SD=4.04), and 70.5% of the cohort comprised women aged >30 years. The mean prepregnancy BMI was 22.48 kg/m² (SD=3.31), and overweight or obese participants (BMI≥24 kg/m²) accounted for 30.0% of the sample. Among the 437 participants, 416 (95.2%) were Han Chinese, and 266 (60.9%) were primiparous. Regarding educational level, 51.9% of the participants had a bachelor’s degree and above. Most participants continued working (63.6%), lived with their partners (73.2%) and exercised at least once per week (74.4%) during pregnancy. Furthermore, 8.5% of the participants had a history of GDM, and 32.75% had a family history of diabetes. For dietary patterns, 249 (118) participants were inclined towards animal (plant)-based dietary patterns. No significant differences were observed between participants with animal-based or plant-based dietary patterns in maternal age, ethnicity, educational level, average monthly personal income, living conditions and exercise frequency during pregnancy, prepregnancy BMI, parity, history of GDM, and family history of diabetes. However, more women with GDM inclined towards plant-based dietary patterns did not work during pregnancy than those inclined towards animal-based dietary patterns (p=0.012).

Table 2

Maternal characteristics in plant-based dietary and animal-based dietary patterns among pregnant women with GDM, n (%)

Characteristics	All participants (n=437)	Animal-based dietary pattern (n=249)	Plant-based dietary pattern (n=188)	P value
Age (year)				0.542
<25	13 (3.0)	5 (2.0)	8 (4.3)
25–29	116 (26.5)	65 (26.1)	51 (27.1)
30–34	212 (48.5)	125 (50.2)	87 (46.3)
≥35	96 (22.0)	54 (21.7)	42 (22.3)
Ethnicity				0.185
Han	416 (95.2)	234 (94.0)	182 (96.8)
Minority	21 (4.8)	15 (6.0)	6 (3.2)
Educational level				0.193
Middle school or below	42 (9.6)	23 (9.2)	19 (10.1)
High school	58 (13.3)	28 (11.2)	30 (16.0)
Junior college degree	110 (25.2)	58 (23.3)	52 (27.7)
Bachelor degree and above	227 (51.9)	140 (56.2)	87 (46.3)
Average monthly personal income (CNY)				0.285
Lowest (<3000)	8 (1.8)	4 (1.6)	4 (2.1)
Low (3000–5999)	123 (28.1)	65 (26.1)	58 (30.9)
Medium (6000–8999)	163 (37.3)	97 (39.0)	66 (35.1)
High (9000–11999)	82 (18.8)	53 (21.3)	29 (15.4)
Highest (≥12 000)	61 (14.0)	30 (12.0)	31 (16.5)
Working during pregnancy				0.012
No	159 (36.4)	78 (31.3)	81 (43.1)
Yes	278 (63.6)	171 (68.7)	107 (56.9)
Living condition during pregnancy				0.276
With partner	320 (73.2)	177 (71.1)	143 (76.1)
With partner and parents	117 (26.8)	72 (28.9)	45 (23.9)
Exercise frequency during pregnancy				0.140
No	112 (25.6)	72 (28.9)	40 (21.3)
1–2 times/week	171 (39.1)	95 (38.2)	76 (40.4)
3–5 times/week	116 (26.5)	58 (23.3)	58 (30.9)
6–7 times/week	38 (8.7)	24 (9.6)	14 (7.4)
Prepregnancy BMI (kg/m²)				0.170
<18.5	49 (11.2)	31 (12.4)	18 (9.6)
18.5–23.9	257 (58.8)	135 (54.2)	122 (64.9)
24.0–27.9	107 (24.5)	68 (27.3)	39 (20.7)
≥28	24 (5.5)	15 (6.0)	9 (4.8)
Parity				0.843
Primiparous	266 (60.9)	153 (61.4)	113 (60.1)
Multiparous	171 (39.1)	96 (38.6)	75 (39.9)
History of GDM				0.418
No	400 (91.5)	229 (92.0)	171 (91.0)
Yes	37 (8.5)	20 (8.0)	17 (9.0)
Family history of diabetes				0.759
No	294 (67.3)	166 (66.7)	128 (68.1)
Yes	143 (32.7)	83 (33.3)	60 (31.9)

BMI, body mass index; GDM, gestational diabetes mellitus.

Dietary patterns in relation to plasma glucose levels

In this study, all plasma glucose values were normally distributed. The mean fasting plasma glucose, 1 hour plasma glucose and 2-hour plasma glucose during OGTT of women with GDM inclined towards animal-based dietary patterns were 4.81, 10.25, 8.82 mmol/L (SD=0.49, 1.59, 1.44), respectively, and for those inclined towards plant-based dietary patterns the respective values were 4.80, 10.01, 8.50 mmol/L (SD=0.51, 1.45, 1.54) (figure 1). The results of a linear regression model, constructed to assess the association between dietary patterns and plasma glucose levels during the OGTT, are shown in table 3. In the crude model, the 2-hour plasma glucose level among women with GDM inclined towards plant-based dietary patterns decreased by 0.314 mmol/L (95% CI (−0.596 to –0.032)), compared with those inclined towards animal-based dietary patterns. After adjusting for maternal characteristics, the plant-based dietary pattern remained significant in relation to a lower 2-hour plasma glucose level in women with GDM (β=−0.288; 95% CI (−0.568 to –0.008)). Although the interaction between dietary patterns and fasting or 1-hour plasma glucose during the OGTT was not significant, the fasting and 1-hour plasma glucose levels of women with GDM inclined towards plant-based dietary patterns were all lower than those inclined towards animal-based dietary patterns.

Table 3

Associations between dietary patterns and plasma glucose levels during oral glucose tolerance testing among pregnant women with GDM (n=437)

Dietary pattern* (Refer to animal-based dietary pattern)	Fasting plasma glucose		1-hour plasma glucose		2-hour plasma glucose
	β (95% CI)	P value	β (95% CI)	P value	β (95% CI)	P value
Crude model	−0.006 (−0.101 to 0.089)	0.900	−0.235 (−0.526 to 0.055)	0.112	−0.314 (−0.596 to 0.032)	0.029
Adjusted model†	−0.019 (−0.112 to 0.074)	0.685	−0.251 (−0.542 to 0.039)	0.090	−0.288 (−0.568 to 0.008)	0.044
Sensitivity analysis‡	−0.019 (−0.111 to 0.074)	0.692	−0.252 (−0.543 to 0.038)	0.088	−0.289 (−0.569 to 0.009)	0.043

*Plant-based dietary pattern refer to animal-based dietary pattern.

†Adjusted model was adjusted for age, ethnicity, education, average monthly personal income, working during pregnancy, living condition, exercise frequency, prepregnancy BMI, parity, history of GDM, family history of diabetes.

‡Sensitivity analysis added an interaction term (age and prepregnancy BMI) based on the adjusted multivariate linear regression model.

BMI, body mass index; CI, confidence interval; GDM, gestational diabetes mellitus.

Figure 1. Plasma glucose during OGTT between plant-based and animal-based dietary pattern. OGTT, oral glucose tolerance test.

Sensitivity analysis

In the sensitivity analysis, after adding an interaction term (age and prepregnancy BMI) in the adjusted multivariate linear regression model, the results remained similar, indicating that the plant-based dietary patterns were related to lower 2-hour plasma glucose among GDM women (β=−0.289; 95% CI (−0.569 to –0.009) (table 3).

Stratified analyses

The multivariate linear regression models for the factors of plasma glucose during the OGTT are presented in online supplemental table 2-4. There were significant differences for fasting plasma glucose during OGTT in living condition during pregnancy, prepregnancy BMI and family history of diabetes. Significant factors influencing the 1-hour plasma glucose levels during the OGTT included age and history of GDM. In addition, the 2-hour plasma glucose levels during the OGTT varied with age and parity. Based on the results above, we conducted stratified analyses by maternal age, living conditions during pregnancy, prepregnancy BMI, parity, history of GDM and family history of diabetes to explore the association between plasma glucose in women with GDM and different dietary patterns; all were adjusted for other maternal characteristics. The results are shown in table 4.

Table 4

Stratified analyses for plasma glucose levels in relation to animal-based and plant-based dietary patterns among pregnant women with GDM*†

Characteristics	n (%)	Fasting plasma glucose		1 hour plasma glucose		2 hour plasma glucose
		β (95% CI)	P value	β (95% CI)	P value	β (95% CI)	P value
Age (year)
<30	129 (29.5)	0.097 (−0.097 to 0.290)	0.323	−0.040 (−0.669 to 0.588)	0.899	−0.359 (−0.987 to 0.268)	0.259
≥30	308 (70.5)	−0.059 (−0.167 to 0.048)	0.280	−0.386 (−0.721 to −0.051)	0.024	−0.328 (−0.642 to −0.015)	0.040
Living condition
With partner	320 (73.2)	−0.028 (−0.138 to 0.081)	0.612	−0.346 (−0.700 to 0.007)	0.055	−0.257 (−0.594 to 0.080)	0.134
With partner and parents	117 (26.8)	−0.002 (−0.180 to 0.176)	0.985	0.012 (−0.511 to 0.487)	0.962	−0.375 (−0.891 to 0.891)	0.145
Prepregnancy BMI (kg/m²)
<24	306 (70.0)	0.019 (−0.089 to 0.128)	0.727	−0.232 (−0.567 to 0.103)	0.174	−0.372 (−0.713 to −0.031）	0.033
≥24	131 (30.0)	−0.070 (−0.255 to 0.114)	0.451	−0.366 (−0.964 to 0.231)	0.227	−0.103 (−0.602 to 0.396)	0.685
Parity
Primipara	266 (60.9)	−0.004 (−0.130 to 0.122)	0.953	−0.242 (−0.630 to 0.147)	0.222	−0.229 (−0.602 to 0.143)	0.227
Multipara	171 (39.1)	−0.033 (−0.177 to 0.110)	0.646	−0.229 (−0.680 to 0.222)	0.317	−0.400 (−0.834 to 0.034)	0.071
History of GDM
No	400 (91.5)	−0.001 (−0.096 to 0.095)	0.986	−0.221 (−0.523 to 0.082)	0.152	−0.273 (−0.566 to 0.021)	0.068
Yes	37 (8.5)	−0.075 (−0.462 to 0.312)	0.694	−0.005 (−1.023 to1.013)	0.992	−0.405 (−1.602 to 0.792)	0.492
Family history of diabetes
No	294 (67.3)	−0.056 (−0.158 to 0.047)	0.287	−0.096 (−0.431 to 0.239)	0.573	−0.342 (−0.667 to −0.018)	0.039
Yes	143 (32.7)	0.056 (−0.134 to 0.265)	0.516	−0.599 (−1.168 to −0.029)	0.039	−0.239 (−0.811 to 0.332)	0.409

*Animal-based dietary pattern was reference.

†Multiple linear regression model was adjusted for age, ethnicity, education, average monthly personal income, working during pregnancy, living condition, exercise frequency, prepregnancy BMI, parity, history of GDM, family history of diabetes, except for the corresponding stratification variable.

BMI, body mass index; GDM, gestational diabetes mellitus.

Approximately 70.5% of women with GDM were aged >30 years. Hence, we conducted a stratified analysis by maternal age (<30, ≥30). When comparing women with GDM inclined towards plant-based versus animal-based dietary patterns, the stratified analyses results showed significant reductions in 1-hour plasma glucose (β=−0.386; 95% CI (−0.721 to –0.051)) and 2-hour plasma glucose (β=−0.328; 95% CI (−0.642 to –0.015) in the age ≥30 years subgroup. For prepregnancy BMI, we observed a significant inverse association between plant-based dietary patterns and 2-hour plasma glucose only in the prepregnancy BMI<24 kg/m² subgroup (β=−0.372; 95% CI (−0.713 to –0.031)). In participants with a family history of diabetes, 2-hour plasma glucose was lower among those inclined towards a plant-based versus an animal-based dietary pattern (β=−0.342; 95% CI (−0.667 to –0.018)). We observed the same trend towards an association between higher intake of plant-based dietary and lower 1-hour plasma glucose in participants without a family history of diabetes (β=−0.599; 95% CI (−1.168 to –0.029)). In the subgroup analyses, living conditions during pregnancy, parity and GDM history modified the association between 2-hour plasma glucose levels and dietary patterns among all participants. However, the trend remained: women with GDM inclined towards plant-based dietary patterns had lower 2-hour plasma glucose levels during the OGTT.

Discussion

Although a few studies have revealed the protective effect of plant-based diets on the development of GDM and suggest the consumption of legumes, nuts, fruits and vegetables,^7–9 it is unclear whether plant-based diets have different effects on plasma glucose levels among women with GDM. In this study, we identified six maternal dietary patterns classified as plant-based and animal-based and examined their associations with plasma glucose during OGTT in women with GDM. For these women, we found that higher adherence to plant-based dietary patterns was associated with lower plasma glucose levels, especially 2-hour plasma glucose levels. After adjusting for covariates in the regression model, the association between the plant-based dietary patterns and 2-hour plasma glucose levels remained statistically significant. Furthermore, sensitivity analysis results confirmed the robustness and reliability of the association between plant-based dietary patterns and lower 2-hour plasma glucose levels among women with GDM. Interestingly, the association of plant-based dietary patterns with 2-hour plasma glucose appeared to be more pronounced in women ≥30 years old and those with prepregnancy BMI<24 kg/m².

Plant-based dietary patterns during pregnancy were associated with lower 2-hour plasma glucose levels during the OGTT in women with GDM. However, no significant associations were observed between fasting plasma glucose, 1-hour plasma glucose and plant-based dietary patterns. This may be because the association between 2-hour plasma glucose and incident diabetes was stronger than that between fasting plasma glucose and diabetes, indicating that 2-hour plasma glucose is a stronger risk predictor for diabetes.³⁶ Our results align with Hu et al’s,¹¹ which showed an association between high vegetable, fruit and rice intakes (a plant-based diet) and lower glucose levels. Therefore, plant-based diets undeniably have a beneficial effect on plasma glucose levels in women with GDM. Furthermore, our findings may help establish dietary guidelines for regulating plasma glucose levels.

Certain biological mechanisms may explain our findings; however, these mechanisms need further exploration. Plant-based foods are rich in dietary fibres, antioxidants, polyunsaturated fatty acids and micronutrients, all of which are associated with weight loss, improved insulin sensitivity, reduced inflammation and improved gut microbial composition.¹⁵ The consumption of dietary fibres, such as fruits or vegetables, may increase the viscosity of stomach contents and slow down food digestion and glucose absorption, thereby decreasing postprandial plasma glucose levels.³⁷ The intake of dietary antioxidants has beneficial effects on glucose metabolism by mitigating oxidative stress, which interferes with cellular glucose uptake.³⁸ Supplementing with micronutrients, such as sodium, potassium and magnesium, might help in regulating glucose parameters and improving insulin sensitivity.³⁹ In addition, plant-based foods usually contain less saturated fats and animal proteins, which may lead to high plasma concentrations of branched-chain amino acids, thereby affecting insulin sensitivity.⁴⁰ Hence, a plant-based diet is recommended as a healthy eating pattern for women with GDM. Plant-based dietary patterns do not require completely eliminating animal-based foods, but rather promote more plant-based food consumption and less intake of animal foods.¹³

In the stratified analysis, the protective effect of plant-based dietary patterns on 1-hour plasma glucose or 2-hour plasma glucose was stronger among women aged ≥30 years or older and those with prepregnancy BMI<24 kg/m². Older women are more likely to have a lower insulin sensitivity and a higher incidence of GDM.^{24 41} Our findings have implications for management strategies aimed at improving plasma glucose in women with GDM aged >30 years. In addition, the protective effect of plant-based dietary patterns was significant in the low prepregnancy BMI group (<24 kg/m²), but not in the overweight and obese group. Consistent with a previous study, the positive association between plant protein intake and insulin sensitivity disappeared when BMI was considered, which suggests that BMI is a better predictor of insulin sensitivity than dietary intake in pregnancy.⁴²

Regarding its strengths, the study focused on dietary patterns rather than individual nutrients or dietary components, and this allowed uncovering practical information and guidance for healthy eating in pregnant women with GDM. Moreover, the abundant covariates adjusted for in the study allowed us to mitigate the effects of potential confounders. However, this study had several limitations. First, information on dietary food intake in the cohort was self-reported, which may have led to inevitable measurement errors and recall bias. Nevertheless, the FFQ used in this study has been shown to have reasonable validity for assessing food intake among pregnant women.³⁰ Moreover, we provided each participant with a visual aid booklet containing coloured images of various food portion sizes to reduce measurement errors. Second, it is difficult to replicate the results of different studies using a data-driven approach to derive dietary patterns. Third, the observational study design may limit the power of the causal relationships between dietary patterns and plasma glucose levels. Further randomised controlled trials are required to confirm these findings.

In conclusion, our study suggests that plant-based dietary patterns during pregnancy are associated with lower plasma glucose levels during the OGTT. Increasing the intake of plant-based diets might be more beneficial for plasma glucose management in older women with GDM and those with lower prepregnancy BMI values. Our findings provide insights for dietary guidance during pregnancy to improve the plasma glucose levels of women with GDM.

Data availability statement

Data are available upon reasonable request. All data relevant to the study are available upon reasonable request to the correspongding author via email.

Ethics statements

Patient consent for publication

Consent obtained directly from patient(s).

Ethics approval

This study involves human participants and was approved by the Ethical Committee of Fujian Maternity and Child Health Hospital (No: 2021KR041). Participants gave informed consent to participate in the study before taking part.

Footnote

XQC and QZ contributed equally.

Contributors The authors’ contributions to the manuscript are as detailed below: XQC and QZ contributed to the study design and conduct, data analysis, manuscript drafting and manuscript revision. YPL participated in the study design, data collection, data analysis and interpretation. XMJ contributed to the study design, data interpretation and manuscript revision. XXG participated in the data collection and interpretation and manuscript revision. Y-QP participated in the study design and manuscript revision. JL and RL contributed to the data collection, data analysis and interpretation. XMJ is responsible for the overall content as the guarantor. All authors have approved the final version to be published and agreed to be accountable for all aspects of the work.

Funding This study was supported by Joint Funds for the Innovation of Science from Fujian Province (No. 2020Y9133), Nursing Research Fund of Fujian Maternity and Child Health Hospital (YCXH 22-20) and Startup Fund for scientific research of Fujian Medical University (2022QH1190).

Competing interests None declared.

Patient and public involvement Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research.

Provenance and peer review Not commissioned; externally peer reviewed.

Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.

References

¹ US Preventive Services Task Force, Davidson KW, Barry MJ, et al. Screening for gestational diabetes: US preventive services task force recommendation statement. JAMA 2021; 326: 531–8. doi:10.1001/jama.2021.11922

² International Diabetes Federation. IDF Diabetes Atlas. 10th ed. Brussels, Belgium: International Diabetes Federation, 2021.

³ International Diabetes Federation. China diabetes report 2000-2045; 2021. Available: https://diabetesatlas.org/data/en/country/42/cn.html

⁴ Muche AA, Olayemi OO, Gete YK. Effects of gestational diabetes mellitus on risk of adverse maternal outcomes: a prospective cohort study in Northwest Ethiopia. BMC Pregnancy Childbirth 2020; 20: 73. doi:10.1186/s12884-020-2759-8

⁵ Sweeting A, Wong J, Murphy HR, et al. A clinical update on gestational diabetes mellitus. Endocr Rev 2022; 43: 763–93. doi:10.1210/endrev/bnac003

⁶ Maslova E, Rytter D, Bech BH, et al. Maternal protein intake during pregnancy and offspring overweight 20 Y later. Am J Clin Nutr 2014; 100: 1139–48. doi:10.3945/ajcn.113.082222

⁷ Bao W, Bowers K, Tobias DK, et al. Prepregnancy dietary protein intake, major dietary protein sources, and the risk of gestational diabetes mellitus: a prospective cohort study. Diabetes Care 2013; 36: 2001–8. doi:10.2337/dc12-2018

⁸ Hassani Zadeh S, Boffetta P, Hosseinzadeh M. Dietary patterns and risk of gestational diabetes mellitus: A systematic review and meta-analysis of cohort studies. Clin Nutr ESPEN 2020; 36: 1–9. doi:10.1016/j.clnesp.2020.02.009

⁹ Chen L, Hu FB, Yeung E, et al. Prepregnancy consumption of fruits and fruit juices and the risk of gestational diabetes mellitus: a prospective cohort study. Diabetes Care 2012; 35: 1079–82. doi:10.2337/dc11-2105

¹⁰ Moeller SM, Reedy J, Millen AE, et al. Dietary patterns: challenges and opportunities in dietary patterns research an experimental biology workshop, April 1, 2006. J Am Diet Assoc 2007; 107: 1233–9. doi:10.1016/j.jada.2007.03.014

¹¹ Hu J, Oken E, Aris IM, et al. Dietary patterns during pregnancy are associated with the risk of gestational diabetes mellitus: evidence from a Chinese prospective birth cohort study. Nutrients 2019; 11: 405. doi:10.3390/nu11020405

¹² Hao Y, Qu L, Guo Y, et al. Association of pre-pregnancy low-carbohydrate diet with maternal oral glucose tolerance test levels in gestational diabetes. BMC Pregnancy Childbirth 2022; 22: 734. doi:10.1186/s12884-022-05059-2

¹³ Chen Z, Qian F, Liu G, et al. Prepregnancy plant-based diets and the risk of gestational diabetes mellitus: a prospective cohort study of 14,926 women. Am J Clin Nutr 2021; 114: 1997–2005. doi:10.1093/ajcn/nqab275

¹⁴ Millen BE, Abrams S, Adams-Campbell L, et al. The 2015 dietary guidelines advisory committee scientific report: development and major conclusions. Adv Nutr 2016; 7: 438–44. doi:10.3945/an.116.012120

¹⁵ Zamani B, Milajerdi A, Tehrani H, et al. Association of a plant-based dietary pattern in relation to gestational diabetes mellitus. Nutr Diet 2019; 76: 589–96. doi:10.1111/1747-0080.12512

¹⁶ Kidane R, Eshete T, Sintayehu T, et al. Dietary patterns and hypertensive disorders among pregnant women attending Antenatal and delivery care in public hospitals of Jimma zone, Southwest Ethiopia. J Multidiscip Healthc 2022; 15: 1645–56. doi:10.2147/JMDH.S373749

¹⁷ Qian F, Liu G, Hu FB, et al. Association between plant-based dietary patterns and risk of type 2 diabetes: A systematic review and meta-analysis. JAMA Intern Med 2019; 179: 1335–44. doi:10.1001/jamainternmed.2019.2195

¹⁸ Quek J, Lim G, Lim WH, et al. The Association of plant-based diet with cardiovascular disease and mortality: A meta-analysis and systematic review of prospect cohort studies. Front Cardiovasc Med 2021; 8: 756810. doi:10.3389/fcvm.2021.756810

¹⁹ Shahdadian F, Ghiasvand R, Abbasi B, et al. Association between major dietary patterns and Polycystic ovary syndrome: evidence from a case-control study. Appl Physiol Nutr Metab 2019; 44: 52–8. doi:10.1139/apnm-2018-0145

²⁰ He J-R, Yuan M-Y, Chen N-N, et al. Maternal dietary patterns and gestational diabetes mellitus: a large prospective cohort study in China. Br J Nutr 2015; 113: 1292–300. doi:10.1017/S0007114515000707

²¹ Mak JKL, Pham NM, Lee AH, et al. Dietary patterns during pregnancy and risk of gestational diabetes: a prospective cohort study in Western China. Nutr J 2018; 17: 107. doi:10.1186/s12937-018-0413-3

²² Kim SY, England L, Wilson HG, et al. Percentage of gestational diabetes mellitus attributable to overweight and obesity. Am J Public Health 2010; 100: 1047–52. doi:10.2105/AJPH.2009.172890

²³ Kyozuka H, Murata T, Isogami H, et al. Preconception dietary inflammatory index and risk of gestational diabetes mellitus based on maternal body mass index: findings from a Japanese birth cohort study. Nutrients 2022; 14: 4100. doi:10.3390/nu14194100

²⁴ Carolan M, Davey M-A, Biro MA, et al. Maternal age, Ethnicity and gestational diabetes mellitus. Midwifery 2012; 28: 778–83. doi:10.1016/j.midw.2011.08.014

²⁵ Zhou Z, Chen G, Fan D, et al. Size and shape of associations of OGTT as well as mediating effects on adverse pregnancy outcomes among women with gestational diabetes mellitus: population-based study from Southern Han Chinese. Front Endocrinol (Lausanne) 2020; 11: 135. doi:10.3389/fendo.2020.00135

²⁶ Lean SC, Derricott H, Jones RL, et al. Advanced maternal age and adverse pregnancy outcomes: A systematic review and meta-analysis. PLoS One 2017; 12. doi:10.1371/journal.pone.0186287

²⁷ Weschenfelder F, Lohse K, Lehmann T, et al. Circadian rhythm and gestational diabetes: working conditions, sleeping habits and lifestyle influence insulin dependency during pregnancy. Acta Diabetol 2021; 58: 1177–86. doi:10.1007/s00592-021-01708-8

²⁸ Schoenaker DAJM, Soedamah-Muthu SS, Callaway LK, et al. Pre-pregnancy dietary patterns and risk of gestational diabetes mellitus: results from an Australian population-based prospective cohort study. Diabetologia 2015; 58: 2726–35. doi:10.1007/s00125-015-3742-1

²⁹ International Association of Diabetes and Pregnancy Study Groups Consensus Panel, Metzger BE, Gabbe SG, et al. International Association of diabetes and pregnancy study groups recommendations on the diagnosis and classification of hyperglycemia in pregnancy. Diabetes Care 2010; 33: 676–82. doi:10.2337/dc09-1848

³⁰ Zhang H, Qiu X, Zhong C, et al. Reproducibility and relative validity of a semi-quantitative food frequency questionnaire for Chinese pregnant women. Nutr J 2015; 14: 56. doi:10.1186/s12937-015-0044-x

³¹ Li Y, Zhou X, Zhang Y, et al. Association of maternal dietary patterns with birth weight and the mediation of gestational weight gain: A prospective birth cohort. Front Nutr 2021; 8: 782011. doi:10.3389/fnut.2021.782011

³² Chen CM. Overview of obesity in Mainland China. Obes Rev 2008; 9 Suppl 1: 14–21. doi:10.1111/j.1467-789X.2007.00433.x

³³ Zhou X, Chen R, Zhong C, et al. Maternal dietary pattern characterised by high protein and low carbohydrate intake in pregnancy is associated with a higher risk of gestational diabetes mellitus in Chinese women: a prospective cohort study. Br J Nutr 2018; 120: 1045–55. doi:10.1017/S0007114518002453

³⁴ David LA, Maurice CF, Carmody RN, et al. Diet rapidly and reproducibly alters the human gut Microbiome. Nature 2014; 505: 559–63. doi:10.1038/nature12820

³⁵ Wang DH, You DF, Huang LL, et al. Sensitivity analysis method for unmeasured confounding interference in observational study. Zhonghua Liu Xing Bing Xue Za Zhi 2019; 40: 1470–5. doi:10.3760/cma.j.issn.0254-6450.2019.11.023

³⁶ Rathmann W, Strassburger K, Heier M, et al. Incidence of type 2 diabetes in the elderly German population and the effect of clinical and lifestyle risk factors: KORA S4/F4 cohort study. Diabet Med 2009; 26: 1212–9. doi:10.1111/j.1464-5491.2009.02863.x

³⁷ Dahl WJ, Stewart ML. Position of the Academy of nutrition and Dietetics: health implications of dietary fiber. J Acad Nutr Diet 2015; 115: 1861–70. doi:10.1016/j.jand.2015.09.003

³⁸ Du H, Li L, Bennett D, et al. Fresh fruit consumption in relation to incident diabetes and diabetic vascular complications: A 7-Y prospective study of 0.5 million Chinese adults. PLoS Med 2017; 14: e1002279. doi:10.1371/journal.pmed.1002279

³⁹ Veronese N, Watutantrige-Fernando S, Luchini C, et al. Effect of magnesium supplementation on glucose metabolism in people with or at risk of diabetes: a systematic review and meta-analysis of double-blind randomized controlled trials. Eur J Clin Nutr 2016; 70: 1463. doi:10.1038/ejcn.2016.209

⁴⁰ Chen Z, Franco OH, Lamballais S, et al. Associations of specific dietary protein with longitudinal insulin resistance, Prediabetes and type 2 diabetes: the Rotterdam study. Clin Nutr 2020; 39: 242–9. doi:10.1016/j.clnu.2019.01.021

⁴¹ Ma X, Becker D, Arena VC, et al. The effect of age on insulin sensitivity and insulin secretion in first-degree relatives of type 1 diabetic patients: a population analysis. J Clin Endocrinol Metab 2009; 94: 2446–51. doi:10.1210/jc.2008-2687

⁴² Allman BR, Diaz Fuentes E, Williams DK, et al. Obesity status affects the relationship between protein intake and insulin sensitivity in late pregnancy. Nutrients 2019; 11: 2190. doi:10.3390/nu11092190