Pre-eclampsia is a pregnancy-specific hypertension syndrome after 20 weeks’ gestation, often coming with proteinuria and oedema. This disorder is a major cause of fetal and maternal morbidity and mortality worldwide, which probably affects 3–5 % of all pregnant women(Reference Ananth, Keyes and Wapner1). Although endothelial dysfunction and immunologic aberrations have been considered as important contributors to pre-eclampsia, the definite pathophysiology is not yet fully understood(Reference Rana, Lemoine and Granger2). Until now, delivery is still the only effective treatment for pre-eclampsia, while therapeutic drugs are not routinely recommended except for low-dose aspirin for high-risk women(3). Given limited treatment options, modifications of dietary patterns have emerged as a potential intervention to prevent the development of pre-eclampsia(Reference Ikem, Halldorsson and Birgisdóttir4–Reference Schoenaker, Soedamah-Muthu and Callaway6).
Previous researches have investigated the potential effects of single nutrient supplementation on pre-eclampsia development. Unfortunately, fish oil(Reference Zhou, Yelland and McPhee7), folic acid(Reference Wen, White and Rybak8), antioxidants such as vitamins C and E(Reference Sheikhi, Sharifi-Zahabi and Paknahad9), and vitamin D(Reference Palacios, Kostiuk and Peña-Rosas10) studies have failed to provide anticipated clinical benefits. Ca supplementation seemingly exhibits protection from pre-eclampsia, with the greatest effect among women with low-baseline Ca intake(Reference Hofmeyr, Lawrie and Atallah11). These findings suggest the complexity of nutritional intervention to reduce pre-eclampsia risk, especially in consideration of that food and nutrients are generally consumed in various characteristic combinations that synergistically influencing diseases(Reference Jacobs and Steffen12). Taken together, evidences estimating the overall effects of dietary patterns on pre-eclampsia risk might provide a better understanding regarding the relationship between modifiable dietary factors with pre-eclampsia.
Recently, several studies have assessed the effect of overall dietary pattern on the risk of pre-eclampsia(Reference Ikem, Halldorsson and Birgisdóttir4,Reference Schoenaker, Soedamah-Muthu and Callaway6,Reference Brantsaeter, Haugen and Samuelsen13) A Norway study found that women with high scores on a dietary pattern characterised by vegetables and plant foods are less likely to suffer from pre-eclampsia than those without(Reference Brantsaeter, Haugen and Samuelsen13). Moreover, a meta-analysis mainly synthesising data from developed countries showed that dietary pattern with a higher intake of fruits, vegetables, whole grains and fish is associated with a decreased likelihood of pre-eclampsia(Reference Kibret, Chojenta and Gresham5). Similar evidences in Chinese population, who are thought to have entirely different dietary habits from western population(Reference Wang, Lin and Bloomgarden14), are still limited. Only a post hoc analysis of a cluster randomised controlled trial conducted in northwestern China showed a positive association between vegetable dietary pattern and reduced risk of pre-eclampsia(Reference Mi, Wen and Li15). However, considering the small number of incident pre-eclampsia cases (n 19) in the above study, along with the lack of adjustment for confounding factors, further researches are warranted to clarify the relationship between different dietary patterns and pre-eclampsia risk in this specific population.
Therefore, we conducted a matched case–control study to investigate the associations between dietary patterns derived from the principal components factor analysis and the odds of pre-eclampsia in a central province of China.
Methods
Study population
This was a 1:1 matched case–control study conducted between March 2016 and February 2019 in the First Affiliated Hospital of Zhengzhou University, China. Eligible cases with singleton pregnancy and pre-eclampsia were included. The diagnosis of pre-eclampsia was confirmed by the presence of both gestational hypertension and proteinuria according to the Diagnosis and Treatment Guideline of Hypertensive Disorders in Pregnancy (2015)(16). Subjects were excluded if they had any of the following: (1) patients with gestational hypertension alone, eclampsia or pregnancy complicated by chronic hypertension; (2) patients with heart disease, malignancy, hyperthyroidism, immune system diseases, chronic renal insufficiency or other endocrine system diseases and (3) patients with epilepsy, depression or other mental/cognitive dysfunction.
For each participating case, we attempted to enrol one matched healthy control according to age (±3 years) and gestational weeks (±1 week), with the same inclusion and exclusion criteria applying, except for those with a history of pre-eclampsia. All of the controls were hospitalised during the same period. All the participants provided signed informed consent before being interviewed. The details of study flow chart are presented in Fig. 1. This study was conducted according to the guidelines laid down in the Declaration of Helsinki, and all procedures involving human subjects/patients were ethically approved by the First Affiliated Hospital of Zhengzhou University Ethics Committee (No. Scientific research-2016-LW-34).
Study flow chart for the inclusion and exclusion process.
Data collection
The individual information was collected by trained interviewers (masters of public health) using a structured questionnaire on socio-demographic characteristics, lifestyle habits, pre-pregnancy weight, history of disease, menstrual and childbearing. Participants’ weight and height were measured by a trained researcher. The detailed information has been described in previous study(Reference Cao, Liu and Zhao17).
Dietary assessment and food grouping
Dietary consumption information was assessed using a valid seventy-nine-item FFQ(Reference Zhang and Ho18), to evaluate the food intake during the last 3 months before recruitment. For each food item, four possible frequencies (never, per month, per week and per day) and the amount of consumption each time in gram or millilitre are available. Photographs of food portion sizes were used to help estimate the amount of food consumption. The intake of energy (kcal/d) and key nutrients from each food were calculated according to the Chinese Food Composition Tables(Reference Yang19,Reference Yuexin and Yaguang20) . To reduce the complexity of the data, the seventy-nine food items were aggregated into twenty-eight groups according to culinary usage and the similarities of the nutritional composition to identify dietary patterns.
Statistical analyses
Demographic characteristics and other potential pre-eclampsia risk factors were compared using paired t tests (continuous variables) or paired χ 2 tests (categorical variables) between cases and controls. The residual method was used to adjust total energy intake for each food item(Reference Willett and Stampfer21). Factor analysis with factor loadings extracted using the principal component method and varimax/orthogonal rotation was used to derive the dietary patterns for the twenty-eight standardised, energy-adjusted groups. Four factors were retained based on an inspection of the scree plots, and those factors were rotated by orthogonal transformation (varimax procedure) to facilitate their interpretability. Variables that have loadings of 0·25 or greater are considered to contribute significantly to the dietary pattern(Reference Schulze, Hoffmann and Kroke22). Factor scores are calculated for each derived patterns by summarising all the food groups(Reference Wirfält, Drake and Wallström23). Factors are numbered and given temporary labels according to the food groups that loaded highly on the dietary pattern.
Conditional logistic regressions were used to analyse the association between dietary pattern scores and pre-eclampsia. For each factor, the bottom quartiles group (Q1) was defined as the reference group. Two multivariate logistic regression models were introduced to calculate OR with their corresponding 95 % CI: model 1: without any adjustment; model 2: adjusted for age, gestational week, education, income, pre-pregnancy BMI, passive smoking during pregnancy, folic acid supplement and physical activity.
Subgroup analyses were conducted to investigate whether the associations of dietary patterns differed across various subgroups classified by BMI, physical activity level and family history of hypertension or other vascular disease (including hypertension, coronary artery disease, cerebrovascular diseases, gestational hypertension, pre-eclampsia and eclampsia). All of the statistical analyses were performed using SPSS software (version 22.0), and a two-sided P value < 0·05 was considered statistically significant.
Results
The demographic and other characteristics are shown in Table 1. A total of 440 pre-eclampsia cases with a mean age of 30·88 (sd 5·03) years old were included. In addition, 440 eligible age- (±3 years) and gestational week- (±1) matched controls (age: 31·03 (sd 4·85) years) recruited from the same hospital were also enrolled. Compared with the controls, the cases were more likely to have a lower educational level, a lower serum Ca and vitamin D level but a higher pre-pregnancy BMI and total energy intake. The family history of hypertension or other vascular disease was more frequently seen in the cases group. Cases and controls did not differ in age, gestational weeks, physical activity, multivitamin use, Ca supplementation, vitamin D supplementation, folic acid supplementation or passive smoking.
Demographics, lifestyle characteristics and pre-eclampsia risk factors of the study population (Mean values and standard deviations; numbers and percentages)
MET, metabolic equivalent.
* Pre-pregnancy BMI was defined as pre-pregnancy body weight (kg)/square of height (m2).
† Physical activities included daily occupational, leisure time and household chores, evaluated by metabolic equivalent hours per day.
‡ Passive smoking was defined as exposure to second-hand smoke for 1 or more cigarettes daily for at least 6 consecutive months during the past year.
Four main factors were retained from the factor analysis. The factor loadings associated with each dietary pattern are shown in Table 2. Each pattern was labelled according to the food groups with high absolute loadings. The first factor, named the high fruit-vegetable pattern, was characterised by a high intake of fruits, vegetables, and molluscs and shellfish. The second factor, named as the high protein pattern, was characterised by a high intake of poultry without skin, high fat milk, mushroom and algae, nuts, and soyabean and its products. The third factor, named as high fat-grain pattern, represented a high intake of poultry with skin, cooking oil, red meat and whole grain. The fourth factor, named as the high salt-sugar pattern, was characterised by a high intake of soft drinks, processed vegetable and processed meat. Factors 1–4 accounted for 7·97, 7·52, 5·08 and 5·06 % of the variability, respectively, and 25·63 % of the variance in food intake overall.
Varimax-rotated food group factor loading scores *
* Dietary patterns were determined by principal components factor analysis. Factor loadings with absolute values ≥ 0·25 were listed in the table among twenty-eight food groups.
After adjustments for age, gestational week, education, income, pre-pregnancy BMI, passive smoking during pregnancy, folic acid supplement and physical activity, the high fruit-vegetable pattern was associated with a decreased odds of pre-eclampsia, with OR for comparison between the highest v. lowest quartile scores of 0·71 (95 % CI 0·55, 0·92) (P trend = 0·013). Significant association was also observed between the high protein dietary pattern and the odds of pre-eclampsia (Q4 v. Q1, OR 0·72, 95 % CI 0·54, 0·95, P trend = 0·011). Whereas high fat-grain dietary pattern showed a numerical U-shaped association with pre-eclampsia, the lowest odds of pre-eclampsia was observed in the third quartile (Q3 v. Q1, OR 0·75, 95 % CI 0·57, 0·98, P trend = 0·111). No significant association was observed for the relationship of high salt-sugar pattern (P trend = 0·335) with pre-eclampsia (Table 3).
Association of quartiles of dietary patterns with the odds of pre-eclampsia (Numbers, odds ratios and 95 confidential intervals)
Model 1: without any adjustment.
Model 2: adjusted for age, gestational week, education, income, pre-pregnancy BMI, passive smoking, folic acid supplementation and physical activity.
Table 4 shows the results of subgroup analyses stratified by BMI, physical activity and family history of hypertension or other vascular disease. The associations of dietary patterns and the odds of pre-eclampsia were similar for most strata. However, stronger association between high fruit-vegetable pattern and the odds of pre-eclampsia ((Q4 v. Q1, OR 0·58, 95 % CI 0·38, 0·86, P trend = 0·010) for participants with a below-median level of physical activity; (Q4 v. Q1, OR 0·87, 95 % CI 0·62, 1·24, P trend = 0·488) for participants with a median or above level of physical activity (P interaction = 0·031)) seemed to exist among participants with a below-median level of physical activity. Similarly, those participants also have a stronger association between high protein dietary pattern and prevalent pre-eclampsia ((Q4 v. Q1, OR 0·51, 95 % CI 0·33, 0·81, P trend = 0·001) for participants with a below-median level of physical activity; (Q4 v. Q1, OR 0·87, 95 % CI 0·61, 1·26, P trend = 0·442) for participants with a median or above level of physical activity (P interaction = 0·022)). Besides, an interaction effect between high protein pattern and family history of hypertension or other vascular disease was also observed ((Q4 v. Q1, OR 0·55, 95 % CI 0·38, 0·79, P trend = 0·001) for participants without family history of hypertension or other vascular disease; (Q4 v. Q1, OR 1·05, 95 % CI 0·68, 1·64, P trend = 0·969) for participants with family history of hypertension or other vascular disease (P interaction = 0·041)).
Stratified analyses for the association between dietary patterns and the odds of pre-eclampsia (Numbers, odds ratios and 95 confidential intervals)
* Adjusted for age, gestational week, education, income, passive smoking, folic acid supplementation and physical activity.
† Adjusted for age, gestational week, education, income, pre-pregnancy BMI, passive smoking, folic acid supplementation and physical activity.
‡ Adjusted for age, gestational week, education, income, pre-pregnancy BMI, passive smoking and folic acid supplementation.
Discussion
In this first matched case–control study of Chinese pregnant women, we investigated the associations between dietary patterns and prevalent pre-eclampsia by the means of factor analysis. We found that high fruit-vegetable dietary pattern and high protein dietary pattern were significantly associated with a lower odds of pre-eclampsia, even adjusting for confounding factors. Moreover, we also reported the potential modified effect of physical activity and family history of vascular disease on the association of dietary patterns with pre-eclampsia.
Increasing studies exploring the associations between dietary patterns and pre-eclampsia risk have shown that high fruit-vegetable dietary pattern was a protective factor for pre-eclampsia(Reference Brantsaeter, Haugen and Samuelsen13,Reference Mi, Wen and Li15,Reference Hillesund, Øverby and Engel24,Reference Hajianfar, Esmaillzadeh and Feizi25) . The Norwegian Mother and Child Cohort Study with more than 70 000 pregnancy women found that the pattern characterised with high consumption of vegetables, plant foods and vegetable oils was associated with reduced risk of pre-eclampsia(Reference Brantsaeter, Haugen and Samuelsen13). The Danish National Birth Cohort found an inverse association between a seafood diet pattern featuring high fish and vegetables and pre-eclampsia development(Reference Ikem, Halldorsson and Birgisdóttir4). Although living with different dietary habits(Reference Wang, Lin and Bloomgarden14), our study conducted in central China also supports the hypothesis that high fruit-vegetable diet might be beneficial for prevention of pre-eclampsia. Comparing to a previous post hoc analysis(Reference Mi, Wen and Li15) conducted in northwestern China, we revealed a smaller effect size of high fruit-vegetable dietary pattern on the decreased odds of pre-eclampsia, this could be partly attributed to different study design and younger participants in their study.
In the present study, the high fruit-vegetable dietary pattern characterised by high intake of fruits, vegetables, molluscs and shellfish is similar to traditional Mediterranean diet, which has been evidenced to reduce risk of several chronic disease(Reference Romagnolo and Selmin26,Reference Chauveau, Aparicio and Bellizzi27) . Previous studies have demonstrated that this dietary pattern can reduce inflammatory markers level and improve insulin resistance(Reference Berg, Lappas and Strandhagen28,Reference Esmaillzadeh, Kimiagar and Mehrabi29) , which were significantly involved in the pathophysiology of pre-eclampsia(Reference Ives, Sinkey and Rajapreyar30). In addition, diets rich in fruit and vegetables have been reported to be inversely associated with plasma homocysteine level(Reference Duthie, Duthie and Russell31), a marker promoting pre-eclampsia development(Reference Maged, Saad and Meshaal32,Reference Sun, Qian and Zhang33) .
Interventional trials have failed to build a definite relationship between protein supplementation and pre-eclampsia in pregnant women(Reference Xu, Shatenstein and Luo34). A balanced protein-energy supplementation (protein content < 25 % of total energy content) also showed no effects on reducing pre-eclampsia risk in a meta-analysis(Reference Kramer and Kakuma35). In those included studies, however, alternate treatment allocation and a number of participants lost to follow-up had largely weaken their findings. Interestingly, we found a potential protective role of high protein dietary pattern in pre-eclampsia development among Chinese pregnant women. Different from protein supplementation, the dietary protein sources in this study are more diverse. Some specific components like fruit, nuts and sea fish in the high protein dietary pattern have been found to decrease the odds of pre-eclampsia(Reference Xu, Shatenstein and Luo34). More importantly, a relatively high protein intake might reduce the energy demand of carbohydrate and fat, contributing to a balanced protein-energy dietary model(Reference Clausen, Slott and Solvoll36). Besides, it should be noted that a family history of hypertension or other vascular disease might attenuate the benefits of high protein intake. Considering the remarkable relationship between family history of vascular disease and pre-eclampsia(Reference Ness, Markovic and Bass37,Reference Qiu, Williams and Leisenring38) , it is reasonable to speculate that dietary factors alone may not be sufficient to influence the pre-eclampsia development in pregnant women under high-risk.
Our study found a numerical but not significant U-shaped association between high fat-grain dietary pattern and prevalent pre-eclampsia, suggesting a probable threshold effect in the present analysis. In contrast, the Generation R Study and Danish National Birth Cohort study found that the traditional Northwest European dietary pattern characterised as high intake of red meat and potato was associated with increased odds of pregnancy hypertensive disorders(Reference Ikem, Halldorsson and Birgisdóttir4,Reference Timmermans, Steegers-Theunissen and Vujkovic39) . Considering that a non-linear relationship exists between total fat intake and CVD in general population(Reference Zhu, Bo and Liu40), further studies are still warranted to find the optimum dose of fat intake to prevent pre-eclampsia in Chinese pregnant women.
Strengths and limitations
The present study has several strengths. First, the investigation was the first matched case–control design with a large number of pre-eclampsia cases to find that both high fruit-vegetable dietary pattern and high protein dietary pattern were significantly associated with lower odds of pre-eclampsia. Additionally, multivariable adjustments for demographics and lifestyle factors enhance the robustness of the results. These findings add important information from Chinese population on the role of diet in pre-eclampsia and shed light on future strategy of dietary management during pregnancy. However, several limitations should be simultaneously acknowledged. First, the results derived from a case–control study make it impossible for us to investigate the causal relationships between dietary pattern and pre-eclampsia development. Second, given relatively small sample size, the results’ interpretations regarding the associations should be generalised with caution. Third, our findings may be limited by recall bias during the interview for the questionnaire and selection bias due to selection of hospital-based controls. Moreover, cases were more likely to change their dietary habits after diseases. To make the measurements more accurate, trained interviewers used photographs of food portion sizes to administer the FFQ through face-to-face interview. And to decrease the recall bias, we only collected the dietary intake of the pregnant women during the 3 months before delivery. Finally, it is well known that FFQ could bring a certain degree of measurement error even though we used the FFQ which has been previously validated(Reference Zhang and Ho18).
Conclusion
Our findings indicate that high fruit-vegetable dietary pattern and high protein dietary pattern might protect against pre-eclampsia in Chinese pregnant women. Further prospective cohort studies with larger sample size are needed to verify these findings.
Acknowledgements
None.
This study was supported by the National Natural Science Foundation of China (grant no. 81602852).
Y. L. and X. Z. conceived and designed the study; Y. Z., W. F., Y. C. and D. D. conducted the research; Q. J. L., F. F. Z. and W. F. D. performed the statistical analysis. Y. Z. and Y. C. drafted the manuscript. All authors revised the manuscript for important intellectual content and approved the final version.
There are no conflicts of interest to declare.
