Dietary carotenoid intake and risk of developing preeclampsia: a hospital-based case–control study

Preeclampsia (PE) is a major pregnancy complication with a worldwide incidence rate of 2.61%, as reported by the World Health Organization [1]. PE occurs in 4–5% of pregnancies worldwide [2] and leads to 75,000 maternal deaths each year [3]. It is also one of the main causes of maternal, foetal and neonatal deaths in low- and middle-income countries [4]. The main clinical manifestations of PE are hypertension, proteinuria and oedema after 20 weeks of gestation, which in cases of severe PE may be accompanied by systemic multiple organ damage; heart failure; haemolysis, elevated liver enzymes and low platelets syndrome; pulmonary oedema; placental abruption; and systemic small vessel spasm. Furthermore, PE substantially increases the risk of adverse pregnancy outcomes, such as foetal growth restriction, premature delivery and stillbirth [5]. There are currently no effective treatments for PE, with premature delivery being almost ineluctable in cases of severe PE. Therefore, identifying risk factors that can help prevent the occurrence of PE is important to reduce morbidity and mortality. The current known risk factors for PE include genetic susceptibility, family history, number of pregnancies, maternal age, maternal smoking, pre-pregnancy body mass index (BMI), the use of assisted reproductive technology and maternal comorbidity condition, such as diabetes, chronic kidney disease and systemic lupus erythematosus [6, 7]. Considering dietary intake can be easily controlled and given, studies on the relationship between dietary factors (energy, nutrients, foods or overall dietary patterns) and PE have attracted more and more attention [8].

The pathophysiology of PE is not clear, but an imbalance between antioxidants and pro-oxidants is known to contribute to the development of the condition [2]. Carotenoids are yellow, orange and red fat-soluble pigments that are widely found in microorganisms, plants, animals and the human body [9]. Carotenoids serve as one of main antioxidant defence systems in the human body [10] and thus play an important role in protection from oxidative stress [11‐13]. A recent review noted that the association between carotenoid subclasses and the risk of developing PE has been inconclusive [14]. A meta-analysis of 58 studies, which comprised cohort studies, case–control studies, cross-sectional studies and randomised controlled trials (RCTs), found there was limited and non-definitive evidence that women with PE had low concentrations of carotenoids [15]. Another meta-analysis of RCTs found that oral antioxidant supplements, including lycopene, could not effectively prevent the occurrence of PE [16]. The differences between these meta-analyses’ findings may be partly attributable to differences between the dosages, intervention start times, sample sizes and maternal health status in studies, which demonstrates that more research is required in this field.

Carotenoids can also affect the growth and health of infants, especially their visual and cognitive development [17]. In addition, a recent review recommended that women should consume a diet rich in carotenoids during pregnancy, although there is no current recommendation on carotenoid dietary intake [14]. The few studies that have measured dietary carotenoid intake during pregnancy show that intake varies between study populations [18‐20]. There are scarce data from observational studies showing the association between the risk of developing PE and the intake of dietary carotenoids and related compounds, we assessed this relationship in pregnant Chinese women via a hospital-based case–control study.

To the best of our knowledge, this was the first study to assess the relationship between the risk of developing PE and the intake of dietary carotenoids and related compounds in China. After adjusting for potential confounding factors, we found that a high intake of total carotenoids, β-carotene, β-cryptoxanthin, lycopene and lut-zea was associated with a reduced risk of developing PE. The highest quartile intakes of dietary carotenoids, except that of α-carotene, were associated with a 45% or greater reduction in the risk of developing PE than the other quartiles, which indicated a significant dose-response trend. The carotenoid profiles were considered to be relatively complete in this study, and represented approximately 95% of the carotenoids in the blood [27]. We included six of the most studied categories: α-carotene, β-carotene, β-cryptoxanthin, lycopene, and lut-zea.

Our results for total carotenoids, α-carotene, β-carotene and lutein were consistent with those of several previous studies that have reported that the total carotenoid concentrations in the sera of patients with mild and severe PE were significantly lower than those in the sera of healthy pregnant women [28, 29]. Our results for β-carotene were similar to those of a case–control study conducted by Yusuf, et al. [30] in Jordan, which used three 24-h dietary records to compare dietary carotenoid intake of a PE and a control groups. It should be noted that only correlation analysis (Chi-square test) was performed in their study, but further correlation strength (OR) analysis was not. In addition, dietary recording may lead to changes in eating behavior in the process of recording food intake, suggesting that bias may exist. In contrast, FFQ covers a longer time span, so it may better reflect habitual intake than dietary record, and may be more suitable for epidemiological studies on the relationship between dietary exposure and chronic diseases. A case–control study conducted in Zimbabwe (n = 173/186) showed that higher concentrations of β-carotene in the blood were inversely associated with the risk of developing PE [25]. Cohen et al. [31] found an inverse association between lycopene, lutein and total carotenoid concentrations and the risk of developing PE, but lutein was the only carotenoid that was significantly associated after adjusting for confounders. Zhang et al. [32] did not find any relationships between plasma α-carotene concentrations and the risk of developing PE.

The current study has some inconsistencies with previous studies on β-cryptoxanthin and lycopene. For example, Zhang et al. [32] did not find any associations between the risk of developing PE and the intake of β-cryptoxanthin and lycopene (n = 125/179). Notably, the blood samples in that study were collected only when the clinical manifestations of PE became obvious. In addition, the gestational age at the time of blood collection from the case and control groups in their study was different (36.0 ± 0.3 weeks vs. 37.3 ± 0.3 weeks, P <  0.05). Similarly, two RCTs [33, 34] that have been conducted in India did not find that lycopene intake reduced the incidence rate of PE (n = 44/54 and 159/159, respectively). However, in the current study, we found that a high lycopene intake had an unambiguously protective effect against the development of PE (lycopene intakes in the case and control groups were 2.18 mg and 2.93 mg, respectively). Our results are similar to those of another RCT (n = 116/135) conducted in India [35], which found that fewer pregnant women developed PE in the intervention group supplemented with 4 mg lycopene per day from 16 to 20 weeks compared to the placebo group (intervention vs. placebo: 8.6% vs. 17.7%, P = 0.04). Subsequently, blood samples from 50 PE pregnant women and 50 healthy pregnant women in this Indian study revealed that those with PE had significantly lower blood concentrations of lycopene than those without PE [36]. Finally, a meta-analysis conducted in 2018 of RCTs of oral antioxidant therapy for the prevention and treatment of PE found that oral lycopene supplementation of 2 mg/d or 4 mg/d from the second trimester to delivery did not prevent the occurrence of PE (relative risk [RR] = 0.67, 95% CI: 0.44–1.04, P = 0.08) [16]. Many factors could be responsible for these differences between study results, such as the study design, the treatment dose, the season, and pregnant women’s ethnicity and health status (high or low PE risk). Previous studies have found that pregnant women who are younger [12], smoking [19], and getting pregnant in the spring and summer [20] consume fewer carotenoids. In view of the results of this current study, future large-scale studies are required on the dietary carotenoid intake of pregnant women with PE.

Although most of the studies only measured blood concentrations and did not collect dietary data, they indirectly confirm our results, as the human body cannot synthesise carotenoids and thus the carotenoids in blood are derived from the diet. The Norwegian Mother and Child Cohort Study (MoBa) found that the plasma carotenoid concentrations of pregnant women were significantly correlated with food consumption, especially of fruits (r = 0.24; p <  0.01) and vegetables (r = 0.32; p <  0.01), which lends validity to our method [37]. In their study, strong correlations between carotenoids and food intake were found between a-carotene and carrots (r = 0.50; p < 0.01) and cooked vegetables (r = 0.39; p < 0.01). A meta-analysis [38] of RCTs on the effect of changes in fruit and vegetable intake on carotenoid blood concentration showed that four common carotenoids (α-carotene, β-carotene, β-cryptoxanthin and lutein) may be used as biomarkers to objectively measure general fruit and vegetable intake.

Currently, the pathogenesis of PE is thought to involve two stages: reduced placental perfusion and secondary maternal multisystem impairment [2]. Placental ischaemia and reperfusion injury caused by insufficient remodelling of the uterine spiral artery is considered to be a cause of oxidative stress [39‐41], which has also been confirmed by animal [42] and in vitro experiments [43]. The injury of vascular endothelial function caused by oxidative stress and inflammatory media leads to the decrease of vasodilators and the increase of vasoconstrictors, thus promoting vasospasm [44, 45]. It is clear that oxidative stress, a basic pathophysiological mediator of PE, plays an important role in the occurrence of PE. In truth, oxidative stress plays a key role in the development of both normal and defective placentas [2, 46]. However, there appears to be an imbalance between antioxidant and pro-oxidant mechanisms in PE. Some studies have confirmed that pregnant women with PE have higher concentrations of markers of oxidative damage and lower concentrations of antioxidants and lower total antioxidant capacity than healthy pregnant women [36, 47, 48]. The human body has various antioxidants, both enzymatic and non-enzymatic, to counter the effects of oxidants. Among non-enzymatic antioxidants, nutrients such as vitamin C, vitamin E and carotenoids play a crucial role in the human antioxidant system [10]. Previous studies have found that the intake of antioxidant nutrients affects oxidative stress [49, 50], with an insufficient intake increasing the risk of higher levels of oxidative stress and leading to poor obstetric outcomes [51, 52]. Two Cochrane reviews evaluated the effectiveness of vitamin C and E alone or in combination in the prevention of PE, but no benefit was found for routine supplementation of vitamin C and/or E [53, 54]. Carotenoids contain many double bonds that can quench singlet oxygen and scavenge free radicals and oxides, which gives them significant antioxidant activity [27]. The demand for micronutrients becomes more sensitive during pregnancy and inadequate intake may affect the mother and foetus [55]. Therefore, it is important to fully understand the role of dietary carotenoids in the occurrence of PE.

In addition to antioxidant nutrients, vitamin D supplementation has been proposed for the prevention of PE, although evidence has been inconsistent [56]. Our previous research observed a negative relationship between dietary VD intake and PE risk [57]. Vitamin D is a regulator of inflammation and may affect the occurrence of PE by regulating maternal immune response and reducing the concentrations of proinflammatory cytokines. Furthermore, Yusuf, et al. [30] found a strong positive association exists for the intake of fat and negative association for the vegetables and fruits with PE. A systematic literature review and meta-analysis confirmed that high intake of fruits and vegetables had a protective effect on PE [8]. Brantsæter et al. [58] found that “vegetable dietary pattern” characterized by vegetables, plant-based foods and vegetable oils could reduce the risk of PE, whereas a dietary pattern characterized by a high consumption of processed meat, sweet drinks, and salty snacks increases the likelihood of PE. The possible reason is plant foods are rich in micronutrients (phytochemicals, antioxidants, vitamins and minerals) and dietary fibre. Moreover, a review pointed out that high intake of carotenoids may be related to a healthier diet and lifestyle, which may be beneficial in themselves [14]. Considering the role of these dietary related factors, a sensitivity analysis was performed on the basis of general confounding factors. But the results indicated that a high intake of total carotenoids, β-carotene, β-cryptoxanthin and lut-zea was still associated with a reduced risk of developing PE, which may reflect that carotenoids intake can play a role as a predictor in the risk of PE, independent of the effects of other nutritional factors.

Some limitations of the study should be acknowledged. First, the causality of the studied association remained uncertain due to the inherent limitation of the retrospective study design of case–control study. Second, according to the rare disease assumption in the case-control study, OR might not well estimate RR (relative ratio) for the low incidence (but not rare) disease of PE in China (about 3%) [59]. The First Affiliated Hospital of Zhengzhou University is the most authoritative comprehensive hospital in Henan Province and has treated enough PE patients in this area. We recruited almost all incident cases which admitted in the target hospitals on workdays, but some cases admitted on holidays might be missed. For hospital-based case-control studies, it’s really hard to exactly identify the source population of the cases and controls. We compared the distribution of the source of the study population (Zhengzhou and other regions) in the case group and the control group, and found that the proportion of the two groups from Zhengzhou was about 20%, suggesting that the two groups were comparable. Thus, in this case, we believe OR may be an acceptable choice for estimating of the incidence rate ratio. Third, the data on dietary carotenoid intake were based on the recall of the participants, which means that recall bias was unavoidable. To minimise recall bias, we took the following measures: (1) participants’ dietary intakes were assessed in face-to-face interviews with trained investigators; (2) detailed food pictures were used to help the participants recall their eating habits; and (3) we investigated dietary intake in only the 3 months immediately preceding delivery, during which period the diet is relatively stable compared to other periods of pregnancy. This meant that our study may be unable to explain the effect of diet in early pregnancy on the occurrence of PE, but given that it is often less than 3 months from the beginning to delivery of PE patients, we believe that this study reflects the relationship between the risk of developing PE and the intake of dietary carotenoids and related compounds to a certain degree. Finally, instead of detecting serum carotenoids, which can be traumatic, low-cooperative and expensive, we investigated dietary carotenoid intake in a non-invasive and low-cost manner using an FFQ. As mentioned above, considering the consistency between dietary and circulating carotenoid levels and the significance of measuring carotenoid levels from dietary sources, this study may guide pregnant women’s dietary behavior and contribute to the prevention of PE.

Our results indicate that high intakes of total carotenoids, β-Carotene, β-cryptoxanthin, lycopene and lut-zea may be associated with a reduction in the risk of developing PE. Larger-scale studies are needed to verify this relationship. In the future, we recommend that pregnant women consume carotenoid-rich foods during pregnancy to prevent the occurrence of PE.