Association between Maternal Zinc Status, Dietary Zinc Intake and Pregnancy Complications: A Systematic Review

Adequate zinc stores in the body are extremely important during periods of accelerated growth. However, zinc deficiency is common in developing countries and low maternal circulating zinc concentrations have previously been associated with pregnancy complications. We reviewed current literature assessing circulating zinc and dietary zinc intake during pregnancy and the associations with preeclampsia (PE); spontaneous preterm birth (sPTB); low birthweight (LBW); and gestational diabetes (GDM). Searches of MEDLINE; CINAHL and Scopus databases identified 639 articles and 64 studies were reviewed. In 10 out of 16 studies a difference was reported with respect to circulating zinc between women who gave birth to a LBW infant (≤2500 g) and those who gave birth to an infant of adequate weight (>2500 g), particularly in populations where inadequate zinc intake is prevalent. In 16 of our 33 studies an association was found between hypertensive disorders of pregnancy and circulating zinc; particularly in women with severe PE (blood pressure ≥160/110 mmHg). No association between maternal zinc status and sPTB or GDM was seen; however; direct comparisons between the studies was difficult. Furthermore; only a small number of studies were based on women from populations where there is a high risk of zinc deficiency. Therefore; the link between maternal zinc status and pregnancy success in these populations cannot be established. Future studies should focus on those vulnerable to zinc deficiency and include dietary zinc intake as a measure of zinc status.


Introduction
Adequate maternal nutrition, particularly before and during pregnancy, is imperative to the health of both the mother and child [1,2]. Poor nutrition in pregnancy may lead to inappropriate nutrient partitioning between the mother and fetus, which can be deleterious to the health of both [3]. Each year, 3.5 million deaths in women and children are attributed to undernutrition [4]. Zinc deficiency is predicted to be responsible for 1% of all deaths globally and 4.4% of deaths in children aged 6 months to 5 years [5]. The World Health Organization (WHO) prioritized minimizing zinc deficiency in developing nations as part of the Millennium Development Goal 1: to eradicate extreme poverty and hunger [6]. Therefore, understanding the effects of zinc deficiency on pregnancy and fetal growth is very important.
Zinc is an essential component of over 1000 proteins including antioxidant enzymes, metalloenzymes, zinc-binding factors and zinc transporters. These are required for a variety of biological processes including carbohydrate and protein metabolism, DNA and RNA synthesis, cellular replication and differentiation, and hormone regulation [7][8][9][10]. The importance of zinc to the growth of the fetus is demonstrated by the active transport of zinc across the placenta into the fetal circulation resulting in higher cord blood concentrations compared to those in the maternal circulation [11][12][13][14]. Rodent models of severe maternal zinc deficiency show increased rates of fetal loss and congenital malformations in the surviving fetuses [15] as well as reduced fetal growth [16][17][18], lower implantation rates and impaired placental growth [19], all highlighting the teratogenic effects of zinc deficiency in pregnancy.
Diet is the main factor that determines zinc status [20]. In the United States and Australia, an additional 2-4 mg zinc per day is recommended for pregnant women compared to non-pregnant women [21,22]. It is widely acknowledged that many pregnant women do not meet this recommendation [23][24][25], particularly in developing countries where diets are often plant-based. Grains and legumes contain a significant amount of phytic acid and phytate binding of zinc limits its absorption in the small intestine, contributing to zinc deficiency [22]. Estimates based on bioavailability of zinc, physiological requirements and predicted zinc absorption suggest the prevalence of zinc deficiency to range from 4% (European countries including the United Kingdom, Sweden, Germany and France) to 73% in Bangladesh, India and Nepal [26]. A more recent evaluation, based on similar estimates, also predicted inadequate zinc intakes in over 25% in populations in Southeast Asia and Africa [27].
A recent Cochrane review assessed the effects of zinc supplementation versus no supplementation (with or without placebo) on the success of pregnancy in 21 randomized controlled trials (RCTs) [28]. It was concluded that zinc supplementation reduced the risk of spontaneous preterm birth (sPTB) by 14% (RR: 0.86, 95% CI: 0.76-0.97; 16 RCTs) but there was no effect on other outcomes such as stillbirth/neonatal death, birthweight and pregnancy-induced hypertension [28]. However, this review did not include the effects of zinc supplementation on reducing the risk of gestational diabetes (GDM) and analysis of maternal circulating zinc concentrations provides evidence that low maternal zinc may be associated with GDM, as well as preeclampsia (PE), gestational hypertension (GH), sPTB and infant birthweight [24,29]. The association between serum zinc and PE has been reviewed recently [30] but there has been no extensive review that has assessed maternal zinc concentrations with respect to a range of pregnancy complications. Here, we review the current literature based on observational studies assessing the association between maternal zinc status and a number of pregnancy complications in order to determine whether maternal circulating zinc or dietary zinc intake are important factors associated with pregnancy outcome.

Eligibility Criteria
Studies included human prospective cohorts, case-control, longitudinal and cross-sectional studies assessing maternal circulating zinc concentrations and pregnancy complications including PE, eclampsia, GH, GDM, small for gestational age (SGA), intrauterine growth restriction (IUGR; <10th percentile), low-birthweight (LBW; ≤2500 g) and sPTB. Only studies that measured maternal circulating zinc during pregnancy or at delivery and/or dietary zinc intake at these times were included. Studies that assessed zinc concentrations in placenta, amniotic fluid, in offspring (post-natally), cord blood only and breast milk were excluded. There were no restrictions imposed on age of women included in the studies or on any other population characteristic such as race or body mass index (BMI). Given the heterogeneity of the observational strategies, a meta-analysis was not possible.

Information Sources and Search
The search strategy and procedure was guided by the PRISMA statement [31]. Potential studies were located through electronic databases (Ovid Medline (1946-present), CINAHL (1937-present) and Scopus (1995-present)), as well as manual searches of references in review articles and relevant articles known by the authors. Limits included full text articles written in English and published in academic journals. The last search was performed on 25 August 2016. Search terms and MeSH headings in the title, abstract, and index terms, were initially identified in Medline and subsequent key words were used for the remaining databases (Appendix B). Briefly, the search included the following: zinc; dietary zinc; zinc intake; plasma zinc; serum zinc; preeclampsia; eclampsia; gestational hypertension; gestational diabetes mellitus; fetal macrosomia; small for gestational age; intrauterine growth restriction; low birthweight; preterm birth.

Data Collection
An independent search of the literature was performed in April 2015 and again in August 2016. Titles and abstracts were examined independently by two of the authors who documented reasons for excluding full text articles. Any differences between the two reviewers were clarified; a third reviewer resolved any disagreements. If an article appeared in duplicate from two or three of the databases, only the search containing the most relevant and useful information was included. For each eligible study, the following data was extracted: author, year and country of publication; inclusion/exclusion criteria; sample size; zinc measure including sample type, collection time during pregnancy and method of analysis and pregnancy outcome. Most studies did not report on exclusion/inclusion criteria; these were therefore not included in the results table. Values determining zinc status were all converted to µg/L for easier comparisons between studies (Appendix B). Figure 1 outlines the literature search and selection of studies. We identified 635 citations after searching Medline (OVID), CINAHL and Scopus databases. A further seven were added by the authors. After screening the title and abstract, 116 full text papers were read. Of these, 67 studies met the inclusion criteria, including 29 on SGA/LBW (Table 1), 34 on hypertensive disorders of pregnancy (Table 4), 11 on sPTB (Table 3) and 9 on GDM (Table 4). Eleven studies assessed multiple pregnancy outcomes and are included in the relevant pregnancy outcome tables. Table 5 summarizes all included studies and whether there was a positive, negative or no association between zinc status and the pregnancy complication. The included studies were tabulated based on those that measured dietary zinc intake, then those that measured serum/plasma zinc. Globally, the average percentage of people affected by inadequate zinc intake is estimated to be 17.3% [27]. As dietary consumption of zinc is most influential on zinc status, studies that measured circulating zinc were further categorized based on whether they sampled from countries where inadequate zinc intake has been predicted to affect <17% or ≥17% of the population. We did not limit the studies to a specific period during gestation when zinc was measured and this information was not provided in eight studies [32][33][34][35][36][37][38][39]. However, zinc concentrations decline across gestation due to a combination of factors including hemodilution and increased fetal demand [40,41] and this made direct comparison of the studies difficult.
The association between maternal circulating zinc and birthweight was assessed in 14 studies based on women where inadequate dietary zinc intake was predicted to affect ≥17% of the population [57][58][59][60][61][62][63][64][65][66][67][68][69], of which 7 reported a significant association (Table 1) [57,[59][60][61][66][67][68]. All three of the studies based on women from Africa reported serum/plasma zinc on average 72-333 µg/L lower in women who gave birth to a LBW infant compared to those who gave birth to an appropriate weight infant [57,59,67]. In another study, the risk of delivering a LBW infant was also reported to be 3-fold greater in women with serum zinc levels ≤392.2 µg/L compared to those with levels above this figure (3.07, 95% CI: 1.07-8.97) [67]. Conversely, two other studies reported serum zinc to be 40-172 µg/L higher in women who gave birth to a LBW infant compared to those who gave birth to an appropriate weight infant [60,66]. A further four studies, also based on women from India, reported no association between circulating zinc levels and birthweight [62][63][64]69] and this was also reported in two studies of Turkish women [58,65]. However, univariate analysis and small sample size in these studies may not provide an accurate assessment of the effects of maternal circulating zinc and birthweight.

Hypertensive Disorders of Pregnancy
Only one study assessed dietary zinc intake and the association with hypertensive disorders (Table 4) and found no significant differences in dietary zinc intake between 13 women who developed a hypertensive disorder in pregnancy and 44 whose pregnancies remained uncomplicated [70].
Thirteen studies analyzed serum/plasma zinc in women who developed a hypertensive disorder of pregnancy in women residing in countries where inadequate zinc intake is estimated to be low (<17%) ( Table 4). Three studies reported mean serum/plasma zinc to be on average 120-1200 µg/L lower in women who developed PE compared to women whose pregnancies remained uncomplicated [49,71,72] and included one study that reported a reduction in risk of PE with serum levels above 1360 µg/L after adjusting for maternal age, height and weight before pregnancy (aOR: 0.005, 95% CI: 0.001-0.07) [71]. A further two studies reported circulating zinc to be lower in women who developed severe PE (blood pressure BP ≥ 160/110) compared to women whose pregnancies remained uncomplicated [73,74]. The remaining eight studies, whose sample sizes ranged from 10-271 women with PE/GH and 10-2038 women with an uncomplicated pregnancy, reported no difference in maternal zinc status between women with a hypertensive disorder of pregnancy and those without [38,47,54,[75][76][77][78][79].

Spontaneous Preterm Birth
The literature search identified four studies which measured dietary zinc intakes during pregnancy and sPTB with varying conclusions (Table 3) [43,44,94,95]. Two of these studies, which analyzed 5738 and 818 women respectively, determined that low zinc intake (≤6 mg/day which is ≤54% of the recommended 11 mg/day [21]) was associated with a more than 2-fold increase in the risk of delivering preterm (aOR: 2.3, 95% CI: 1.2-4.5 and aOR: 1.85 95% CI: 1.09-3.12, respectively), after adjusting for factors such as ethnicity, pre-pregnancy BMI, smoking, alcohol and multivitamin consumption [44,94]. If delivery date was calculated by last menstrual period (LMP), zinc intake below 9 mg/day was associated with a 2.75-fold increased risk in delivering <32 weeks gestation (aOR: 2.75, 95% CI: 1.31-5.77) [44]. However, another study reported no association between low dietary zinc intake (less than the median) and the risk of sPTB (OR: 1.1, 95% CI: 0.7-1.7) [43] but mean zinc intake of the women in this study was 14 mg/day, higher than the recommended 11 mg/day, indicating that low zinc intake was not prevalent within this studied population.
When separated based on estimates of inadequate zinc intake, there were three studies which assessed whether there was an association between circulating zinc and sPTB in low-risk populations (Table 3). While two showed no significant difference between serum/plasma zinc levels during gestation in women who gave birth preterm and those who gave birth at term [48,54], one study which recruited 3081 women in China found a 2.4-fold increase risk of PTB with serum levels <767 µg/L (aOR: 2.41, 95% CI: 1.57, 3.70) [96].
The association between maternal circulating zinc and sPTB was determined in four studies on populations with inadequate zinc intake ≥17%, all of which sampled women in India (Table 4) [61,63,64,69]. Two of the studies reported serum/plasma zinc to be higher in women who delivered preterm compared to those who delivered at term (average 98-1991 µg/L increase) [63,64]. However, no difference in circulating zinc measured at delivery was reported in the remaining two studies [61,69].

Gestational Diabetes Mellitus
Two studies looked at the association between dietary zinc intake and GDM (Table 4) [97,98]. One collected data at 24-28 weeks gestation, and found an 11% reduction in the risk of gestational hyperglycaemia with every 1 mg/day increase in dietary zinc intake (aOR: 0.89, 95% CI: 0.82-0.96) [97]. The second, which sampled women at 14-20 weeks' gestation, found no association between maternal dietary zinc intakes below 50% of the recommended daily allowance and GDM (OR: 1.4, 95% CI: 0.6-2.9) [98]. Differences between the studies included when dietary zinc was measured (early versus late second trimester) as well as ethnicity (Italian versus Iranian in which, genetic and cultural differences are likely).
Of the five studies which assessed the association between circulating zinc and GDM in countries where inadequate zinc intake is estimated to be <17%, two, both studying Italian women, reported a significant difference in serum/plasma zinc in women who developed GDM compared to women whose pregnancies remained uncomplicated (Table 4) [47,97]. However, while one study reported that serum zinc was negatively associated with the risk of hyperglycemia in pregnancy (aOR: 0.94, 95% CI: 0.91-0.96) [97], the other found that there was in increase in serum zinc in women with GDM compared to women whose pregnancy remained uncomplicated (GDM mean (SD): 766.6 (117.6) vs. uncomplicated: 627.5 (150) µg/L, p < 0.001) [47]. Both studies sampled women at similar times during pregnancy and used atomic absorption spectrometry to quantitate zinc. The remaining three studies found no difference in circulating zinc [35,38,99] however, given the small sample size of women with GDM in these studies (n = 5-46), it is likely they were underpowered and not suitable for the chosen statistical tests.
There were two studies that sampled women from countries where inadequate zinc intake was estimated to be ≥17% and assessed the association between maternal circulating zinc and GDM (Table 4) [98,100]. Neither study reported a difference in serum zinc in early pregnancy or at delivery in women with GDM compared to those whose pregnancies remained uncomplicated.

Discussion
This systematic review assessed whether maternal circulating zinc levels and/or dietary zinc intake were associated with a number of pregnancy complications. Overall, the evidence regarding the association between maternal zinc status and PE/GH, LBW/SGA, sPTB and GDM is weak and heterogeneity between the studies made comparisons difficult. However, systematic analysis of the available literature indicated some trends between maternal zinc status and infant birthweight as well as the development of severe PE (BP ≥160/110 mmHg).
There is consistent evidence in animal models that maternal dietary zinc deficiency during pregnancy reduces fetal growth [16][17][18][19]. From the studies that measured maternal zinc intake during pregnancy reviewed here, a possible relationship between low zinc intake (≤54% of the recommended 11 mg/day) and decreased infant birthweight may exist in human populations. Both food frequency questionnaires and 24 h recalls are limited by the preparedness of the participants to accurately record their diets, the food composition tables used and their ability to capture variations within diets [102]. This may explain the conflicting results between studies which assessed dietary zinc intake and the association with infant birthweight, sPTB and GDM. However, three of the four studies that measured dietary zinc intake in pregnancy and recorded infant birthweight reported a significant reduction in maternal zinc status in those who delivered a LBW/SGA infant [42,44,45]. The relationship between infant birthweight and maternal serum/plasma zinc is less clear. Plasma measures of zinc are considered preferable over serum as erythrocytes can be a source of zinc contamination within serum samples [22]. However, plasma zinc only accounts for approximately 0.1% of total body zinc [103], is heavily influenced by confounding factors like stress, infection and hormones [101,[104][105][106][107] and does not directly correlate with dietary zinc intake [108]. This limits how useful measuring circulating zinc is as a biomarker for health and disease. When studies on LBW/SGA that measured maternal circulating zinc were separated based on populations where inadequate zinc intake is predicted to be ≥17%, 7 of the 13 studies reported a difference in serum/plasma zinc between women who delivered LBW/SGA infant and those whose infants were of an appropriate weight. Given the lack of suitable alternatives, particularly in studies of pregnant women, determining zinc status by measuring serum/plasma zinc can still be informative about the importance of zinc to pregnancy, especially if measured in conjunction with dietary zinc intakes.
Other maternal factors such as age, BMI, smoking status and alcohol consumption in pregnancy not only influence pregnancy outcome but also circulating zinc [109,110]. BMI is a significant factor in influencing the risk for developing PE and GH [111,112]. However, only 11 of the 32 studies on PE/GH [33,36,54,71,75,79,80,82,84,88,92] reported on BMI, making it difficult to comment on whether differences in BMI may be influencing the outcomes of the studies included in this review. Despite this, there may be a relationship between maternal circulating zinc levels and the severity of PE. Mean maternal zinc concentrations in women with severe PE (ranging from 388 to 410 µg/L) [73,74,83] were well below 562.1 µg/L, which is the defined zinc deficiency cut-off [26,113]. In women with mild PE and those with uncomplicated pregnancies, mean maternal zinc concentrations ranged between 684-831 µg/L [37,39,83] and 630-1022 µg/L [37,39,74,83] respectively. A current leading hypothesis relating to the development of PE is increased placental oxidative stress [114]. Zinc itself has antioxidant capabilities and is an integral structural component of superoxide dismutase, a first line defense antioxidant [115] which has reduced activity in cell lines, animal models and human studies of zinc deficiency [116][117][118][119][120]. Hence, it is possible in pregnancies complicated by PE, that low maternal zinc concentration (<562.1 µg/L) may reduce the potential to combat rises in free radical production and increase the severity of the complication.
Zinc levels in maternal circulation decrease across gestation; this is thought to be due to a combination of increased maternal blood volume and fetal demands [40,[121][122][123], and therefore comparisons between studies which measured zinc in maternal serum or plasma early in pregnancy versus late should be interpreted with caution. Overall, regardless of pregnancy outcome, the majority (31 out of 59 studies which measured maternal circulating zinc) collected samples during labor or at delivery. Physiologically, parturition results in huge changes to maternal hormonal profile with rises in estrogen, oxytocin and prostaglandin required to initiate labor [124]. Furthermore, there is an increase in the production of inflammatory cytokines and a withdrawal of anti-inflammatory cytokines within the gestational tissues [125]. Infection and inflammation decrease plasma zinc [104] and use of the contraceptive pill, which raises estrogen and progesterone levels, also decreases circulating zinc [101,105]. Given that pregnancy itself is likely to confound zinc status, this has implications for interpreting studies that have measured serum/plasma zinc at delivery. In addition, how zinc may be associated with a pregnancy outcome needs to be measured before the pregnancy complication has manifested. Only five studies of 6795 pregnant women in total measured either circulating zinc or dietary zinc intake prior to 20 weeks gestation [44,53,54,79,98]. All found no significant difference in maternal zinc status during this time period between women who developed a pregnancy complication and those who did not, indicating that zinc status in early pregnancy may not be associated with adverse pregnancy outcomes.
Due to the additional demands associated with pregnancy and fetal growth, pregnant women are more vulnerable to multiple nutrient deficiencies [126] and this is potentially another cofounding factor when assessing the association between maternal zinc status and pregnancy outcome. This is because nutrients can interact with each other in both a positive (e.g., vitamin A and zinc [127]) and negative manner (e.g., calcium or iron and zinc [128,129]). A number of studies reviewed here measured serum/plasma concentrations of other nutrients as well as zinc, including copper [35,97], iron [75,98], selenium [51,88], magnesium [95,99] and lead [78]. While circulating zinc levels were not different for the pregnancy outcomes studied in these articles, those of other micronutrients were. Serum copper concentrations were found to be higher in women with GDM or those who delivered an SGA infant when compared to women with an uncomplicated pregnancy in two studies [35,97]. Furthermore, serum iron was higher in women with PE and GDM compared to women whose pregnancies were uncomplicated [75,98]. Two other studies found selenium to be lower in the serum of women with PE or those who delivered an SGA infant compared to women with an uncomplicated pregnancy [51,88]. Therefore, it is important to consider other nutritional factors that may influence pregnancy outcome as well as micronutrient ratios in order to fully understand the importance of micronutrient status on pregnancy success.
Finally, the lack of studies identified in this review analyzing truly zinc deficient women, nor those in populations at high risk of zinc deficiency, is a major limitation in determining the effects of zinc on pregnancy outcome. Only 8 of the 64 studies reported mean circulating zinc below 562.1 µg/L [49,50,60,72,74,83,87,91] and there were very few studies based on women in countries where inadequate zinc intake is predicted to be prevalent like South-East Asia and parts of Africa [26,27]. The majority of studies were based on populations in the United States and Europe where zinc deficiency is estimated to only affect 3.9%-12.7% of the population [26]. Therefore, there is the potential that the results from this review may be skewed given the lack of evidence based on women living in areas predicted to be at high risk of zinc deficiency.

Conclusions
The current review has explored the connection between maternal zinc status and pregnancy complications including hypertensive disorders of pregnancy, infant birthweight, spontaneous preterm birth (sPTB) and gestational diabetes mellitus (GDM). While it appears that there may be a relationship between maternal dietary zinc intake and infant birthweight and the development of severe PE, there is little evidence to suggest an association between zinc and sPTB or GDM. However, heterogeneity in the studies identified in this review reflects real uncertainty in the evidence linking zinc deficiency and pregnancy complications and therefore this warrants further study, particularly in developing countries whose populations are at increased risk of zinc deficiency. If we are to continue to reduce preventable deaths of newborns and children under the age of five [6], understanding the importance of micronutrients like zinc in child development, particularly in utero, will greatly increase the likelihood of success. Future studies need to focus on women more vulnerable to zinc deficiency in pregnancy in order to fully determine the effects of zinc status on pregnancy outcome.