1. Introduction
Arsenic, cadmium, mercury, and lead are established or potentially neurotoxic and genotoxic metals that lead to serious health problems in the case of human exposure [
1,
2,
3,
4]. Excessive exposure to these toxic metals can irreversibly damage normal infant development [
5,
6,
7,
8]. Recently, even low-level toxic metal exposure, unrelated to environmental exposure in high-risk residential areas, has been reported to adversely affect the normal physical growth of fetuses, infants and/or older children [
9,
10,
11,
12].
Among the four toxic metals, Pb is the one that is best known to impede children’s growth, as it accumulates in the long bones and leads to bone damage [
8,
13]. Both prenatal and postnatal low-level Pb exposure (<5.0 µg/dL) has been reported to be inversely associated with post-birth growth of Korean infants and children [
12,
14,
15]. However, few studies have investigated the dose–response association of low-level postnatal Pb exposure and the subsequent postnatal growth of children in late infancy, which is generally regarded as a high-risk period for Pb exposure [
16,
17]. For As, Cd, or Hg, prenatal low-level exposure has been reported to have unfavorable effects on intrauterine or post-birth growth in some studies [
9,
10,
18,
19,
20]. However, evidence of this is inconsistent in the current literature [
11,
13,
21]. Moreover, studies on the effects of low-level postnatal As, Cd, or Hg exposure on the postnatal growth of infants or children are scarce as well. In addition, most studies investigated the effect of a single metal; only a few studies have simultaneously evaluated and compared the effect of low levels of multiple toxic metals on the growth of infants or children in the same population [
22,
23,
24]. Factors such as different levels of exposure to toxic metals, differing amounts of materials that counteract or augment the harmful action of the toxic metals, or incomplete adjustment for other growth-affecting factors might contribute to the mixed results among previous studies [
11,
12,
25,
26].
The weaning period, also known as the period of complementary feeding (CF), is critical to the physical and mental development of children [
27]; this may also be a critical period for the examination of the association between toxic metal exposure and children’s growth and development. During this period, infants are exposed to toxic metals through dietary intake, putting contaminated non-food objects into their mouths, and indirect smoking [
17]. Among these, diet may be the most important source of toxic metal exposure among the general population [
28]. Both breastmilk and solid food can be important sources of toxic metal exposure during infancy.
All four toxic metals have been reported to be transmitted to babies by breastfeeding, although the degree of excretion of toxic metals in breastmilk seems different for different toxic metals [
29]. Formula and its mixing water have also been reported as sources of toxic metal exposure [
30,
31]. The exposure pattern to toxic metals through the intake of contaminated solid foods may be similar to that of adults living in the same area. In Korea, the main source of complementary food usually includes rice-based food, red meat, fish and vegetables. However, rice can have elevated As levels [
32]. Korean rice contained a similar amount of As to that of U.S. rice or Taiwan rice, although the level is distinctively lower than that of Bangladeshi rice [
33,
34,
35]. Rice has been reported as an important source of As exposure among Korean immigrants in the US [
36]. Fish intake has also been reported to be associated with increased As exposure in Korean adults and increased Hg exposure in Korean children [
25,
37]. According to a previous study on Korean children, grains, fish, shellfish, and seaweeds were important sources of Cd exposure; sources of Pb exposure were fruits, vegetables, and grains, rather than meats [
38]. As, Cd, and Hg levels were higher among Korean adults than in those reported in western countries in a recent Korean national environmental health survey [
39].
In addition, during the weaning period, infants are more likely to lack essential trace elements, which may serve to counteract the effects of toxic minerals [
27,
40]. For example, iron deficiency is the most common during this period. Childhood iron deficiency causes not only anemia but also retarded physical development [
41]. In addition, if iron deficiency occurs during the first year of life, then permanent neural system disorders could arise [
42]. Iron has also been reported to be involved in the metabolism of toxic metals, and iron deficiency has been reported to increase the absorption of Cd, Pb, and other toxic metals [
43].
Nevertheless, population-based epidemiological studies examining the relationships among diet intake, iron status, and the toxic metal burden in weaning-age infants are limited. In addition, few studies have evaluated blood levels of toxic metals in this age group, especially in Korea. Previously, we investigated postnatal Hg exposure and its relationships with anthropometry and dietary factors in healthy Korean weaning-age infants [
25]. In the present study, we investigated the exposure levels of four potentially growth-affecting toxic metals (As, Cd, Hg, and Pb) in a different population with similar characteristics. The aim of this study was to explore the relationship between growth, feeding method, diet during the weaning period, iron levels, and the heavy toxic metal burden in weaning-age Korean infants.
4. Discussion
This study shows that Cd, Hg, and Pb exposure in healthy Korean weaning-age infants residing in Seoul is usually below the upper limit suggested by the EPA and CDC of the U.S. [
3,
15,
45]. Median Cd and Hg levels were distinctively lower in our subjects than in a small subset of Korean newborns or school-aged children, and Pb levels were slightly lower in our subjects than in Korean or western school-aged children [
47,
48,
49]. The globally accepted upper normal limit blood level for As has not yet been established [
50]. However, according to the 12 μg/L upper limit suggested by the Mayo Clinic in the U.S. [
51], all infants had acceptable blood As levels. Still, according to the 1.0 μg/L upper limit from the Agency for Toxic Substances and Disease Registry [
3], 66.7% of our subjects had higher levels than desired. Although few studies investigating blood As levels among weaning-age infants are available in the literature, the geometric mean As value of the infants in our study was similar to the cord blood As levels of Nepal infants; blood As levels are distinctively lower than that of adults from high-risk areas such as Bangladesh, and distinctively higher than that of Western adults and children [
52,
53,
54,
55]. This matches the results from a previous Korean national study showing that urine As levels in Korean adults were similar to those from eastern Asia but much higher than those in western adults [
56,
57]. However, because blood and urine in Koreans may often contain a considerable amount of organic As through seafood ingestion (which is regarded as far less toxic than inorganic As), the differences in blood As levels between the infants in our study and the infants in western populations may not equal the differences in toxicity between them [
37,
56].
Among the four toxic metals, only As and Pb levels were negatively associated with anthropometry of Korean weaning-age infants in adjusted linear regression models. As was only associated with current head circumference, whereas Pb was associated with both post-birth increase of weight and current head circumference. A direct comparison of the degree of the adverse influence of As and Pb on post-birth growth through blood levels might not be possible. However, Pb seemed to influence growth restriction factors to a greater degree in an adjusted linear model for current head circumference, although the range of Pb levels was lower in our subjects than in those from previous growth-related studies [
8,
12,
16]. To our knowledge, this is a unique result showing that low level postnatal Pb exposure, mostly <3 µg/dL, may be associated with restricting post-birth growth, including head size increase of children during infancy. This study is also noteworthy because a significant inverse association between low to modest As exposure and post-birth anthropometry was observed among subjects who did not reside in previously known high-risk areas with As-contaminated water exceeding WHO standards [
29,
58]. Compared to the other three toxic metals, investigations on the association of low levels of prenatal or postnatal As exposure and post-birth anthropometry in children are extremely scarce. In addition, this is one of the first reports demonstrating the inverse association between low As and low Pb levels and post-birth growth within the same population in the weaning period, in which potential growth-affecting factors including ID or IDA and feeding type or diet adequacy were adjusted.
The method by which As restricts growth is unclear. Endocrine-disrupting properties leading to insulin-like growth factor 1 (IGF-1) suppression have been suggested as a mechanism of growth impairment [
59]. Our study cannot explain the reason why As was associated with only current head circumference and not post-birth weight gain. It may be associated with the toxic nature of low As exposure itself and/or a specific exposure period during childhood. We cannot exclude the possibility that a smaller head circumference at birth associated with intrauterine As exposure may considerably contribute to the lower head circumference in this study, because brain growth is particularly vulnerable to intrauterine insult [
60]. Our study might be limited in its ability to prove an independent association between postnatal As exposure and current head circumference because the cord-blood As levels were not investigated, although mother’s age at delivery was adjusted, which might be associated with cord-blood As levels. In addition, as the current head circumference of our infants was mostly within normal range, the clinical significance of the observed inverse association between As and growth may need to be determined through further functional study. In a previous Nepal cohort, cord blood As levels, similar to ours, were associated with adverse neurodevelopment at birth, but not at 36 months of age [
54]. Although many studies used the urinary As level as a biomarker for exposure [
61], blood As has the advantage of reflecting the As burden in tissue compartments [
55]. In general, blood As concentrations have been considered to reflect only recent exposure. However, considering that the feeding patterns and diet of weaning-age infants are usually repetitive once established and maintained long term, blood As levels can reflect long-term steady-state exposure [
55]. We did not find an independent association between low-level post-birth Cd or Hg exposure and post-birth anthropometry, which is consistent with the results of a few previous studies [
13,
25]. It may be associated, at least in part, with much lower Cd levels in our subjects compared to newborns or older children in Korea or western countries [
47,
49].
Feeding type and breastfeeding duration were significantly correlated with all four metal levels in our study, which was consistent with the results of previous studies [
29]. The relatively slow growth of breastfed infants compared with formula-fed infants during the first few years of life is well documented [
62]. Consequently, it is important to adjust the type of feeding or amount and duration of breastfeeding to document the association between toxic metals and infant growth. In this study, although post-birth Z-scores were negatively correlated with the duration of breastfeeding, As and Pb levels were chosen as final explanatory variables, rather than feeding type or duration of breastfeeding, in the adjusted regression models for anthropometry, which was contrary to the case of Hg levels in our previous study [
25]. Therefore, breastfeeding should be encouraged, and As and Pb exposure should be reduced as much as possible by controlling the diet and environment of breastfeeding mothers and their babies. In fact, in the case of As, there are also reports that the inorganic As secretion of breastmilk is decreased by increased methylation in lactating mothers; thus, the As concentration in breast milk is very low even in contaminated areas, unlike that of Hg [
29,
63]. Pb levels seemed more affected by IDA than by CF type in our subjects. Given that prolonged breastfeeding is also a well-documented risk factor for ID or IDA [
64,
65], prevention of ID or IDA can be one of the most important precautions for reducing toxic metal exposure in breastfed infants. In this study, ID was also found to be associated with higher Cd levels, which was consistent with the results from previous studies on adults and children [
43,
66].
There was a significant correlation between the adequacy of specific food intake and specific toxic metal levels in weaning-age infants, similar to Korean adults and older children [
37,
38]. Although the amount of food intake was not precisely quantified in this study, the perceived adequacy of rice-based food intake was significantly associated with blood As levels and it tended to be associated with blood Cd levels as well. Infants with fish intake more than once per week also had increased blood As and Hg levels. In general, while As in both rice and breast milk is regarded as highly toxic inorganic As, seafood contains organic As, which does not induce substantial damage to the body because it is rapidly and completely excreted in the urine [
63]. However, attention to the As exposure through seafood intake is also required, as recent studies showed increased levels of dimethylarsinic acid (DMA) in the urine of adult populations after seafood intake, which is far more toxic than organic As [
67].
There are some limitations to this study. It was a single-center, cross-sectional study with a modest number of infants. Toxic metal levels were measured only once. Precise quantification of dietary intake was not performed. While blood is the preferred biomarker for Pb and Hg exposure, urine is the usual biomarker for As and Cd exposure in many studies. Finally, this study did not thoroughly adjust for environmental, genetic, social, and economic factors, which could potentially affect growth. Nevertheless, based on information about the medical insurance status, detailed addresses of the residence area, and smoking history of mothers, it is possible to assume that our subjects may likely not be at risk for a poor nutritional status or environmental exposure to toxic metals. In half of the study participants, the parents’ highest education level was available, which showed that the proportion of highest education levels were higher in our study population than in the general Korean population with similar ages.