Background
Epoxy and vinyl chloride resins are widely used for the inner coating of canned and retort pouch foods. Bisphenol A diglycidyl ether (BADGE) is synthesized when epoxy resins are produced via the condensation of Bisphenol A (BPA) and epichlorohydrin [
1]. Similarly, the reaction of phenol formaldehyde resin with epichlorohydrin produces bisphenol F diglycidyl ether (BFDGE). During storage, food coatings can interact with aqueous and acidic foods, leading to the formation of hydrolysis derivatives, such as BADGE·2H
2O, BADGE·H
2O, BFDGE·2H
2O, and BFDGE·H
2O, that may migrate into the food. Furthermore, BADGE and BFDGE can be added to remove the hydrochloric acid produced during thermal coating, which could result in the formation of chlorine derivatives [
2]. Migration of BADGE, BFDGE, and their derivatives into canned foods and beverages has been reported in several countries. For instance, derivatives of BADGE have been detected in canned vegetables and mixed dishes in France, [
3] while BADGE and derivatives of BADGE have been found in canned fish, vegetables, and sauces in Japan [
4]. Both BADGE and BFDGE have been detected in infant formulas in Canada [
5]. Therefore, food is regarded as a major source of exposure to BADGE and BFDGE [
6]. BADGE, BFDGE, and their derivatives have been detected in adult human body fluids and tissues [
6‐
8]. Furthermore, BADGE is also used in paints, adhesives, floorings, clothing, textiles, and furniture coatings [
9,
10]. Infants characteristically put anything they can reach into their mouths, increasing their potential exposure to BADGE and its toxic effects.
The capacity of BADGE and its derivatives to disrupt endocrine function and induce cytotoxicity and genotoxicity raises significant concerns regarding their safe use [
11‐
13]. Infants expressing BADGE-specific immunoglobulin G (IgG) showed high food-specific immunoglobulin E (IgE) levels, suggesting that BADGE affects the immune system [
14]. Furthermore, BFDGE has been implicated in cytotoxic, mutagenic, and genotoxic effects [
15]. Accordingly, the commission of the European Communities has set a limit on their migration into food, up to 9 mg/kg for the sum of BADGE and its hydrolyzed derivatives (BADGE·H
2O, BADGE·2H
2O) and 1 mg/kg for the sum of BADGE formed as reaction products with HCl (BADGE·HCl, BADGE·2HCl, BADGE·H
2O·HCl) [
16]. The Commission has also prohibited the use or presence of BFDGE. However, there are no regulations regarding the use of BADGEs or BFDGE in medical devices.
In addition to everyday applications, various medical devices, such as intravenous (IV) sets, syringes, and catheters, are made of plastics, such as BPA-based polycarbonate, and epoxy resins are used as adhesives. Because BADGE is synthesized during the production of epoxy resins, medical devices can be a source of exposure to BADGE. Neonates admitted to the neonatal intensive care unit (NICU) require several medical devices for life support, including incubators, ventilators, IV sets, gastroesophageal tubes, and urinary catheters. A previous study on neonates admitted to the NICU reported that the increased use of medical devices was associated with high levels of BPA in their urine, regardless of their oral exposure to BPA [
17]. High levels of urinary BPA have also been observed in patients in pediatric intensive care units, where medical devices are frequently used [
18,
19]. However, there are no reports on the association between the use of medical devices and BADGE concentrations.
The purpose of the present study was to measure serum levels of BADGE, derivatives of BADGE, and BFDGE in infants with a history of hospitalization in the NICU and evaluate their association with medical devices and the living environment. Derivatives of BFDGE were not included owing to a lack of established standards and detection methods.
Results
We summarized the characteristics of the participating infants at the time of first and second serum sample collection (Table
1). The gestational age ranged from 30 to 38 weeks; 90% of the infants were preterm (< 38 weeks), and 80% had low birthweight (< 2500 g). The length of their hospital stays ranged from 5 to 62 days. Among the participants, the first serum samples were collected from three infants while still in the NICU (infant number 1–3). For the remaining participants, the first serum samples were collected between 8- and 39-days post-discharge from the NICU. Infants 2 and 3 required respiratory support through endotracheal intubation, and their first serum samples were collected 53 and 31 days after extubation, respectively.
Table 1
Characteristics of study participants
1 | Boy | 30 | Cesarean | 1513 | 43 | Preterm, low birth weight, TTN | Threatened premature delivery | 62 | 56 | - | 0 | 2.59 | 261 | 6.92 |
2 | Boy | 30 | Vaginal | 1749 | 41 | Preterm, low birth weight, RDS | Nothing | 59 | 56 | 53 | 0 | 2.80 | 212 | 5.76 |
3 | Girl | 32 | Cesarean | 1804 | 42 | Preterm, low birth weight, RDS | Gestational diabetes mellitus | 35 | 34 | 31 | 0 | 2.44 | 229 | 7.41 |
4 | Boy | 33 | Cesarean | 2105 | 47 | Preterm, low birth weight | Threatened premature delivery | 24 | 32 | - | 8 | 2.84 | 214 | 8.13 |
5 | Boy | 34 | Vaginal | 2042 | 44 | Preterm, low birth weight | Type 2 diabetes mellitus, pregnancy-induced hypertension | 25 | 37 | - | 12 | 3.07 | 219 | 8.70 |
6 | Boy | 34 | Vaginal | 2796 | 49 | Preterm, TTN | Threatened premature delivery | 20 | 46 | - | 26 | 3.99 | 228 | 9.05 |
7 | Boy | 35 | Vaginal | 2302 | 46 | Preterm, low birth weight | Premature rupture of membrane | 19 | 47 | - | 28 | 4.07 | 220 | 8.12 |
8 | Boy | 35 | Vaginal | 2382 | 45 | Preterm | Nothing | 24 | 63 | - | 39 | 4.53 | 222 | 6.85 |
9 | Boy | 35 | Cesarean | 2718 | 50 | Preterm, hyperbilirubinemia | Premature rupture of membrane | 28 | 65 | - | 37 | 3.45 | 254 | 7.17 |
10 | Girl | 38 | Cesarean | 3210 | 48 | Hypoglycemia | GDM | 5 | 33 | - | 28 | 4.21 | 215 | 8.20 |
The details of nutritional intake and home environment obtained from the questionnaire were described (Table
2). Families of infants 3, 4, 5, and 6 answered that they use products containing epoxy resin at the time of the first or second questionnaire. All those families used printing presses, but only one of those families (that of infant 5) also used adhesives.
Table 2
Feeding details and household environment of infants
1 | Breast milk & formula | 77.2 | Glass | Nothing | Not used | Formula & weaning food | 1 | Everyday | Nothing | Plastic, fabric | Plastic |
2 | Breast milk & formula | 178.6 | Glass | Nothing | Not used | Formula & weaning food | 1 | Nothing | Nothing | Plastic, fabric | Plastic, glass |
3 | Breast milk & formula | 10.3 | Glass | Nothing | Not used | Formula & weaning food | 2 | 1~3/month | Printing presses | Plastic, fabric, wood | Plastic |
4 | Breast milk & formula | 61.7 | Glass | Printing presses | Wood | Formula & weaning food | 2 | Everyday | Nothing | Plastic, wood | Plastic, glass |
5 | Formula only | 214.8 | Plastic | Nothing | Plastic, fabric | Formula & weaning food | 2 | 1~3/month | Adhesives, printing presses | Plastic, fabric | Plastic, glass |
6 | Breast milk only | 0.0 | Not use | Printing presses | Plastic, fabric, wood | Breast milk & Formula & weaning food | 2 | Nothing | Printing presses | Plastic, fabric, wood | Plastic, glass, metal |
7 | Breast milk & formula | 78.6 | Glass | Nothing | Unknown | Formula & weaning food | 1 | 1-3/week | Nothing | Plastic, fabric, wood | Plastic |
8 | Breast milk & formula | 171.0 | Plastic, Glass | Nothing | Plastic, wood | Formula & weaning food | 3 | 4-6/week | Nothing | Wood | Metal, other |
9 | Breast milk & formula | 23.2 | Plastic, Glass | Nothing | Plastic, fabric | Breast milk & weaning food | 1 | 1-3/month | Nothing | Plastic, fabric | Plastic |
10 | Breast milk & formula | 166.3 | Glass | Nothing | Fabric | Breast milk & Formula & weaning food | 2 | 1-3/week | Nothing | Plastic, fabric | Plastic |
The duration of medical device use during the NICU admission showed that all infants received peripheral or central venous infusions (Table
3). Infants 1, 2, 3, and 6 were provided with respiratory support due to respiratory impairment, and infants 2 and 3 underwent invasive ventilation due to respiratory distress syndrome. Infants with lower gestational age were exposed to more medical devices for longer durations.
Table 3
Duration of medical device use (days) on infants in the NICU
1 | 0 | 14 | 48 | 23 | 24 | 0 | 40 |
2 | 0 | 11 | 38 | 13 | 16 | 2 | 32 |
3 | 0 | 7 | 17 | 4 | 0 | 2 | 16 |
4 | 0 | 9 | 14 | 0 | 0 | 0 | 9 |
5 | 0 | 8 | 10 | 0 | 0 | 0 | 10 |
6 | 0 | 8 | 6 | 3 | 0 | 0 | 4 |
7 | 5 | 0 | 7 | 0 | 6 | 0 | 8 |
8 | 0 | 5 | 6 | 0 | 0 | 0 | 8 |
9 | 8 | 0 | 3 | 0 | 0 | 0 | 4 |
10 | 4 | 0 | 0 | 0 | 0 | 0 | 3 |
In all serum samples, only BADGE·2H
2O was quantified, while BADGE·H
2O and BADGE were at their lower limit of detection (LLOD, < 0.09 ng/mL), and BFDGE was at its lower limit of quantitation (LLOQ, < 0.39 ng/mL, Table
4). The highest BADGE·2H
2O concentration in the first serum sample was in infant number 3 (157.58 ng/mL), one of the infants who received invasive ventilation. Concentrations of BADGE·2H
2O tended to be lower in the second serum sample collected than in the first (
P = 0.0593). However, the BADGE·2H
2O concentration in the second sample collected from infant 5 was higher than in the first sample (122.85 ng/mL). The median excretion speed of BADGE·2H
2O per date was − 0.06 ng/mL/day (range: -0.80–0.47) (Table
4).
Table 4
Concentrations of BADGEs and BFDGE in serum samples and excretion speed
1 | 25.24 | LLOD | LLOD | LLOQ | 0.86 | LLOD | LLOD | LLOQ | -0.12 |
2 | 2.43 | LLOD | LLOD | LLOQ | 1.04 | LLOD | LLOD | LLOQ | -0.01 |
3 | 157.58 | LLOD | LLOD | LLOQ | 2.56 | LLOD | LLOD | LLOQ | -0.80 |
4 | 2.54 | LLOD | LLOD | LLOQ | 1.73 | LLOD | LLOD | LLOQ | 0.00 |
5 | 36.78 | LLOD | LLOD | LLOQ | 122.85 | LLOD | LLOD | LLOQ | 0.47 |
6 | 2.30 | LLOD | LLOD | LLOQ | 1.52 | LLOD | LLOD | LLOQ | 0.00 |
7 | 32.29 | LLOD | LLOD | LLOQ | 3.78 | LLOD | LLOD | LLOQ | -0.16 |
8 | 11.02 | LLOD | LLOD | LLOQ | 7.04 | LLOD | LLOD | LLOQ | -0.03 |
9 | 21.03 | LLOD | LLOD | LLOQ | 2.48 | LLOD | LLOD | LLOQ | -0.10 |
10 | 25.60 | LLOD | LLOD | LLOQ | 2.45 | LLOD | LLOD | LLOQ | -0.13 |
median (min-max) | 23.13 (2.30–157.58) | | | | 2.47 (0.86–122.85) | | | | -0.06 (-0.80–0.47) |
The concentrations of BADGE·2H
2O at the second sample collection grouped by frequency of commercial baby food consumption were compared (Table
5). There was no significant difference in BADGE·2H
2O concentrations in the samples obtained at the second collection between the group that consumed commercial baby food less than once per week and the group that consumed it more often than once per week (
P = 0.9168).
Table 5
Differences in BADGE·2H2O concentration by frequency of commercial baby food
<1/week (infant number 2,3,5,6,9) | 1.04–122.85 |
≥1/week (infant number 1,4,7,8,10) | 0.86–7.04 |
Discussion
The objective of this study was to assess the levels of BADGE, derivatives of BADGE, and BFDGE in infants previously hospitalized in the NICU and examine their potential correlation with medical devices and living conditions. In the present study, the serum levels of BADGE, derivatives of BADGE, and BFDGE were measured twice, at approximately 1–2 months of age and again at approximately 7 months of age, in 10 infants with a history of NICU admission. Only BADGE·2H2O was present at levels above the threshold value of the testing method at both sampling times, whereas BADGE·H2O, BADGE, and BFDGE were not. There have been no previous reports measuring the serum concentrations of BADGE, derivatives of BADGE, and BFDGE in infants. Additionally, there have been no previous reports evaluating the exposure to BADGE via medical devices. Consequently, this is the first report examining the relationship between medical device use and living environment at home and serum BADGE·2H2O concentrations in infants with a history of NICU hospitalization.
While no reports of measuring BADGEs (BADGE and its derivatives) or BFDGEs (BFDGE and its derivatives) in infants have been published, urinary BPA has been measured [
16]. Urinary BPA levels were significantly higher in infants in the NICU who used four or more medical devices within three days than in infants who used three or fewer medical devices [
17]. Although breast milk and formula BPA concentrations were also measured, no significant differences were found, and the infants’ urinary BPA concentrations did not differ according to the feeding method. Furthermore, urinary BPA levels were significantly high in pediatric intensive care unit patients who were endotracheally intubated or who underwent hemodialysis [
18]. These reports suggested that invasive medical procedures increase chemical exposure from medical devices. In the present study, the first sample in infants 1–3 was collected while they were in the NICU. Therefore, the concentration of the first samples of those three infants reflected the chemical exposure in the NICU. Of the three, infant 3 had the highest BADGE·2H
2O concentration, followed by infant 1, while infant 2 had the lowest BADGE·2H
2O concentration. All three infants were fed breast milk and formula, and no correlation was observed between the amount of formula and BADGE·2H
2O concentration. Infants 2 and 3 were both on ventilators for two days; however, the BADGE·2H
2O concentration was remarkably high for only infant 3. The first serum sample from infant 2 was collected 53 days after extubation, whereas the sample from infant 3 was collected 31 days after extubation, which is a discrepancy in timing that may have contributed to the difference in BADGE·2H
2O concentrations observed. These findings may suggest that invasive ventilation is associated with increased serum BADGE concentrations, similar to a previously reported association between BPA and invasive ventilation [
17]. Although infant 3 had been extubated for 31 days prior to the sampling, his BADGE·2H
2O concentration was still clearly higher than that of the other infants and may have been even higher during endotracheal intubation management. There are no previous reports on the daily excretion rate of BADGE·2H
2O. The daily excretion rate calculated in the present study was affected by new or sustained exposures between the first and second sample collection. Therefore, it is important to note that this excretion rate does not simply represent the rate at which BADGE·2H
2O is excreted from the body.
There was no significant difference, but nine of the ten infants had lower BADGE·2H
2O concentrations in the second sample than in the first, which may be due to the potential influence of the insertion of various medical devices during the NICU admission and immaturity of metabolism and excretion. In humans, BADGE is complexly hydrolyzed, oxidized, and conjugated to produce several derivatives, which are excreted in urine and feces [
22]. Epoxide hydrolases, working primarily in the liver, enzymatically hydrolyze BADGE to produce the first by-product, BADGE·H
2O. Further hydrolysis by the same enzyme produces BADGE·2H
2O, which is further oxidized by monooxygenase to other oxidation byproducts [
22]. Many cytochrome P450s that function as monooxygenases have low or no activity at birth, and their activity gradually increases over the first 3 months of life [
23]. These suggest that delayed biotransformation of BADGE·2H
2O in infants at 1–2 months of age occurs and may induce its accumulation. In addition, it is well known that renal function in infants is immature. Glomerular filtration rate (GFR) at birth is 19.6 ml/min per 1.73 m
2 and gradually increases to 59.4 ml/min per 1.73 m
2 by 4 weeks of age [
24]. Subsequently, GFR increases until about 2 years of age, when it reaches the adult values [
25]. Therefore, the lower excretion of BADGE·2H
2O in urine in early infancy might be one of the causes of higher serum concentration in the first sample.
Several reports have assessed the exposure to BADGEs and BFDGEs in both adults and children. The urinary levels of BADGE and its derivatives (BADGE·H
2O, BADGE·2H
2O, BADGE·H
2O·HCl) observed in 57 adult volunteers from the U.S. and China, as well as 70 Chinese children, indicated that BADGE·2H
2O was the predominant substance detected among them [
26]. Furthermore, BADGE·2H
2O was detected with the highest frequency, exceeding 90%, among the nine BADGE derivatives (BADGE·H
2O, BADGE·2H
2O, BADGE·H
2O·HCl, BADGE·2HCl, BADGE, BFDGE·2H
2O, BFDGE·H
2O, BFDGE·2HCl, and BFDGE) in the serum and urine of 181 Chinese children and adolescents [
27]. Hydrolysis of BADGE·H
2O to BADGE·2H
2O is rapid, but oxidation of BADGE·2H
2O to other byproducts by monooxygenase is slow [
22]. Therefore, BADGE·2H
2O possibly accumulates in the body and is the predominant BADGE metabolite. The serum and urinary concentrations of BADGE·2H
2O in 181 Chinese children and adolescents were up to 38.441 ng/mL and 8.902 ng/mL, respectively, indicating that concentrations of BADGE·2H
2O are high in human blood [
27]. Compared to serum levels of BADGE·2H
2O in Chinese children and adolescents, those of some infants in the present study were markedly higher (157.58 ng/mL in infant 3 and 122.85 ng/mL in infant 5). These previous reports indicate a high detection rate of BADGE·2H
2O, which was similar to the results of the present study; however, unlike in previous reports, other BADGEs and BFDGE were not detected in infants, which may be due to differences in the sources of exposure and metabolism.
Furthermore, urinary levels of BADGE·2HCl, BADGE·2H
2O, BADGE·H
2O, BADGE·HCl, and BFDGE·2H
2O were negatively correlated with age, suggesting that younger individuals may have a higher risk of exposure [
27]. Sealants, paints, adhesives, furniture coatings, textile fillers, packaging materials, dental fillers, and coatings inside cans all contain BADGE [
28]. Moreover, BADGE and its derivatives have also been detected in the air and indoor dust [
29‐
31]. Infants who live closer to the floor and have higher respiratory rates than adults are more prone to ingesting dust particles. Only one infant in our study, infant 5, had a markedly higher concentration of BADGE·2H
2O at the second time of sample collection than at the first. The family of infant 5 was the only one who indicated in the second questionnaire that they used adhesives. It is unclear who in the family was using the adhesives and how, but it may be related to the significantly high BADGE·2H
2O concentrations in infant 5. Since BADGE·2H
2O is formed by the hydrolysis of BADGE in the body [
22], it is expected that the infant’s BADGE·2H
2O would be elevated due to exposure to some BADGE.
Several studies demonstrated the toxicity of BADGE and its derivatives. For instance, BADGE·2HCl and BFDGE·2HCl exhibit antiandrogenic effects by binding to the androgen receptors in vitro; [
13]. BADGE·2H
2O and BADGE 2Cl induce breast cancer cell proliferation via an estrogenic mechanism, without binding to estrogen receptors [
32]. Moreover, BADGE and BFDGE were found to induce morphological changes and cell detachment from the substrate in Caco-2 cells derived from human intestinal epithelial cells [
33]. As infants are vulnerable to damage due to being in a state of remarkably rapid growth and development, it is necessary to assess the impact of BADGE and BFDGE on children. Based on our results and previous findings, we suggest that younger children face a greater risk of exposure to BADGE·2H
2O. Future studies should increase the number of cases and determine the effects of BADGE·2H
2O on infants.
The absence of a control group is a limitation of the present study, making it unclear whether the medical equipment used during NICU admission caused the BADGE exposure. Further, it is of considerable importance that this study was able to investigate the relationship between the differences in medical device use and BADGE·2H
2O concentrations in infants with the same background history of NICU admission. There is uncertainty regarding the human placental transfer of BADGE or its derivatives, with a report that BADGE·2H
2O was detected in only 1 sample out of 14 umbilical cord blood samples [
7]. The placental transfer of BADGE·2H
2O may have affected its concentrations in infants, but this was unknown because maternal BADGE·2H
2O concentrations were not measured in this study. In addition, the concentration of BADGE in breast milk and formula and that released from baby bottles was not measured; thus, oral exposure was not evaluated. Similarly, BADGEs in needles, syringes, spits, and other medical devices used for blood collection have not yet been evaluated. Finally, the present study did not investigate whether commercial baby food packaging is indeed a source of exposure to BADGEs because information about the packaging material of commercial baby food was not available.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.