Exposed and vulnerable: sources and health implications of chemical exposures in neonatal, pediatric, and cardiac intensive care units

Abstract

Purpose of review:

Exposures to endocrine disrupting chemicals (EDCs) in early life have demonstrable adverse implications on child health and development. Yet, there is a dearth of studies evaluating the potential exposures to EDCs, such as bisphenols, parabens, phthalates, and volatile organic compounds (VOCs), in hospital-based settings among children who are critically ill and/or particularly vulnerable. This narrative review seeks to provide up-to-date evidence on the sources and magnitude of exposure to EDCs in neonatal-, pediatric-, and cardiac intensive care units (NICUs/PICUs/CICUs) as well as resulting health impacts.

Recent findings:

Thirty-three studies were included in this review. The most frequently studied and characterized EDCs in NICUs/PICUs/CICUs were phthalates followed by terephthalates and alternative plasticizers. Evaluations of health outcomes resulting from such exposures were scarce, and few studies assessed health outcomes after hospital discharge.

Summary:

EDCs are pervasive in NICU/PICU/CICU settings and pediatric exposure levels are much higher than in other environments. However, the existing evidence has multiple limitations that should be addressed in future work. Specifically, studies evaluating a more expansive array of chemicals, including contemporary and emerging replacements for legacy compounds, are needed, as are studies that consider chemical mixtures. Additionally, few studies evaluated the health impacts of chemical exposures, and their mixtures, in NICU/PICU/CICU settings, especially long-term health outcomes observed after hospital discharge. Such studies could be invaluable in supporting policy as well as development of medical products without toxic chemicals.

Introduction

There are more than 84,000 chemicals allowed for use in commerce in the U.S., comprising more than 7.2 trillion pounds of chemicals produced in 2019 alone [1]. Of concern are endocrine disrupting chemicals (EDCs), which are capable of interfering with normal hormonal processes of the human body [2], such as parabens, phthalates, organophosphate esters (OPEs), volatile organic compounds (VOCs), and bisphenols.

Children are particularly susceptible to environmental chemical exposures for myriad reasons, including dynamic physiological growth and development, immature organ systems, and behaviors (e.g., hand-to-mouth), and they generally have higher overall body burden of exposures due to their smaller body weight and size [3]. Additionally, exposures during critical periods of development, i.e., “windows of susceptibility,” may have greater impacts on health over the life-course. Children born prematurely and/or of low birthweight, or who are born with severe illnesses or complications requiring complex medical care, such as those in neonatal-, pediatric-, or cardiac intensive care units (NICUs/PICUs/CICUs), may be even more susceptible to environmental chemical exposures. Previous studies have estimated that up to 13% of all live born neonates (~468,000 neonates) [4] as well as over 300,000 children (ages 0–17) require intensive care in the U.S, annually [5,6]. The critical care environment is highly unique, and for the pediatric patient, represents a period of heightened susceptibility and vulnerability. Indeed, admissions to intensive care units may result in periods of heightened exposure to chemical contaminants due to the increased reliance on plastics for the manufacturing of routine (e.g., intravenous tubing) [7] and lifesaving devices (e.g., extracorporeal membrane oxygenation [ECMO]) in intensive care medicine [***8], as well as the use of EDCs as excipients [9] and preservatives [10] in medications. As these children are at the nexus of susceptibility, greater exposure potentials, and elevated risk of subsequent and potentially life-long morbidities, characterizing the exposure landscape in NICU/PICU/CICU settings and their health implications are of great importance.

Therefore, the aim of our review was to summarize the existing epidemiologic evidence regarding environmental chemical exposures that occur during admission to in intensive care units and their impacts on health outcomes for this unique and highly vulnerable pediatric population.

Methods

We identified articles for review using a targeted literature search in PubMed and Scopus databases between November 2023 and May 2024. We limited our search to EDCs of primary concern including per- and polyfluoroalkyl substances (PFAS); phthalates, terephthalates, and alternative plasticizers; benzophenones; bisphenols; OPEs; parabens; and VOCs. Our search strategy included exposure terms: “PFAS,” “perfluoroalkyl,” “polyfluoroalkyl,” “bisphenol,” “phenol,” “benzophenone,” “paraben,” “phthalate,” “terephthalate,” “TOTM,” “TEHTM,” “DINCH,” “alternative plasticizer,” “EDC,” “endocrine-disruptor,” “endocrine-disrupting,”, “endocrine disruptor,” “endocrine disrupting,” “organophosphate,” “OPE,” “OPFR,” “volatile organic compounds,” “volatile organic chemicals,” “VOC,” “cyclohexanone,” and “chemical exposure” as well as setting/environment terms: “NICU,” “PICU,” “neonatal intensive care unit,” “pediatric intensive care unit,” “cardiac intensive care unit,” and “CICU.” We restricted the search to studies published in peer-reviewed journals, written in English, and published between January 2018 – May 2024. We also used a “snowballing” approach in which additional studies were selected from references of articles identified by the above searches. Overall, we identified 33 studies that met the above criteria.

Results

Thirty-one of the 33 included studies quantitatively characterized exposure to at least one chemical by evaluation of: administered medications (n = 3), medical devices/equipment/supplies used in the NICU/PICU/CICU (n = 4), biospecimens collected from NICU/PICU/CICU patients (n = 20), or both medical devices/equipment/supplies and biospecimens (n = 4) (Table 1). Of the remaining two studies, one evaluated exposure as presence/absence of chemical on product labeling (i.e., “DEHP-free”) [11]; the other study used gas chromatography tandem mass spectrometry (GC-MS/MS) to quantify concentrations of chemicals in medical products, but it did not ultimately report quantitative results in the published article [***8].

Table 1.

Summary of chemical classes and matrices used for exposure assessment

References	Matrix	Bisphenols	Parabens	Phthalates	Terephthalates	Alternative plasticizers	VOCs
Al-Saleh et al. 2023 [20]	Total parenteral nutrition; urine			X
Ayar et al. 2021 [25]	Urine	X
Ayar et al. 2022 [26]	Urine	X
Bembea et al. 2022 [40]	Blood						X
Bernard et al. 2021 [37]	Urine			X	X	X
Bernard et al. 2023 [38]	Urine			X	X	X
Busgang et al. 2022 [34]	Urine			X
Buyukeren 2023 [28]	Urine	X
Cleys et al. 2023 [39]	Hair			X	X	X
Eckert et al. 2020 [17]	Blood products; blood; urine			X		X
El-Metwally et al. 2018 [43]	Urine						X
Everett et al. 2020 [41]	Blood						X
Gaynor et al. 2019 [29]	Urine	X		X
Gaynor et al. 2024 [42]	Urine						X
Guerrelli et al. 2024 [18]	Blood products; blood			X
Iacobelli et al. 2023 [13]	Medications		X
Iribarne-Duran et al. 2019 [12]	Medical products	X	X
Jenkins et al. 2021a [7]	Medical products; urine			X
Jenkins et al. 2021b [11]	Medical products			X
Kuwamura et al. 2024 [24]	Blood	X
Malarvannan et al. 2019 [8]	Medical products			X	X	X
Panneel et al. 2023 [19]	Total parenteral nutrition			X	X	X
Pinguet et al. 2019 [36]	Urine			X	X	X
Prazad et al. 2021 [23]	Breast milk (infused through neonatal enteral feeding system)						X
Saito et al. 2021 [14]	Medications		X
Sridharan et al. 2020 [15]	Medications		X
Strømmen et al. 2021 [27]	Urine	X	X
Stroustrup et al. 2018 [32]	Urine			X
Stroustrup et al. 2020 [33]	Urine			X
Stroustrup et al. 2023 [35]	Urine			X
Vanhorebeek et al. 2022 [30]	Blood			X
Wang and Kannan 2023 [21]	Medications; medical products			X
Weng et al. 2022 [31]	Urine			X

Across all 33 studies, 107 unique compounds were evaluated including: five bisphenols; four parabens; 14 parent phthalate compounds and 22 metabolites (representing exposure to 11 unique parent compounds); one parent terephthalate compound and four of its metabolites; 10 parent alternative plasticizer compounds and 13 metabolites (representing exposure to three unique parent compounds); and one parent VOC compound and 33 metabolites (representing exposure to 16 unique parent compounds) (Table 2). No studies evaluated exposure to benzophenones, OPEs, or PFAS.

Table 2.

Parent compounds and/or metabolites included across studies. Note: for parent compounds listed with its metabolites, bolding indicates that the parent compound itself was measured in at least one study.

Parent compound and metabolite(s)	Abbreviation	References
Bisphenols
Bisphenol A diglycidyl ether	BADGE	[24]
Bisphenol A (2,3-dihydroxypropyl) ether	BADGE∙H2O	[24]
Bisphenol A bis(2,3-dihydroxypropyl) ether	BADGE∙2H2O	[24]
Bisphenol F diglycidyl ether	BFDGE	[24]
Bisphenol A	BPA	[12], [25–29]
Parabens
Butyl paraben	BuP	[12–13], [27]
Ethyl paraben	EtP	[12–14], [27]
Methyl paraben	MeP	[12–15], [27]
Propyl paraben	PrP	[12–15], [27]
Phthalates, terephthalates, alternative plasticizers
Phthalates
Butyl benzyl phthalate	BBzP	[19–21]
Monobenzyl phthalate	MBzP	[20], [32], [34–35], [39]
Diallyl phthalate	DAP	[21]
Di-isobutyl phthalate	DiBP	[21]
Mono-isobutyl phthalate	MiBP	[20], [32], [34–35], [39]
Di-n-butyl phthalate	DnBP	[19–21]
Mono-n-butyl phthalate	MnBP	[20], [32], [34–35], [39]
Dicyclohexyl phthalate	DCHP	[21]
Di-isodecyl phthalate	DiDP	[8], [19]
Mono-2-(propyl-6-carboxy-hexyl) phthalate	cx-MPHxP	[39]
Mono-2-(propyl-6-hydroxy-heptyl) phthalate	OH-MPHxP	[39]
Mono-2-(propyl-6-oxo-heptyl) phthalate	oxo-MPHxP	[39]
Di-2-ethylhexyl phthalate	DEHP	[7–8], [11], [17–21], [30]
Mono-(2-ethyl-5-carboxypentyl) phthalate	MECPP	[17–18], [20], [29], [32], [34], [36–39]
Mono-(2-ethyl-5-hydroxyhexyl) phthalate	MEHHP	[7], [17–18], [20], [29], [31–32], [34], [36], [39]
Mono-2-ethylhexyl phthalate	MEHP	[7], [17–18], [20], [29], [31–32], [34], [36], [38–39]
Mono-(2-ethyl-5-oxohexyl) phthalate	MEOHP	[7], [17], [20], [29], [31–32], [34], [36], [39]
Mono-(2-carboxymethyl hexyl) phthalate	MMCHP	[32], [36]
Diethyl phthalate	DEP	[19–21]
Monoethyl phthalate	MEP	[20], [32], [34–35], [39]
Di-n-hexyl phthalate	DnHP	[21]
Di-isononyl phthalate	DiNP	[8], [19]
Mono-isononyl phthalate	MiNP	[32], [36]
Mono-carboxy-isononyl phthalate	cx-MiNP	[36–37], [39]
Mono-hydroxy-isononyl phthalate	OH-MiNP	[36–37], [39]
Mono-oxo-isononyl phthalate	oxo-MiNP	[36–37]
Di-isopropyl phthalate	DiPP
Mono-isopropyl phthalate	MiPP	[32]
Dimethyl phthalate	DMP	[19], [21]
Monomethyl phthalate	MMP	[32]
Di-n-octyl phthalate	DnOP	[21]
Mono-(3-carboxypropyl) phthalate	MCPP	[34–35]
Mono-n-octyl phthalate	MOP	[32]
Di-2-propylheptyl phthalate	DPHP	[8]
Dipentyl phthalate	DPP	[19]
Monopentyl phthalate	MPP	[32]
Terephthalates
Di-2-ethylhexyl terephthalate	DEHTP	[8], [19]
Mono-(2-ethyl-5-carboxypentyl) terephthalate	cx-MEPTP	[36], [38–39]
Mono-(2-ethylhexyl) terephthalate	MEHTP	[36], [39]
Mono-(2-ethyl-5-hydroxyhexyl) terephthalate	OH-MEHTP	[36], [39]
Mono-(2-ethyl-5-oxo-hexyl) terephthalate	oxo-MEHTP	[36]
Mono-(2-caboxyl-methyl-hexyl) terephthalate	cx-MMHTP	[36]
Alternative plasticizers
Acetyl tributyl citrate	ATBC	[8], [19]
Butylated hydroxytoluene	BHT	[8]
Di-2-ethylhexyl adipate	DEHA	[8], [19]
Mono-(2-ethylhexyl) adipate	MEHA	[36], [39]
Mono-(2-ethyl-5-hydroxyhexyl) adipate	OH-MEHA	[39]
Mono-(2-ethyl-5-oxohexyl) adipate	oxo-MEHA	[39]
Di-n-hexyl azelate	DnHA	[8]
Di-n-hexyl sebacate	DnHs	[8]
Di-isononyl cyclohexane-1,2-dicarboxylate	DINCH	[8]
Cyclohexane-1,2-dicarboxylic-mono-isononyl ester	MINCH	[36], [39]
Cyclohexane-1,2-dicarboxylic-mono-carboxyisooctyl ester	cx-MINCH	[36], [39]
Cyclohexane-1,2-dicarboxylic-mono-hydroxyisononyl ester	OH-MINCH	[36], [39]
Cyclohexane-1,2-dicarboxylic-mono-(7-hydroxy-4-methyl) octyl ester	oxo-MINCH	[36]
Ethanone-2m2-dimethoxy-1,2-diphenyl	DMPA	[8]
2,4-diphenyl-4-methyl-1-pentene	α-MSD	[8]
Tris-(2-ethylhexyl) trimellitate/Trioctyl trimellitate	TEHTM/TOTM	[8], [17], [19]
Di-2-ethylhexyl trimellitate	DEHTM	[17], [39]
Mono-(2-ethylhexyl) trimellitate	MEHTM	[17], [36]
Mono-(2-ethyl-5-carobyxpentyl) trimellitate	cx-MEHTM	[38]
Mono-(2-ethyl-5-hydroxyhexyl) trimellitate	OH-MEHTM	[17]
Mono-(2-ethyl-5-oxohexyl) trimellitate	oxo-MEHTM	[17]
Mono-(2-carboxymethylhexyl) trimellitate	cx-MMHTM	[17]
2,2,4-trimethyl-1,2-pentanediol di-iso-butyrate	TXIB	[8]
Volatile organic compounds
Acrolein
N-acetyl-S-(2-carboxyethyl)-L-cysteine	2CoEMA	[42–43]
N-acetyl-S-(3-hydroxypropyl)-L-cysteine	3HPMA	[42–43]
Acrylamide
N-acetyl-S-(2-carbamoylethyl)-L-cysteine	2CaEMA	[42–43]
N-acetyl-S-(2-carbamoyl-2-hydroxyethyl)-L-cysteine	2CaHEMA	[42–43]
Acrylonitrile
N-acetyl-S-(1-cyano-2-hydroxyethyl)-L-cysteine	1CyHEMA	[43]
N-acetyl-S-(2-cyanoethyl)-L-cysteine	2CyEMA	[42–43]
Benzene
Ethylbenzene	EB	[23]
1,2,4-trimethylbenzene	124-TMB	[23]
N-acetyl-S-(phenyl)-L-cysteine	PhMA	[42–43]
Trans,trans-muconic acid	MUCA	[42–43]
1,3-butadiene
N-acetyl-S-(3,4-dihydroxybutyl)-L-cysteine	34HBMA	[42–43]
N-acetyl-S-(4-hydroxy-2-buten-1-yl)-L-cysteine	4HBeMA	[42–43]
Butane
1-butanol	BuOH	[23]
Carbon disulfide
2-thioxothiazolidine-4-carboxylic acid	TTCA	[42]
Crotonaldehyde
N-acetyl-S-(3-hydroxypropyl-1-methyl)-L-cysteine	HPMMA	[42]
Cyanide
2-aminothiazoline-4-carboxylic acide	2ATCA	[42–43]
Cyclohexanone	CHX	[23], [40–41]
3,3,5-Trimethylcyclohexanone	335TMC	[23]
Cis 1,2-cyclohexanediol	C12C	[40–41]
Trans 1,2-cyclohexanediol	T12C	[40–41]
1,3-cyclohexanediol	13C	[41]
Cis 1,4-cyclohexanediol	C14C	[40]
Trans 1,4-cyclohexanediol	T14C	[40–41]
N,N-dimethylformamide
N-acetyl-S-(N-methylcarbamoyl)-L-cysteine	MCaMA	[42–43]
Ethylene oxide
N-acetyl-S-(2-hydroxyethyl)-L-cysteine	2HEMA	[42–43]
Propylene oxide
N-acetyl-S-(2-hydroxypropyl)-L-cysteine	2HPMA	[42–43]
Toluene
N-acetyl-S-(benzyl)-L-cysteine	BzMA	[42–43]
Styrene
N-acetyl-S-(1&2-phenyl-2-hydroxyethyl)-L-cysteine	2HPhEMA	[43]
Mandelic acid	MADA	[42–43]
Phenylglyoxylic acid	PhGA	[42–43]
Xylene
2-methylhippuric acid	2MHA	[43]
3-methylhippuric acid + 4-methylhippuric acid	3MHA+4MHA	[42–43]
a-Xylene	a-Xyl	[23]
m,p-Xylene	m,p-Xyl	[23]

Sources of chemical exposures in NICU/PICU/CICU populations

Twelve studies characterized environmental exposure to various industrial chemicals by estimating doses from medications or quantitatively assaying medical devices, medical equipment, and/or medical supplies commonly used in NICUS/PICU/CICUs (Table 1).

Bisphenols.

Only one study, Iribarne-Duran et al. (2019), tested for bisphenols in commonly used devices, equipment, and supplies in the NICU [12]. Authors tested 52 unused items from a NICU in Spain (though products were manufactured in 18 different countries), including feeding syringes and feeding tubes, wound dressings, diapers, stopcocks, and various intravenous fluid bags. Bisphenol A (BPA) was detected in 60% of all items tested. Overall, the items with the highest BPA concentrations were three-way stopcocks (7,050 ng/g), patterned transparent film dressings (688.1 ng/g), gastro-duodenal feeding tubes (301 ng/g), and sterile gloves (141 ng/g) (Supplemental Table S1).

Parabens.

Four studies evaluated parabens; one study quantified parabens in commonplace NICU items, and three estimated daily intake based on reported medications administered to the patient (Table 1). Iribarne-Duran et al. (2019) analyzed four parabens: methyl paraben (MeP), ethyl paraben (EtP), propyl paraben (PrP), and butyl paraben (BuP), and found at least one of them in 87% of all 52 items tested [12]. When considering route of exposure, MeP was the most frequently detected compound for oral feeding items (e.g., 63.7 ng/g for small dummy/pacifier nipple, 64.8 ng/g for gastro-duodenal feeding tube), dermal contact products (e.g., 481 ng/g for light therapy protection glasses, 208 ng/g for patterned transparent film dressing), and intravenous/parenteral items (e.g., 106 ng/g for intravenous infusion system extension set). BuP was the most frequently detected congener in respiratory devices, though not necessarily at the highest concentrations (compared to other congeners). The nasal cannula tested, for example, had a concentration of 22.3 ng/g for BuP but a concentration of 76.2 ng/g was found for EtP.

Of the three studies that estimated daily intakes of parabens from administered medications, two were exclusively among neonates admitted to the NICU [13,14], and one was among all patients admitted to an intensive care unit (adult, pediatric, or neonatal) [15]. Sources of paraben exposure via medications differed across studies, but major sources included enema solutions, vitamin K₂ oral syrup, gentamicin, sodium feredetate, and paracetamol/acetaminophen. Parabens are used as antimicrobials in several personal care products and pharmaceuticals. Similarly, wet sanitary wipes used on infants and children contain total paraben concentrations as high as 3.22 mg/g [16]. Sridharan et al. (2020) reported mean (standard deviation [SD]) estimated daily intake (all routes together) of MeP 0.16 (0.09) mg/kg-body weight for neonates and 0.28 (0.36) mg/kg-body weight for pediatric patients, which was significantly greater than those for adult patients (0.04 [0.06] mg/kg-body weight) (Supplemental Table S1) [15]. Conversely, the mean (SD) estimated daily intake of PrP was lowest for neonates (0.018 [0.01] mg/kg-body weight) compared to pediatric patients (0.03 [0.04] mg/kg-body weight) and adults (0.04 [0.1] mg/kg-body weight) (Supplemental Table S2). For the two neonatal-exclusive studies, the sum of all parabens was used to estimate exposure. The median (range) of estimated daily intake of parabens (all routes together) in Saito et al. (2021) was estimated to be 0.03 (0.02, 0.50) mg/kg-body weight [14], whereas for Iacobelli et al. (2023) the estimated daily intake was 2.1 (1.0, 3.1) mg/kg-body weight [13]. Across the three studies, mean (SD) cumulative paraben intake over the duration of hospital stay was as high as 80.3 (84.6) mg/kg-body weight [13], representing a median (IQR) of 3.0 (2.0, 7.0) days that exceeded acceptable daily intakes of parabens suggested by the European Food and Safety Authority (EFSA) (10 mg/kg-body weight/day).

Phthalates, terephthalates, and alternative plasticizers.

Seven studies evaluated phthalates, terephthalates, and/or alternative plasticizers in blood products, medications, medical devices/equipment/supplies, and/or parenteral nutrition solutions.

One study [17] reported that an alternative plasticizer, tris-(2-ethylhexyl) trimellitate (TEHTM/TOTM), was not detected in packaged red blood cells or fresh frozen plasma samples administered to neonates undergoing cardiopulmonary bypass (Supplemental Table S2). However, di-2-ethylhexyl phthalate (DEHP) and mono-2-ethylhexyl phthalate (MEHP, a metabolite) were detected in all samples, resulting in a single surgery-related chemical dose at levels ranging from 6x to 25x greater than the current tolerable daily intake (TDI) suggested by the EFSA (50 μg/kg-body weight) and up to 2x greater than the TDI suggested by the U.S. Food and Drug Administration (FDA) (600 μg/kg-body weight) (Supplemental Table S2) [17]. Another study that evaluated blood products showed that washing red blood cells prior to administration significantly lowers the concentrations of DEHP and MEHP [18]. For example, mean (SD) DEHP concentrations were 57.3 (25.5) μM prior to washing and 13.4 (4.9) μM after washing (Supplemental Table S2).

From three studies that assessed exposures via parenteral nutrition solutions routinely administered to neonates in NICUs [19,20], predominant phthalates detected included diethyl phthalate (DEP), benzyl butyl phthalate (BBzP), di-n-butyl phthalate (DnBP), and DEHP. DEHP was detected at the highest concentrations in one study (mean: 561 ng/mL before passing through infusion set, mean: 642 ng/mL after passing through infusion set) [19], while the other study observed BBzP at the highest concentrations (median: 8.4 ng/mL from leftover solutions in bag) [20]. Panneel et al. (2023) noted large differences in phthalate composition and concentrations depending on whether the nutrition solution was crystalloid- or lipid-based, finding a greater array of phthalates, and higher concentrations, for lipid-based solutions (Supplemental Table S2) [19]. For example, DnBP was infrequently detected (<50% detection) in crystalloid-based solutions while mean concentrations in lipid-based solutions were 44 ng/mL before infusion to 23.4 ng/mL after infusion (Supplemental Table S2) [19]. Panneel et al. (2023) analyzed for di-2-ethylhexyl terephthalate (DEHTP) and several alternative plasticizers (including TEHTM/TOTM), and found that acetyl tributyl citrate (ATBC) had the highest quantified concentrations of all assayed analytes (including phthalates), with mean concentrations ranging from 306 ng/mL before infusion and 6,024 ng/mL after infusion (Supplemental Table S2) [19]. As with phthalates, DEHTP and alternative plasticizers were infrequently detected in crystalloid-based solutions whereas they were frequently detected in lipid-based solutions. For example, DEHTP was detected in <50% of crystalloid-based solutions while the mean concentrations in lipid-based solutions ranged from 55.4 ng/mL before infusion to 30.6 ng/mL after infusion (Supplemental Table S2) [19]. Jenkins et al. (2021a) reported that soybean lipid emulsions (used in preparation of total parenteral nutrition) had no detectable DEHP from its original container, but they found the highest median DEHP concentrations (relative to other fluids, including saline) after infusion through standard intravenous infusion sets labeled as containing DEHP (median: 12,000 μg/L) (Supplemental Table S2) [7].

Two studies quantified phthalate concentrations in an array of medications and/or medical devices/equipment/supplies obtained from NICUs/PICUs, including over 170 products of diverse uses (e.g., first aid supplies, intravenous infusion sets, medical creams/liquids) [8,21]. DEHP was the most frequently detected phthalate in both studies, including from products labeled as “DEHP-free”; Malarvannan et al. (2019) additionally tested for DEHTP and alternative plasticizers, noting that di-2-ethylhexyl adipate (DEHA) and TEHTM/TOTM were the next most frequently detected plasticizers after DEHP (Supplemental Table S2) [***8]. Across studies, there were notable differences in phthalate composition and concentration by product type (Supplemental Table S2). For example, Wang and Kannan (2023) analyzed medical creams/liquids from a NICU and reported high abundance of diethyl phthalate (DEP), while all of the other product types were dominated by DEHP [***21]. For example, zinc oxide creams used as a topical skin product contained DEP concentrations as high as 2050 μg/g, and DEHP concentrations as high as 4.48 μg/g. These authors estimated that the median daily exposure doses for neonates from all direct sources of DEHP exposure (i.e., nipples, dummies/pacifiers, milk bottles, tubings, catheters, connectors, extensions, and cannulas) was 1.96 μg/kg-body weight while the maximum was 1,130 μg/kg-body weight, the latter of which is ~2x greater than the TDI suggested by the U.S. FDA [22]. Importantly, the daily exposure dose calculations were based on single-/incidental-use events, with most patients having frequent and/or long-term interactions with such medical products.

Volatile organic compounds.

Only one study assessed VOC exposures in NICU/PICU/CICU settings [23]. Significantly greater concentrations of cyclohexanone and trimethylcyclohexanone (a metabolite of cyclohexanone) were found in breast milk of mothers or donor milk after infusion through an enteral feeding system. This result highlights enteral feeding systems as a potential source of exposure, including formula- or nutrition solution-feedings. In maternal breast milk, for example, mean (SD) pre-infusion cyclohexanone concentrations were below the limit of detection compared to 1.14 mg/kg (2.23) at four hours of infusion (Supplemental Table S3) [23].

Biomonitoring of chemical exposures in NICU/PICU/CICU populations

Bisphenols.

One study quantified bisphenols in blood while five studies used urine (Table 1).

Blood.

Ten neonates who were treated in the NICU had substantially higher concentrations of bisphenol A bis(2,3-dihydroxypropyl) ether (BADGE∙2H₂O) in blood samples collected shortly after (or during) their NICU stay compared to that at age 7 months (median: 25.1 ng/mL vs. 2.47 ng/mL) (Supplemental Table S1) [24].

Urine.

Among the studies that used urine samples, the patient populations ranged from all pediatric patients receiving care in a PICU for at least one week [25,26], very low birth weight neonates receiving care in a NICU [27], neonates with and without a diagnosis of transient tachypnea of the newborn (TTN) [28], and neonates undergoing cardiac operations involving cardiopulmonary bypass [29].

In both studies of PICU patients, urinary concentrations of BPA were generally higher at one-week post-admission and PICU discharge compared to the first day of PICU admission (Supplemental Table S1) [25,26], and concentrations increased with a greater number of medical devices [25,26]. For example, mean urinary BPA concentrations among those with four or more medical devices at a time (n = 68) was 220 μg/g-creatinine (cr) compared to 145 μg/g-cr for those with fewer than four devices (n = 47) [26]. Further, among 117 participants, BPA concentrations were highest among those who received hemodiafiltration/hemodialysis procedures; mean (SD) urinary BPA concentrations among patients receiving continuous veno-venous hemodiafiltration was 481 μg/g-cr compared to 114 μg/g-cr for those without [25]. From the study of very low birthweight neonates (n = 50), median urinary concentrations of BPA were modestly higher in the first week of life compared to the fifth week (11.3 ng/mL vs. 7.27 ng/mL) (Supplemental Table S1) [27]. In the study of neonates with (n = 32) and without TTN (n = 33), urinary concentrations of BPA were significantly higher among those with TTN (median: 8.0 μg/g-cr, IQR: 3.6, 26.1) compared to those without TTN (median: 2.0 μg/g-cr, IQR: 1.0, 3.9) (Supplemental Table S1) [28]. Moreover, in a study of 18 neonates undergoing cardiac operations involving cardiopulmonary bypass, there were no differences in preoperative, specific gravity-corrected urinary BPA concentrations relative to the postoperative period (Supplemental Table S1) [29]. However, geometric mean BPA concentrations were more than 10x higher in neonates (13.9 μg/g-cr) than those of their mothers (1.3 μg/g-cr) in urine samples collected at the same time [29].

Parabens.

Only one study evaluated exposure to parabens [27]. In this study of very low birthweight neonates (n = 50), urinary concentrations of four parabens (MeP, EtP, PrP, and BuP) were quantified in the first and fifth week of life, with median concentrations of MeP being substantially higher in the first week relative to the fifth week (541 ng/mL vs. 261 ng/mL), and concentrations of MeP were higher than the other parabens at both timepoints (Supplemental Table S1).

Phthalates, terephthalates, and alternative plasticizers.

Fifteen studies evaluated NICU/PICU/CICU exposure to plasticizers using blood (n = 3), urine (n = 12), and hair (n = 1); one of these studies used both blood and urine samples (Table 1).

Blood.

In a comparison of two pediatric cohorts of PICU patients, the Tight Glucose Control Study (enrolled 2004 to 2007, n = 216) and the PEPaNIC Study (enrolled 2012 to 2015, n = 334), average plasma concentrations of MECPP, MEHHP, and MEOHP (assessed on the last day of PICU stay) were significantly lower in PEPaNIC participants than those in the Tight Glucose Control Study, even after accounting for confounders (e.g., age, sex, height, weight, indication for PICU stay, and length of PICU stay), suggesting that exposure to DEHP in the PICU had decreased over time [30].

The other two studies that used blood as a matrix were of pediatric patients undergoing cardiac operations. Eckert et al. (2020) did not detect TEHTM or its primary metabolites, mono-(2-ethylhexyl) trimellitate (MEHTM) and di-2-ethylhexyl trimellitate (DEHTM), in preoperative blood samples from 21 infants, nor did they detect DEHP and its primary metabolite (MEHP) in the same samples [17]. However, in postoperative samples, median concentrations were 18.4 μg/L, 1046 μg/L, and 158 μg/L for TEHTM, DEHP, and MEHP, respectively (Supplemental Table S2) [17]. Guerrelli et al. (2024) analyzed perioperative blood samples from 110 children for DEHP and DEHP metabolites, and found notable differences in concentrations between postoperative and preoperative samples, and differences depended on whether the operation involved cardiopulmonary bypass and, if so, the priming solution used [18]. For example, postoperative day 0 DEHP concentrations (mean [SD]) were highest when cardiopulmonary bypass was used and primed with red blood cells (3.79 μmol [2.24]) and lowest when cardiopulmonary bypass was not used during surgery (0.92 μmol [0.28]) (Supplemental Table S2) [18].

Urine.

Of the 12 studies that used urine as the biological matrix, three were of children with congenital heart defects, two of which specifically assessed exposures relative to cardiac operations. Eckert et al. (2020) did not detect TEHTM metabolites in any urine samples collected before or after cardiac surgery from 21 infants. However, DEHP metabolites were detected in all urine samples at both timepoints, and concentrations were significantly higher in the postoperative samples (Supplemental Table S2) [17]. For example, median (IQR) concentrations of mono-(2-ethyl-5-carboxypentyl) phthalate (MECPP) were ~27x higher in postoperative samples (242 [9.4, 1073] μg/L) compared to preoperative samples (9.1 [2.2, 89.8] μg/L) (Supplemental Table S2). Gaynor et al. (2019) similarly observed that concentrations of DEHP metabolites were substantially higher in postoperative samples relative to preoperative samples collected from 18 neonates, with MECPP, mono-(2-ethyl-5-hydroxyhexyl) phthalate (MEHHP), and mono-(2-ethyl-5-oxohexyl) phthalate (MEOHP) detected in all preoperative samples, and these three metabolites (along with MEHP) were detected in all postoperative samples [29]. Concentrations of MECPP were particularly high in both pre- and postoperative samples, compared to the other metabolites, and especially high in postoperative samples where the geometric mean concentration was 1166 μg/L, compared to 58.5 μg/L in preoperative samples (Supplemental Table S2) [29]. Among 16 pediatric patients with congenital heart defects undergoing catheterization in the PICU, Weng et al. (2022) quantified DEHP metabolites in all urine samples collected at initial hospitalization, pre-cardiac catheterization, post-catheterization, and at hospital discharge. Concentrations of MEHHP and MEOHP were quantified at the highest concentrations at all timepoints. The geometric mean concentrations of MEHHP and MEOHP at pre-cardiac catheterization were 30.9 μg/g-cr and 30.5 μg/g-cr, respectively (Supplemental Table S2) [31].

Five studies exclusively focused on preterm infants in the NICU. In Al-Saleh et al. (2023), all urinary phthalate concentrations in samples collected from 33 neonates were significantly higher in post-admission samples than in samples collected at NICU admission, and concentrations were the highest for MECPP. For example, the median concentration of MECPP at NICU admission was 220.5 μg/g-cr compared to 1861 μg/g-cr post-admission (prior to discharge) (Supplemental Table S2) [20]. The remaining four studies included participants of the NICU-HEALTH cohort, either alone or in tandem with other NICU cohorts (as part of the DINE cohort of the Environmental Influences on Child Health Outcomes [ECHO] program).

In NICU-HEALTH samples collected from 64 very preterm infants between 2011–2013, mono-isopropyl phthalate (MiPP), mono-(2-carboxymethyl hexyl) phthalate (MMCHP), monopentyl phthalate (MPP), mono-isononyl phthalate (MiNP), and mono-n-octyl phthalate (MOP) were infrequently detected in urine samples collected between NICU admission and 34 weeks postmenstrual age. All other phthalate metabolites were detected in 89% (mono-3-carboxypropyl phthalate [MCPP]); 99% (MBzP); or 100% (monomethyl phthalate [MMP], MEP, MECPP, MnBP, mono-isobutyl phthalate [MiBP], MEHHP, MBzP, MEHP, and MEOHP) of samples, with MECPP quantified at the highest specific gravity-adjusted concentrations (median: 49.7 ng/mL) (Supplemental Table S2) [32]. In a separate study of the same neonates, authors estimated associations between specific medical equipment bundles (e.g., invasive respiratory support) and estimated ∑DEHP concentrations, finding that (after accounting for co-exposures to medical equipment bundles) those receiving respiratory support (especially non-invasive) had substantially higher ∑DEHP concentrations than those without respiratory support [33]. A study of 101 NICU-HEALTH participants enrolled from 2011 to 2020 also observed high detection of phthalate metabolites, with the highest concentration found for MECPP. Specifically, among males, the median MECPP concentration was 354 ng/mL whereas those for females was 440 ng/mL (Supplemental Table S2) [34]. Lastly, among 360 preterm infants from two cohorts (Prematurity and Respiratory Outcomes Program [PROP] and NICU-HEALTH) contributing to the larger DINE ECHO cohort, the highest concentrations for DEHP metabolites, reported as ∑DEHP (molar sum of MECPP, MEHHP, MEHP, and MEOHP concentrations) were found (Supplemental Table S2) [35].

Three studies of neonates admitted to the NICU were conducted as part of the Assessment and Risk Management of Medical Devices in plasticized polyvinyl chloride (ARMED) project. Urine specimens collected from 104 neonates after 24 hours from NICU admission were analyzed for five DEHP metabolites, four di-isononyl phthalate (DiNP) metabolites, four di-isononyl cyclohexane-1,2-dicarboxylate (DINCH) metabolites, one DEHA metabolite, five DEHTP metabolites, and three TEHTM/TOTM metabolites [36]. All metabolites were detected in at least one urine sample, though detection rates were <50% for MiNP; mono-hydroxy-isononyl phthalate (OH-MiNP); cyclohexane-1,2-dicarboxylic-mono-isononyl ester (MINCH); cyclohexane-1,2-dicarboxylic-mono-hydroxyisononyl ester (OH-MINCH); cyclohexane-1,2-dicarboxylic-mono-(7-hydroxy-4-methyl) octyl ester (oxo-MINCH); and mono-(2-ethylhexyl) terephthalate (MEHTP). MECPP was found at the highest concentrations (median: 193 ng/mL), followed by MEHP (median: 22.9 ng/mL), and mono-(2-ethyl-5-carboxypentyl) terephthalate (cx-MEPTP) (median: 3.54 ng/mL) (Supplemental Table S2) [36]. The same study sample of 104 neonates was evaluated in Bernard et al. (2021), but all samples from the first six days post-NICU admission were utilized. Authors compared phthalate concentrations across patient and treatment characteristics, finding notable differences. For example, urine samples collected from patients who received ECMO had a mean ∑DEHP concentration (molar sum of MEHP, MEHHP, MEOHP, and MECPP concentrations) of 23,155 ng/μmol-cr, which was ~36x higher than the concentrations found in all other neonates (640.27 ng/μmol-cr) [37]. Bernard et al. (2023) quantified DEHP, DEHTP, and TEHTM/TOTM metabolites in 508 urine samples collected daily from 97 neonates admitted to NICUs at two centers from 2018 to 2019 [38]. DEHP metabolites were detected in all urine samples while detection rates of DEHTP and TEHTM/TOTM metabolites varied. As examples, mono-(2-ethyl-5-hydroxyhexyl) terephthalate (OH-MEHTP) was detected in 100% of samples whereas MEHTP was detected in 65%, and 96% of samples had detectable 2-MEHTM whereas only 12% of samples had detectable 2,4-DEHTM [38]. Generally, carboxylated derivates of DEHP and DEHTP were quantified at higher concentrations than hydroxylated or monoester derivatives, whereas primary monoester metabolites were detected at higher levels for TEHTM/TOTM. In a subset of neonates with urine samples collected after NICU discharge (but still being cared for in a neonatal nursery), urinary concentrations of DEHP and TEHTM/TOTM metabolites were significantly lower at hospital discharge compared to the last day in the NICU. For example, in one of the centers, median (IQR) concentrations of ∑DEHP (sum of all DEHP metabolite concentrations) dropped from 420 (151, 697) ng/mL at last NICU day to 17 (10, 30) ng/mL at hospital discharge [38].

Hair.

In the only study that used hair as a biological matrix, authors quantified 23 phthalate, terephthalate, and alternative plasticizer metabolites in samples collected from 45 neonates in the nursery and NICU, as well as their mothers [***39]. All metabolites of DEP, DiNP, BBzP, DiNP, DINCH, and di-isodecyl phthalate (DiDP) were detected in <50% of samples, while metabolites of DEHP (MEHP), DnBP (MnBP), DiBP (MiBP), DEHA (MEHA), DEHTP (MEHTP), and TEHTM/TOTM (1,4+2,4-DEHTM) were detected in ≥ 50% of samples. MEHTP was quantified at the highest concentrations across all samples but was substantially higher in NICU neonates compared to the other groups. For example, the median (IQR) concentration was 964 (322, 1926) ng/g in NICU neonates compared to 133 (60, 345) ng/g in non-NICU neonates (Supplemental Table S2) [***39]. Additionally, concentrations of MEHTP were significantly correlated with days of non-invasive ventilation (Spearman’s ρ = 0.38) as well as the use of a gastric tube (Spearman’s ρ = 0.42) [***39].

Volatile organic compounds.

Four studies quantified VOCs in blood (n = 2) or urine (n = 2) of children admitted to the NICU/PICU/CICU (Table 1). Study populations included preterm infants resting in cribs or incubators, pediatric patients receiving ECMO, and neonates with congenital heart defects undergoing cardiac operations involving cardiopulmonary bypass.

Blood.

Both studies that used blood samples solely quantified cyclohexanone and its metabolites [40,41]. Bembea et al. (2022) noted minimal detection of cyclohexanone on day one of ECMO (13% of 90 participants) but universal detection of metabolites (1,2- and 1,4-cyclohexanediol) (Supplemental Table S3). Median day one concentrations of cyclohexanone metabolites (molar sum of 1,2- and 1,4-cyclohexanediol) were significantly higher for those supported by infant circuits (5.7 ng/μL) than those supported by pediatric circuits (3.1 ng/μL) [40]. Everett et al. (2020) found that, among 85 neonates, the median (IQR) serum concentrations of cyclohexanone were significantly higher in samples collected immediately after cardiopulmonary bypass cessation (1744 μg/L [1469, 2291]) compared to those collected in the preoperative period (572 μg/L [389, 974]) and at 12 hours post-operation (146 μg/L [77, 278]) (Supplemental Table S3) [41]. Conversely, concentrations of cyclohexanone metabolites were the highest at 12 hours postoperation (Supplemental Table S3).

Urine.

In the two studies that used urine, VOCs were detected in every sample. Gaynor et al. (2024) characterized exposure to VOCs among 10 neonates with congenital heart defects undergoing cardiac operations [42]. Authors detected a median concentration of 14 VOC metabolites per urine sample, with BzMA (a metabolite of toluene and benzyl alcohol), 34HBMA (a metabolite of 1,3 butadiene), and 3HPMA (a metabolite of acrolein) detected in all daily urine samples (Supplemental Table S3), and geometric mean BzMA concentrations (20.3 ng/mL) at the 90^th percentile of National Health and Nutrition Examination Survey (NHANES) participants ages 3 to 5 years in the 2015–2016 cycle [42]. El-Metwally et al. (2018) reported that all preterm neonate urine samples had detectable concentrations of six VOC metabolites, including BzMA and 3HPMA (Supplemental Table S3) [43]. Authors noted significant differences in urinary metabolite concentrations comparing those in incubators (n = 40) vs. cribs (n = 40), with 11 metabolite concentrations (from 10 parent compounds) higher in the incubator group and two metabolites (both of 1,3 butadiene) higher in the crib group. Median (IQR) BzMA concentrations, for example, were 106 ng/mL (18.7, 388) in the incubator group and 63.1 ng/mL (5.30, 479) in the open crib group (Supplemental Table S3) [43]. Notably, when compared to NHANES 2011–2012 participants ages 6 to 11 years, geometric mean BzMA concentrations were 7x greater in the crib group and 12x greater in the incubator group [43].

Health implications of chemical exposures in medical settings

Ten studies evaluated associations of NICU/PICU/CICU-based chemical exposures with health outcomes. These outcomes were related to endpoints of neurodevelopment (n = 4), cardiometabolism (n = 2), pulmonary function (n = 2), physical growth and development (n = 1), and mortality/other (n = 1) (Table 3).

Table 3.

Summary of findings from studies evaluating health effects of chemical exposures in NICU/PICU/CICU.

Reference	Study population	Outcome domain	Health impact observed
Bisphenols
Buyukeren 2023 [28]	32 newborns admitted to NICU and 33 newborns not admitted to NICU	Pulmonary	Newborns with greater urinary BPA concentrations were more likely to have severe TTN
Phthalates
Guerrelli et al. 2024 [18]	110 pediatric patients undergoing cardiac operations	Cardiometabolism	Higher blood ∑DEHP concentration associated with: higher postoperative blood anion gap higher postoperative blood sugar concentrations higher postoperative blood sodium lower postoperative blood potassium concentrations lower postoperative blood calcium concentrations higher postoperative blood lactate concentrations higher postoperative blood pH higher risk for postoperative complications
Jenkins et al. 2021b [11]	31 premature infants with bronchopulmonary dysplasia	Cardiometabolism	Use of DEHP-free IV fluids in NICU associated with lower incidence of hypertension compared to time-periods using DEHP-containing IV fluids
Busgang et al. 2022 [34]	101 very preterm neonates	Physical growth	Shorter latency to first growth spurt per doubling in urinary concentration of: MBzP, MECPP, MEHHP, MEOHP Higher growth rate per doubling in urinary concentration of: MECPP, MEHHP, MEOHP
Stroustrup et al. 2018 [32]	64 premature and very low birth weight neonates	Neurodevelopment	Higher attention and regulation scores per 10 ng/mL increase in ∑DEHP concentration
Vanhorebeek et al. 2022 [30]	550 pediatric patients admitted to PICU	Neurodevelopment	Being in a PICU cohort with higher DEHP exposure was not associated with measures of attention evaluated four years later
Al-Saleh et al. 2023 [20]	33 preterm neonates admitted to NICU and required TPN solution	Neurodevelopment	Higher gross motor scores at age 2 months per unit increase in urinary concentration of: ∑DEHP, MBzP, and MEP Lower fine motor scores at age 2 months per unit increase in urinary concentration of: MECPP and MEHP Higher problem-solving scores at age 2 months per unit increase in urinary concentration of: MECPP, MEHP, MEOHP, MEP, and MnBP Lower problem-solving scores at age 2 months per unit increase in urinary concentration of MiBP
Stroustrup et al. 2023 [35]	360 extremely or very preterm neonates admitted to NICU	Pulmonary	Higher urinary ∑DEHP concentrations between 26–30 and 34–36 weeks corrected-gestational age associated with higher risk of BPD diagnosis Higher urinary MBzP concentrations between 34–36 weeks corrected-gestational age associated with higher risk of BPD diagnosis Higher urinary MEP concentrations between 34–36 weeks corrected-gestational age associated with higher risk of BPD diagnosis
Everett et al. 2020 [41]	85 neonates with congenital heart defects receiving a cardiac operation involving cardiopulmonary bypass	Neurodevelopment	An IQR increase in geometric mean of perioperative blood cyclohexanone concentrations associated with lower cognitive and language composite scores at age 12 months
Bembea et al. 2022 [40]	90 pediatric patients admitted to PICU and requiring ECMO	Unfavorable outcome at hospital discharge	Higher risk of unfavorable outcome per doubling in blood ∑cyclohexanediol metabolites concentration

Abbreviations: ASQ-III, Ages and Stages Questionnaire – 3^rd edition; BPD, bronchopulmonary dysplasia; BSID-III, Bayley’s Scales of Infant and Toddler Development – 3^rd edition; IQR, interquartile range; PCPC, Pediatric Cerebral Performance Category; TTN, severity of tachycardia of newborn.

^aMolar sum (∑) of DEHP metabolites

^bMolar sum (∑) of cyclohexanone metabolites

Bisphenols.

In the only study evaluating bisphenols, preterm neonates with TTN had higher urinary BPA concentrations compared to those without TTN, and having a creatinine-standardized BPA concentration of ≥2.65 μg/g-cr was predictive of TTN diagnosis (area under the curve = 83%) [28].

Phthalates, terephthalates, and alternative plasticizers.

Seven studies reported health outcomes from phthalates exposures in the NICU/PICU/CICU; five of these studies exclusively included premature infants. None of the studies involved associations of terephthalates or alternative plasticizers with health outcomes. Outcome domains included neurodevelopment (n = 3), cardiometabolism (n = 2), physical growth (n = 1), and pulmonary function (n = 1). Below, we highlight findings from these studies by outcome domain.

Neurodevelopment.

Stroustrup et al. (2018) observed that, among premature infants admitted to the NICU and tested for neurodevelopment using the NICU Network Neurobehavioral Scales (NNNS) prior to discharge, those with higher urinary concentrations of ∑DEHP (molar sum of MEHP, MEHHP, MEOHP, and MECPP concentrations) had higher attention and regulation domain scores (indicating a positive association) [32]. Using weighted quantile sums (WQS) regression to evaluate the effect of the phthalate mixture on NNNS domain scores, higher phthalate exposure index values were associated with improved performance on habituation (driven by MECPP, MEP, and MMP), attention (driven by MEOHP and MEHHP), and regulation (driven by MEOHP and MEP) scores; as well as reduced performance on handling (driven by MECPP, MEP, and MMP), non-optimal reflexes (driven by MEOHP, MCPP, and MEP), and excitability scores (driven by MEOHP, MCPP, and MEP) [32]. Among preterm neonates receiving total parenteral nutrition, Al-Saleh et al. (2023) observed that higher average urinary phthalate concentrations while in the NICU (i.e., average of daily urine phthalate concentrations), for multiple metabolites, were associated with differential neurodevelopment scores at age 2 months, as assessed by the Ages and Stages Questionnaire, 3^rd edition (ASQ-III) [20]. Specifically, higher urinary concentrations of MEP and MBzP were each associated with higher gross motor scores, while MnBP, MECPP, MEOHP, and ∑DEHP (molar sum of MEHP, MEHHP, MEOHP, and MECPP concentrations) were each associated with higher problem-solving scores [20]. Vanhorebeek et al. (2022) found no significant differences in attention measures assessed four years post-PICU discharge when comparing those recruited from a cohort with higher urinary phthalate concentrations relative to those with lower phthalate concentrations [30].

Cardiometabolism.

A study of pediatric patients undergoing cardiac operations found that higher postoperative blood concentrations (reported as “equivalent levels”) of phthalates (sum of DEHP, MEHP, MECPP, and MEHHP molar concentrations) were associated with multiple postoperative cardiometabolic endpoints, including higher blood anion gap, hyperglycemia, hypernatremia, hypokalemia, hypocalcemia, higher blood lactate concentrations, higher blood pH, and higher risk of postoperative complications (e.g., cardiac arrest, acute kidney injury and renal insufficiency) [18]. In a study of premature, very low birthweight infants, authors conducted a natural experiment to determine the effects of DEHP on hypertension incidence among those with bronchopulmonary dysplasia. In the baseline period where DEHP-containing intravenous fluids were used, incidence of hypertension was 7.7%, dropping to 1.4% when the NICU switched to DEHP-free fluids, and returning to 10.1% when DEHP-containing fluids returned to use [11].

Physical growth.

In a study of very preterm neonates admitted to the NICU, higher postnatal urinary phthalate concentrations for multiple metabolites were associated with altered physical growth trajectories in early life [34]. Specifically, higher concentrations of MBzP, MECPP, MEHHP, and MEOHP were each associated with shorter latency time to first growth spurt as well as higher growth rates, as assessed from birth up to corrected-gestational age 40 weeks (based on postmenstrual age) [34].

Pulmonary.

Among extremely or very preterm neonates admitted to the NICU, higher urinary ∑DEHP (molar sum of MEHP, MEHHP, MEOHP, and MECPP concentrations), MBzP, and MEP concentrations between 26–30 weeks postmenstrual age (∑DEHP) or 34–36 weeks postmenstrual age (∑DEHP, MBzP, and MEP) were associated with higher risk of bronchopulmonary dysplasia diagnosis at 36 weeks postmenstrual age [35].

Volatile organic compounds.

Only two studies assessed health impacts of VOC exposure in the NICU/PICU/CICU, both of which focused on cyclohexanone and its cyclohexanediol metabolites. Everett et al. (2020) reported that among neonates with congenital heart defects who underwent a cardiac operation involving cardiopulmonary bypass, an IQR increase in geometric mean perioperative serum cyclohexanone concentrations was associated with lower cognitive and language domains of the Bayley’s Scaled of Infant and Toddler Development – 3^rd edition (BSID-III) during follow-up at age 12 months [41]. Bembea et al. (2022) found that a doubling in plasma cyclohexanediol (molar sum of 1,2- and 1,4-cyclohexanediol metabolites) concentration on day one of ECMO was associated with greater risk of dying before hospital discharge or being discharged with an worse neurofunctional status (compared to baseline), using the Pediatric Cerebral Performance Category score [40].

Summary of findings: critical knowledge gaps and implications for research and beyond

Recent studies show that neonates and children are exposed to a tremendous array of EDCs in NICU, PICU, and CICU settings. Sources of exposure varied by chemical class. However, fluids and/or medications given for intravenous therapies (e.g., sodium chloride solutions, red blood cells, parenteral nutrition solutions), as well as the equipment sets used for their administration, were common sources shared across multiple chemical classes, including bisphenols [12], parabens [13–15], phthalates [7,19–21], terephthalates [19], alternative plasticizers [19], and VOCs [23]. These sources may be particularly important since the leached chemicals contained in these fluids are 100% bioavailable given the route of intravenous administration, implying a greater received exposure dose per administration. Fluids which are lipid-rich, such as lipid emulsions as well as red blood cells, are also particularly prone to accumulation of lipophilic chemicals, like phthalates [21,44]. This was evidenced by higher phthalate concentrations in blood samples from children receiving cardiopulmonary bypass where RBCs were used as the priming solution over crystalloid solutions [18], as well as in parenteral nutrition that included lipid emulsions [7,19,21]. For example, when using a parenteral feeding system that contained DEHP, a single administration of soybean-based 20% lipid emulsion for parenteral nutrition, prepared for administration to a 3.0 kg neonate (at a standard flow rate of 1.8 mL/hour over 24 hours), was shown to result in a median dose of 173 μg/kg-body weight per administration, which is over 3x the TDI set by the EFSA [7]. However, it is again important to note that this is just one exposure source during a single event in the NICU. There are often multiple simultaneous exposure sources to these chemicals over multiple use-events and days, implying potentially large cumulative exposures during the duration of stay in intensive care unit [7,21]. For example, over the course of NICU stay for 14 very low birthweight neonates, one study estimated a mean cumulative DEHP exposure of 230,207 μg, resulting in a cumulative dose of 182,369 μg/kg-body weight, equating to a daily dose of 2306 μg/kg-body weight, which is ~46x higher than the EFSA TDI and ~3.8x higher than the U.S. FDA TDI [7]. Furthermore, certain treatments or procedures are likely sources of substantial exposures, e.g., ECMO, cardiopulmonary bypass, and transfusions, since blood has high surface area contacts with medical devices/equipment during their use [17,18,29,40–42].

While certain chemicals were frequently included in studies (over half included assessment of DEHP exposure), only a fraction evaluated contemporary substitutions for these legacy chemicals. Terephthalates (e.g., DEHTP) and alternative plasticizers (e.g., ATBC, DINCH, TEHTM/TOTM, DEHA) are increasingly used as alternatives to phthalates (especially DEHTP as an alternative to DEHP) [45]. ATBC, for example, is often used in blood bags and tubing given its anti-coagulant properties, while DEHA is also used in blood bags due to its desirable low temperature plasticizing properties [46]. Yet, from the recent studies included in this review, only five studies assessed exposures to terephthalates (all focused solely on DEHTP) [8,19,36–39] and only six studies evaluated at least one alternative plasticizer [8,17,19,36–39]. None of these studies assessed the health impacts of exposure to these phthalate alternatives, which is concerning since some compounds have much higher migration potential than DEHP (e.g., ATBC and DEHA) [47]. For bisphenols, similarly, no studies evaluated exposures to BPA replacements (e.g., bisphenol F [BPF], bisphenol S [BPS]), and only one study assessed other emerging bisphenols (e.g., BADGE, BADGE∙2H₂O) [24]. While OPEs are increasingly utilized as plasticizers [48–50], in addition to their use as flame retardants [51], no studies have tested for OPEs in medical products or biospecimens of NICU/PICU/CICU populations. This is true for PFAS as well; most medical products are manufactured with polytetrafluoroethylene (PTFE) [52], a fluoropolymer that is considered biologically inert, legacy PFAS (e.g., perfluorooctanoic acid [PFOA]) are often used in tandem with PTFE during the manufacturing process [22], and therefore could be present in the final product as an unintended contaminant. Importantly, an argument justifying the continued use of PTFE in medical products is that they provide enhanced safety to the final products, e.g., higher durability of PTFE grafts over other materials; however, a recent review of PTFE prepared for use by the U.S. FDA found limited evidence to support this claim [53]. Taken together, future studies should include exposure assessment for terephthalates, alternative plasticizers, OPEs, PFAS, and other chemicals such as methyl siloxanes, in an effort to better understand the broader exposure landscape in NICUs and PICUs.

Health impacts of intensive care unit chemical exposures were scarce, with just 10 studies including any assessment of health outcomes and only three studies considering health outcomes after hospital discharge [20,30,41]. Comparisons across studies is a challenge given the large heterogeneity in patient populations, methods and timing of exposure assessment, and the specific chemicals evaluated in each study. However, this lack of congruence across studies highlights the critical need for additional studies to not only characterize exposures in NICU/PICU/CICU settings and populations, but for studies to evaluate the immediate-, short-, and long-term implications of exposures as well. A focus on long-term health impacts is particularly relevant since early life chemical exposures have been shown to affect health trajectories across the life-course [54], including into adulthood [55].

The paucity of studies characterizing exposures and health impacts of newly introduced or increasingly utilized chemicals, particularly those serving as replacements to existing chemicals, is concerning. Still, current evidence supports the European Union’s Medical Devices Regulation (2017/745) which was submitted to application in 2021 and would require manufacturers to present justification and labelling for products containing greater than 0.1% of EDCs (by weight). However, robust, human-relevant toxicological assessments for many chemicals, including for alternative plasticizers currently used in manufacturing of medical products, are extremely limited [56]. As such, manufacturers aiming for compliance with new regulations may introduce new chemicals into the marketplace that ultimately have comparable or even more toxicity than those they are replacing (i.e., “regrettable substitutions”). BPF and BPS, as examples, have comparable endocrine-disrupting potentials to BPA [57]. A limitation of the studies is whether devices from different manufacturers, or different lots from the same manufacturer, vary substantially with respect to levels of chemicals of concern. Without testing devices for chemicals of concern, a manufacturer may not even be aware of where chemicals of concern are present in their process. Therefore, it is vital and prudent that all stakeholders, including healthcare professionals, researchers, and manufacturers, work together to produce and evaluate a comprehensive evidence base to inform paths forward.

While the studies included in this review provide crucially needed evidence to better understand the exposure landscape in NICUs/PICUs/CICUs and their health implications, they are not without limitations. Several of the included studies had relatively small sample sizes (e.g., <50 participants), which can limit statistical power but also hinder efforts to control for all necessary confounders in statistical analyses, the latter of which may be particularly relevant depending on the heterogeneity of the pediatric population being studied. Additionally, since many of these studied chemicals are EDCs, evaluation of sex-specific associations of chemicals with health outcomes may be an important consideration, and to do so would therefore require an even larger sample size. Relatedly, single biospecimen collections, especially for non-persistent pollutants with short half-lives (e.g., phthalates), may not adequately reflect average exposures over the duration of hospital stay. Repeated sample collections may therefore be helpful in reducing exposure measurement error bias, and they can also be informative of intra-individual variation in exposure over the duration of stay in the intensive care unit as well as indicate their primary sources. As an example, several studies in this review conducted repeated sample collections around a key event, such as before and after cardiac operations involving cardiopulmonary bypass [17,18,29,41,42]. As evidenced by the studies in this review, children are exposed to multiple chemicals simultaneously during their stay in an intensive care unit, and these chemical mixtures likely change over their stay depending on provided treatments as well as operations. Another source of potential exposures is through provision of nutrients like breast milk, as pumping and storage of breast milk is common in an in-patient setting, even if the source of breast milk is mother’s own milk. Therefore, storage containers and the methods of warming of breast milk may increase leaching of certain chemicals. However, very few studies in this review characterized exposure mixtures and connected these to health outcomes [33–35], and no studies evaluated mixtures of chemicals from multiple chemical classes. It is therefore imperative that future studies incorporate formal evaluations of chemical mixtures in relation to health outcomes, and that they also consider the broader chemical landscape in their analyses and not just a single chemical class. Finally, as aforementioned, few studies evaluated health outcomes beyond hospital discharge. As such, it is difficult to elucidate the long-term impacts resulting from chemical exposures incurred while in an intensive care unit (or hospital, broadly); therefore, evaluation of longer-term follow-up should be a priority of future research.

Conclusions

This review identified the several key knowledge gaps that exist within understanding the chemical exposure landscape of children admitted to NICUs/PICUs, including the potential impacts of such exposures on health outcomes across the spectrum of domains and in ranging timeframes (e.g., short-term and long-term). Exposure assessment that considers multiple chemical classes, as well as increasingly utilized substitutions/replacements for legacy compounds, should be a high priority, especially to better understand the effects of chemical mixtures on health outcomes among some of our most vulnerable populations of children.