Carcinogenic and Tobacco Smoke-Derived Particulate Matter Biomarker Uptake and Associated Healthcare Patterns among Children

Abstract

Background:

The objective was to assess the associations of child tobacco smoke exposure (TSE) biomarkers (urinary cotinine, NNAL, and nicotelline N-oxides) and parent-reported smoking and child TSE patterns with total hospital visits, pediatric emergency department (PED) visits, urgent care (UC), revisits, and hospital admissions among 0–9-year-olds.

Methods:

A convenience sample of PED/UC patients (N=242) who presented to a large, U.S. children’s hospital who had baseline urine samples assayed for the TSE biomarkers of interest were included. Biomarker levels were log-transformed, and linear and Poisson regression models were built.

Results:

The geometric means of child cotinine, creatinine-adjusted NNAL, and N-oxide levels were 11.2ng/ml, 30.9pg/mg creatinine, and 24.1pg/ml, respectively. The mean (SD) number of daily cigarettes smoked by parents was 10.2 (6.1) cigarettes. Each one-unit increase in log-NNAL levels was associated with an increase in total UC visits (aRR=1.68, 95%CI=1.18–2.39) among 0–9-year-olds, while controlling for the covariates. Each one-unit increase in child log-NNAL/cotinine ratio (x10³) values was associated with an increase in total hospital visits (aRR=1.39, 95%CI=1.10–1.75) and UC visits (aRR=1.56, 95%CI=1.14–2.13) over 6-months.

Conclusion:

Systematic screening for child TSE should be conducted during all hospital visits. The comprehensive assessment of TSE biomarkers should be considered to objectively measure young children’s exposure.

Introduction

Involuntary exposure to tobacco smoke is recognized as a major health harm.¹ There are thousands of toxic pollutants in tobacco smoke, which include an estimated 70 known or probable carcinogens.² Consequently, young children are most vulnerable to TSE due to their age-specific behaviors, physiology, and body size.³ The dangers of TSE during childhood are well-documented as TSE causes acute and chronic illnesses such as respiratory infections, ear infections, and asthma.¹ Nearly 4-in-10 U.S. children ages 3–11 years are exposed to tobacco smoke as biochemically measured by cotinine, the major proximate nicotine metabolite.⁴ Disparities exist with higher cotinine levels observed among certain child groups including Black/African American children and those with low socioeconomic backgrounds.⁵ While not much research has examined TSE biomarkers in younger children (e.g., <3 years), one study of 0–9-year-olds observed a negative association between child age and cotinine levels, with 0–1-year-olds having the highest exposure levels.⁶ Concerning the degree of exposure, other research reported that over 1-in-10 low-income infants had cotinine levels indicative of active smoking.⁷

Prior studies report that children with higher cotinine levels have higher healthcare utilization patterns.^6,8 However, cotinine measurements may lead to underestimated exposure to the strongest carcinogenic tobacco-specific nitrosamine (TSNA), 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone (NNK), which is metabolized to 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanol (NNAL) found in urine.⁹ Research indicates that 6–11-year-olds exposed to tobacco smoke have three times the amount of estimated NNK exposure when compared to adults.¹⁰ Children who live with parental smokers and are exposed to tobacco smoke in their homes have the highest NNAL levels, followed by those not exposed in their homes.¹¹ Research also reports 6–12-year-olds with parents who smoke a higher number of cigarettes have higher cotinine and NNAL levels.¹²

The NNAL/cotinine ratio is a promising biomarker to distinguish thirdhand smoke (THS) exposure from secondhand smoke (SHS) exposure; children’s overall TSE consists of exposure to THS and SHS.¹³ THS is tobacco smoke residue that remains in environments where tobacco has been smoked long after active smoking has dissipated, reacts with ambient oxidants to yield potential harmful byproducts, re-emits back to the gas phase, and resuspends into the air.¹³ Compared to those who actively smoke tobacco, the NNAL/cotinine ratio is expected to be higher among those exposed to THS, especially young children.¹³ This is due to age-specific behaviors (e.g., floor time playing) and greater inhalation, ingestion, and dermal exposure to THS pollution found in their environments. Research indicates that when compared to adults, children have similar urinary cotinine levels, but have two-fold higher urinary NNAL levels and 1.5 times higher NNAL/cotinine ratio levels.¹⁴

Carcinogenic biomarker uptake in infants and children, including NNK exposure, is important for assessment of potential hazards,^15–17 but limited work has assessed urinary NNAL among this population. There is a gap in the literature on the assessment of the alkaloid nicotelline, a tobacco smoke-derived particulate matter biomarker that can be measured as N-oxides in urine.¹⁸ Similar to the NNAL/cotinine ratio, it is posited that the nicotelline N-oxides/cotinine ratio would be higher among those exposed to THS. Therefore, this study will begin to address this gap in child TSE research by examining the uptake of several biomarkers and the use of electronic medical record (EMR) data to assess healthcare utilization patterns among a sample of racially diverse, low-income children ages 0–9 years who live with a tobacco smoker.

The primary study objective was to assess the associations of TSE biomarkers (i.e., urinary cotinine, NNAL, and nicotelline N-oxides) and ratios (i.e., NNAL/cotinine (x10³), N-oxides/cotinine (x10³)) with total hospital visits, pediatric emergency department (PED) visits, urgent care (UC), revisits, and hospital admissions among PED/UC patients ages 0–9 years over 6-months. The associations between biomarker patterns and healthcare utilization patterns were also examined among developmental age groups of 0–4 years (infants and toddlers) and 5–9 years (school-aged children). We hypothesized that children with higher levels of urinary cotinine, NNAL, NNAL/cotinine ratio (x10³), N-oxides, and N-oxides/cotinine ratio (x10³) would have higher pediatric visit patterns than children with lower levels. To identify potential TSE-related disparities among children of smokers, another objective was to examine biomarker levels with child characteristics and parent-reported smoking and child TSE patterns. We posited that children with higher biomarker levels, measured in individual and ratio form, would be younger, non-Hispanic Black, and of lower income.

Methods

Participants and Procedures

A convenience sample of PED and UC patients who presented to a large, U.S. urban midwestern children’s hospital were initially recruited for Healthy Families (NIH R01HD03354), a randomized controlled trial designed to facilitate parental tobacco cessation; study protocol details are available.¹⁹ Briefly, PED/UC patients were eligible if they presented with a TSE-related complaint (e.g., difficulty breathing) documented by the U.S. Surgeon General,²⁰ were practitioner-confirmed clinically stable, were accompanied by a parent who smoked tobacco products, and were not tobacco users themselves. Patients who were ineligible had parents who were currently taking smoking cessation medications or exclusively used smokeless tobacco or electronic cigarettes (e-cigarettes). A sub-sample of children had urine samples collected as part of Healthy Families Phase II (R01ES027815) that leveraged the experimental design of Healthy Families to collect additional biomarkers of SHS and THS exposure; study protocol details are available.²¹ This study’s analytic sample consists of 242 patients 0–9 years old who participated in Healthy Families Phase II and had data available on urinary biomarkers, parental assessments, and healthcare visits. The Cincinnati Children’s Hospital Medical Center’s institutional review board approved all study procedures.

Measures

Urine Samples and Analytical Chemistry

Clinical research staff obtained urine samples from children during their baseline PED/UC visit; samples were stored at −80 degrees Celsius. Analysis of urinary cotinine, NNAL, and nicotelline N-oxides was conducted using a modification of the original liquid chromatography tandem mass spectrometry (LC/MS-MS) and sample preparation methods that provided greater sensitivity (i.e., lower limit of quantification (LOQ)) for children exposed to SHS and THS).^18,22 The LOQs for cotinine, NNAL, and N-oxides were 0.05ng/mL, 0.25pg/mL, 1.37pg/mL, respectively. NNAL concentrations were analyzed with creatinine (LOQ=0.05) normalization. Most of the analytic sample (n=241) had urine analyzed for cotinine with all children having detectable levels. A sub-sample of children had urine analyzed for NNAL (n=118) and N-oxides (n=116). All participants with NNAL measured had detectable levels, and only two children had N-oxides that were below the LOQ, which were imputed for analysis as LOQ$/\sqrt{2}$. NNAL and N-oxides were also assessed in ratio form with cotinine as the denominator (x10³).

Child EMR Data

Data were extracted from children’s EMRs on PED/UC revisits within 30-days following their baseline visit. Revisits 30-days post discharge from the baseline PED/UC visit was the ideal timeframe to identify revisits that may have been potentially preventable.^6,23,24 Data were also extracted on total hospital visits, PED visits, UC visits, and inpatient hospital admissions over 6-months following their baseline visit. Total hospital visits was defined as total PED visits, UC visits, and inpatient hospital admissions over 6-months.

Child EMR data extracted included possible TSE-related past medical histories (PMHs; asthma, bronchiolitis, pneumonia, and prematurity), and the following baseline variables: treatment location (PED, UC) and visit date to determine seasonality and year; triage level; possible TSE-related chief complaints of cough or congestion, difficulty breathing or wheezing, and ear pain; discharge diagnosis (viral or other infectious disease, bacterial disease, pulmonary disease, and allergic, inflammatory, or other disease); and disposition (discharged to home and admitted to hospital). Child sociodemographic information (age, sex, race, ethnicity, and insurance type) was also obtained from EMRs. Child race included White, Black, and Other/Multiracial categories (i.e., Asian, other race not listed, and multiple races), and insurance type included commercial insurance and public insurance or self-pay.

Parent Assessment Data

Parents reported their smoking patterns and child TSE patterns including the daily number of cigarettes smoked by the parent and cumulative child TSE, defined as the total number of cigarettes smoked around the child in the past week by all smokers (e.g., grandparent) in any location (e.g., home, car). Parental assessment data also included sociodemographic information (parent sex and education level) and housing type (single-family home and multi-unit home or apartment).

Data Analysis

Biomarker data underwent logarithmic transformations to address skewed distributions, and therefore, we report medians (Mdns) and interquartile ranges (IQRs). We fitted linear regression models to explore the associations between child, parent, and baseline PED/UC visit characteristics as the explanatory variables and log-transformed biomarkers and biomarker ratios (x10³) as the response variables. Then, we used Poisson regression models to assess the associations between log-transformed biomarkers, in individual and ratio form, as the explanatory variables with healthcare utilization outcomes as the response variables, while adjusting for child sociodemographics (age, sex, race, ethnicity, insurance), PMHs, baseline visit season and year, and parent sex. We fitted similar Poisson regression models to assess the associations between TSE biomarkers and healthcare utilization patterns among infants and toddlers ages 0–4 years and school-aged children ages 5–9 years. All statistical analyses were performed using R version 4.0.5.,²⁵ with the type I error set at 0.05 (two-tailed).

Results

Child TSE Biomarker Levels

Descriptive statistics of child TSE biomarker levels in individual form are presented in Table 1. The geometric means (GeoMs) of the child NNAL/cotinine ratio (x10³) and N-oxides/cotinine ratio (x10³) were 1.0 and 0.8, respectively.

Table 1.

Descriptive Statistics for Child Urinary Cotinine, NNAL, and N-oxides

	Overall (N=242)
Biomarker	n (%)a
Cotinine (ng/mL)
n	241
Geometric Mean (95%CI)	11.2 (9.5–13.0)
Median (IQR)	12.2 (4.7–25.9)
Range	0.3–415.4
NNAL (pg/mg creatinine)
n	118
Geometric Mean (95%CI)	30.9 (24.8–38.5)
Median (IQR)	36.3 (14.1–74.3)
Range	1.3–489.3
N-oxides (pg/mL)
n	116
Geometric Mean (95%CI)	24.1 (19.0–30.5)
Median (IQR)	26.2 (9.9–67.7)
Range	1.0–371.7

Abbreviations: CI, confidence interval; IQR, interquartile range.

Child Characteristics based on Child TSE Biomarker Levels

The overall mean (SD) age of children was 4.1 (2.9) years and 43.4% were female (Table 2). The majority of children were of non-Hispanic origin (97.1%), and Black (61.6%) followed by White (28.1%). Most children had public insurance or were self-pay (95.9%), and about two-thirds of children lived in multi-unit homes or apartments (62.8%). There were negative correlations found between child age and log-NNAL (r=−0.51, p<0.001) and log-N-oxides (r=−0.24, p=0.015). Black children (Mdn=1.8, p<0.001) had significantly lower N-oxides/cotinine ratio (x10³) values than white children (Mdn=4.6). Children who had public insurance or were self-pay (Mdn=12.2ng/ml, p=0.022) had higher cotinine levels than children who had commercial insurance (Mdn=3.6ng/ml). No differences were found between child sex, ethnicity, or housing type with TSE biomarkers in individual and ratio form.

Table 2.

Child Characteristics by Median Child Urinary Cotinine, NNAL, and N-oxides in Individual and Ratio Form

	Overall (N=242)	Cotinine (ng/mL) (n=241)		NNAL (pg/mg creatinine) (n=118)		NNAL/Cotinine Ratio (×103) (n=117)		N-oxides (pg/mL) (n=116)		N-oxides/Cotinine Ratio (×103) (n=115)
Characteristic	n (%)a	Median (IQR)	p-valueb	Median (IQR)	p-valueb	Median (IQR)	p-valueb	Median (IQR)	p-valueb	Median (IQR)	p-valueb
Child Age, M (SD) c	4.1 (2.9)	-	0.127	-	<0.001	-	0.332	-	0.015	-	0.404
0–4 years	132 (54.5)	15.0 (5.3–28.6)	0.255	48.9 (22.6–118.5)	0.083	3.8 (1.6–5.8)	0.795	42.5 (10.3–89.7)	0.161	2.2 (1.0–5.6)	0.862
5–9 years	110 (45.5)	11.5 (4.7–23.0)		33.7 (13.3–57.7)		2.9 (1.6–4.4)		23.6 (7.7–53.9)		2.1 (1.2–4.6)
Child Sex
Male	137 (56.6)	12.7 (5.5–25.8)	Ref	43.9 (16.97–99.2)	Ref	3.1 (1.3–6.1)	Ref	36.3 (10.0–68.5)	Ref	2.1 (1.2–5.4)	Ref
Female	105 (43.4)	11.5 (4.7–26.6)	0.898	32.4 (18.1–52.6)	0.131	3.1 (1.8–4.9)	0.631	23.0 (9.2–61.4)	0.406	2.3 (1.1–4.6)	0.748
Child Race
White	68 (28.1)	9.5 (3.8–21.1)	Ref	43.9 (25.5–97.8)	Ref	4.1 (3.1–5.4)	Ref	50.8 (15.0–86.4)	Ref	4.6 (2.6–7.6)	Ref
Black	149 (61.6)	12.5 (6.0–28.7)	0.061	30.4 (14.8–55.7)	0.107	2.8 (1.3–4.6)	0.078	23.2 (8.5–58.5)	0.086	1.8 (0.9–4.3)	<0.001
Other	19 (7.8)	19.8 (5.5–28.6)	0.337	69.2 (35.3–106.4)	0.771	1.7 (1.4–6.4)	0.233	50.5 (9.7–88.3)	0.584	1.9 (1.3–2.7)	0.079
Unknown	6 (2.5)	17.5 (8.5–33.2)	0.351	53.2 (35.3–53.8)	0.644	1.7 (1.6–3.4)	0.429	70.2 (38.8–71.5)	0.948	1.3 (1.1–2.1)	0.096
Child Ethnicity
Non-Hispanic	235 (97.1)	12.0 (4.7–27.1)	Ref	36.4 (17.7–76.4)	Ref	3.1 (1.6–5.4)	Ref	26.8 (10.3–68.9)	Ref	2.2 (1.2–5.0)	Ref
Hispanic	4 (1.7)	14.2 (7.9–21.3)	0.807	34.7 (26.1–43.4)	0.740	1.4 (1.3–1.5)	0.357	14.7 (11.0–18.3)	0.476	0.9 (0.9–0.9)	0.233
Unknown	3 (1.2)	13.9 (7.5–18.7)	0.519	22.6 (13.3–32.0)	0.152	3.0 (2.3–3.7)	0.915	26.8 (14.2–39.3)	0.298	2.7 (2.1–3.2)	0.892
Child Insurance Type
Commercial	10 (4.1)	3.6 (1.3–15.0)	Ref	11.1 (9.7–61.0)	Ref	3.3 (3.1–9.7)	Ref	6.2 (5.5–6.3)	Ref	1.4 (1.3–7.5)	Ref
Public/Self-pay	232 (95.9)	12.2 (5.2–26.2)	0.022	36.5 (17.7–72.9)	0.245	3.1 (1.5–5.3)	0.400	26.9 (10.8–68.0)	0.206	2.2 (1.2–4.6)	0.876
Housing Type
Single-Family	90 (37.2)	13.3 (3.0–25.8)	Ref	40.7 (16.7–78.9)	Ref	3.1 (1.6–4.6)	Ref	23.5 (7.4–74.9)	Ref	2.2 (1.2–5.4)	Ref
Multi-Unit Home or Apartment	152 (62.8)	11.6 (6.1–26.2)	0.209	35.2 (18.3–70.8)	0.477	3.1 (1.6–5.3)	0.705	32.7 (11.2–64.9)	0.566	2.2 (1.2–4.4)	0.624

Abbreviations: M, mean; SD, standard deviation; IQR, interquartile range; Ref, reference category; No., number.

^aPercent refers to column percent unless otherwise noted.

^bP-values correspond with linear regression results, and boldface indicates statistical significance (p<0.05) unless noted otherwise.

^cP-values refer to Pearson correlation results, and boldface indicates statistical significance (p<0.05).

Parent Characteristics based on Child TSE Biomarker Levels

Concerning parent-reported characteristics, 89.3% of parents were female (Table 3). The mean (SD) number of daily cigarettes smoked by parents was 10.2 (6.1) cigarettes (range=1.0–30.0), and mean (SD) number of weekly cigarettes smoked by all smokers in any location around the child (i.e., cumulative child TSE) was 10.2 (22.4) cigarettes (range=0–224). Children with mothers enrolled into the study had higher cotinine levels (Mdn=12.6ng/ml, p=0.019) and N-oxide levels (Mdn=30.8pg/mL, p=0.011) compared to children with fathers enrolled into the study (Mdn=7.1ng/mL; Mdn=6.0pg/mL, respectively). There were positive correlations between number of daily cigarettes smoked by parents and child log-cotinine (r=0.19, p=0.005), log-NNAL (r=0.25, p=0.036), and log-N-oxides (r=0.26, p=0.023). There was a significant positive correlation between cumulative child TSE and child cotinine (r=0.17, p=0.022). No differences were found between parent e-cigarette use and TSE biomarkers in individual and ratio form.

Table 3.

Parent Characteristics and Child TSE Patterns by Median Child Urinary Cotinine, NNAL, and N-oxides in Individual and Ratio Form

	Overall (N=242)	Cotinine (ng/mL) (n=241)		NNAL (pg/mg creatinine) (n=118)		NNAL/Cotinine Ratio (×103) (n=117)		N-oxides (pg/mL) (n=116)		N-oxides/Cotinine Ratio (×103) (n=115)
Characteristic	n (%)a	Median (IQR)	p-valueb	Median (IQR)	p-valueb	Median (IQR)	p-valueb	Median (IQR)	p-valueb	Median (IQR)	p-valueb
Parent Sex
Male	26 (10.7)	7.1 (2.2–21.2)	Ref	20.7 (10.0–55.2)	Ref	3.1 (1.8–6.4)	Ref	6.0 (4.8–14.6)	Ref	1.6 (1.1–2.2)	Ref
Female	216 (89.3)	12.6 (5.5–27.3)	0.019	36.8 (18.3–74.1)	0.209	3.1 (1.4–5.2)	0.554	30.8 (11.3–68.2)	0.011	2.2 (1.2–5.1)	0.460
Parent Current E-Cigarette Use
No	234 (96.7)	11.6 (4.7–25.9)	Ref	36.2 (17.6–70.9)	Ref	3.1 (1.6–5.3)	Ref	24.3 (9.5–66.7)	Ref	2.1 (1.2–4.6)	Ref
Yes	8 (3.3)	19.6 (15.6–27.5)	0.216	119.1 (74.9–141.8)	0.196	4.5 (2.7–6.5)	0.812	88.3 (61.3–94.7)	0.164	6.6 (4.0–10.6)	0.149
Parent No. Cigarettes/Day, M (SD) c	10.2 (6.1)	-	0.005	-	0.036	-	0.741	-	0.023	-	0.141
No. Cigarettes/Week Smoked Around Child by All Smokers in Any Location, M (SD) c	10.2 (22.4)	-	0.022	-	0.960	-	0.879	-	0.602	-	0.746

Abbreviations: M, mean; SD, standard deviation; IQR, interquartile range; Ref, reference category; No., number.

^aPercent refers to column percent unless otherwise noted.

^bP-values correspond with linear regression results, and boldface indicates statistical significance (p<0.05) unless noted otherwise.

^cP-values refer to Pearson correlation results, and boldface indicates statistical significance (p<0.05).

Baseline PED/UC Visit Characteristics based on Child TSE Biomarker Levels

The majority of children had treatment at a UC site (70.7%) and were triaged as having a low-to-moderate acuity level (81.4%) (Table 4). Nearly one-fourth had a PMH of a respiratory condition (24.4%) and 9.1% had a PMH of prematurity. The top chief complaint was cough or congestion (35.9%) followed by difficulty breathing or wheezing (22.3%) and ear pain (16.9%). Most children were discharged to home (90.9%), and half (49.6%) had a discharge diagnosis of viral or other infectious disease.

Table 4.

Baseline PED/UC Visit Characteristics by Median Child Urinary Cotinine, NNAL, and N-oxides in Individual and Ratio Form

	Overall (N=242)	Cotinine (ng/mL) (n=241)		NNAL (pg/mg creatinine) (n=118)		NNAL/Cotinine Ratio (×103) (n=117)		N-oxides (pg/mL) (n=116)		N-oxides/Cotinine Ratio (×103) (n=115)
PED/UC Visit Characteristic	n (%)a	Median (IQR)	p-valueb	Median (IQR)	p-valueb	Median (IQR)	p-valueb	Median (IQR)	p-valueb	Median (IQR)	p-valueb
Visit Year
2016	74 (30.6)	10.8 (4.0–21.6)	Ref	42.4 (21.1–67.2)	Ref	3.6 (1.7–5.4)	Ref	34.3 (14.1–53.0)	Ref	2.0 (1.4–4.3)	Ref
2017	96 (39.7)	11.8 (5.6–30.2)	0.081	32.4 (16.9–65.9)	0.512	3.1 (2.0–6.9)	0.515	37.9 (13.7–68.9)	0.655	3.6 (1.5–6.2)	0.202
2018	47 (19.4)	14.1 (5.3–30.5)	0.090	42.5 (26.7–89.5)	0.389	3.3 (1.4–5.3)	0.249	23.3 (8.3–95.7)	0.725	2.5 (1.2–4.0)	0.562
2019	25 (10.3)	15.3 (7.5–26.6)	0.091	30.0 (11.1–54.4)	0.744	1.7 (1.0–3.6)	0.140	20.6 (7.7–72.9)	0.654	1.2 (0.7–1.9)	0.084
Visit Season
Summer (June-August)	60 (24.8)	9.8 (3.8–24.0)	Ref	32.43 (20.0–71.0)	Ref	4.1 (1.9–6.7)	Ref	28.9 (13.1–58.8)	Ref	3.9 (1.5–6.8)	Ref
Fall (September-November)	58 (24.0)	13.9 (5.3–23.5)	0.472	38.88 (25.3–79.6)	0.663	3.6 (2.3–5.3)	0.432	36.9 (11.3–65.9)	0.677	2.3 (1.6–4.0)	0.210
Winter (December-February)	61 (25.2)	14.0 (5.7–28.1)	0.163	30.71 (16.4–47.2)	0.594	3.0 (1.2–3.7)	0.144	42.4 (9.6–85.1)	0.845	2.8 (0.7–4.6)	0.087
Spring (March-May)	63 (26.0)	13.6 (6.3–28.2)	0.386	52.0 (13.9–119.1)	0.362	1.9 (1.2–4.5)	0.008	18.8 (7.7–62.6)	0.453	1.6 (0.9–2.7)	0.001
Arrival Location
UC	171 (70.7)	11.8 (5.4–25.4)	Ref	34.6 (15.7–58.0)	Ref	3.2 (1.7–5.4)	Ref	27.9 (11.2–64.9)	Ref	2.2 (1.3–5.1)	Ref
PED	71 (29.3)	13.0 (4.3–28.5)	0.633	45.6 (22.3–95.8)	0.154	2.9 (1.1–4.5)	0.057	20.0 (7.7–78.5)	0.908	1.6 (1.0–4.2)	0.099
Triage Level
Low/Moderate Acuity (3–5)	70 (81.4)	12.6 (4.3–27.6)	Ref	45.6 (23.5–81.9)	Ref	3.2 (1.6–5.8)	Ref	21.7 (10.3–84.4)	Ref	2.0 (1.2–5.5)	Ref
High Acuity (2)	16 (18.6)	19.4 (4.3–45.9)	0.322	73.2 (21.6–150.8)	0.450	1.1 (0.3–1.6)	0.007	21.3 (7.0–45.4)	0.717	1.0 (0.5–1.3)	0.101
Child PMH of Respiratory Condition c
No	183 (75.6)	13.0 (5.1–28.1)	Ref	40.7 (24.2–89.2)	Ref	3.3 (1.9–5.7)	Ref	32.7 (14.2–69.0)	Ref	2.2 (1.2–5.3)	Ref
Yes	59 (24.4)	10.3 (4.5–21.0)	0.242	22.7 (13.9–51.7)	0.006	2.2 (1.3–4.4)	0.244	20.6 (5.8–67.5)	0.044	1.8 (1.0–3.9)	0.459
Child PMH of Prematurity
No	220 (90.9)	11.6 (4.8–26.2)	Ref	35.5 (17.7–71.0)	Ref	3.1 (1.6–5.4)	Ref	27.9 (10.6–70.4)	Ref	2.2 (1.2–5.0)	Ref
Yes	22 (9.1)	14.5 (5.4–20.3)	0.671	39.7 (26.0–84.4)	0.171	2.4 (1.1–3.7)	0.176	23.5 (6.1–59.2)	0.360	1.3 (0.6–1.9)	0.043
Chief Complaint of Cough/Congestion
No	155 (64.1)	11.6 (4.4–24.8)	Ref	34.4 (17.5–71.0)	Ref	3.1 (1.4–5.1)	Ref	22.8 (8.4–68.0)	Ref	2.2 (1.2–4.9)	Ref
Yes	87 (35.9)	13.9 (6.4–29.0)	0.451	43.9 (21.3–99.4)	0.083	3.2 (1.7–5.3)	0.525	44.4 (14.5–66.7)	0.234	2.0 (1.2–4.3)	0.837
Chief Complaint of Difficulty Breathing/Wheezing
No	188 (77.7)	11.6 (4.9–24.9)	Ref	34.8 (17.6–70.9)	Ref	3.2 (1.6–5.3)	Ref	27.9 (10.8–69.7)	Ref	2.2 (1.2–5.3)	Ref
Yes	54 (22.3)	13.2 (4.6–33.9)	0.333	46.2 (22.0–85.9)	0.399	2.0 (1.2–4.8)	0.532	20.0 (6.2–52.9)	0.455	1.4 (0.9–3.7)	0.136
Chief Complaint of Ear Pain
No	201 (83.1)	12.7 (4.7–25.7)	Ref	36.2 (17.5–69.7)	Ref	2.9 (1.4–5.0)	Ref	28.9 (8.2–67.5)	Ref	2.1 (1.2–4.3)	Ref
Yes	41 (16.9)	11.4 (5.0–28.3)	0.971	42.5 (24.6–88.5)	0.257	3.4 (2.6–6.7)	0.178	23.0 (14.3–61.9)	0.712	3.7 (0.9–6.4)	0.361
Disposition
Discharge to Home	220 (90.9)	12.2 (5.2–25.9)	Ref	36.0 (17.6–70.9)	Ref	3.1 (1.6–5.4)	Ref	26.8 (10.5–67.3)	Ref	2.2 (1.2–4.9)	Ref
Admit	22 (9.1)	11.8 (3.5–27.6)	0.788	78.1 (22.6–292.0)	0.124	2.6 (0.7–4.0)	0.045	11.3 (6.5–79.8)	0.490	1.0 (0.5–1.8)	0.064
Discharge Diagnosis
Allergic/Inflammatory or Other	18 (7.4)	9.9 (5.1–21.8)	Ref	42.5 (19.7–63.2)	Ref	3.4 (1.2–4.6)	Ref	9.5 (4.5–21.4)	Ref	1.6 (1.1–4.2)	Ref
Viral/Other Infectious Disease	120 (49.6)	13.6 (4.7–25.8)	0.986	36.0 (20.4–58.8)	0.662	3.5 (1.9–5.6)	0.175	32.7 (10.9–66.4)	0.016	2.5 (1.3–4.7)	0.114
Bacterial	49 (20.3)	11.1 (5.9–28.34	0.692	25.9 (10.7–86.4)	0.799	3.1 (1.2–4.7)	0.368	39.2 (14.2–92.1)	0.026	2.5 (1.1–5.7)	0.162
Pulmonary	55 (22.7)	12.8 (4.5–27.0)	0.858	45.6 (19.3–105.7)	0.455	1.8 (1.3–4.0)	0.783	22.4 (8.21–64.8)	0.062	1.5 (0.9–3.5)	0.723

Abbreviations: PED, pediatric emergency department; UC, urgent care; IQR, interquartile range; Ref, reference category; PMH, past medical history.

^aPercent refers to column percent.

^bP-values correspond with linear regression results, and boldface indicates statistical significance (p<0.05).

^cRespiratory conditions include asthma, bronchiolitis, and pneumonia.

Children who had their baseline visit during spring had lower NNAL/cotinine ratio (x10³) values (Mdn=1.9, p=0.008) and N-oxides/cotinine ratio (x10³) values (Mdn=1.6, p=0.001) compared to children who had their baseline visit during summer (Mdn=4.1; Mdn=3.9, respectively). Children with high triage acuity (Mdn=1.1, p=0.008) had lower NNAL/cotinine ratio (x10³) values than children with low-to-moderate triage acuity (Mdn=3.2). Children with a PMH of a respiratory condition had lower NNAL (Mdn=22.7pg/mg creatinine, p=0.022) and lower N-oxides (Mdn=20.6pg/mL, p=0.044) than children without this PMH (Mdn=40.7pg/mg creatinine; Mdn=32.7pg/mL, respectively). Children with a PMH of prematurity had lower N-oxides/cotinine ratio (x10³) values (Mdn=1.3, p=0.043) than children without this PMH (Mdn=2.2). Children admitted to the hospital during their baseline visit had lower NNAL/cotinine ratio (x10³) values (Mdn=2.6, p=0.045) than children who were discharged to home (Mdn=3.1). Children with a bacterial diagnosis (Mdn=39.2pg/mL, p=0.026) and viral or other infectious disease diagnosis (Mdn=32.7pg/mL, p=0.016) at baseline had significantly higher N-oxides than children with an allergic, inflammatory, or other disease diagnosis (Mdn=9.5pg/mL). No other differences were found between PED/UC visit characteristics and child TSE biomarkers.

Child TSE Biomarker Levels and Parent-Reported Smoking and Child TSE Patterns based on Healthcare Visit Patterns among 0–9-Year-Olds

Regarding healthcare visit patterns, children had an overall mean (SD) of 0.9 (0.1) total hospital visits over 6-months, and 0.2 (0.03) PED/UC revisits within 30-days. Children had an overall mean of 0.4 (0.1) PED visits, 0.5 (0.1) UC visits, and 0.1 (0.02) hospital admissions over 6-months following their baseline PED/UC visit.

Poisson regression results indicated that each one-unit increase of child log-NNAL/cotinine ratio (x10³) values was associated with an increase in total number of hospital visits over 6-months (adjusted relative risk [aRR]=1.39, 95%CI=1.10–1.75, p=0.005) following the baseline PED/UC visit, while controlling for the covariates (Table 5). Each one-unit increase in child log-NNAL levels (aRR=1.68, 95%CI=1.18–2.39, p=0.004) and NNAL/cotinine ratio (x10³) values (aRR=1.56, 95%CI=1.14–2.13, p=0.005) were associated with an increase in total UC visits over 6-months following the baseline visit. For each one-unit increase in the number of daily cigarettes smoked by parents, children were at reduced risk of having a PED/UC revisit within 30-days (aRR=0.93, 95%CI=0.88–0.99, p=0.033). For each one-unit increase in the number of weekly cigarettes smoked around the child by all smokers in any location, children were at increased risk of having a higher number of hospital admissions over 6-months (aRR=1.02, 95%CI=1.00–1.04, p=0.029) following the baseline visit. Figure 1 presents the scatterplots for the five significant associations of child TSE biomarkers and parent-reported smoking and child TSE patterns with healthcare visits over 6-months among 0–9-year-olds.

Table 5.

Child Urinary Cotinine, NNAL, and N-oxides in Individual and Ratio Form and Parent-Reported Smoking and Child TSE Patterns by Healthcare Visits over 6-Months

	Total Hospital Visits	Total Revisits, Within 30-Days	Total PED Visits	Total UC Visits	Total Hospital Admissions
	aRR (95% CI)a	aRR (95% CI)a	aRR (95% CI)a	aRR (95% CI)a	aRR (95% CI)a
Children 0–9 Years Old (N=242)
Cotinine (ng/mL)	0.97 (0.87–1.08)	1.06 (0.84–1.33)	1.06 (0.90–1.24)	0.91 (0.78– 1.05)	1.30 (0.83–2.04)
NNAL (pg/mg creatinine)	1.23 (0.96–1.58)	1.13 (0.66–1.93)	0.90 (0.62– 1.32)	1.68 (1.18– 2.39) **	-
NNAL/Cotinine Ratio (×103)	1.39 (1.10–1.75) **	1.11 (0.68–1.81)	1.26 (0.88– 1.80)	1.56 (1.14– 2.13) **	-
N-oxides (pg/mL)	0.95 (0.78–1.15)	1.08 (0.67–1.75)	1.05 (0.78– 1.41)	0.90 (0.69– 1.17)	-
N-oxides/Cotinine Ratio (×103)	1.19 (0.93–1.52)	1.04 (0.57–1.89)	1.32 (0.88– 1.99)	1.15 (0.84– 1.58)	-
Parent No. Cigarettes/Day	0.98 (0.96–1.01)	0.93 (0.88–0.99) *	0.99 (0.96– 1.03)	0.98 (0.94– 1.01)	1.07 (0.98– 1.17)
No. Cigarettes/Week Smoked Around Child by All Smokers in Any Location	1.00 (0.99–1.01)	1.01 (1.00–1.02)	1.00 (1.00, 1.01)	1.00 (0.99– 1.01)	1.02 (1.00–1.04) *
Infants and Toddlers 0–4 Years Old (n=132)
Cotinine (ng/mL)	0.96 (0.85–1.08)	1.06 (0.82–1.37)	1.11 (0.92–1.34)	0.86 (0.72–1.02)	1.71 (0.99–2.97)
NNAL (pg/mg creatinine)	0.94 (0.64–1.39)	3.04 (0.15–63.52)	0.68 (0.31–1.52)	1.05 (0.66–1.68)	-
NNAL/Cotinine Ratio (×103)	1.50 (0.98–2.30)	1.57 (0.53–4.59)	1.66 (0.78–3.54)	1.41 (0.83–2.39)	-
N-oxides (pg/mL)	0.72 (0.49–1.06)	1.66 (0.36–7.69)	0.68 (0.24–1.91)	0.63 (0.39–1.03)	-
N-oxides/Cotinine Ratio (×103)	1.25 (0.73–2.12)	1.46 (0.31–6.83)	1.59 (0.47–5.36)	0.98 (0.51–1.89)
Parent No. Cigarettes/Day	0.96 (0.93–0.99) *	0.93 (0.86–0.99) *	0.98 (0.94–1.03)	0.93 (0.89–0.98) **	1.07 (0.96–1.18)
No. Cigarettes/Week Smoked Around Child by All Smokers in Any Location	1.00 (0.99–1.01)	0.98 (0.95–1.01)	1.00 (0.99–1.02)	0.99 (0.98–1.01)	1.08 (1.01–1.16) *
School-Aged Children 5–9 Years Old (n=110)
Cotinine (ng/mL)	1.09 (0.86–1.38)	1.08 (0.56–2.08)	1.06 (0.73–1.54)	1.08 (0.79–1.48)	0.54 (0.14–2.10)
NNAL (pg/mg creatinine)	1.88 (1.13–3.13) *	0.46 (0.12–1.77)	0.83 (0.42–1.67)	5.64 (2.15–14.82) ***	-
NNAL/Cotinine Ratio (×103)	1.30 (0.82–2.07)	0.59 (0.19–1.81)	0.95 (0.47–1.90)	1.76 (0.93–3.36)	-
N-oxides (pg/mL)	1.08 (0.80–1.44)	0.67 (0.28–1.59)	0.95 (0.60–1.50)	1.23 (0.80–1.88)	-
N-oxides/Cotinine Ratio (×103)	0.94 (0.62–1.44)	0.49 (0.13–1.89)	1.00 (0.54–1.84)	1.02 (0.55–1.88)	-
Parent No. Cigarettes/Day	1.03 (0.98–1.09)	1.02 (0.88–1.18)	1.05 (0.97–1.15)	1.02 (0.95–1.09)	1.19 (0.92–1.54)
No. Cigarettes/Week Smoked Around Child by All Smokers in Any Location	1.00 (0.99–1.01)	1.04 (0.99–1.09)	1.01 (1.00–1.02)	0.99 (0.98–1.01)	-

Abbreviations: PED, pediatric emergency department; UC, urgent care; aRR, adjusted relative risk; CI, confidence interval; No., number.

^aPoisson regression results adjusted for child age, sex, race, ethnicity, insurance type, past medical histories, parent sex, and visit year and season.

Boldface indicates statistical significance.

^***p<0.001

^**p<0.01

^*p<0.05.

Figure 1.

Scatterplots for the significant associations of child TSE biomarker levels in individual and ratio form and parent-reported smoking and child TSE patterns with healthcare visits over 6-months among 0–9-year-olds. Results are adjusted for child age, sex, race, ethnicity, insurance type, past medical histories, parent sex, and visit year and season.

[a] NNAL/Cotinine Ratio (x10³) and Total Hospital Visits

[b] Parent Daily Number of Cigarettes Smoked and Total PED/UC Revisits within 30-days

[c] NNAL (pg/mg creatinine) and Total UC Visits

[d] NNAL/Cotinine Ratio (x10³) and Total UC Visits

[e] Number of Weekly Cigarettes Smoked Around the Child by All Smokers in Any Location and Total Hospital Admissions

Child TSE Biomarker Levels and Parent-Reported Smoking and Child TSE Patterns based on Healthcare Visit Patterns among 0–4-Year-Olds

The sub-sample of infants and toddlers 0–4 years old (n=132) had an overall mean (SD) of 1.2 (0.1) total hospital visits over 6-months, and 0.3 (0.1) revisits to the PED/UC within 30-days. This younger child age group had an overall mean of 0.6 (0.1) PED visits, 0.7 (0.1) UC visits, and 0.1 (0.03) hospital admissions over 6-months following their baseline visit.

For each one-unit increase in the number of daily cigarettes smoked by parents, 0–4-year-olds were at reduced risk of having a higher number of total hospital visits over 6-months (aRR=0.96, 95%CI=0.93–0.99, p=0.013), PED/UC revisits within 30-days (aRR=0.93, 95%CI=0.86–0.99, p=0.036), and UC visits over 6-months (aRR=0.93, 95%CI=0.89–0.98, p=0.007) following the baseline visit (see Table 5). For each one-unit increase in the number of weekly cigarettes smoked around the child by all smokers in any location, younger children were at increased risk of having a higher number of hospital admissions over 6-months (aRR=1.08, 95%CI=1.01–1.16, p=0.034).

Figure 2 presents the scatterplots for the four significant associations of parent-reported smoking and child TSE patterns with healthcare visits over 6-months among 0–4-year-olds.

Figure 2.

Scatterplots for the significant associations of parent-reported smoking and child TSE patterns with healthcare visits over 6-months among 0–4-year-olds. Results are adjusted for child age, sex, race, ethnicity, insurance type, past medical histories, parent sex, and visit year and season.

[a] Parent Daily Number of Cigarettes Smoked and Total Hospital Visits

[b] Parent Daily Number of Cigarettes Smoked and Total PED/UC Revisits within 30-days

[c] Parent Daily Number of Cigarettes Smoked and Total UC Visits

[d] Number of Weekly Cigarettes Smoked Around the Child by All Smokers in Any Location and Total Hospital Admissions

Child TSE Biomarker Levels and Parent-Reported Smoking and Child TSE based on Healthcare Visit Patterns among 5–9-Year-Olds

The sub-sample of school-aged children 5–9 years-old (n=110) had an overall mean (SD) of 0.6 (1.0) total hospital visits over 6-months, and 0.1 (0.03) revisits to the PED/UC within 30-days. This older child age group had an overall mean of 0.3 (0.1) PED visits, 0.4 (0.1) UC visits, and 0.03 (0.02) hospital admissions over 6-months following their baseline visit.

Each one-unit increase in log-NNAL levels among older children was associated with an increase in total hospital visits (aRR=1.88, 95%CI=1.13–3.13, p=0.015) and UC visits (aRR=5.64, 95%CI=2.15–14.82, p<0.001) over 6-months following their baseline visit (see Table 5). Figure 3 presents the scatterplots for the two significant associations of child NNAL levels with healthcare visits over 6-months among 5–9-year-olds.

Figure 3.

Scatterplots for the significant associations of child NNAL levels in individual form with healthcare visits over 6-months among 5–9-year-olds. Results are adjusted for child age, sex, race, ethnicity, insurance type, past medical histories, parent sex, and visit year and season.

[a] NNAL (pg/mg creatinine) and Total Hospital Visits

[b] NNAL (pg/mg creatinine) and Total UC Visits

Discussion

This study assessed the associations of several TSE biomarkers with healthcare utilization 6-months following the baseline PED/UC visit of young, racially diverse children who were exposed to tobacco smoke. As posited, for every one-unit log-NNAL value increase, 0–9-year-olds were at 68% excess risk of having a higher number of total UC visits over 6-months. Notably among the school-aged group, for each one-unit log-NNAL increase, 5–9-year-olds were at 88% and 564% excess risk of having a higher number of total hospital visits and UC visits over 6-months, respectively. Our findings greatly expand upon prior research that found higher cotinine levels²⁶ and EMR documentation of TSE²⁷ increase children’s risks of ED visits and/or hospitalizations.

In this study, higher child NNAL/cotinine ratio (x10³) values were associated with an increased risk of having a higher number of total hospital visits and UC visits over 6-months. Specifically, for each one-unit log-NNAL/cotinine ratio (x10³) value increase, 0–9-year-olds were at 39% and 56% excess risk of having a higher number of total hospital visits and UC visits, respectively. However, contrary to our hypothesis, no differences were found between cotinine in individual form and healthcare visits over the past 6-months. Our findings support the notion that when infants, toddlers, and school-aged children are exposed to tobacco smoke, they have higher NNK exposure compared to nicotine exposure. Thus, cotinine measurement may lead to the underestimation of child TSE levels and associated carcinogenic risk.⁹ One possible explanation is that urinary cotinine has a shorter half-life of about 16 hours compared to urinary NNAL with a lengthier half-life of 10–45 days.²⁸ NNAL has a long window of detection of up to 6–12 weeks post TSE, and is sensitive to intermittent exposure.²⁹ Thus, intermittent or nondaily TSE will lead to a much faster decrease in cotinine compared with NNAL, resulting in an increase in the ratio of NNAL/cotinine.

Concerning cotinine and NNAL in individual form and NNAL in ratio form with cotinine (x10³) among 0–9-year-olds who lived with tobacco smokers in the current study, the GeoMs were 11.2ng/mL, 30.9pg/mg creatinine and 1.0, respectively. Thus, children in this study had disproportionately high levels of nicotine biomarker uptake measured via cotinine and carcinogenic biomarker uptake measured via creatinine-adjusted NNAL when compared to other studies. Specific to child NNAL levels, two studies of U.S. children ages 6–11 years who participated in the National Health and Nutrition Examination Survey 2007–2008³⁰ and 2011–2012¹⁰ waves had urinary creatinine-adjusted NNAL GeoMs of 13.9pg/mg creatinine (Mdn=2.1) and 2.4 pg/mg creatinine (Mdn=1.8), respectively. This translates to our study, with child baseline visits conducted from 2016–2019, having 122% and 1,188% higher urinary creatinine-adjusted NNAL levels compared to U.S. child averages from 2007–2008 and 2011–2012, respectively. U.S. child TSE trends have decreased over time, but remain high in certain populations including children who are Black and of lower socioeconomic status,^4,5 aligning with the results of the current study’s sample. Further, a study of 101 9–11-year-olds who lived with tobacco smokers reported GeoMs of cotinine, NNAL, and NNAL/cotinine ratio (x10³) as 1.9ng/mg creatinine, 14.5pg/mg creatinine, and 5.7, respectively.¹⁴ Thus, the current study reports higher GeoMs of cotinine and NNAL in individual form, but a lower GeoM of NNAL/cotinine ratio (x10³), likely due to our sample of children living with tobacco smokers and having higher SHS exposure as evidenced by the higher mean cotinine denominator. Additionally, we found higher Mdn cotinine levels of 12.2ng/ml compared to raw Mdn levels ranging from 0.69–1.83ng/ml among a sample of 0–4-year-olds with varying maternal smoking statuses and a mean number of 3.5 daily cigarettes reported by mothers who were current smokers.⁷ Further, a study that assessed Korea National Health and Nutrition Examination Survey 2016–2017 data found that 301 6–12-year-olds who lived with parents who were nonsmokers had the lowest GeoM NNAL levels of 0.8pg/mg creatinine compared to children who lived with parents who were tobacco smokers with no SHS exposure at home (GeoM=1.3pg/mg creatinine) and SHS exposure at home (GeoM=3.2pg/mg creatinine).¹¹ Additional research that used Korea National Health and Nutrition Examination Survey 2016–2018 data found that 847 6–12-year-olds with parents who were nonsmokers had significantly lower median urinary NNAL and cotinine concentrations (Mdn=0.8pg/mL and Mdn=0.2ng/mL, respectively) relative to concentrations of children with parents who were tobacco smokers (Mdn=1.4pg/mL and Mdn=0.4ng/mL, respectively).¹² Comparatively, 0–9-year-olds’ Mdn urinary creatinine-adjusted NNAL and cotinine levels of 36.3pg/mg creatinine and 12.2ng/ml in the current study, respectively, are disproportionately higher than 6–12-year-olds who lived with either nonsmokers or smokers. More research on younger children (e.g., 0–5-year-olds) is needed so that TSE biomarker levels can be compared with younger children of nonsmokers compared to younger children of smokers.

Interestingly, no differences were found based on the proposed tracer for particulate matter, N-oxides, in individual and ratio form with healthcare utilization patterns. One potential reason is because of the estimated elimination half-life of urinary N-oxides is about two hours.¹⁸ Another potential explanation for our varying findings on NNAL and N-oxides revolves around THS. As THS ages over time, nicotine concentrations will decrease more rapidly compared to gaseous and particulate matter found in THS-contaminated environments that may absorb into indoor surfaces. In turn, this leads to THS largely contributing to NNK exposure especially via the pathways of dermal contact, inhalation, and ingestion over time due to nicotine reacting with nitric oxide.³¹ The high child cotinine levels observed in this study may also explain the findings that children with lower NNAL/cotinine ratio (x10³) values, indicative of higher SHS exposure, had a spring visit, high triage acuity, and were admitted to the hospital from their baseline PED/UC visit. Additionally, children with higher N-oxide levels had bacterial and viral or other infectious disease diagnoses, but children with lower N-oxide levels had a PMH of prematurity, PMH of a respiratory condition, and a spring visit. Children with a PMH of a respiratory condition also had lower NNAL levels. It is well known that prematurity is a negative prenatal TSE-related outcome,¹ and TSE can affect infants’ lung development that can lead to short- and long-term respiratory complications (e.g., asthma).³² Therefore, it is speculated that greater nicotine and carcinogenic exposures can lead to increased chronic illness or illness severity, and particulate matter exposure can lead to increased illness susceptibility among young children. This needs to be explored in larger studies.

Parents’ daily smoking patterns positively correlated with urinary cotinine, NNAL, and N-oxide levels. Cumulative child TSE also positively correlated with urinary cotinine levels. For each one-unit increase in the number of weekly cigarettes smoked around the child by all smokers in any location, 0–9-year-olds were at increased risk of having a higher number of total hospital admissions over 6-months. Infants and toddlers 0–4 years old with higher cumulative TSE were also at increased risk of having a higher number of hospital admissions over 6-months. Interestingly, however, for each one-unit increase in the number of daily cigarettes smoked by parents, 0–9-year-olds were at reduced risk of having PED/UC revisits within 30-days, and 0–4-year-olds were at reduced risk of having PED/UC revisits as well as total hospital visits and total UC visits over 6-months. There are several explanations for these varying findings, which are likely due to the use of parent-reports that may potentially underestimate children’s TSE due to the lack of awareness of children’s exposure to SHS and/or THS and perceived stigma and social desirability biases inherent in the healthcare setting.^33–35

It is important to note that the current study’s sample was drawn from a PED/UC that typically cares for children of lower socioeconomic backgrounds (e.g., 96% had public insurance or were self-pay) and about one-in-five have a TSE-related illness (e.g., otits media).³⁶ Despite similar socioeconomic backgrounds, some TSE-related inequities were identified based on high biomarker levels. Children with public insurance or were self-pay, a proxy for low socioeconomic status, had higher cotinine levels. Black children had lower N-oxides/cotinine ratio (x10³) values, which is likely due to higher cotinine levels (Mdn=12.5) indicative of higher smoking around the child or potentially the varying nicotine metabolism when compared to White children,³⁷ irrespective of not detecting a significant difference. Further, while younger PED/UC patients have higher cotinine levels,⁶ our study found no correlation between child age and cotinine, but found a negative correlation between child age and NNAL. The observed exposure to NNAL in young children is concerning especially since a prior review of the literature indicates that the most consistently elevated carcinogenic-derived biomarker in those exposed to tobacco smoke is NNAL, which supports that TSE can cause lung cancer in nonsmokers.³⁸

Limitations

The current study has several strengths including the measurement of several TSE biomarkers and an EMR review of healthcare utilization patterns. However, the limitations should be discussed. This study involved a convenience sample of young PED/UC patients at one U.S. large, children’s hospital. Generalizability to other settings and populations may be limited, and causal associations could not be ascertained. EMR data may have had missing data, however, use of EMRs increased overall data quality by eliminating parent recall bias. At the time of data extraction, we only had data available at 6-months following the baseline visit. Additionally, we were unable to assess hospital visits of other types (e.g., primary care, pediatric subspecialty). Therefore, future studies should consider assessing longer timepoints such as 12-months following baseline visits and assessing primary care visits. We used biomarkers with varying half-lives to comprehensively measure TSE among children of tobacco smokers. Future research should consider comparing biomarker levels with children of nonsmokers.

Conclusions

This study provides biochemically validated evidence that it is prudent to consider carcinogenic-derived biomarkers that may differentiate between SHS and THS exposure when examining child TSE. Higher NNAL and NNAL/cotinine ratio (x10³) values at baseline were associated with increased risk of higher healthcare utilization, especially UC visits, prospectively over 6-months among a sample of racially diverse, clinically ill children who lived with a tobacco smoker and sought care at the PED/UC. It is important to systematically screen for child TSE including SHS and THS exposure during all hospital visits, and consider biomarkers of exposure to objectively measure TSE. Parent reports of number of daily cigarettes and cumulative TSE may not accurately reflect child exposure, especially among those with intermittent TSE or higher THS exposure patterns. Initiating pediatric TSE reduction efforts at the hospital, especially in the UC and inpatient settings, has been endorsed as feasible,^39,40 and could potentially reduce child TSE-associated health consequences and healthcare utilization patterns. TSE reduction interventions are critically needed during pediatric visits from infancy to the school-age years, especially for PED/UC patients’ families who may not be receiving these prevention efforts in other preventive healthcare settings.

Impact:

• Higher levels of cotinine, a widely used tobacco smoke exposure biomarker, have been associated with higher healthcare utilization patterns among children.

• Less is known on the associations of carcinogenic and tobacco smoke-derived particulate matter biomarker uptake with child healthcare utilization patterns.

• This study assessed the associations of several biomarkers with healthcare utilization patterns among pediatric emergency department patients ages 0–9 years who lived with tobacco smokers.

• Higher urinary NNAL biomarker levels, in individual and ratio form with cotinine, increased children’s risk for urgent care visits over 6-months.

• Higher parent-reported cumulative child tobacco smoke exposure increased children’s risk for hospital admissions.