Low dose blood BTEX are associated with pulmonary function through changes in inflammatory markers among US adults: NHANES 2007-2012

doi:10.21203/rs.3.rs-2359772/v1

Download PDF

Research Article

Low dose blood BTEX are associated with pulmonary function through changes in inflammatory markers among US adults: NHANES 2007-2012

https://doi.org/10.21203/rs.3.rs-2359772/v1

This work is licensed under a CC BY 4.0 License

Journal Publication

published 02 May, 2023

Read the published version in Environmental Science and Pollution Research →

You are reading this latest preprint version

The effects of blood benzene, toluene, ethylbenzene, and xylenes (BTEX) on lung function among general adults remain unknown. We enrolled 5,519 adults with measured blood BTEX concentrations and lung function from the US National Health and Nutrition Examination Survey 2007–2012. Weighted linear models were fitted to assess the associations of BTEX with lung function and inflammation parameters (white blood cell five-part differential count and C-reactive protein). The mediating effect of inflammation between BTEX and lung function was also examined. Blood BTEX concentrations decreased yearly from 1999 and were extremely low from 2007–2012. Benzene and toluene exerted the greatest influence on lung function in terms of forced vital capacity (FVC), forced expiratory volume in the first second (FEV1), calculated FEV1:FVC ratio, peak expiratory flow rate (PEFR), and forced mid expiratory flow (FEF_25%–75%). Both ethylbenzene and all xylene isomers had no effects on FVC but reduced FEV1, FEV1:FVC ratio, PEFR, and FEF25–75%. Weighted quantile analyses demonstrated that BTEX mixture was associated with decreases in FVC, FEV1, FEV1:FVC ratio, PEFR, and FEF25–75%, with benzene weighted most heavily for all lung function parameters. BTEX also increased the levels of inflammation indicated by white blood cell five-part differential count and C-reactive protein, and increased levels of inflammation also reduced lung function. From multiple mediation analysis, inflammation mediated the effects of benzene on FEV1 and PEFR, the effects of toluene on FEV1, and the effects of ethylbenzene on FEV1 and PEFR. Low-dose exposure to BTEX was associated with reduced pulmonary function both in large and small airways. Inflammation could be involved in this pathogenesis.

aromatic hydrocarbon

low dose

lung function

inflammation parameters

mediation effect

large airway function

small airway function

Volatile organic compounds (VOCs) are air pollutants that vaporise easily under normal atmospheric conditions of temperature and pressure. Benzene, toluene, ethylbenzene, and meta-xylene (m-xylene), para-xylene (p-xyline), and ortho-xylene (o-xylene) (collectively known as BTEX), as the main chemical components of aromatic hydrocarbons, account for a considerable proportion of the total VOCs in the urban environment (Zalel et al. 2008). BTEX occur naturally in crude oil and are used as solvents in industrial products, as additives in gasoline, and as intermediates in the synthesis of organic compounds (Bolden et al. 2015). They originate from a constellation of anthropogenic activities, both outdoor (i.e. fuel combustion and solvent use) (de Blas et al. 2012; Gallego et al. 2008) and indoors (i.e. cleaning, smoking, cooking, and off-gassing from building materials) (Gallego et al. 2008; Schneider et al. 2001). Typically, the BTEX concentrations indoors are higher than those outdoors (Jung et al. 2011), contributing substantially to individual exposure level (Su et al. 2011). Airborne inhalation is the predominant route of BTEX exposure for the general population, although ingestion of contaminated water and food also accounts for a small contribution (Bolden et al. 2015).

Exposure to BTEX can affect the respiratory system, including the airways and lungs that are the preferential target organs for gas and aerosols as well as immune function; long-term exposure to BTEX can also result in carcinogenesis (Bolden et al. 2015; Latif et al. 2019). Lung function is an early and critical indicator for predicting cardiorespiratory morbidity and mortality (Sin et al. 2005). Spirometry is a routinely used clinical pulmonary function test that has been recognised as the gold standard for diagnosing chronic obstructive pulmonary disease (COPD) and asthma according to international guidelines (Mirza et al. 2018). Several studies have demonstrated that exposure to VOCs reduces lung function both in the occupational (Du et al. 2020; Mukherjee et al. 2016; Ojo et al. 2017; Singh et al. 2017; Weichenthal et al. 2012) and general population (Cakmak et al. 2014; Chen et al. 2018; Elliott et al. 2006). Only two studies have focused exclusively on BTEX but have reported inconsistent results (Chen et al. 2018; Kwon et al. 2018). Most studies have assessed VOC exposure (including BTEX) based on atmospheric monitoring data (Cakmak et al. 2014; Chen et al. 2018; Du et al. 2020; Mukherjee et al. 2016; Ojo et al. 2017; Weichenthal et al. 2012); however, these data are vulnerable to exposure misclassification, because they do not account for indoor sources, the factors influencing individual intake, and BTEX metabolism.

Moreover, bioaccumulation in humans resulting from repeated long-term exposure cannot be detected through data monitoring (Ashley and Prah 1997). Consequently, blood VOC concentrations integrating exposures from all sources may more accurately reflect chronic exposure than estimations from air VOC levels and can be used to calculate internal dose (Ashley et al. 1994; Ashley and Prah 1997; Sexton et al. 2005). BTEX that accumulate in adipose tissue are stable and are slowly released into the bloodstream. Blood BTEX are therefore less susceptible to degradation after exposure than BTEX accumulated elsewhere in the body (Ashley and Prah 1997; Nise et al. 1989).

Additionally, both mice and human studies have revealed that VOC exposure may promote airway inflammation (Kwon et al. 2018; Vaughan Watson et al. 2021; Wang et al. 2012), which in turn may be involved in the pathogenesis of reduced pulmonary function. Although one cohort study investigated the effects of exposure to ambient BTEX on inflammatory markers and lung function, a significant association was observed only between BTEX and inflammation (Kwon et al. 2018). The effects of BTEX on lung function through regulation of the inflammatory response remain unclear. Because blood BTEX concentrations have decreased since 2005 (Jain 2017), no large-scale studies have been conducted on the association between low-dose blood BTEX and lung function in the general population thus far. White blood cells (WBC) count as a reflection of inflammation that plays a vital role in detecting human’s health is easily detected (Jackson and Miller 2020; Zhu et al. 2021). Therefore, the present study investigated the effects of low concentrations of blood BTEX on lung function through blood inflammatory parameters measured using white blood cell (WBC) five-part differential count and C-reactive protein (CRP) in a representative sample of the US population. The data were extracted from three combined cycles of the National Health and Nutrition Examination Survey (NHANES) spanning the years 2007 to 2012.

Study population

The NHANES was a series of studies conducted to monitor the health and nutritional status of a noninstitutionalised, civilian population in the United States and comprised both interviews and physical examinations. These studies adopted a complex, multistage, probability sampling design to select participants representative of the US population and released the results in 2-year cycles. In our study, we pooled three independent cycles (2007–2008, 2009–2010, and 2011–2012), which were selected based on the availability of blood and spirometry variables. Participants aged ≥ 18 years who had at least one detectable BTEX component in their blood were considered for inclusion in the study (n = 7,543). After exclusion of participants who had low-quality measurement data (see 2.3 Spirometry test) for lung function (n = 2,024), our study sample comprised 5,519 participants. Various subsets of the study sample were used to generate results for each specific VOC, because each VOC had a different pattern of missing values.

VOCs in the whole-blood samples

In accordance with the NHANES Laboratory/Medical Technologist Procedures Manual (LPM), whole-blood specimens were processed to determine the blood VOC concentrations. Details on the measurement of blood VOCs have been described in NHANES LPM for Volatile Organic Compounds (https://wwwn.cdc.gov/nchs/data/nhanes/2009-2010/labmethods/VOCWB_F_MET_Volatile_Organic_Compounds_(VOCs)_Blood.pdf). In brief, blood concentrations of BTEX, namely benzene, toluene, ethylbenzene, m-xylene/p-xylene, and o-xylene, constituted the exposure variables. For benzene, ethylbenzene, and o-xylene, more than half of the population had analytic results below the lower limit of detection (LOD); in this case, data were replaced with LOD/square root of 2 according to NHANES LPM. A total of 7,543 participants had detection results for at least one of the BTEX components.

Spirometry test

Spirometry measures the amount and speed of air that a person can exhale after taking the deepest possible breath and meets the recommendations of the American Thoracic Society (ATS) (Miller et al. 2005). Participants who experienced concurrent chest pain or another physical problem upon forceful expiration; had a recent heart attack, stroke, tuberculosis exposure; or had recently hemoptysis were excluded from testing. Details of the spirometry procedure have been described in Spirometry Procedures Manual (https://www.cdc.gov/nchs/data/nhanes/nhanes_11_12/spirometry_procedures_manual.pdf). According to ATS criteria, a satisfactory exhalation manoeuvre is one for which all spirometry quality curve attributes are acceptable. Thus, examinees whose effort quality attributes were labelled as follows were excluded: 1) the curves reflected a considerable time to peak flow or a nonrepeatable peak flow; 2) the curves indicated less than 6 s of exhalation or no plateau; or 3) the curves reflected coughing or excessive hesitation, as detected through the large and extrapolated volume. Among the participants who performed forceful exhalation at least three times and satisfied the criteria of an acceptable manoeuvre, we only included the participants who met the ATS data collection standards in terms of the forced expiratory volume at the first second (FEV1) and forced vital capacity (FVC) quality grades “A” and “B”; among three acceptable exhalation manoeuvres, the two highest values for the FEV1 and FVC must agree within a margin of 150 mL. The FEV1 and FVC repeatability goals were set to obtain results of an optimal quality, and we thus used these standards as our primary outcome measures (Enright 2010). Another study using lung function as the outcome also adopted these criteria (Elliott et al. 2006). Accordingly, 2,024 participants with low-quality spirometry test results were excluded, and 5,519 participants who had two reproducible manoeuvres constituted our study sample. The following spirometric parameters were employed in our analyses and were extracted from the single optimal overall value of two reproducible curves: 1) FEV1 (mL), FVC (mL), peak expiratory flow rate (PEFR; mL/s), maximum midexpiratory flow rate (FEF_25%–75%; mL/s), and the calculated FEV1:FVC ratio.

Blood inflammatory markers

In our analyses, we restricted inflammatory parameters to WBC and five-part differential count and CRP in blood to ensure consistency with the blood source of VOC measurement. Blood specimens were collected and processed according to the instructions of the NHANES LPM. The Beckman Coulter MAXM blood analyser was employed in mobile examination centres (MECs) in combination with an automatic diluting and mixing device for sample processing; this method produced a WBC count for the blood specimens and provided a distribution of blood cells for all participants. The WBC differential is calculated using volume, conductivity, and light scatter technology. Latex-enhanced nephelometry was used to quantify CRP. CRP values were only available from two NHANES cycles (2007–2008 and 2009–2010). Detailed analytical methods are described in the NHANES LPM.

Covariates

Different survey components were used in this analysis to obtain the potential covariates, namely age (Because the relationship between continuous age and each lung function parameter was non-linear, we transformed continuous age into 5 dummy variables to represent the 6 different age groups: 18 ≤ age < 30; 30 ≤ age < 40; 40 ≤ age < 50; 50 ≤ age < 60; 60 ≤ age < 70; age ≥ 70), sex (male/female), race or ethnicity (Mexican American/non-Hispanic White/non-Hispanic Black/other race), body mass index (BMI, kg/m2), occupation with COPD risk (yes/no), smoking status (no/former/current smoker), alcohol consumption (yes/no), physical activity intensity (mild/moderate/vigorous physical activity), asthma (yes/no), emphysema (yes/no), chronic bronchitis (yes/no), hypertension (yes/no), diabetes (yes/no), different survey cycles (Because we included three survey cycles from NHANES and blood concentrations of BTEX demonstrated temporal trend, with decreasing pattern from 2007 to 2012, we treated three survey cycles with 2 dummy variables). The variables of health conditions were all extracted from the health questionnaire.

Occupation with or without COPD risk was classified based on the variable “ocd392” that documented the jobs participants engaged for the longest time. Smoking status was created by three variables “smq020”, “smq040” and “smq680”: non-smokers were those who reported never smoked 100 cigarettes in lifetime and did not use tobacco/nicotine last 5 days; current smokers were those who reported smoking at least 100 cigarettes in lifetime, and at the time of the interview reported smoking either every day or some days or used tobacco/nicotine last 5 days; former smokers were those who smoked at least 100 cigarettes in lifetime but did not smoke at the time of the interview and not use tobacco/nicotine last 5 days. Physical activity intensity was created by two variables named “moderate recreational activity” (paq665) and “vigorous recreational activity” (paq650); then those people who reported neither moderate nor vigorous activity were recorded by us with a third variable named ‘mild activity’ for statistics of this study.

Ethics

The National Center for Health Statistics (NCHS) Research Ethics Review Board approved the NHANES, in which written informed consent had been obtained from the survey respondents.

Statistical analyses

According to the NCHS guidelines (https://wwwn.cdc.gov/nchs/nhanes/tutorials/module3.aspx), because whole-blood VOCs were measured in a one-half subsample of people attending an MEC exam, 6-year sample weights were calculated according to VOC subsample weights (‘wtsvoc6y’) when multiple survey cycles were combined; 2-year sample weights were calculated according to VOC subsample weights (‘wtsvoc2y’) when single survey cycles were used. Each participant was assigned a sample weight accounting for the complex survey design, nonresponse, poststratification considerations. P values of demographic parameters, inflammatory markers, and lung function parameters among different NHANES survey cycles were obtained using a weighted linear regression model for continuous variables and weighted chi-square test for categorical variables.

In the regression models with BTEX as the continuous dependent or independent variables, natural logarithm transformation was performed to normalise the data approximately with reference to an earlier NHANES study (Elliott et al. 2006). To investigate the effects of BTEX on pulmonary function, the weighted multiple linear models were fitted. Because BTEX concentrations were subjected to natural logarithm transformation, effect estimates are reported as the absolute changes in lung functions associated with an interquartile range (IQR) increase of each VOC on the natural log scale. The following four models were defined a priori: model 1, no adjustment; model 2, adjusted for demographic factors (age treated as dummy variables, sex, and race) and BMI; model 3, further adjusted for occupational exposure with or without risk of COPD, lifestyle factors (smoking status, alcohol consumption, and physical activity intensity); and model 4, further adjusted for health factors, namely asthma, emphysema, chronic bronchitis, hypertension, diabetes, and survey cycles treated as dummy variables. Sensitivity analyses were conducted to test the robustness of the associations by (i) excluding the outliers of VOC concentrations above the 99th percentile and below the limit of detection (LOD) to reduce the effect of extremely high values on the results of the association between BTEX and lung function both in the models treating VOCs as continuous variables or as tertiles/quartiles ; (ii) excluding participants younger than age 25 years or 30 years old to eliminate the potential effects of vigorous lung function in their early twenties or twenties, respectively; (iii) using two-pollutant models with all the covariates adjusted to control for potential confounding from coexposure of two VOC species; and (iv) using a weighted quantile sum (WQS) regression model to evaluate the effects of mixed exposure relative to air pollutants that highly correlate with each other (Carrico et al. 2015). Data were randomly separated into a training set (40%) and validation set (60%), and the coefficient direction for mixed BTEX was constrained to be negative through use of a single-pollutant model. After applying bootstrapping 1,000 times in the training set, we obtained the weight of each VOC for each outcome in the validation set, which was between 0 to 1 and summed to 1.

To facilitate interpretation of the association between BTEX and lung function, which was obtained using logarithmic transformation, we categorised non-transformed BTEX into tertiles/quartiles to conduct the association analysis with all the covariates adjusted. Participants were assigned to quartiles for toluene and m-/p-xylene that had more than 75% of measurements above the LOD; for benzene, ethylbenzene, and o-xylene that had a large proportion of measurements below the LOD, we categorized participants into three groups: the low-exposure group with concentrations < LOD, and the median-, and high-exposure groups that were equally divided among detectable samples (Sun et al. 2021). The effect estimate for each tertile/quartile is reported as the absolute change in lung function parameters, with the participants in the first tertile/quartile of each VOC concentration serving as the reference. A trend test was performed to test the hypothesis of an ordered relationship across the tertile/quartile of each VOC.

To investigate the health effects of BTEX on the inflammatory markers (WBC five-part differential count) measured in the blood specimens, the weighted linear models adjusted for all covariates were fitted. The effect estimates are reported as the absolute changes in cell count associated with an IQR increase of each VOC on the natural log scale. The health effects on lung function resulting from inflammatory parameters were also examined using weighted linear models, and the effect estimates are reported as the absolute changes in lung function parameters with an IQR increase determined using the WBC and five-part differential count. Sensitivity analysis was conducted using CRP as an inflammatory marker to test the association between VOC exposure and CRP as well as the association between CRP and lung function. With all the covariates mentioned in model 4 adjusted, we examined the mediating effect of the inflammatory markers between BTEX and the lung function parameters by using a “mediation” package, which is applied when processing a set of mediator combinations rather than a single mediator.

A two-sided P value of < 0.05 indicated statistical significance. All analyses were conducted in the R software version 3.5.3.

Characteristics of study participants and BTEX distributions

Procedure of participants selection was described in Figure S1. Of the 5,519 participants who had undergone both blood BTEX tests and spirometry tests in the NHANES study from 2007 to 2012, the mean age was 43.9 years, and 48.6% were male. The sample was ethnically diverse, with the non-Hispanic White accounting for the majority (69.3%). Most of the participants consumed alcohol (74.1%) and more than a quarter of them were currently smokers. Other characteristics are summarised in Table 1.

Table 1

The essential characteristics and pulmonary function measures of participants in the National Health and Nutrition Examination Survey (NHANES, 2007–2012) a
	Total (N = 5519)	NHANES 2007–2008 (N = 1699)	NHANES 2009–2010 (N = 1820)	NHANES 2011–2012 (N = 2000)	P value ^b
Age, years	43.8 (15.7)	43.8 (15.7)	43.9 (15.5)	43.7 (15.8)	0.957
Male, %	2684 (48.6)	827 (48.7)	892(51.0)	965 (48.2)	0.932
BMI, kg/m²	28.6 (6.6)	28.7 (6.8)	28.7 (6.4)	28.5 (6.6)	0.630
Race, %					0.003
Mexican Americans	459 (8.3)	150 (8.8)	162 (8.9)	147 (7.3)
Non-Hispanic White	3824 (69.3)	1218 (71.7)	1253 (68.8)	1353 (67.7)
Non-Hispanic Black	559 (10.1)	153 (9.0)	178 (9.8)	228 (11.4)
Other race	677 (12.3)	179 (10.5)	227 (12.5)	271 (13.6)
Occupation with risk of chronic obstructive pulmonary disease, %	882 (16.0)	278 (16.3)	292 (16.1)	312 (15.6)	0.039
Current smoking status, % ^c					0.531
No smoking	2765 (50.1)	831 (48.9)	921 (50.6)	1013 (50.7)
Former smoking	1160 (21.0)	359 (21.1)	369 (20.3)	432 (21.6)
Current smoking	1458 (26.4)	469 (27.6)	493 (27.1)	496 (24.8)
At least 12 alcohol drinks/1 year, %	4089 (74.1)	1211 (71.2)	1330 (73.1)	1548 (77.4)	< 0.001
Physical activity, % ^c					0.001
Mild physical activities	2252 (40.8)	742 (43.7)	782 (43.0)	728 (36.4)
Moderate physical activities	1674 (30.3)	490 (28.8)	561 (30.8)	623 (31.2)
Vigorous physical activities	1593 (28.9)	468 (27.5)	477 (26.2)	648 (32.4)
Health status, %
Ever been told with asthma	790 (14.3)	235 (13.8)	247 (13.6)	308 (15.4)	0.314
Ever been told with emphysema	48 ( 0.9)	15 (0.9)	16.0 (0.9)	17 (0.8)	0.875
Ever been told with chronic bronchitis	256 ( 4.6)	83 (4.9)	65 ( 3.6)	108 (5.4)	0.144
Ever been told with hypertension	1437 (26.0)	457 (26.9)	451 (24.8)	529 (26.4)	< 0.001
Ever been told with diabetes	472 ( 8.5)	129 ( 7.6)	143 ( 7.8)	200 (10.0)	< 0.001
Inflammatory markers, 10³ cell/µL
White blood cell	7.1 (2.1)	7.3 (2.1)	7.1 (2.1)	7.0 (2.0)	< 0.001
Lymphocyte	2.1 (0.7)	2.2 (0.7)	2.1 (0.7)	2.1 (0.6)	< 0.001
Monocyte	0.5 (0.2)	0.6 (0.2)	0.5 (0.2)	0.5 (0.2)	< 0.001
Neutrophils	4.2 (1.6)	4.3 (1.6)	4.2 (1.6)	4.1 (1.6)	0.053
Eosinophils	0.2 (0.2)	0.2 (0.1)	0.2 (0.2)	0.20 (0.2)	0.149
Basophils	0.04 (0.06)	0.04 (0.06)	0.04 (0.07)	0.04 (0.06)	0.904
Pulmonary functions measures
FEV1 (mL)	3265.1 (906.4)	3281.1 (903.2)	3262.9 (879.7)	3253.5 (932.9)	0.776
FVC (mL)	4176.4 (1096.1)	4195.3 (1107.7)	4173.6 (1070.1)	4162.9 (1109.8)	0.792
FEV1:FVC ratio	0.8 (0.1)	0.8 (0.1)	0.8 (0.1)	0.8 (0.1)	0.865
PEFR (mL/sec)	8417.4 (2174.1)	8339.5 (2190.4)	8447.6 (2113.8)	8456.0 (2213.4)	0.355
FEF_{25 − 75%} (mL/sec)	3067.7 (1290.9)	3095.7 (1280.6)	3062.7 (1265.6)	3048.5 (1322.2)	0.677
^a Volatile organic compounds (VOCs) subsample weights and weights for combined NHANES survey cycles were constructed in making estimates. ^b P values were obtained from weighted linear regression model for continuous variables and Chi-squared test for categorical variables, respectively. Abbreviations: FEV1 = Forced Expiratory Volume in the 1st Second, mL. FVC = Forced Vital Capacity, mL. PEFR = Peak Expiratory Flow Rate, mL/sec. FEF_{25 − 75%} = Forced Mid-expiratory Flow Rate, mL/sec.

Benzene, toluene, ethylbenzene, m-/p-xylene, and o-xylene were detected in 31.7%, 92.6%, 40.6%, 84.8%, and 45.7% of the total study population, respectively (Table 2). As previously reported in the NHANES population, the median blood concentrations of BTEX demonstrated a decreasing trend from 2005 to 2012 (Jain 2017). We verified this falling-off by extending the survey cycle from 1999 to 2012 (Figure S2).

Table 2

Concentrations of BTEX tested in whole blood specimens in the National Health and Nutrition Examination Survey (NHANES, 2007–2012), restricted to participants with pulmonary function data a
				Total	NHANES 2007–2008	NHANES 2009–2010	NHANES 2011–2012
VOCs	LOD	Total No. ^b	%>LOD	Median (25th, 75th percentile) ^c	Median (25th, 75th percentile) ^c	Median (25th, 75th percentile) ^c	Median (25th, 75th percentile) ^c	P value ^d
Benzene	0.024 (ng/mL)	5153	31.7	0.017 (0.017, 0.039)	0.017 (0.017, 0.049)	0.017 (0.017, 0.042)	0.017 (0.017, 0.031)	< 0.001
Toluene	0.025 (ng/mL)	2925	92.6	0.079 (0.045, 0.174)	0.084 (0.050, 0.187)	0.065 (0.038, 0.160)	NA	< 0.001
Ethylbenzene	0.024 (ng/mL)	5151	40.6	0.017 (0.017, 0.039)	0.017 (0.017, 0.039)	0.017 (0.017, 0.051)	0.017 (0.017, 0.029)	< 0.001
m-/p-Xylene	0.034 (ng/mL)	5121	84.8	0.070 (0.042, 0.132)	0.073 (0.045, 0.133)	0.075 (0.043, 0.155)	0.063 (0.039, 0.110)	< 0.001
o-Xylene	0.024 (ng/mL)	5058	45.7	0.017 (0.017, 0.038)	0.024 (0.017, 0.039)	0.017 (0.017, 0.044)	0.017 (0.017, 0.032)	< 0.001
Abbreviations: LOD, limit of detection. ^a BTEX subsample weights and weights for combined NHANES survey cycles were constructed in making estimates. ^b The number of people with testing values for each VOC, including the values more than or less than the limit of detection. Testing values below LOD was replaced with LOD/√2. The different sample size for each VOC was attributed to the incomplete detecting of BTEX for everyone. ^c Due to the skewed distribution of BTEX, median, the first quartile, and the third quartile were chosen to describe the distribution. ^d P values of the median concentrations of each VOC among three NHANES survey cycles were obtained from Kruskal-Wallis test.

Association between blood BTEX and lung function

Regarding the health effects of BTEX, benzene and toluene affected all five lung function parameters. An IQR increment of benzene on the natural log scale (0.83 ng/L) was associated with a change of − 23.8 mL (− 46.5, − 1.1) in FVC, -57.7 mL (-76.3, -39.1) in FEV1, − 0.9% (− 1.2, − 0.7) in FEV1:FVC ratio, − 184.5 mL/s (− 236.4, − 132.6) in PEFR, and − 137.5 mL/s (− 170.7, − 104.2) in FEF_25%–75%. An IQR increment of toluene on the natural log scale (1.47 ng/L) was associated with a change of -44.6 mL (-82.1, -7.0) in FVC, -64.0 mL (-94.4, -33.7) in FEV1, -0.7% (-1.1, -0.3) in FEV1:FVC ratio, -95.0 mL/s (-180.4, -9.6) in PEFR, and − 149.0 mL/s (-203.9, -94.0) in FEF_25%–75%. Both ethylbenzene and xylene isomers significantly reduced FEV1, FEV1:FVC ratio, PEFR and FEF_25%–75% (Table 3). In sensitivity analysis, to reduce the bias caused by the outliers of BTEX, we excluded the measurements above the 99th percentile and below the LOD, the effect sizes on lung functions from each VOC were more significant and larger except o-xylene (Table S2). To eliminate the potential effects of vigorous lung function in early twenties or twenties, after exclusion participants younger than age 25 years or 30 years old, the effect estimates of each VOC basically remained except for xylenes on FEV1 and on PEFR (Table S2). We then categorised non-transformed concentrations of BTEX to estimate their effects on lung function. Table 4 presents the changes in lung function for each tertile/quartile of VOC compared with those in the participants in the lowest category. The participants in the highest tertile/quartile of each VOC had higher significant decrements of lung function compared with participants in the lowest group. Tests for linear trends across quintiles were basically significant except for the effects of BTEX on part of outcomes. In this scenario, the linear concentration-response association of BTEX with some outcomes disappeared when pollutant was categorised into tertiles/quartiles, which may be attributed to a loss of information and an increase in uncertainty caused by use of category variable (Guo et al. 2018). In sensitivity analysis that exclude the outlier of VOC with measurements above the 99th percentile and below the LOD, the associations remained (Table S3).

Table 3

The expected absolute change of spirometry parameter (95% CI) as each VOC increases from the 25th to 75th percentile on the natural log scale a
Spirometry parameters	VOCs	Model 1 ^b		Model 2 ^b		Model 3 ^b		Model 4 ^b
Spirometry parameters	VOCs	Difference (95% CI)	P value	Difference (95% CI)	P value	Difference (95% CI)	P value	Difference (95% CI)	P value
FVC (mL)	Benzene	49.6 (23.7, 75.5)	< 0.001	-25.3 (-41.1, -9.5)	0.002	-27.4 (-50.1, -4.7)	0.018	-23.8 (-46.5, -1.1)	0.040
	Toluene	84.3 (33.3, 135.3)	0.001	-47.5 (-78.7, -16.2)	0.003	-44.8 (-82.2, -7.4)	0.019	-44.6 (-82.1, -7.0)	0.020
	Ethylbenzene	63.3 (33.6, 92.9)	< 0.0001	-12.9 (-31, 5.2)	0.163	-6.8 (-27.4, 13.8)	0.519	-3.4 (-24.1, 17.3)	0.748
	m-/p- Xylene	100.5 (66.1, 134.8)	< 0.0001	-9 (-30.2, 12.1)	0.402	-0.2 (-23.5, 23.0)	0.983	1.6 (-21.6, 24.8)	0.893
	o-Xylene	57.1 (26.0, 88.2)	< 0.001	-10.3 (-29.3, 8.7)	0.286	-1.8 (-21.6, 18.1)	0.861	1.0 (-18.8, 20.9)	0.919
FEV1 (mL)	Benzene	-16.3 (-37.8, 5.2)	0.137	-72.1 (-85.1, -59.1)	< 0.0001	-62.7 (-81.4, -44)	< 0.0001	-57.7 (-76.3, -39.1)	< 0.0001
	Toluene	-10.2 (-52.1, 31.8)	0.635	-95.8 (-121.4, -70.2)	< 0.0001	-66.4 (-96.9, -35.8)	< 0.0001	-64.0 (-94.4, -33.7)	< 0.0001
	Ethylbenzene	-1.2 (-25.7, 23.3)	0.924	-54.0 (-68.9, -39.1)	< 0.0001	-31.1 (-48, -14.3)	< 0.001	-28.3 (-45.1, -11.5)	< 0.001
	m-/p- Xylene	22.1 (-6.4, 50.5)	0.129	-50.3 (-67.7, -32.9)	< 0.0001	-22.8 (-41.9, -3.8)	0.019	-21.1 (-40, -2.3)	0.028
	o-Xylene	4.3 (-21.5, 30.1)	0.743	-37.3 (-53.0, -21.7)	< 0.0001	-16.5 (-32.8, -0.3)	0.046	-14.8 (-30.9, -1.4)	0.035
FEV1: FVC ratio (%)	Benzene	-1.4 (-1.6, -1.2)	< 0.0001	-1.3 (-1.5, -1.2)	< 0.0001	-1.0 (-1.3, -0.8)	< 0.0001	-0.9 (-1.2, -0.7)	< 0.0001
	Toluene	-1.9 (-2.2, -1.5)	< 0.0001	-1.5 (-1.8, -1.2)	< 0.0001	-0.8 (-1.2, -0.4)	< 0.0001	-0.7 (-1.1, -0.3)	< 0.001
	Ethylbenzene	-1.2 (-1.4, -1.0)	< 0.0001	-1.1 (-1.3, -0.9)	< 0.0001	-0.6 (-0.8, -0.4)	< 0.0001	-0.6 (-0.8, -0.4)	< 0.0001
	m-/p- Xylene	-1.4 (-1.6, -1.1)	< 0.0001	-1.1 (-1.3, -0.9)	< 0.0001	-0.5 (-0.7, -0.3)	< 0.0001	-0.5 (-0.7, -0.3)	< 0.0001
	o-Xylene	-1.0 (-1.2, -0.7)	< 0.0001	-0.7 (-0.9, -0.5)	< 0.0001	-0.3 (-0.5, -0.1)	< 0.001	-0.4 (-0.6, -0.2)	< 0.001
PEFR (mL/sec)	Benzene	-134.5 (-185.9, -83.1)	< 0.0001	-263.3 (-299.8, -226.8)	< 0.0001	-199.3 (-251.6, -146.9)	< 0.0001	-184.5 (-236.4, -132.6)	< 0.0001
	Toluene	-32.6 (-133.4, 68.2)	0.527	-268.6 (-340.8, -196.5)	< 0.0001	-108.5 (-194.3, -22.8)	0.013	-95.0 (-180.4, -9.6)	0.029
	Ethylbenzene	-51.8 (-110.6, 7.0)	0.084	-186.5 (-228.1, -144.8)	< 0.0001	-84.0 (-131, -36.9)	< 0.001	-79.1 (-125.9, -32.3)	< 0.001
	m-/p- Xylene	22.7 (-45.7,91.1)	0.516	-171 (-219.7, -122.3)	< 0.0001	-52.1 (-104.4, 0.2)	0.051	-55.6 (-108.5, -2.6)	0.040
	o-Xylene	-5.0 (-66.9,56.9)	0.874	-125.3 (-169.4, -81.1)	< 0.0001	-38.3 (-83.5, 6.9)	0.097	-43.5 (-89.0, -2.2)	0.032
FEF_{25 − 75%} (mL/sec)	Benzene	-130.6 (-161, -100.1)	< 0.0001	-171.4 (-194.6, -148.2)	< 0.0001	-145.8 (-179.3, -112.3)	< 0.0001	-137.5 (-170.7, -104.2)	< 0.0001
	Toluene	-189.5 (-249.2, -129.9)	< 0.0001	-220.6 (-267, -174.3)	< 0.0001	-152.0 (-207.5, -96.5)	< 0.0001	-149.0 (-203.9, -94.0)	< 0.0001
	Ethylbenzene	-105.0 (-139.7, -70.3)	< 0.0001	-135.3 (-161.8, -108.9)	< 0.0001	-81.9 (-111.9, -51.9)	< 0.0001	-79.0 (-108.9, -49.1)	< 0.0001
	m-/p- Xylene	-107.5 (-148, -66.9)	< 0.0001	-139.4 (-170.4, -108.3)	< 0.0001	-78.6 (-112.5, -44.6)	< 0.0001	-77.0 (-110.6, -43.3)	< 0.0001
	o-Xylene	-79.8 (-116.5, -43.1)	< 0.0001	-91.3(-119.2, -63.4)	< 0.0001	-47.1(-76.1, -18.2)	0.001	-46.7 (-75.5, -17.9)	0.001
Abbreviations: FEV1 = Forced Expiratory Volume in the 1st Second. FVC = Forced Vital Capacity. PEFR = Peak Expiratory Flow Rate. FEF_{25 − 75%} = Forced Mid-expiratory Flow Rate, ml/sec. ^a The concentrations of each VOC were log transformed due to the skewed distribution. ^b Model 1, no adjustment; model 2, adjusted for demographical factors (age, sex, race) and body-mass index; model 3, further adjusted for occupational exposure, lifestyle factors (smoking status, alcohol consumption, and physical activity intensity); and model 4, further adjusted for health factors, including asthma (yes or no), emphysema (yes or no), chronic bronchitis (yes or no), hypertension (yes or no), and diabetes (yes or no), created dummy variables for different age groups ( 18 ≤ age < 30; 30 ≤ age < 40; 40 ≤ age < 50; 50 ≤ age < 60; 60 ≤ age < 70; age ≥ 70) and for survey cycle (cycle = 2007–2009; cycle = 2009–2010; cycle = 2011–2012).

Table 4

The expected absolute change of spirometry parameter (95% CI) a by tertiles or quartiles of each VOC.
Models	Difference (95% CI)
Models	FVC (ml)	P value	FFV1 (mL)	P value	FEV1:FVC ratio (%)	P value	PEFR (mL/sec)	P value	FEF_25-75% (mL/sec)	P value
Benzene (ng/mL)
T1 (≤ 0.017)	0		0		0		0		0
T2 (0.018–0.115)	22.7 (-27.7, 73.0)	0.378	5.9 (-38.1, 49.8)	0.794	0.1 (-0.5, 0.6)	0.817	-41.3 (-164.2, 81.5)	0.510	-38.3 (-116.8, 40.2)	0.339
T3 (> 0.115)	-80.0 (-131.0, -29.0)	0.002	-172.0 (-230.5, -113.5)	< 0.0001	-3.0 (-3.7, -2.24)	< 0.0001	-536.4 (-699.9, -372.8)	< 0.0001	-439.2 (-543.8, -334.7)	< 0.0001
P trend ^b	0.014		< 0.0001		< 0.0001		< 0.0001		< 0.0001
Toluene (ng/mL)
Q1 (≤ 0.046)	0		0		0		0		0
Q2 (0.047–0.080)	8.7 (-58.6, 76.0)	0.800	20.1 (-34.2, 74.4)	0.467	0.1 (-0.6, 0.8)	0.729	16.3 (-136.1, 168.8)	0.834	16.6 (-81.7, 114.9)	0.741
Q3 (0.081–0.199)	-16.1 (-84.6, 52.4)	0.645	2.1 9 (-53.2, 57.4)	0.939	0.3 (-0.4, 1.0)	0.402	8.6 (-145.9, 163.1)	0.913	-19.6 (-119.7, 80.6)	0.702
Q4 (> 0.199)	-93.8 (-175.4, -12.1)	0.025	-131.0 (-196.9, -65.0)	< 0.001	-1.5 (-2.3, -0.7)	< 0.001	-192.6 (-377.6, -7.7)	0.041	-321.7 (-441.1, -202.4)	< 0.0001
P trend ^b	0.038		0.001		0.010		0.104		< 0.0001
Ethylbenzene (ng/mL)
T1 (≤ 0.017)	0		0		0		0		0
T2 (0.018–0.053)	36.0 (-11.8, 83.7)	0.140	29.2 (-9.5, 68.0)	0.140	0.2 (-0.3, 0.6)	0.491	58.6 (-49.3, 166.6)	0.287	14.2 (-54.6, 82.9)	0.687
T3 (> 0.053)	-7.8 (-66.3, 50.8)	0.795	-103.3 (-150.8, -55.7)	< 0.0001	-2.3 (-2.9, -1.7)	< 0.0001	-301.5 (-433.8, -169.2)	< 0.0001	-316.5 (-400.8, -232.2)	< 0.0001
P trend ^b	0.811		0.002		< 0.0001		< 0.001		< 0.0001
m-/p-Xylene (ng/mL)
Q1 (≤ 0.043)	0		0		0		0		0
Q2 (0.044–0.070)	-15.2 (-66.0, 35.6)	0.558	-2.2 (-43.5, 39.1)	0.916	0.2 (-0.3, 0.7)	0.376	-12.1 (-126.7, 102.4)	0.836	-14.7 (-88.3, 58.9)	0.695
Q3 (0.071–0.133)	4.4 (-46.5, 55.4)	0.865	-6.2 (-47.6, 35.2)	0.769	-0.2 (-0.7, 0.3)	0.501	-6.0 (-120.7, 108.8)	0.919	-65.3 (-139.1, 8.5)	0.083
Q4 (> 0.133)	-5.6 (-63.7, 52.5)	0.850	-64.0 (-111.2, -16.8)	0.008	-1.4 (-2.0, -0.8)	< 0.0001	-142.1 (-272.7, -11.5)	0.033	-243.4 (-327.5, -159.3)	< 0.0001
P trend ^b	0.968		0.020		< 0.0001		0.074		< 0.0001
o-Xylene (ng/mL)
T1 (≤ 0.017)	0		0		0		0		0
T2 (0.018–0.041)	-14.5 (-60.8, 31.7)	0.538	-33.2 (-70.8, 4.3)	0.083	-0.6 (-1.1, -0.1)	0.011	-69.0 (-174.3, 36.2)	0.199	-95.4 (-162.4, -28.5)	0.005
T3 (> 0.041)	14.7 (-36.3, 65.7)	0.571	-44.2 (-85.7, -2.8)	0.036	-1.2 (-1.7, -0.7)	< 0.0001	-131.9 (-248.0, -15.7)	0.026	-179.2 (-253.1, -105.3)	< 0.0001
P trend ^b	0.708		0.023		< 0.0001		0.022		< 0.0001
Abbreviations: FEV1 = Forced Expiratory Volume in the 1st Second. FVC = Forced Vital Capacity. PEFR = Peak Expiratory Flow Rate. FEF_{25 − 75%} = Forced Mid-expiratory Flow Rate ml/sec. ^a The expected absolute change of lung functions was estimated with model 4. ^b P trend test was performed across each VOC tertile/quartile with the corresponding tertile/quartile treated as a numeric variable with an ordinal variable code as 1–3/1–4.

In the two-pollutant models with adjustment for co-exposure to other VOCs, the associations of BTEX with spirometric parameters were not consistent with the results in single-pollutant model (Figure S3), which may have resulted from the high correlation between each BTEX (Table S1). In the multipollutant analyses, a quartile increase in the WQS index of the BTEX mixture was significantly associated with a change of -56.1 mL (-93.4, -13.1) in FVC, − 87.6 mL (− 107.3, − 46.1) in FEV1, − 1.1% (− 1.4, − 0.6) in FEV1:FVC ratio, − 157.1 mL/s (− 226.2, − 51.1) in PEFR, and − 186.6 mL/s (− 232.0, − 114.0) in FEF_25%–75% (Table 5). In the BTEX mixture, benzene exerted the strongest association with all five outcomes with weights from 0.458 to 0.698, suggesting that benzene was dominant in the association between the BTEX mixture and lung function.

Table 5

Weighted quartile sum regression of respirometer parameter and one-quartile change in blood BTEX mixture a
	Mixture weights					Abs. Change (95% CI)	P value
Spirometry parameter	Benzene	Toluene	Ethylbenzene	m-/p-Xylene	o-Xylene	Abs. Change (95% CI)	P value
FVC (mL)	0.653	0.179	0.015	0.012	0.141	-56.1 ( -93.4, -13.1)	0.009
FEV1 (mL)	0.698	0.036	0.014	0.018	0.234	-87.6 (-107.3, -46.1)	< 0.0001
FEV1:FVC ratio (%)	0.458	0.006	0.033	0.116	0.387	-1.1 (-1.4, -0.6)	< 0.0001
PEFR (mL/sec)	0.684	0.040	0.013	0.037	0.227	-157.1 ( -226.2, -51.1)	0.002
FEF_{25 − 75%} (mL/sec)	0.626	0.007	0.026	0.142	0.200	-186.6 ( -232.0, -114.0)	< 0.0001
Abbreviations: FEV1 = Forced Expiratory Volume in the 1st Second. FVC = Forced Vital Capacity. PEFR = Peak Expiratory Flow Rate. FEF_{25 − 75%} = Forced Mid-expiratory Flow Rate, ml/sec. ^a The expected absolute change of lung functions was estimated with model 4.

Association between VOC and inflammatory markers

The associations between VOCs and inflammatory markers as measured using the WBC and five-part differential count in fully adjusted models are depicted in Fig. 1. We observed that an IQR increase of benzene on the natural log scale was significantly associated with an increment of 0.44 × 10³ counts/µL (0.38, 0.51) for WBCs, 0.09 × 10³ counts/µL (0.07, 0.12) for lymphocytes, 0.02 × 10³ counts/µL (0.01, 0.03) for monocytes, 0.32 × 10³ counts/µL (0.27, 0.37) for neutrophils, and 0.004 × 10³ counts/µL (0.002, 0.007) for basophils. Ethylbenzene was also associated with these five indicators. Exposure to high levels of toluene increased the numbers of WBCs, lymphocytes, neutrophils, and monocytes in blood. Furthermore, m-/p-xylene and o-xylene had the least influence on the inflammatory markers and were only associated with WBC, lymphocyte, and neutrophil counts. Sensitivity analyses (Figure S4) also revealed that benzene, toluene, ethylbenzene, and o-xylene were associated with increased levels of CRP.

Association between inflammatory markers and lung function

Associations between the WBC and five-part differential count and spirometric parameters in the fully adjusted models are detailed in Table 6. An IQR increment of WBCs was associated with a change of − 79.9 mL (− 103.6, − 56.2) in FVC, − 66.4 mL (− 85.8, − 47.0) in FEV1, − 163.3 mL/s (− 217.4, − 109.2) in PEFR, and − 53.3 mL/s (− 88.1, − 18.6) in FEF_25%–75%. An increase in neutrophils count was also significantly associated with a reduction in the aforementioned four lung function parameters; an IQR increment of neutrophils was associated with a change of − 51.7 mL (− 75.4, − 28.0) in FVC, − 48.6 mL (− 67.9, − 29.2) in FEV1, − 135.5 mL (− 189.3 − 81.3) in PEFR, and − 48.3 mL (− 82.9, − 13.6) in FEF_25%–75%. Increased lymphocytes and monocytes counts were associated with reductions in FVC, FEV1, and PEFR, whereas an increased eosinophil count decreased all five lung function parameters. Sensitivity analysis of the association between CRP and lung function also demonstrated that CRP increase had detrimental effects on lung function (Table S4).

Table 6

The expected absolute change of spirometry parameter (95% CI) a as each hematological parameter increases from the 25th to 75th percentile on the natural log scale b
	Differences of spirometry parameters (95% CI)
Hematological parameter	FVC (mL)	P value	FEV1 (mL)	P value	FEV1:FVC ratio (%)	P value	PEFR (mL/sec)	P value	FEF_{25 − 75%} (mL/sec)	P value
White blood cells (10³ /µL)	-79.9 (-103.6, -56.2)	< 0.0001	-66.4 (-85.8, -47.0)	< 0.0001	-0.1(-0.4, 0.1)	0.320	-163.3(-217.4, -109.2)	< 0.0001	-53.3 (-88.1, -18.6)	0.003
Lymphocytes (10³ /µL)	-72.3 (-93.5, -51.2)	< 0.0001	-47.0 (-64.3, -29.6)	< 0.0001	0.2(-0.1, 0.4)	0.131	-86.0 (-134.4, -37.6)	< 0.001	-12.3 (-43.3,18.7)	0.437
Neutrophils (10³ /µL)	-51.7 (-75.4, -28.0)	< 0.0001	-48.6 (-67.9, -29.2)	< 0.0001	-0.2(-0.4, 0.1)	0.128	-135.3 (-189.3, -81.3)	< 0.0001	-48.3 (-82.9, -13.6)	0.006
Monocytes (10³ /µL)	-55.4 (-75.8, -35.1)	< 0.0001	-41.9 (-58.6, -25.2)	< 0.0001	0 (-0.2, 0.2)	0.936	-103.8 (-150.3, -57.3)	< 0.0001	-29.5 (-59.3,0.3)	0.052
Basophils (10³ /µL)	6.8 (-21.9, 35.4)	0.643	1.2 (-22.2, 24.7)	0.918	0 (-0.3, 0.3)	0.931	-34 (-99.3, 31.3)	0.307	3.6 (-38.2,45.4)	0.866
Eosinophils (10³ /µL)	-49.9 (-73.4, -26.4)	< 0.0001	-56.1 (-75.3, -36.9)	< 0.0001	-0.5 (-0.7, -0.2)	< 0.0001	-103.6 (-157.2, -50.0)	< 0.001	-87.4 (-121.7, -53.2)	< 0.0001
Abbreviations: FEV1 = Forced Expiratory Volume in the 1st Second. FVC = Forced Vital Capacity. PEFR = Peak Expiratory Flow Rate. FEF_{25 − 75%} = Forced Mid-expiratory Flow Rate, ml/sec. ^a The expected absolute change of spirometry parameter was estimated with model 4. ^b The concentrations of each VOC were log transformed due to the skewed distribution.

Mediating effects of inflammatory markers between VOC and lung function

The multiple mediating effects of the inflammatory markers (WBC and five-part differential count and CRP) between BTEX and lung function after control for all covariates were examined. As illustrated in Fig. 2, the effects of benzene on FEV1 and PEFR were mediated by the inflammatory indicators: the direct effects of benzene on FEV1 (path c’) after controlling for mediated effects were indicated with a regression coefficient (𝛽) of − 55.837 (P < 0.0001). The total effect (sum of the mediated and direct effect) of benzene on FEV1 (path c) was statistically significant, with a regression coefficient (𝛽) of − 65.653 (P < 0.0001). The mediation effect of WBCs, lymphocytes, neutrophils, and basophils on FEV1 was − 9.816 (path a1 * path b1 + path a2 * path b2 + path a3 * path b3 + path a4 * path b4). Thus, 15% of the total effect of benzene on FEV1 was indirectly caused by the above inflammatory indicators. Similarly, 11.0 %of the total effect of benzene on PEFR was indirectly caused by WBCs, lymphocytes, and neutrophils.

Mediation analysis also revealed that toluene affected FEV1, and ethylbenzene affected FEV1 and PEFR: 12.9% of the total effect of toluene on FEV1 was indirectly caused by WBCs; WBCs, lymphocytes, and neutrophils explained 17.8% and 16.7% of the total effects of ethylbenzene on FEV1 and PEFR, respectively (Fig. 2).

In three combined cycles of the NHANES data, although the concentrations of BTEX decreased from 1999 onward, we determined that low-dose exposure to BTEX was associated with reduced pulmonary function in the covariate-adjusted model, which simultaneously indicated the monotonic exposure–response relationships. The multipollutant analyses identified benzene as the primary agent driving such associations. Furthermore, inflammation could be involved in the pathogenesis of reduced lung function induced through VOCs by association and mediation analyses.

Studies have reported that exposure to air VOCs was associated with loss of pulmonary function among adults (Cakmak et al. 2014; Chen et al. 2018; Du et al. 2020; Mukherjee et al. 2016; Ojo et al. 2017; Singh et al. 2017; Weichenthal et al. 2012) and older population (Mendes et al. 2016; Yoon et al. 2010). Most studies have focused on the health effects of air VOC exposure in workplace environments (Du et al. 2020; Mukherjee et al. 2016; Ojo et al. 2017; Singh et al. 2017; Weichenthal et al. 2012), which limits the possibility for extrapolation to the general population. One large-scale study reported an association between indoor residential VOC exposure and lung function in the general population (Cakmak et al. 2014). Another study integrating exposure to VOC (including BTEX) from both indoor and outdoor sources indicated the detrimental effects of BTEX on lung function (Chen et al. 2018). By contrast, research elucidating the health effects of indoor BTEX on lung function failed to draw significant conclusions (Kwon et al. 2018). Only one study used VOC levels detected in blood as a measure of exposure (Elliott et al. 2006), which differs from exposure measurements relying on atmospheric monitoring data. In summary, the health effects of BTEX on lung function have only been specifically investigated in two studies (Chen et al. 2018; Kwon et al. 2018). In addition, few studies have been conducted among people exposed in nonoccupational settings, and general conclusions cannot be drawn using only studies of occupationally exposed groups.

With respect to the methods for measuring exposure levels, some studies have examined VOCs from a single source, such as indoor (Cakmak et al. 2014; Singh et al. 2017) or outdoor sources (Du et al. 2020; Mukherjee et al. 2016; Ojo et al. 2017; Weichenthal et al. 2012), whereas others have examined a combination of the two (Chen et al. 2018). Because of the multiple exposure routes and differences in individual intake and metabolism of BTEX, internal exposure biomarkers are regarded as more accurate in reflecting chronic and repetitive exposure to VOCs from all sources (Ashley et al. 1994; Ashley and Prah 1997; Sexton et al. 2005). Until now, only one study used blood VOCs as exposure indicators to investigate the effect on lung function, reporting that only 1,4-dichlorobenzene affected lung function (Elliott et al. 2006). Widely used BTEX may remain in the environment and pose potential human health risks, especially for lung function following long-term exposure. Despite this risk, relevant studies remain scarce. Our study verified that even though the blood concentrations of BTEX decreased annually in the general population, they still had significant effects in terms of reduced lung function.

Currently, the mechanism of how VOC affects lung function is unknown. One mechanistic study proposed that the C − O polar bonding of VOCs is prone to highly reactive chemical reactions with amine-rich epithelial intima and mucous membranes and subsequently results in inflammatory responses (Yun et al. 2018). Hence, this molecular basis explains the association of BTEX with inflammation. In mice models, Wang et al. determined that exposure to benzene, toluene, and xylene and formaldehyde at concentrations 100 times the Chinese indoor air quality standard over 2 weeks (2 h per day, 5 days per week) could significantly increase total cells, macrophages, and eosinophils (Wang et al. 2012). Exposure to polyvinylchloride flooring could increase inflammation, as indicated through elevated Th2 cytokine levels and decreased interleukin 12 production in maturing dendritic cells in mice (Wang et al. 2012). Additionally, an epidemiological study revealed that moving to a new building with high concentrations of VOCs including BTEX could lead to high urinary leukotriene E4 levels (Kwon et al. 2018). Watson et al. reported that blood BTEX have negative effects on complete blood count with five-part differential parameters in the NHANES spanning 2005 to 2010 (Vaughan Watson et al. 2021). Oxidative stress is also associated with health and disease status and involved in mechanisms leading to lung function decline (Kwon et al. 2018; Yoon et al. 2010). No studies have been conducted to explore the mediation effect of inflammation induced by VOCs, especially BTEX exposure on lung function decline. Although one cohort study investigated the effects of exposure to ambient BTEX on inflammatory response and lung function simultaneously, a significant association was only observed between BTEX and inflammatory parameters (Kwon et al. 2018), with no report of the mediating role of inflammation between BTEX and lung function. In our study, we not only observed the significant association of blood BTEX with reduced lung function using elevated WBCs and the five-part differential parameters and CRP but also determined that elevated WBCs and the five-part differential parameters negatively affected lung function. CRP is a nonspecific indicator of systemic inflammation, and some evidence implies its contribution to regulating inflammatory processes (Sproston and Ashworth 2018). White blood cells (WBC) count as a reflection of inflammation plays a vital role in detecting human’s health (Jackson and Miller 2020; Zhu et al. 2021).

Only two NHANES survey cycles tested CRP levels; therefore, this information was incorporated into the sensitivity analysis. Subsequently, we verified the role of WBCs and the five-part differential parameters in mediating the effect of benzene, toluene, and ethylbenzene on reduced lung function through the mediation effect model.

VOCs have been linked to pulmonary dysfunction; however, the health effects of low doses of VOCs on lung function and the mechanism underlying these effects remain unclear. A novelty of the present study is its examination of the association between low doses of VOCs and lung function through changes in inflammatory markers. And we evaluated comprehensively the health effects of BTEX on five lung function parameters that are related to both large and small airways. FEV1 is the amount of air moving during the first second of FVC maneuver, representing the movement of air through the larger airway (Goel et al. 2015). FEV1:FVC ratio and PEFR are also thought to reflect the caliber of larger airways, whereas FEF_{25 − 75%} known as mid expiratory flow is more reflection of the caliber of small airways (Goel et al. 2015). We found that the effects of BTEX on FEF_25%−75% appear to be considerably greater than those on FEV1 and FVC but similar to those on PEFR (Table 3), indicating that long-term exposure to BTEX may be both detrimental to large and small airway functions. However, the statistical significance of the associations between each VOC and the pulmonary function parameters may have been the product of chance in a single-pollutant model. Therefore, we conducted two analyses to validate our findings. First, trend tests were performed to test the effect of each tertile/quartile of BTEX on lung function in the single-pollutant model to determine whether an increased blood concentration of each VOC was associated with reduced lung function. Among the VOCs that exhibited a significant association with the pulmonary function parameters, we noted a dose–response effect of these VOCs on the related lung function parameters. Second, in view of the high correlation among the pollutants, we applied a WQS regression model and determined the significant effects of the BTEX mixtures on lung function. Consistent with the results of the single-pollutant model, BTEX mixture was associated with all five lung function parameters in the multipollutant model. In the single-pollutant analysis, benzene was the VOC that affected the most lung function parameters. Similarly, the mixture analyses indicated that benzene was the primary agent driving the associations, with its weights ranging from 0.458 to 0.698. Benzene may have had stronger adverse effects than the other VOCs because it is the most toxic oil component with carcinogenic nature and the slowest degradation rate under most field conditions (Li et al. 2021; McNally et al. 2017). After determining the association between BTEX and lung function, we elucidated the potential mechanism of inflammation linking BTEX exposure to reduced lung function. Although we determined that ethylbenzene, m-/p-xylene, and o-xylene have no effect on certain lung function parameter FVC, all these three components significantly increased WBCs, lymphocytes, and neutrophils. It is reasonable because symptoms present as physiological changes after the associated biological changes have already occurred. Hereby, although the deleterious effects of ethylbenzene, m-/p-xylene, and o-xylene on lung function have not been detected, they already have adverse effects on biological indicators. This finding suggests to some degree that inflammation depends on the pathway through which VOC damages lung function. Furthermore, the mediating effect analysis clearly demonstrated the mediating role of inflammation. Benzene affected the most inflammatory markers (WBCs, lymphocytes, neutrophils, monocytes, basophils, and CRP), and these inflammatory markers mediated the effects of benzene on two outcomes (FEV1 and PEFR). This result is consistent with the aforementioned conclusions and further indicates that benzene is the driving substance in BTEX leading to reduced lung function.

Instead of measuring BTEX exposure through air quality monitoring, we used BTEX in blood as the exposure variable. Because not all pollutants in the environment readily enter organisms, the adverse health effects of their environmental concentrations are far less critical than the amount of chemicals eventually absorbed by the human body (Escher and Hermens 2004). Therefore, the internal dose absorbed into the body and tested through biological monitoring is regarded as a more accurate indicator of exposure (Escher and Hermens 2004). The blood concentrations of VOCs including BTEX in the NHANES population have been decreasing since the late 1980’s, possibly due to the substitutions of low-emission products and environmental management (Jain 2017; Su et al. 2011). Although the concentrations of BTEX in blood are already low, they still had a significant effect on inflammatory markers and lung function. For example, benzene was detected in only 31.7% of our study population but was observed to have strong health side effects. This result is a warning for environmental policy makers: Simply reducing rather than forbidding the usage of some VOCs classified as carcinogens, such as benzene, will result in continued adverse health effects.

This study has several strengths. First, we used a large and nationally representative sample of US adults. Second, we assessed the exposure level of each BTEX component using internal biomarkers, which can reduce exposure misclassification. Third, we evaluated the lung function effects of each BTEX chemical exposure separately and as a mixed exposure. For the modelling of a mixed BTEX substances as exposure variables, WQS was suitable for application in the scenario of highly correlated mixtures and addressed potential confounding by coexposures, which helped the analysis more closely represent the real-world situation (Czarnota et al. 2015). Moreover, we demonstrated the mechanism through which BTEX affect lung function.

However, some limitations should be considered. First, this cross-sectional study relied on a single blood measurement, which limits inferences as to causality. Second, we cannot discount the possibility of exposure misclassification, because the measured blood concentrations of BTEX may change over time and places and not reflect the relevant exposure accurately. However, an experiment with a human model revealed that blood concentrations are directly related to air exposure concentrations, and that such exposure may result in longer periods of elevated blood concentrations (Ashley and Prah 1997). This indicates that blood BTEX can reflect the overall bodily exposure to a certain extent. Thus, the second limitation is not a big issue. Third, exposure to second-hand and environmental tobacco smoke were not fully accounted for in this study and may have affected the results. However, we adjusted for the most significant exposure to cigarettes, namely status as a current smoker, and still observed significant associations between VOCs and health outcomes.

The results of this cross-sectional analysis of a representative sample of US adults suggest that exposure to BTEX is associated with reduced lung function, which is mediated through inflammation. Our results extend those of related studies by demonstrating that, although the exposure dose to BTEX is low, these aromatic hydrocarbons are still harmful to lung function; this finding emphasises the urgency of reducing exposure to BTEX.

BTEX: benzene, toluene, ethylbenzene, and meta-xylene (m-xylene), para-xylene (p-xyline), and ortho-xylene (o-xylene);

VOCs: volatile organic compounds;

FVC: forced vital capacity;

FEV1: forced expiratory volume in the first second;

PEFR: peak expiratory flow rate;

FEF_25%–75%: forced mid expiratory flow

COPD: chronic obstructive pulmonary disease;

WBC: white blood cell;

CRP: C-reactive protein;

LOD: the lower limit of detection.

Acknowledgements

The authors would like to acknowledge the contributions of CDC Division of Laboratory Science, the Tobacco and Volatiles branch. All authors read and approved the final manuscript.

Author contribution

Yansu He conducted formal analysis, drafted and reviewed the manuscript. Yong Lin ran the analysis. Hong Qiu validated the accuracy of data analyses through a technical review. Linying Wu validated the integrity and accuracy of outcome assessment. Kin Fai Ho conceptualized and supervised the study and reviewed and revised the manuscript. All authors have read and approved the final manuscript.

Funding

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Availability of data and materials

The datasets used and analysed during the current study are available from the corresponding author on reasonable request.

Ethics approval and consent to participate

The National Center for Health Statistics (NCHS) Research Ethics Review Board approved the NHANES, in which written informed consent had been obtained from the survey respondents.

Consent for publication

Not applicable.

Competing interests

The authors declare that they have no conflict of interest

Author details

¹ JC School of Public Health and Primary Care, The Chinese University of Hong Kong, HKSAR, China. ² Department of Computer Science and Engineering, The Hong Kong University of Science and Technology, HKSAR, China. ³ Institute of Environment, Energy and Sustainability, The Chinese University of Hong Kong, HKSAR, China. ⁴Department of Respiratory Disease, The First Affiliated Hospital, College of Medicine, Zhejiang University, Hangzhou, China.

Ashley DL, Bonin MA, Cardinali FL, McCraw JM, Wooten JV (1994) Blood concentrations of volatile organic compounds in a nonoccupationally exposed US population and in groups with suspected exposure. Clin Chem 40:1401–1404. https://doi.org/10.1093/clinchem/40.7.1401
Ashley DL, Prah JD (1997) Time dependence of blood concentrations during and after exposure to a mixture of volatile organic compounds. Arch Environ Health 52:26–33. https://doi.org/10.1080/00039899709603796
Bolden AL, Kwiatkowski CF, Colborn T (2015) New Look at BTEX: Are Ambient Levels a Problem? Environ Sci Technol 49:5261–5276. https://doi.org/10.1021/es505316f
Cakmak S, Dales RE, Liu L, Kauri LM, Lemieux CL, Hebbern C, Zhu J (2014) Residential exposure to volatile organic compounds and lung function: results from a population-based cross-sectional survey. Environ Pollut 194:145–151. https://doi.org/10.1016/j.envpol.2014.07.020
Carrico C, Gennings C, Wheeler DC, Factor-Litvak P (2015) Characterization of Weighted Quantile Sum Regression for Highly Correlated Data in a Risk Analysis Setting. J Agric Biol Environ Stat 20:100–120. https://doi.org/10.1007/s13253-014-0180-3
Chen L, Hu G, Fan R, Lv Y, Dai Y, Xu Z (2018) Association of PAHs and BTEX exposure with lung function and respiratory symptoms among a nonoccupational population near the coal chemical industry in Northern China. Environ Int 120:480–488. https://doi.org/10.1016/j.envint.2018.08.004
Czarnota J, Gennings C, Wheeler DC (2015) Assessment of weighted quantile sum regression for modeling chemical mixtures and cancer risk. Cancer Inf 14:159–171. https://doi.org/10.4137/CIN.S17295
de Blas M, Navazo M, Alonso L, Durana N, Gomez MC, Iza J (2012) Simultaneous indoor and outdoor on-line hourly monitoring of atmospheric volatile organic compounds in an urban building. The role of inside and outside sources. Sci Total Environ 426:327–335. https://doi.org/10.1016/j.scitotenv.2012.04.003
Du M, Hall GL, Franklin P, Musk AB, Mullins BJ, de Klerk N, Elliott NSJ, Sodhi-Berry N, Brims F, Reid A (2020) Association between diesel engine exhaust exposure and lung function in Australian gold miners. Int J Hyg Environ Health 226:113507. https://doi.org/10.1016/j.ijheh.2020.113507
Elliott L, Longnecker MP, Kissling GE, London SJ (2006) Volatile organic compounds and pulmonary function in the Third National Health and Nutrition Examination Survey, 1988–1994. Environ Health Perspect 114:1210–1214. https://doi.org/10.1289/ehp.9019
Enright P (2010) FEV1 and FVC repeatability goals when performing spirometry. Prim Care Respir J 19:194. https://doi.org/10.4104/pcrj.2010.00031
Escher BI, Hermens JL (2004) Internal exposure: linking bioavailability to effects. Environ Sci Technol 38:455a–462. https://doi.org/10.1021/es0406740. a
Gallego E, Roca FX, Guardino X, Rosell MG (2008) Indoor and outdoor BTX levels in Barcelona City metropolitan area and Catalan rural areas. J Environ Sci (China) 20:1063–1069. https://doi.org/10.1016/s1001-0742(08)62150-6
Goel A, Goyal M, Singh R, Verma N, Tiwari S (2015) Diurnal Variation in Peak Expiratory Flow and Forced Expiratory Volume. J Clin Diagn Res 9. https://doi.org/10.7860/jcdr/2015/15156.6661. Cc05-07
Guo C, Zhang Z, Lau AKH, Lin CQ, Chuang YC, Chan J, Jiang WK, Tam T, Yeoh E-K, Chan T-C, Chang L-Y, Lao XQ (2018) Effect of long-term exposure to fine particulate matter on lung function decline and risk of chronic obstructive pulmonary disease in Taiwan: a longitudinal, cohort study. Lancet Planet Health 2:e114–e125. https://doi.org/10.1016/s2542-5196(18)30028-7
Jackson AJ, Miller BJ (2020) Meta-analysis of total and differential white blood cell counts in schizophrenia. Acta Psychiatr Scand 142:18–26. https://doi.org/10.1111/acps.13140
Jain RB (2017) Detection rates, trends in and factors affecting observed levels of selected volatile organic compounds in blood among US adolescents and adults. Environ Toxicol Pharmacol 56:21–28. https://doi.org/10.1016/j.etap.2017.08.031
Jung KH, Artigas F, Shin JY (2011) Personal, indoor, and outdoor exposure to VOCs in the immediate vicinity of a local airport. Environ Monit Assess 173:555–567. https://doi.org/10.1007/s10661-010-1404-9
Kwon JW, Park HW, Kim WJ, Kim MG, Lee SJ (2018) Exposure to volatile organic compounds and airway inflammation. Environ Health 17:65. https://doi.org/10.1186/s12940-018-0410-1
Latif MT, Abd Hamid HH, Ahamad F, Khan MF, Mohd Nadzir MS, Othman M, Sahani M, Abdul Wahab MI, Mohamad N, Uning R, Poh SC, Fadzil MF, Sentian J, Tahir NM (2019) BTEX compositions and its potential health impacts in Malaysia. Chemosphere 237:124451. https://doi.org/10.1016/j.chemosphere.2019.124451
Li N, Zhai Z, Zheng Y, Lin S, Deng Y, Xiang G, Yao J, Xiang D, Wang S, Yang P, Yang S, Xu P, Wu Y, Hu J, Dai Z, Wang M (2021) Association of 13 Occupational Carcinogens in Patients With Cancer, Individually and Collectively, 1990–2017. JAMA Netw Open 4:e2037530. https://doi.org/10.1001/jamanetworkopen.2020.37530
McNally K, Sams C, Loizou GD, Jones K (2017) Evidence for non-linear metabolism at low benzene exposures? A reanalysis of data. Chem Biol Interact 278:256–268. https://doi.org/10.1016/j.cbi.2017.09.002
Mendes A, Papoila AL, Carreiro-Martins P, Bonassi S, Caires I, Palmeiro T, Aguiar L, Pereira C, Neves P, Mendes D, Botelho MA, Neuparth N, Teixeira JP (2016) The impact of indoor air quality and contaminants on respiratory health of older people living in long-term care residences in Porto. Age Ageing 45:136–142. https://doi.org/10.1093/ageing/afv157
Miller MR, Hankinson J, Brusasco V, Burgos F, Casaburi R, Coates A, Crapo R, Enright P, van der Grinten CP, Gustafsson P, Jensen R, Johnson DC, MacIntyre N, McKay R, Navajas D, Pedersen OF, Pellegrino R, Viegi G, Wanger J (2005) Standardisation of spirometry. Eur Respir J 26:319–338. https://doi.org/10.1183/09031936.05.00034805
Mirza S, Clay RD, Koslow MA, Scanlon PD (2018) COPD Guidelines: A Review of the 2018 GOLD Report. Mayo Clin Proc 93:1488–1502. https://doi.org/10.1016/j.mayocp.2018.05.026
Mukherjee AK, Chattopadhyay BP, Roy SK, Das S, Mazumdar D, Roy M, Chakraborty R, Yadav A (2016) Work-exposure to PM(10) and aromatic volatile organic compounds, excretion of urinary biomarkers and effect on the pulmonary function and heme-metabolism: A study of petrol pump workers and traffic police personnel in Kolkata City, India. J Environ Sci Health A Tox Hazard Subst Environ Eng 51:135–149. https://doi.org/10.1080/10934529.2015.1087740
Nise G, Attewell R, Skerfving S, Orbaek P (1989) Elimination of toluene from venous blood and adipose tissue after occupational exposure. Br J Ind Med 46:407–411. https://doi.org/10.1136/oem.46.6.407
Ojo TO, Onayade AA, Akinyemi PA, Adesanmi AJ (2017) Environmental Working Conditions, Lung Function and Total Serum Bile Acids of Spray Painters Exposed to Organic Solvents in Ile-Ife, Nigeria. J Health Pollut 7:2–10. https://doi.org/10.5696/2156-9614-7-13.2
Schneider P, Gebefügi I, Richter K, Wölke GW, Schneille J, Wichmann HE, Heinrich J (2001) Indoor and outdoor BTX levels in German cities. Sci Total Environ 267:41–51. https://doi.org/10.1016/s0048-9697(00)00766-x
Sexton K, Adgate JL, Church TR, Ashley DL, Needham LL, Ramachandran G, Fredrickson AL, Ryan AD (2005) Children's exposure to volatile organic compounds as determined by longitudinal measurements in blood. Environ Health Perspect 113:342–349. https://doi.org/10.1289/ehp.7412
Sin DD, Wu L, Man SF (2005) The relationship between reduced lung function and cardiovascular mortality: a population-based study and a systematic review of the literature. Chest 127:1952–1959. https://doi.org/10.1378/chest.127.6.1952
Singh A, Kesavachandran CN, Kamal R, Bihari V, Ansari A, Azeez PA, Saxena PN, Ks AK, Khan AH (2017) Indoor air pollution and its association with poor lung function, microalbuminuria and variations in blood pressure among kitchen workers in India: a cross-sectional study. Environ Health 16:33. https://doi.org/10.1186/s12940-017-0243-3
Sproston NR, Ashworth JJ (2018) Role of C-Reactive Protein at Sites of Inflammation and Infection. Front Immunol 9:754. https://doi.org/10.3389/fimmu.2018.00754
Su FC, Mukherjee B, Batterman S (2011) Trends of VOC exposures among a nationally representative sample: Analysis of the NHANES 1988 through 2004 data sets. Atmos Environ (1994) 45: 4858–4867. https://doi.org/10.1016/j.atmosenv.2011.06.016
Sun Y, Chen C, Mustieles V, Wang L, Zhang Y, Wang YX, Messerlian C (2021) Association of Blood Trihalomethane Concentrations with Risk of All-Cause and Cause-Specific Mortality in U.S. Adults: A Prospective Cohort Study. Environ Sci Technol 55:9043–9051. https://doi.org/10.1021/acs.est.1c00862
Vaughan Watson C, Naik S, Lewin M, Ragin-Wilson A, Irvin-Barnwell E (2021) Associations between select blood VOCs and hematological measures in NHANES 2005–2010. J Expo Sci Environ Epidemiol 31:366–376. https://doi.org/10.1038/s41370-019-0192-8
Wang F, Li C, Liu W, Jin Y (2012) Effect of exposure to volatile organic compounds (VOCs) on airway inflammatory response in mice. J Toxicol Sci 37:739–748. https://doi.org/10.2131/jts.37.739
Weichenthal S, Kulka R, Bélisle P, Joseph L, Dubeau A, Martin C, Wang D, Dales R (2012) Personal exposure to specific volatile organic compounds and acute changes in lung function and heart rate variability among urban cyclists. Environ Res 118:118–123. https://doi.org/10.1016/j.envres.2012.06.005
Yoon HI, Hong YC, Cho SH, Kim H, Kim YH, Sohn JR, Kwon M, Park SH, Cho MH, Cheong HK (2010) Exposure to volatile organic compounds and loss of pulmonary function in the elderly. Eur Respir J 36:1270–1276. https://doi.org/10.1183/09031936.00153509
Yun BH, Guo J, Turesky RJ (2018) Formalin-Fixed Paraffin-Embedded Tissues-An Untapped Biospecimen for Biomonitoring DNA Adducts by Mass Spectrometry. Toxics. 6https://doi.org/10.3390/toxics6020030
Zalel A, Yuval, Broday DM (2008) Revealing source signatures in ambient BTEX concentrations. Environ Pollut 156:553–562. https://doi.org/10.1016/j.envpol.2008.01.016
Zhu B, Feng X, Jiang C, Mi S, Yang L, Zhao Z, Zhang Y, Zhang L (2021) Correlation between white blood cell count at admission and mortality in COVID-19 patients: a retrospective study. BMC Infect Dis 21:574. https://doi.org/10.1186/s12879-021-06277-3

Supplementaryinformation.docx

Download PDF

Journal Publication

published 02 May, 2023

Read the published version in Environmental Science and Pollution Research →

Editorial decision: Major Revision
13 Feb, 2023
Reviewers agreed at journal
05 Jan, 2023
Reviewers invited by journal
04 Jan, 2023
Editor invited by journal
02 Jan, 2023
Editor assigned by journal
21 Dec, 2022
First submitted to journal
19 Dec, 2022

You are reading this latest preprint version

Low dose blood BTEX are associated with pulmonary function through changes in inflammatory markers among US adults: NHANES 2007-2012

Status:

Journal Publication

Version 1

Abstract

Figures

Background

Methodology

Study population

VOCs in the whole-blood samples

Spirometry test

Blood inflammatory markers

Covariates

Ethics

Statistical analyses

Results

Characteristics of study participants and BTEX distributions

Association between blood BTEX and lung function

Association between VOC and inflammatory markers

Association between inflammatory markers and lung function

Mediating effects of inflammatory markers between VOC and lung function

Discussion

Conclusions

Abbreviations

Declarations

References

Supplementary Files

Status:

Journal Publication

Version 1