Associations of furan exposure with the prevalence and mortality in asthma: A prospective cohort study

doi:10.21203/rs.3.rs-4867643/v1

Download PDF

Research Article

Associations of furan exposure with the prevalence and mortality in asthma: A prospective cohort study

https://doi.org/10.21203/rs.3.rs-4867643/v1

This work is licensed under a CC BY 4.0 License

You are reading this latest preprint version

Background

Few studies have explored the role of furan exposure plays in aggravating asthma.

Objective

To access the relationship of furan exposure to asthma.

Methods

This is a prospective cohort study, involving 7,047 adults over 20 years old from the National Health and Nutrition Examination Survey 2007–2012. Blood furan levels were employed to quantify furan exposure. Multivariate survey-weighted regressions were utilized to analyze the associations between furan exposure, the prevalence of asthma. Mediation analyses for furan exposure and asthma prevalence were conducted. Multiple Cox regression was employed to evaluate the association between furan exposure and asthma prognosis.

Results

Asthmatics have higher blood furan levels than non-asthmatics (P < 0.001). After adjusting for covariates, log10-transformed blood furan levels (LBFL) were independently associated with an increased risk of asthma prevalence (adjusted odds ratio [aOR] = 2.40, 95% confidence interval [CI] = 1.21–4.78, P = 0.014). There was a significant positive linear relationship between LBFL and risk of asthma (P for linear = 0.0003). In mediation analyses, FEV₁ was identified as mediators in the above relationships, with mediated proportions of 32.73%. Longitudinally, multiple Cox regression analysis demonstrated that LBFL were positively correlated with respiratory mortality in asthma (HR = 27.88, 95% CI = 4.19-185.69, P < 0.0001).

Conclusions

Exposure to furan revealed a positive association with greater odds of asthma, and lung function was identified as an important mediator. An elevated LBFL also is associated with an increased health care use, worse HQL, and prognosis of asthma.

Asthma is a chronic inflammatory respiratory disease marked by airway hyperresponsiveness and remodeling, leading to dyspnea and wheezing. ¹. The multifaceted risk factors associated with asthma encompass environmental variables such as air pollution², occupational dust³, smoking⁴, and environmental chemicals exposure⁵. Great deal of research revealed that exposure to environmental chemicals associated with the rising frequency of asthma, exacerbation and increased healthcare use caused by asthma^6–8. Thus, it is imperative to investigate the relationship between environmental pollutants and asthma in order to support public health and environmental protection policies so lowering the burden of asthma.

Furan (C4H4O) is a volatile and carcinogenic heterocyclic compound in various thermally processed foods, such as cooking, baking, canning, and sterilization at 150 to 200°C⁹. Previous studies have reported that furan can penetrate biofilms. For instance, the intestine and lungs and be rapidly absorbed and extensively metabolized¹⁰. Furan can enter the human body through ingestion and inhalation when preparing food¹⁰. Tǎbǎran et al. found that lung club cells of mice exposed to furan showed necrosis with airway denudation, regeneration and partial repair¹¹. Meanwhile, a recent study has showed that the link between blood furan and chronic obstructive pulmonary disease (COPD) is mediated by inflammation¹². To date, the relationship between furan and asthma has never been investigated.

In this study, we aimed to explore the association between furan exposure, prevalence and mortality in asthma among adults. Furthermore, we investigated whether furan exposure is associated with lung function, participants-reported health outcomes, and participants-reported health-related quality of life (HQL).

Study design and participants

NHANES is a population-based survey carried out annually by the National Center for Health Statistics (NCHS) since 1999, a division of the Centers for Disease Control and Prevention (CDC). The NHANES included interviews, physical exams, and laboratory tests. The study protocol received approval from the NCHS Research Ethics Board, and all participants gave their written consent. Details of NHANES has been documented elsewhere (https://www.cdc.gov/nchs/nhanes/index.htm)^13,14.

Ultimately, we obtained the data from the NHANES 2007–2012. These cycles were chosen because participants underwent the spirometry test in these waves. The total sample size of the NHANES 2007–2012 was 30,442. Individuals under 18 years old (n = 12,729), those missing data for asthma (n = 17), and those without data for blood furan (n = 10,649) were excluded after merging databases. Thus, 7,047 individuals satisfied the inclusion criteria for final analysis (Fig. 1).

Exposure Definitions

Serum blood furan levels were measured using a protocol outlined in a previously published study¹⁵. In brief, the approach detects specific volatile organic compounds (VOCs) in whole blood at low parts-per-trillion levels. Since non-occupationally exposed individuals have blood VOC concentrations within this range, this method is useful for measuring levels. Laboratory measurements were carried out using centralized quality assurance and control techniques. The minimum detectable concentration of blood furan was 0.0250 (unit: ng/mL).

Log10 transformations were carried out where required to achieve a normal data distribution of blood furan level. In addition, participants were assigned into two groups according to the median of log10-transformed blood furan levels (LBFL): LBFL ≤ -1.752 ng/mL is assigned to furan ^Low group while LBFL > -1.752 ng/mL is assigned to furan ^High group.

Primary outcome

There are two primary outcomes in this study: the prevalence of asthma and respiratory mortality in asthma.

Prevalence of asthma

Asthma is defined by a self-reported positive answer to the questionnaire: “Has a doctor or other health professional ever told you that you had asthma?”.

Respiratory mortality in asthma

Another primary outcome was respiratory mortality in asthmatics, including chronic lower respiratory diseases (J40 - J47), influenza and pneumonia (J10 - J18). The information about mortality status and cause of death was obtained by NHANES-linked National Death Index (NDI) on December 31, 2019.

Secondary outcome

The secondary outcomes of this study were lung function (absolute value of FEV₁[L], FEV₁% predicted, absolute value of FVC[L], FVC % predicted, obstructive, and restrictive spirometry pattern) and participants-reported health outcomes, including asthma-related health outcomes (asthma attack, emergency room [ER] visits for asthma), respiratory symptoms (wheezing, phlegm production, dry cough, and exertional dyspnea), and participants-reported health-related quality of life (HQL) (days when physical health was poor, days when mental health was poor, inactive days because of poor health).

Lung function

Lung function was assessed in participants ranging in age from 6 to 79 years. Detailed information about spirometry protocol and procedure is described elsewhere¹⁶. Spirometry testing was performed at the NHANES Mobile Examination Center following a standardized protocol, measuring forced expiratory volume in the first second (FEV₁) and forced vital capacity (FVC). Predicted spirometry values were determined using normative reference equations from NHANES III data¹⁷. Percent (%) predicted FEV₁, FVC, and FEV₁/FVC were calculated using Global Lung Initiative equations based on the participant’s age, gender, height, and race/ethnicity¹⁷. Detailed information on spirometry administration is available in the NHANES spirometry procedures manual¹⁸. Moreover, obstructive spirometry was determined as FEV₁/FVC < 0.70, whereas restrictive spirometry was characterized as FVC < 80% of predicted and FEV₁/FVC ≥ 0.70¹⁹.

Asthma-related health outcomes

Asthma-related health outcomes included asthma attack and ER visits for asthma in the past year. An asthma attack was identified by answering “yes” to the question. Detailed questions are described in Luyster FS et al’s study²⁰.

Respiratory symptoms

Respiratory symptoms consisted of wheezing, phlegm production, cough, and exertional dyspnea. Respiratory symptoms listed above were identified by answering “yes” to the question. Specific questions are described in Wu et al’s study¹⁴.

HQL

HQL was assessed by questions about mental health and physical health in the past 30 days. Specific questions are described in Luyster FS et al’s study²⁰.

Demographic and Clinical Data

The data of demographic and clinical on study participants was collected using questionnaires and laboratory tests, consisting of age, gender (male/female), race/ethnicity, educational attainment, marital status, poverty-income ratio (PIR) level, body mass index (BMI), serum cotinine (ng/ml), alcohol drinking, chronic disease (cancer/malignancy, coronary heart disease, diabetes mellitus, and hypertension) and count of white blood cell (WBC) count.

Race/ethnicity was classified by Mexican American, other Hispanic, non-Hispanic black and white, and others race. Educational attainment was categorized as college or more, high school, middle school or lower. Marital status was classified as married/partnered, divorced/separated, or widowed/single. The PIR was ranged from 0 to 5, calculated based on guidelines and adjusted for family size, year, and state. A value of zero means no income and five means income at least five times the federal poverty level. As a metabolite of nicotine, serum cotinine is utilized to measure recent engagement in tobacco use²¹. Alcohol drinking was defined as how many drinks did you have over the past 12 months on average annually, which classified as ≥ 4 drinks/day, 1–3 drinks/day, and never drinking. Chronic disease included bronchitis, emphysema, cancer/malignancy, coronary heart disease, diabetes mellitus, and hypertension defined as self-reported physician diagnosis.

Statistical analysis

Categorical variables are reported as frequencies and proportions, whereas continuous variables are expressed as means ± standard deviation (SD) or median (Quartiles [Q] 1, 3) based on the data distribution. The difference between two groups for each variable was analyzed appropriately using t-test or Mann Whitney U test for continuous variables and χ² tests for categorical variables.

To explore the link between LBFL and asthma prevalence, we conducted an RCS analysis to evaluated the dose-response association. Multiple logistic regression analyses were used to assess this association, taking into account various covariates. Initially, we employed a crude model without adjusting any covariates. Next, we developed Model I, which adjusted for age, gender, ethnicity, marital status, and educational attainment. Finally, Model II comprised the same variables as Model I, with additional adjustments for BMI, PIR, serum cotinine levels, alcohol drinking in the past year, and diabetes mellitus.

In addition, we performed mediation analyses to determine if inflammation or lung function (represented by FEV₁) mediated the relationship between LBFL and asthma prevalence, adjusting for confounders. In brief, mediation analysis was conducted using the R package "mediation" (version 4.5.0) with 1,000 simulations²². We estimated the average direct effect (ADE), average causal mediation effect (ACME), and average total effect using a model-based inference approach. The proportion of the exposure's effect on the outcome that was mediated was calculated by dividing the ACME by the total effect (ACME + ADE)^22,23. Additionally, we conducted stratified analyses to explore potential moderating effects of age, BMI, gender, ethnicity, marital status, and educational attainment.

Multiple Cox regression analyses were implemented to calculate adjusted hazard ratios (HRs) and 95% CIs for furan exposure and respiratory mortality in patients with asthma. Initially, the crude model was employed. Subsequently, Model I included adjustments for age, PIR, marital status, educational attainment, while Model II adjusted for the same variables as Model I and for serum cotinine, past-year alcohol drinking, coronary heart disease, and hypertension.

All statistical analyses were conducted according to the CDS guidelines (https://wwwn.cdc.gov/nchs/nhanes/tutorials/default.aspx). The statistical analyses included sample weights to account for the complex multi-stage stratified survey design used in NHANES. All statistical analyses were executed using R (version 4.3.1). Statistical significance was assessed at a two-sided P value < 0.05.

Baseline characteristics of participants

A total of 7,047 participants were enrolled in the present study from the NHANES data from 2007 to 2012. Of the study participants, 951 participants diagnosed as asthma and 6,096 participants without asthma. Table 1 listed the sociodemographic and clinical characteristics of the subjects grouped by asthma. As presented in Table 1, asthmatics were more likely to be younger, more female, lower PIR, higher BMI, high serum cotinine, worse lung function, and more diabetes mellitus compared with participants without asthma (all P < 0.05). However, there were no significant differences in educational attainment and alcohol drinking, cancer, coronary heart disease, and high blood pressure were found (all P > 0.05). With respect to blood furan level, there is a significant increased LBFL for asthma patients (21.96% vs. 14.37%, P < 0.001) compare with participants without asthma.

Table 1

Demographics and clinical characteristics of participants group by asthma.
Variable	All (n = 7047)	Healthy controls (n = 6096)	Asthma (n = 951)	P-value
Age	46.00(33.00,59.00)	46.00(34.00,59.00)	44.00(30.00,57.00)	< 0.001
Gender				< 0.001
Female	3576(51.64)	3007(50.27)	569(60.12)
Male	3471(48.36)	3089(49.73)	382(39.88)
Ethnicity				< 0.0001
Mexican American	1123(8.43)	1045(9.11)	78(4.25)
Non-Hispanic Black	1445(10.96)	1213(10.58)	232(13.32)
Non-Hispanic White	3083(67.56)	2627(67.17)	456(69.96)
Other Hispanic	767(5.79)	660(5.76)	107(6.01)
Other Race - Including Multi-Racial	629(7.25)	551(7.38)	78(6.46)
Educational attainment				0.7
College or more	3482(58.65)	2973(58.48)	509(60.36)
High school	1581(22.42)	1368(22.59)	213(21.62)
Middle school or lower	1975(18.78)	1746(18.94)	229(18.01)
Marital status				0.001
Divorced	745(10.19)	618(9.92)	127(11.87)
Living with partner	542(7.88)	460(7.80)	82(8.34)
Married	3663(55.63)	3235(56.77)	428(48.68)
Never married	1239(18.04)	1032(17.15)	207(23.54)
Separated	258(2.64)	220(2.60)	38(2.93)
Widowed	596(5.59)	528(5.75)	68(4.64)
BMI	27.50(24.01,32.00)	27.31(23.94,31.69)	28.59(24.47,33.98)	0.001
PIR	2.91(1.43,5.00)	2.95(1.48,5.00)	2.33(1.20,4.67)	0.01
Serum cotinine	0.04(0.01,4.98)	0.04(0.01, 2.65)	0.06(0.02,88.50)	0.01
Past-year alcohol drinking				0.92
≥ 4 drinks/day	1049(15.68)	912(19.05)	137(19.64)
1–3 drinks/day	2800(45.89)	2420(56.03)	380(55.81)
Never	1762(20.37)	1532(24.92)	230(24.55)
Spirometry
FEV₁ (L)	3.18(2.57,3.84)	3.21(2.61,3.87)	3.03(2.31,3.56)	< 0.0001
FEV₁% predicted	98.11(88.51,107.04)	98.78(90.08,107.60)	92.51(80.34,102.76)	< 0.0001
FVC (L)	4.05(3.33,4.87)	4.06(3.36,4.91)	3.92(3.18,4.64)	0.003
FVC% predicted	101.24(92.16,110.26)	101.67(92.87,110.68)	98.57(88.03,107.73)	< 0.0001
FEV₁/FVC	3.00(2.11,3.85)	3.07(2.20,3.90)	2.57(1.72,3.46)	< 0.0001
Cancer				0.08
No	6403(90.78)	5555(91.26)	848(88.78)
Yes	638(9.07)	537(8.74)	101(11.22)
Coronary heart disease				0.45
No	6764(96.93)	5851(97.17)	913(97.63)
Yes	252(2.76)	219(2.83)	33(2.37)
Diabetes mellitus				< 0.001
Borderline	141(1.93)	122(2.01)	19(1.41)
No	6019(89.20)	5253(89.74)	766(86.08)
Yes	883(8.84)	717(8.25)	166(12.51)
HBP				0.06
No	4564(69.83)	3999(70.50)	565(66.08)
Yes	2475(30.09)	2090(29.50)	385(33.92)
LBFL				< 0.0001
≤ -1.752 ng/mL	6003(84.57)	5256(85.63)	747(78.04)
> -1.752 ng/mL	1044(15.43)	840(14.37)	204(21.96)
Abbreviation: BMI, body mass index; PIR, poverty-income ratio; FEV₁, forced expiratory volume in 1 s; FVC, forced vital capacity; HBP, high blood pressure; LBFL, Log10-transformed blood furan levels.

In addition, 6003 (85.2%) participants were divided into the furan ^High group and 1044 (14.8%) were divided into the furan ^Low group as depicted in Table E1. The patients in the furan ^High group were younger, had an elevated proportion were male, lower BMI, PIR, higher serum cotinine, worse lung function including FEV₁, FEV₁% predicted, FVC, FVC% predicted, and FEV₁/FVC, and more participants with asthma (all P < 0.05).

Primary Outcomes

Furan exposure and asthma prevalence

Compared to the furan ^Low group, participants in the furan ^High group showed a higher risk of asthma in all models (crude model: odds ratio [OR] = 1.68, 95% CI = 1.31–2.15, P < 0.001; model 1: adjusted OR [aOR] = 1.57 95% CI = 1.20–2.06, P = 0.002; model 2: aOR = 1.75, 95% CI = 1.14–2.68, P = 0.012, respectively) (Table 2). In addition, LBFL as continuous variable also suggested a positive association with the prevalence of asthma (crude model: OR = 1.99, 95% CI = 1.39–2.85, P < 0.001; model 1: aOR = 1.86 95% CI = 1.25–2.78, P = 0.003; model 2: aOR = 2.40, 95% CI = 1.21–4.78, P = 0.014) (Table 2).

Table 2

Weighted univariate and multivariate logistic regression analysis of LBFL for the risk of asthma
	Crude Model		Model 1		Model 2
	OR (95% CI)	P-value	OR (95% CI)	P-value	OR (95% CI)	P-value
Log10-transformed blood furan	1.99(1.39,2.85)	< 0.001	1.86(1.25,2.78)	0.003	2.40(1.21,4.78)	0.014
Low (≤ -1.752 ng/mL)	Reference		Reference		Reference
High (> -1.752 ng/mL)	1.68(1.31,2.15)	< 0.001	1.57(1.20,2.06)	0.002	1.75(1.14,2.68)	0.012
Abbreviation: OR, odds ratio; CI, confidence interval; LBFL, Log10-transformed blood furan levels. Crude model lacked adjustments for covariates. Model 1 adjusted for age, gender, ethnicity, marital status, educational attainment. Model 2 adjusted for age, gender, ethnicity, marital status, educational attainment, BMI, PIR, serum cotinine, past-year alcohol drinking, and diabetes mellitus.

Moreover, the RCS model exhibited that LBFL were linearly related to the prevalence of asthma adjusting for age, gender, BMI, ethnicity, marital status, educational attainment, PIR, serum cotinine, past-year alcohol drinking, and diabetes mellitus (P for overall = 0.0003; P for non-linear = 0.1311) (Fig. 2). And ROC analysis showed that the area under roc curve (AUC) value in the fully adjusted model was 0.655 (P < 0.05) (Fig. 3).

Subgroup analyses of factors impacting the association of asthma prevalence with furan exposure

Furthermore, subgroup analysis was employed to further investigate the association between furan exposure and the prevalence of asthma stratified by age, gender, BMI, ethnicity, marital status, and educational attainment in Fig. 4 (Detail data as shown in Table E2). In subgroups analysis, a statistically significant link between exposure to furan and asthma prevalence was only observed in the less than 60 years age group, in those with a BMI between 25-29.9 kg/m², in females, in those were non-Hispanic white and other Hispanic, in married, and in those with middle school or low and college or more (all P < 0.05). In addition, age, gender, BMI, ethnicity, marital status, and educational attainment did not have significant interactions with the furan exposure (P for interaction > 0.05).

Mediation analysis of lung function involved in the effects of furan on asthma

Furthermore, we analyzed the associations of LBFL with inflammatory cells and lung function in adults. There was a substantial positive correlation between LBFL and lymphocytes (β = 0.346, 95% CI = 0.207–0.485, P < 0.0001), neutrophils (β = 1.005, 95% CI = 0.559–1.451, P < 0.0001), and FEV₁ (β = -0.351, 95% CI = -0.500 - -0.202, P < 0.0001), as shown in Table E3. However, no association was observed between furan exposure with monocytes, eosinophils, and basophils (all P > 0.05). Additionally, asthma exhibited a positive association with eosinophils and FEV₁ (β = 0.041, 95% CI = 0.025–0.056, P < 0.0001; β = -0.222, 95% CI = -0.293 - -0.151, P < 0.0001; respectively) (Table E3). Nevertheless, there was no association between asthma with lymphocytes, monocytes, neutrophils and basophils (all P > 0.05).

As Fig. 5 displayed, mediation analyses indicated that the absolute value of FEV₁ partially mediated the association between furan exposure and asthma prevalence, with mediated proportions of 30.93% (P < 0.05), as depicted in Fig. 5.

Furan exposure and respiratory mortality of asthma

Two patients were eliminated due to lack of follow-up among the 951 asthmatic patients (Fig. 1). Additionally, 18 patients died from respiratory illnesses, giving weighted respiratory mortality of 1.90%. Table 3 illustrates the positive correlation found in both models between LBFL and respiratory mortality of asthma. Table 3 shows that LBFL and respiratory mortality in asthma patients were positively correlated, with the Cox regression analysis controlling for age, PIR, marital status, educational attainment, serum cotinine, past-year alcohol drinking, coronary heart disease, and hypertension (hazard ratio [HR] = 27.88, 95% CI = 4.19–185.69, P < 0.0001).

Table 3

Weighted univariate and multivariate cox regression analysis of LBFL and respiratory mortality of asthma.
	Crude Model		Model 1		Model 2
	HR (95% CI)	P-value	HR (95% CI)	P-value	HR (95% CI)	P-value
Log10-transformed blood furan	3.65(1.22, 10.93)	0.021	31.33(7.08, 138.74)	< 0.0001	27.88(4.19, 185.69)	< 0.001
Low (≤ -1.752 ng/mL)	Reference		Reference		Reference
High (> -1.752 ng/mL)	2.52(1.02, 6.22)	0.045	15.07(5.33, 42.61)	< 0.0001	33.88(16.86, 68.10)	< 0.001
Abbreviation: HR, hazard ratio; CI, confidence interval; LBFL, Log10-transformed blood furan levels. Crude model lacked adjustments for covariates. Model 1 adjusted for age, PIR, marital status, educational attainment. Model 2 adjusted for age, PIR, marital status, educational attainment, serum cotinine, past-year alcohol drinking, coronary heart disease, and hypertension.

Secondary outcome

Lung function

Compared with furan ^Low group, those in the furan ^High group had worse FEV₁, FEV₁% predicted, and FVC % predicted (3.08 [2.48,3.64] vs. 3.20 [2.59,3.86] L, P < 0.0001; 91.90 [81.61,101.00] vs. 98.94 [90.22,107.83] %, P < 0.0001; 99.99 [90.15,108.09] vs. 101.52 [92.73,110.49], P = 0.01; respectively) and a higher proportion of obstructive spirometry pattern (26.63% vs. 11.29%, P < 0.0001) (Table E6). But no significant difference was observed in restrictive spirometry pattern (Table E6).

After multivariate adjustment, a one-unit increase in LBFL was significantly associated with 0.35 decrements in absolute value of FEV₁ (β = -0.35, 95% CI = -0.50 - -0.20. P < 0.0001), and with 10.59 decrements in percent predicted FEV₁ (β = -10.59, 95% CI = -14.56 - -6.63. P < 0.0001) after adjusting for age, gender, BMI, ethnicity, marital status, educational attainment, PIR, serum cotinine, bronchitis, emphysema, diabetes mellitus. In addition, after adjusting for all confounders in the master analytical model 2 above, a non-linear correlation was observed between furan exposure and absolute value of FEV₁ (P for overall < 0.0001; P for non-linear = 0.0156) in Figure E1. A linear correlation was observed between furan exposure and percent predicted FEV₁, absolute value of FVC, and percent predicted FVC (P for overall < 0.05; P for non-linear > 0.05) (Figure E1).

Consequently, LBFL was associated with an elevated risk of an obstructive spirometry pattern (aOR = 5.02, 95% CI = 2.97–8.47, P < 0.0001), but not with restrictive spirometry pattern (aOR = 0.84, 95% CI = 0.45–1.56, P = 0.57) (Table 4).

Table 4

Weighted univariate and multivariate regression models of LBFL predicting lung function and patient-reported health outcomes.
Variables	Crude Model		Model 1		Model 2
Variables	β/OR (95% CI)	P-value	β/OR (95% CI)	P-value	β/OR (95% CI)	P-value
Lung function
FEV₁ (L)	-0.26(-0.35, -0.16)	< 0.0001	-0.33(-0.43, -0.24)	< 0.0001	-0.35(-0.50, -0.20)	< 0.0001
FEV₁% predicted	-12.43(-14.76, -10.10)	< 0.0001	-11.01(-13.61, -8.42)	< 0.0001	-10.59(-14.56, -6.63)	< 0.0001
FVC (L)	0.03(-0.11, 0.17)	0.67	-0.08(-0.20, 0.03)	0.14	-0.14(-0.32, 0.04)	0.13
FVC % predicted	-4.16(-6.65, -1.67)	0.002	-3( -5.606, -0.393)	0.03	-3.26( -7.10, 0.59)	0.09
Obstructive	4.82(3.71, 6.26)	< 0.0001	6.17(4.57, 8.32)	< 0.0001	5.02(2.97, 8.47)	< 0.0001
Restrictive	0.93(0.59, 1.49)	0.93	0.86(0.53,1.38)	0.86	0.84(0.45, 1.56)	0.57
Asthma-related health outcomes
Asthma attack in past yr	1.84(0.90, 3.76)	0.09	2.00(0.94,4.27)	0.07	3.25(0.82, 12.80)	0.09
ER visits for asthma in past yr	1.03(0.35, 3.05)	0.96	1.02(0.31, 3.38)	0.97	2.11(0.07, 59.91)	0.65
Respiratory symptoms
Wheezing	4.88(3.68, 6.47)	< 0.0001	4.35(3.23,5.86)	< 0.0001	4.48(2.72, 7.40)	< 0.0001
Phlegm	5.37(3.52, 8.20)	< 0.0001	5.67(3.67, 8.75)	< 0.0001	7.16(3.09, 16.61)	< 0.0001
Cough	7.47(4.75, 11.77)	< 0.0001	6.60(4.24, 10.28)	< 0.0001	7.73(3.58, 16.71)	< 0.0001
Exertional dyspnea	2.58(1.82, 3.66)	< 0.0001	2.45(1.71, 3.51)	< 0.0001	2.90(1.58, 5.30)	0.001
HQL
Days physical health was not good	3.18(1.86, 4.51)	< 0.0001	2.85(1.51, 4.20)	< 0.001	2.61(0.88, 4.33)	0.004
Days mental health was not good	4.05(2.71, 5.39)	< 0.0001	3.37(2.05, 4.69)	< 0.0001	2.23(0.38, 4.09)	0.020
Inactive days due to health	1.85(0.96, 2.73)	< 0.001	1.61(0.74,2.48)	< 0.001	1.04(-0.17, 2.26)	0.09
Abbreviation: ER, emergency room; OR, odds ratio; CI, confidence interval; FEV1, forced expiratory volume in 1 s; FVC, forced vital capacity; LBFL, Log10-transformed blood furan levels; HQL, health-related quality of life. Crude model lacked adjustments for covariates. Model 1 adjusted for age, gender, ethnicity, marital status, educational attainment. Model 1 adjusted for age, gender, ethnicity, marital status, educational attainment, BMI, PIR, serum cotinine, bronchitis, emphysema, diabetes mellitus.

Asthma-related health outcomes

No significant difference was found in asthma attack and ER visits for asthma in past year between furan ^Low group and furan ^High group (all P > 0.05) (Table 4 and E6).

Respiratory symptoms

Compared with participants in the furan ^Low group, those in the furan ^High group had a higher proportion of wheezing (27.59% vs. 10.40%, P < 0.0001), phlegm (19.92% vs. 6.40%, P < 0.0001), cough (26.78% vs. 7.35%, P < 0.0001), and exertional dyspnea (44.65% vs. 28.09%, P < 0.0001) (Table E6).

Similarly, LBFL was associated with higher odds of wheezing (aOR = 4.48, 95% CI = 2.72–7.40, P < 0.0001), phlegm (aOR = 7.16, 95% CI = 3.09–16.61, P < 0.0001), cough (aOR = 7.73, 95% CI = 3.58–16.71, P < 0.0001), and exertional dyspnea (aOR = 2.90, 95% CI = 1.58–5.30, P < 0.0001) in the past 12 months, adjusting for age, gender, BMI, ethnicity, marital status, educational attainment, PIR, serum cotinine, bronchitis, emphysema, diabetes mellitus.

HQL

Participants in the furan ^High group had worse HQL including days when physical health was poor, days when mental health was poor, and inactive days due to poor health (5.60 ± 0.39 vs. 3.23 ± 0.12, P < 0.0001; 6.87 ± 0.46 vs. 3.64 ± 0.10, P < 0.0001; 3.13 ± 0.33 vs. 1.60 ± 0.09, P < 0.0001; respectively), as compared to participants in the furan ^Low group.

Furthermore, LBFL was also associated with more days with poor physical or poor mental health but not of inactive days due to poor health during after adjusted for age, gender, BMI, ethnicity, marital status, educational attainment, PIR, serum cotinine, bronchitis, emphysema, and diabetes mellitus (Table 4).

To our knowledge, we’re first to sought to examine systematic relationship between furan exposure, asthma prevalence, and asthma-related respiratory mortality in this prospective cohort study. We found that a higher LBFL was associated with greater odds of the prevalence and mortality of asthma. Lung function, particularly FEV₁, was identified as a mediator in the relationship between LBFL and asthma prevalence. Additionally, our results indicated that elevated LBFL were associated with worse lung function, increased risk of respiratory symptoms, and lower risk of self-reported HQL after adjusting for covariates. The results indicated that furan as a common environmental pollution may be a notable risk factor for asthma, which highlights the importance of considering environmental exposures in public health and deepens our understanding of the factors contributing to asthma.

Accurately identifying environmental chemical exposure is crucial for understanding its potential impact on lung health⁵. Furan, a volatile organic compound as a potential human carcinogen²⁴. It is common to found in heat-processed foods through heat treatment methods²⁵. Therefore, individuals are at risk of furan toxicity. This is particularly alarming due to the widespread presence of furan in the environment, highlighting its potential role as a public health hazard.

There are some issues that could explain the adverse effects of furan on respiratory health. A previous investigation indicated that bronchiolar club cell enlargement, cytoplasmic vacuolation, and necrosis along with airway denudation and an inflammatory response consisting of bronchiolar wall infiltration by low numbers of many neutrophils and lymphocytes when breathing furan ¹¹. According to Bas et al.'s study, pro-inflammatory cells moved into the bronchioles and alveolar gaps of the rat lungs, indicating that the oral furan administration had harmed the histoarchitectural structures of the lungs²⁶. Furan exposure may cause lung dysfunction through inflammation, altered histoarchitectural characteristics, genomic instability or damage, and loss of redox equilibrium.²⁷. All these studies demonstrated that furan was consumed orally or inhaled causing lung damage through various pathological mechanisms. The exact mechanism of the effect of furan exposure on the lungs of asthmatic patients and mouse models of asthma need further exploration in the future.

Lung function measurement forms a crucial part of the clinical assessment and management of patients with asthma. Individuals exposed to higher levels of furan exhibited notably lower FEV₁, FEV₁% predicted and FVC and FEV₁/FVC, indicative of impaired lung function. Inconsistent with our study, Ahman M et al. found that workers exposed to furan resin sand showed a decrease in FVC and total lung capacity but no fall in any other lung-function variable²⁸. There are two main reasons for the disparity about lung function between the two studies: the difference of enrolled population and the study design of two studies. Hence, further studies required to verify that one possible mechanism for furan-related asthma could be represented by FEV1. Understanding this pathway is crucial for developing targeted interventions aimed at preserving lung function and mitigating the adverse effects on environmental pollutants.

We found that LBFL was positively associated with participants-reported health outcomes including HQL, and respiratory symptom burden. These respiratory symptom burden not only reflect underlying respiratory dysfunction but also significantly impact daily living and overall well-being. The worse HQL indicating that the adverse effects of furan extend beyond clinical measures to affect subjective health perceptions and daily functioning.

Although Sun et al.’s study found that LBFL was associated with respiratory disease mortality in COPD patients¹². Nonetheless, no other studies have investigated the effect of furan exposure on the risk of respiratory disease mortality in participants with asthma. In the study, those in the furan ^High had a higher risk of respiratory disease mortality in participants with asthma compared with the furan ^Low in individuals with asthma, which suggests that a reduction in LBFL may reduce the risk of respiratory disease mortality in participants with asthma.

This study has several limitations. First, our study cannot demonstrate causal relationships between furan exposure and the prevalence of asthma. Further longitudinal studies are necessary to establish causality and monitor changes in lung function and asthma prevalence over time in relation to furan exposure. Second, the diagnosis of asthma depends on self-reported rather than a more reliable and accurate method such as lung function testing, which may lead to recall bias. Third, although we adjusted potential risk factors, the existence of unknown confounders cannot be completely eliminated, and some important covariates, such as medication use (OCS and ICS/LABA), asthma control level, or multi-pollutant adjustment.

In conclusion, this study indicated a positive link of furan exposure with the prevalence and respiratory mortality of asthma. FEV₁ served as a mediator in this relationship underscores the potential of furan to contribute to the asthma burden through direct respiratory impairment. This implies that reducing environmental exposure to furan could potentially decrease asthma prevalence and respiratory mortality, thereby improving overall respiratory health. The evidence provided by this study underscores the need for a comprehensive approach to addressing environmental pollutants like furan and their impact on respiratory health.

ACME: average causal mediation effect

ADE: average direct effect

aOR: adjusted odds ratio

BMI: body mass index

CDC: Centers for Disease Control and Prevention

CI: confidence interval

ER: emergency room

FEV₁: forced expiratory volume in 1 s

FVC: forced vital capacity

HBP: high blood pressure

HQL: health-related quality of life

LBFL: log10-transformed blood furan levels

NCHS: National Center for Health Statistics

NHANES: National Health and Nutrition Examination Survey

OR: odds ratio

PIR: poverty-income ratio

Q: Quartiles RCS: restricted cubic spline

SD: standard deviation

VOCs: volatile organic compounds

WBC: white blood cell

Ethics approval and consent to participate

The study protocol received approval from the NCHS Research Ethics Board, and was performed in accordance with the Declaration of Helsinki. All participants gave their written consent. Details of NHANES has been documented elsewhere (https://www.cdc.gov/nchs/nhanes/index.htm).

Consent for publication

Not applicable.

Availability of data and materials

Publicly available datasets were analyzed in this study. These data can be found here: https://www.cdc.gov/nchs/nhanes/. The code used for these analyses are available from the corresponding author upon reasonable request.

Competing interests

The authors declare that they have no conflicts of interest to disclose.

Funding

This work was supported by the National Natural Science Foundation of China (8217010602).

Author’s contributions

Jiangtao Lin and Shuwen Zhang, conceived the study, performed the data interpretation and manuscript revision, and took accountability for all aspects of the work. Shuwen Zhang planned the work, carried out the data analysis, interpretation and drafted the manuscript. Kunlu Shen, Bingqing Sun and Bowen Liu offered the help of statistical analysis. Chunxiao Li, Xin Hou, Min Xiang, Mengqi Zhou and Jiangtao Lin interpreted the results and contributed to the manuscript revision. All authors approved the final manuscript.

Acknowledgements

The authors are grateful to all volunteers who participated in this study. The graphic abstract was made with assets from Freepik.com (www.freepik.com).

Global Initiative for Asthma. Global strategy for asthma management and prevention. 2023. Accessed January 1 2024. Avaliable at: https://ginasthma.org/wp-content/uploads/2023/07/GINA-2023-Full-report-23_07_06-WMS.pdf
Zheng XY, Ding H, Jiang LN, et al. Association between Air Pollutants and Asthma Emergency Room Visits and Hospital Admissions in Time Series Studies: A Systematic Review and Meta-Analysis. PLoS ONE. 2015;10(9):e0138146.
Bakke P, Eide GE, Hanoa R, Gulsvik A. Occupational dust or gas exposure and prevalences of respiratory symptoms and asthma in a general population. Eur Respir J. 1991;4(3):273–8.
Osborne ML, Pedula KL, O'Hollaren M, et al. Assessing future need for acute care in adult asthmatics: the Profile of Asthma Risk Study: a prospective health maintenance organization-based study. Chest. 2007;132(4):1151–61.
Mattila T, Santonen T, Andersen HR et al. Scoping Review-The Association between Asthma and Environmental Chemicals. Int J Environ Res Public Health 2021;18(3).
Humblet O, Diaz-Ramirez LG, Balmes JR, Pinney SM, Hiatt RA. Perfluoroalkyl chemicals and asthma among children 12–19 years of age: NHANES (1999–2008). Environ Health Perspect. 2014;122(10):1129–33.
Podlecka D, Gromadzińska J, Mikołajewska K, Fijałkowska B, Stelmach I, Jerzynska J. Longitudinal effect of phthalates exposure on allergic diseases in children. Ann Allergy Asthma Immunol. 2020;125(1):84–9.
Pfeffer PE, Mudway IS, Grigg J. Air Pollution and Asthma: Mechanisms of Harm and Considerations for Clinical Interventions. Chest. 2021;159(4):1346–55.
Mogol BA, Gökmen V. Thermal process contaminants: acrylamide, chloropropanols and furan. Curr Opin Food Sci. 2016;7:86–92.
Alizadeh M, Jalal M, Hamed K, et al. Recent Updates on Anti-Inflammatory and Antimicrobial Effects of Furan Natural Derivatives. J Inflamm Res. 2020;13:451–63.
Tǎbǎran AF, O'Sullivan MG, Seabloom DE, et al. Inhaled Furan Selectively Damages Club Cells in Lungs of A/J Mice. Toxicol Pathol. 2019;47(7):842–50.
Sun D, Wang Y, Wang J, Dilixiati N, Ye Q. Inflammation mediates the association between furan exposure and the prevalence and mortality of chronic obstructive pulmonary disease: National Health and Nutrition Examination Survey 2013–2018. BMC Public Health. 2024;24(1):1046.
Shan Z, Rehm CD, Rogers G, et al. Trends in Dietary Carbohydrate, Protein, and Fat Intake and Diet Quality Among US Adults, 1999–2016. JAMA. 2019;322(12):1178–87.
Wu TD, Fawzy A, Brigham E, et al. Association of Triglyceride-Glucose Index and Lung Health: A Population-Based Study. Chest. 2021;160(3):1026–34.
Furan. Report on carcinogens: carcinogen profiles. 2011;12:205–207.
Miller MR, Hankinson J, Brusasco V, et al. Standardisation of spirometry. Eur Respir J. 2005;26(2):319–38.
Hankinson JL, Odencrantz JR, Fedan KB. Spirometric reference values from a sample of the general U.S. population. Am J Respir Crit Care Med. 1999;159(1):179–87.
National Center for Health Statistics (U.S.). National Health and Nutrition Examination Survey, Respiratory Health Spirometry Procedures Manual. Accessed January 1 2024. https://wwwn.cdc.gov/nchs/data/nhanes/2011-2012/manuals/spirometry_procedures_manual.pdf
Cirillo DJ, Agrawal Y, Cassano PA. Lipids and pulmonary function in the Third National Health and Nutrition Examination Survey. Am J Epidemiol. 2002;155(9):842–8.
Luyster FS, Shi X, Baniak LM, Morris JL, Chasens ER. Associations of sleep duration with patient-reported outcomes and health care use in US adults with asthma. Ann Allergy Asthma Immunol. 2020;125(3):319–24.
Zhang W, Peng SF, Chen L, Chen HM, Cheng XE, Tang YH. Association between the Oxidative Balance Score and Telomere Length from the National Health and Nutrition Examination Survey 1999–2002. Oxidative Med Cell Longev. 2022;2022:1345071.
Tingley D, Yamamoto T, Hirose K, Keele L, Imai K. mediation: R Package for Causal Mediation Analysis. J Stat Softw. 2014;59(5):1–38.
Imai K, Keele L, Tingley D, Yamamoto T. Unpacking the Black Box of Causality: Learning about Causal Mechanisms from Experimental and Observational Studies. Am Polit Sci Rev. 2011;105(4):765–89.
Batool Z, Xu D, Zhang X, et al. A review on furan: Formation, analysis, occurrence, carcinogenicity, genotoxicity and reduction methods. Crit Rev Food Sci Nutr. 2021;61(3):395–406.
FDA UJhwcfgdfh. Question and Answers on the Occurrence of Furan in Food. 2004.
Baş H, Pandir D. Protective Effects of Lycopene on Furan-treated Diabetic and Non-diabetic Rat Lung. Biomed Environ Sci: BES. 2016;29(2):143–7.
Owumi SE, Otunla MT, Arunsi UO. A biochemical and histology experimental approach to investigate the adverse effect of chronic lead acetate and dietary furan on rat lungs. Biometals: Int J role metal ions biology Biochem Med. 2023;36(1):201–16.
Ahman M, Alexandersson R, Ekholm U, Bergström B, Dahlqvist M, Ulfvarson U. Impeded lung function in moulders and coremakers handling furan resin sand. Int Arch Occup Environ Health. 1991;63(3):175–80.

No competing interests reported.

OnlineDataSupplementfuranandasthma.docx

Download PDF

Reviewers invited by journal
31 Aug, 2024
Editor assigned by journal
08 Aug, 2024
Submission checks completed at journal
07 Aug, 2024
First submitted to journal
06 Aug, 2024

You are reading this latest preprint version

Associations of furan exposure with the prevalence and mortality in asthma: A prospective cohort study

Status:

Version 1

Abstract

Background

Objective

Methods

Results

Conclusions

Figures

Introduction

Method

Study design and participants

Exposure Definitions

Primary outcome

Prevalence of asthma

Respiratory mortality in asthma

Secondary outcome

Lung function

Asthma-related health outcomes

Respiratory symptoms

HQL

Demographic and Clinical Data

Statistical analysis

Results

Baseline characteristics of participants

Primary Outcomes

Furan exposure and asthma prevalence

Subgroup analyses of factors impacting the association of asthma prevalence with furan exposure

Mediation analysis of lung function involved in the effects of furan on asthma

Furan exposure and respiratory mortality of asthma

Secondary outcome

Lung function

Asthma-related health outcomes

Respiratory symptoms

HQL

Discussion

Conclusion

Abbreviations

Declarations

References

Additional Declarations

Supplementary Files

Status:

Version 1