Lung Damage among Individuals with Biomass Smoke Exposure Characterized by CT Features and Changes in Pulmonary Function

Background Biomass smoke exposure (BSE) is an important etiological factor in chronic obstructive pulmonary disease (COPD). However, few studies have focused on the effects of BSE in the respiratory muscles or lungs. Using a cohort, we selected 98 participants in underdeveloped rural areas: 16 healthy individuals with BSE (“BSE normal”), 19 patients with BSE and COPD (“BSE+COPD”), 13 healthy individuals with cigarette smoke exposure (“CSE normal”), 25 patients with cigarette smoke exposure and COPD (“CSE+COPD”), and 25 healthy controls. Patients with GOLD stage I and II COPD were included. Baseline data (demographic data, BSE or CSE, lung function, and CT ndings) and follow-up lung function data were collected. CT parameters of emphysema, pulmonary small vessels, airway remodeling, pectoralis muscles, and erector spinae muscle were measured.


Abstract Background
Biomass smoke exposure (BSE) is an important etiological factor in chronic obstructive pulmonary disease (COPD). However, few studies have focused on the effects of BSE in the respiratory muscles or lungs.

Methods
Using a cohort, we selected 98 participants in underdeveloped rural areas: 16 healthy individuals with BSE ("BSE normal"), 19 patients with BSE and COPD ("BSE+COPD"), 13 healthy individuals with cigarette smoke exposure ("CSE normal"), 25 patients with cigarette smoke exposure and COPD ("CSE+COPD"), and 25 healthy controls. Patients with GOLD stage I and II COPD were included. Baseline data (demographic data, BSE or CSE, lung function, and CT ndings) and follow-up lung function data were collected. CT parameters of emphysema, pulmonary small vessels, airway remodeling, pectoralis muscles, and erector spinae muscle were measured.

Results
Individuals with BSE were mainly women (32/35, 91.43%). Compared with the CSE+COPD group, the BSE+COPD group demonstrated slow lung function decline, increased lower lung emphysema, and narrower airway lumen and airway wall thickening in moderate and small airways (all P<.05). Compared with healthy controls, the CSE normal and BSE normal groups exhibited signi cant reduction of pulmonary small vessel area and obvious airway remodeling in small airways (P<.05). Compared with the BSE normal group, the BSE+COPD group showed signi cantly more severe emphysema and airway remodeling, as well as reduced left pectoralis major muscle area (all P<.05).

Conclusions
Healthy individuals with BSE had reduced pulmonary small vessel area and obvious airway remodeling; patients with BSE and COPD showed more severe emphysema, airway remodeling, and pectoralis major muscle change. More investigations are needed regarding interventions for BSE. Background Chronic obstructive pulmonary disease (COPD) is associated with exposure to large amounts of toxic particles or gases. Biomass smoke exposure (BSE) and cigarette smoke exposure (CSE) are important risk factors for COPD [1,2]. In recent decades, considerable progress has been made in the treatment of COPD, but its incidence and mortality remain high [1,2], which suggests that COPD-related environmental and occupational factors require further improvement [3]. Among these factors, the absence of interventions for BSE is particularly obvious [4]. Nearly 3 billion people worldwide use biomass as energy for heating and cooking, which results in substantial BSE [1]. Individuals affected by BSE mainly comprise women who use biomass for cooking in underdeveloped rural areas; these individuals tend to have relatively low socioeconomic statuses [5][6][7][8].
To the best of our knowledge, there has been insu cient attention to BSE. A few studies have performed analyses of clinical symptoms and the CT ndings of total proportions of emphysema and small airway disease caused by BSE [9][10][11][12]. There is a lack of comprehensive studies regarding pulmonary lesions caused by BSE, how BSE affects the pulmonary small vessels,the respiratory muscles, the development of emphysema and airway remodeling. The dangers of BSE have received less publicity in social media, compared with the dangers of CSE [12][13][14][15]. For example, among users of TikTok (a popular social media service in China), there are currently 198 accounts that promote smoking cessation-related content; by contrast, only one professional account addresses the risks of biomass smoke pollution. Similarly, the e cacies of clean energy replacement plan are sometimes disappointing; these plans are in uenced by factors such as economics, education, and women's social status in the family [5,6,8,16].These data suggest that the task of reducing BSE cannot be accomplished by individuals or families, requires societal mobilization for effective outcomes [17].
Computed tomography (CT)-derived evaluation of the basic structural changes of COPD (e.g., airway remodeling, emphysema, pulmonary vascular changes, and respiratory muscles) can re ect the severity of lung lesions, thus facilitating objective assessment and prediction of disease progression [18][19][20][21][22]. Our study focused on populations with BSE in economically underdeveloped rural areas to comprehensively investigate CT ndings in patients with and without COPD and exposures to cigarette smoke or biomass fuels. This study aimed to clarify the features of BSE-related lung injury compared to CSE induced injury that results in COPD vs those without COPD.

Study design
The Ethics Committee of Guangzhou Institute of Respiratory Health approved the study protocol. Participants were selected from among the rural population in an underdeveloped mountainous area of northern Guangdong, China, using an observatory research subgroup data from COPD community screening database (from 2014 to 2015) [10,23,24]. No intervention was performed on the whole participants and no medication was prescribed to patients with COPD. BSE exposure population and CSE population were followed for a year. Healthy control individuals were not followed. Written informed consent was obtained from all participants. Detailed information regarding recruitment, clinical management, and follow-up was described in previous studies [10,23,24].

Population
This study included individuals with BSE or CSE (including otherwise healthy individuals and patients with COPD) and healthy controls, who were over > 40 years of age, had complete clinical data, pulmonary function assessments, and lung HRCT data. The study excluded patients who had lung cancer, asthma, interstitial lung disease, pulmonary infarction, pneumonia, or pleural effusion,based on medical history or CT scan ndings. The healthy control group included healthy individuals with normal lung function, without a history of CSE or BSE, without a history of pulmonary disease or respiratory symptoms, and without visible emphysema or low attenuation areas less than a threshold of -950 Houns eld units (LAA-950%) <5% in CT scans.

Patient Groups
According to their history of CSE or BSE and whether they had been diagnosed with COPD, 98 eligible participants were allocated into ve groups: BSE+COPD (n=19), BSE normal (n=16), CSE+COPD (n=25),

Data collected
Baseline records of participants were collected, including demographic information, lung function, history of BSE and/or CSE, and high-resolution CT ndings; follow-up lung function data were also collected. The BSE and CSE indexes were calculated as previously described [10]. The BSE index is de ned as the cumulative exposure of biomass,which is calculated by multiplying the average number of hours per day in the kitchen with the number of years of cooking with biomass.Cigarette smoking index was expressed as pack-years.
Annual reductions in lung function (FEV 1 and FVC) were calculated. High-resolution CT ndings were assessed to determine the progression of emphysema and airway remodeling, as well as damage to pulmonary small vessels, pectoralis major muscle, pectoralis minor muscle, and erector spinae muscle.

CT measurement of emphysema
Emphysema was detected using the chest imaging platform/parenchyma module of 3D slicer software, using a Houns eld unit threshold of −950 (i.e., %LAA-950). The (upper third)/(lower third) ratio of LAA-CT measurement of bronchial inner diameter and bronchial wall thickness Using the Airway Inspector module of 3D slicer software [19], the third, fourth, fth, sixth generation of airway bronchi of the right upper lobe apical bronchus (i.e., RB1) and right lower lobe posterior bronchus (i.e., RB10) were identi ed. Then, the inner diameter, bronchial wall thickness, and bronchial wall area% were detected. CT measurement of cross-sectional area (CSEA)% of pulmonary small vessels In accordance with the method described by Matsuoka et al [19,26], pulmonary small vessels were de ned as "circular" blood vessels that were perpendicular to the cross-sectional plane and exhibited a cross-sectional area of <5 mm 2 . The upper, middle, and lower slices were captured 1 cm above the top of the aortic arch, 1 cm below the carina, and 1 cm below the right lower pulmonary vein, respectively. CSEA <5% was calculated as the percentage of small vessel CSEA to lung CSEA in each slice. The total CSEA <5% was calculated as the percentage of total small vessel CSEA to total lung CSEA of three slices. CSEA% detection was performed as shown in Figure 2. First, the segmentation and modeling of the lungs were conducted using the Parenchyma Analysis and edit modules in the 3D slicer software. The segmented lung areas were lled with -200 to achieve a uniformly gray color. Second, chest vessels were segmented using the VTMK module in 3D slicer software. Pulmonary vessels were segmented by clipping chest vessels with the lung model. Third, in the above three slices, images of pulmonary vessels and gray lungs were saved separately. Image-Pro Plus 6.0 software (Media Cybernetics, Inc.) was used to measure the areas of gray lungs and small pulmonary vessels. Fourth, pulmonary vessels with "Rounders" of 0.9-1 (i.e.,circular blood vessels) and "Areas" of ≤5 mm 2 were segmented as pulmonary small vessels (CSEA <5 ). Fifth, CSEAs of gray lungs and small vessels were measured using the count/size module in Image-Pro Plus software for each slice. Finally, CSEA <5% in each slice and total CSEA <5% were calculated using the formula described above.

CT measurement of pectoralis and erector spinae muscles
The area and density of muscles were measured using the body composition module of 3D slicer software, as shown in Figure 3. The pectoralis major and minor muscles were measured at the rst slice above the aorta, while the erector spinae muscle was measured at the lower slice of the T12 thoracic vertebra [21,27].

Statistical analysis
Data are presented as mean ± standard deviation, median (quartile), or numbers. Comparisons of continuous variables among groups or between multiple groups were evaluated by one-way analysis of variance with a post hoc test (least signi cant difference method) , Kruskal-Wallis test or Mann-Whitney U test . Comparisons of categorical variables among groups were performed using Fisher's exact test.
Pearson's correlation coe cients were used to assess associations between lung function decline and CT index. Statistical analyses were conducted using IBM SPSS Statistics, version 23.0 (IBM Corp., Armonk, NY, USA). Two-tailed P-values < .05 were considered statistically signi cant.

Baseline characteristics
Supplementary Table 1 describes the baseline characteristics of participants in each group. This study focused on economically underdeveloped mountainous rural areas, where individuals with BSE were mainly women who cook with biofuel, while smokers were mainly men. The BIOFUEL-index for BSE COPD and the smoking index for CSE COPD had no signi cantly difference with BSE normal group and CSE normal groups (P>.05). There were no signi cant differences in baseline lung function (FEV 1 % and FEV 1 /FVC) and GOLD stage between the BSE+COPD and CSE+COPD groups (both P>.05).
Our previous study demonstrated that BSE COPD patients presented similar respiratory symptom (cough, dyspnoea and wheezing) and similar overall COPD assessment scores (CAT, mMRC and combined COPD assessment) to CSE COPD patients and weaker activity tolerance (six-minute walk distance) [10].This study discovered that lung function decline (FEV 1 and FEV 1 /FVC) was slower in the BSE+COPD group than in the CSE+COPD group (both P<.05), suggesting that patients with BSE and COPD may exhibit slow progression of disease and experience greater bene t from early active intervention.

CT features of individuals with BSE
Patients with BSE and COPD exhibited prominent emphysema in the lower lung Like patients in the CSE+COPD group, those in the BSE+COPD group had substantial changes in emphysema, and there was no signi cant difference in the total percentage of emphysema between the two groups (P>.05). However, lower lung emphysema was more obvious in the BSE+COPD group than in the CSE+COPD group (P<.05). See Supplementary Table 2.
Pulmonary small vessel areas were signi cantly reduced in otherwise healthy individuals with BSE Compared with the healthy control group, both the BSE normal and CSE normal groups showed reduced pulmonary small vessel areas (P<.05); this was consistent among the upper, middle, and lower parts of the lung (P<.05). Compared with the BSE normal and CSE normal groups, the pulmonary small vessel areas in BSE+COPD and CSE+COPD groups were further reduced, but these differences were not statistically signi cant difference (all P>.05). Finally, there was no signi cant difference in the reduction of pulmonary small blood vessel areas between the BSE+COPD and CSE+COPD groups (P>.05). See Supplementary Table 2.
More serious airway remodeling in medium and small airways (grades III-VI bronchi) was observed in the BSE+COPD group Small airway remodeling was present in otherwise healthy individuals with BSE, while patients with BSE and COPD exhibited serious medium and small airway remodeling.Compared with the CSE normal group, small airway remodeling (i.e., narrow airway and thickened airway wall) was obvious in grades IV-VI bronchi in the BSE normal group (all P<.05). Compared with the CSE+COPD group, more serious airway remodeling in medium and small airways (grades III-VI bronchi) was observed in the BSE+COPD group (all P<.05). See Supplementary Table 2.
Left pectoralis major muscle area signi cantly decreased in patients with BSE and COPD The pectoralis major muscle area and density were considerably greater in the BSE normal group than in the healthy control group (all P<.05); similar ndings were present in the CSE normal group (all P<.05). Compared with the BSE normal group, the left pectoralis major muscle area signi cantly decreased in the BSE+COPD group (P<0.05). Compared with the CSE normal group, the CSE+COPD group tended to show reduced areas of pectoralis and erector spinae muscles, but these differences were not statistically signi cant (all P>.05).See Supplementary Table 2. CT presentation and lung function changes Lung function decline was relatively slow in the BSE+COPD group; there were no correlations between CT indexes and annual reduction of FEV 1 (all P>.05;Supplementary Table 3). However, the FEV 1 reduction was relatively rapid in the CSE+COPD group; this reduction was negatively correlated with the right pectoralis minor muscle area (r=-0.68, P<.01; Supplementary Table 3).

Discussion
CSE is the most common cause of COPD. The harmfulness of smoking is well-known among members of the public, especially because of social media campaigns and the development of smoking cessation clinics [14,15,28,29]. In contrast, the risks of BSE are not well-known among the public and healthrelated investments regarding BSE are insu cient.
This real-world study was conducted in an underdeveloped mountainous area. The composition of the study population is representative of the local area.The study speci cally focused on describing the lung damage caused by BSE, with CSE as a reference exposure.
The limited evidence regarding lung damage caused by BSE has mainly emphasized emphysema and air trapping [10][11][12], which does not provide su cient insights to guide BSE intervention. In this study, we used 3D slicer software to make an overall assessment of lung damage,including airway remodeling in medium and small airways, emphysema, pulmonary small vessels, and respiratory muscles. Qualitative and quantitative determinations of lung damage caused by BSE were performed following three-dimensional and two-dimensional imaging of lung lesions. The resulting representative lung lesion images from individuals with BSE will presumably be useful for further scienti c research and public health-related interventions.
Different from previous reports [10,12], we discovered that airway damage was more serious in the BSE+COPD group than in the CSE+COPD group, such that it involved substantial remodeling of small and medium airways. Furthermore, pectoralis major muscle area decreased in the BSE+COPD group, compared with the BSE normal group; a corresponding difference was not observed between CSE+COPD and CSE normal groups. The impacts of BSE on respiratory muscles, compared with CSE, require further investigation.
An interesting nding was that both the BSE+COPD and BSE normal groups had signi cant reductions in pulmonary small vessel area. To the best of our knowledge, this is the rst study to show changes in pulmonary small vessel area among individuals with BSE. The reduction of pulmonary small vessel area is associated with acute exacerbation of COPD [30]and serves as an independent risk factor for mortality [31].Additionally, BSE is associated with elevated risks of hypertension, coronary heart disease, and stroke [32]. Therefore, in addition to pulmonary small vessels, BSE might also be involved in other cardiovascular events.
The present study showed that lung function decline was slightly slower in the BSE+COPD group, which was consistent with the results reported by Salvi et al [33]. Since efforts to improve cooking fuels and kitchen ventilation will reduce indoor PM 2.5 concentrations, delay lung function decline, and lower the risk of respiratory disease [24,34], early intervention for BSE patients may bene t more.
Phenotyping studies regarding COPD have found that speci c lung CT features are closely related to the progression of COPD. Speci c types of emphysema, airway remodeling, and changes in pulmonary small vessel and chest muscle areas, all of which can be reversed by active intervention, are associated with the progression of COPD [18, 19,21,22,30,31,35,36]. Therefore, we examined correlations between these pathological changes and the annual reduction of FEV 1 , with the aim of identifying a target to delay lung function decline in patients with BSE and COPD. However, we did not nd an association between any CT index and reduction of FEV 1 in the BSE+COPD group. This may have been in uenced by the small number of participants in the BSE+COPD group, therefore producing no statistically signi cant ndings.
This study had some limitations. First, the number of participants was relatively small, and there were no follow-up lung CT data for any participants. This is just a re ection of the insu cient attention to BSE and limited relevant research funding. Second, this study did not speci cally recruit women in the CSE+COPD group or men in the BSE+COPD group; therefore, future studies should focus on sex differences between BSE+COPD and CSE+COPD groups.

Conclusions
In summary, this study focused on individuals with BSE living in an underdeveloped rural area. It showed that otherwise healthy individuals exposed to BSE had reduced pulmonary small vessel area and obvious airway remodeling, while patients with BSE and COPD had severe emphysema, airway remodeling, and muscle change, as well as relatively slow pulmonary function decline. More investigations are thus needed regarding interventions for BSE.

Availability of data and materials
The datasets used and analyzed during the current study are available from the corresponding author on reasonable request.
Ethics approval and consent to participate   Sample computed tomography (CT) scans used to determine pulmonary small vessels a CT image in middle slice of lung. b Segmented lungs shaded in grey. c Pulmonary vessels shaded in green. d pulmonary small vessels shaded in red.