Chronic obstructive pulmonary disease associated with biomass fuel use in women: a systematic review and meta-analysis.

INTRODUCTION: Chronic obstructive pulmonary disease (COPD) is a major and growing cause of morbidity and mortality worldwide. The global prevalence of COPD is growing faster in women than in men. Women are often exposed to indoor pollutants produced by biomass fuels burning during household activities.
METHODS: We conducted a meta-analysis to establish the association between COPD and exposure to biomass smoke in women.Following Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines, we searched MEDLINE and Scopus databases in 31December 2016, with the terms: "wood", "charcoal", "biomass", "solid fuels", "organic fuel", "biofuel", "female", "women", "COPD", "chronic bronchitis", "emphysema", "chronic obstructive pulmonary disease". Studies were eligible if they were case-control or cross-sectional studies involving exposure to indoor biomass smoke, conducted at any time and in any geographic location. Fixed-effects or random-effects meta-analysis was used to generate pooled OR.
RESULTS: 24 studies were included: 5 case-control studies and 19 cross-sectional studies. Biomass-exposed individuals were 1.38 times more likely to be diagnosed with COPD than non-exposed (OR 1.38, 95% CI 1.28 to 1.57).Spirometry-diagnosed COPD studies failed to show a significant association (OR 1.20, 95% CI 0.99 to 1.40). Nevertheless, the summary estimate of OR for chronic bronchitis (CB) was significant (OR 2.11, 95% CI 1.70 to 2.52). The pooled OR for cross-sectional studies and case-control studies were respectively 1.82 (95% CI 1.54 to 2.10) and 1.05 (95% CI 0.81 to 1.30). Significant association was found between COPD and biomass smoke exposure for women living as well in rural as in urban areas.
CONCLUSIONS: This study showed that biomass smoke exposure is associated with COPD in rural and urban women.In many developing countries, modern fuels are more and more used alongside traditional ones, mainly in urban area. Data are needed to further explore the benefit of the use of mixed fuels for cooking on respiratory health, particularly on COPD reduction.

What is the relation between chronic obstructive pulmonary disease (COPD) and biomass smoke exposure for women living in rural and in urban areas?

Significant association was found between COPD and biomass smoke exposure for women living as well in rural as in urban areas.

This study provides detailed analysis of the association between COPD and biomass smoke exposure in women. This study shows that energy poverty continues to be a public health problem both in rural and in urban areas. Designing policies to reduce energy poverty would alleviate the burden of chronic respiratory diseases of women and improve their living conditions.

Introduction

Chronic obstructive pulmonary disease (COPD) is a major and growing cause of morbidity and mortality worldwide.1 WHO estimates COPD to be the tenth leading cause of disability-adjusted life-years in all countries.2 According to WHO, 65 million people suffer from moderate to severe COPD and >3 million people died of COPD in 2015, among which 90% death occurs in low-income and middle-income countries.3 COPD deaths have been increasing these past years, and projections suggest it could become the third leading cause of death by 2030.4 More common in men once upon a time, COPD now affects almost equally men and women.3

Among the various risk factors, the most important is tobacco smoking as it happens in men and women in high-income and middle-income countries.1 5 However, in low-income, middle-income countries, 35% of patients with COPD have developed the disorder after a chronic exposure to indoor smoke from biomass fuels burning.2 6

One-third of the world’s population use biomass fuel, like wood, crop residues such as straw and sticks, dried leaves, twigs, wild grass, animal dung or charcoal, for cooking and/or heating.7 The smoke from these organic fuels increases the incidence of respiratory illness such as COPD.8 WHO has estimated that indoor air pollution from solid fuel use is responsible for 2.6% of the total global burden of disease.9 The use of biomass has contributed to >577 000 premature deaths in Africa and 74 000 in Latin Americas, in 2012.9 Exposure to indoor pollutants produced by biomass fuels burning is particularly high among women and young children,10 leading to 2 million deaths each year.11 Many of these deaths occur either in children under 5 years of age, primarily due to acute lower respiratory infection such pneumonia or in adult women due to COPD.9

The prevalence of COPD is two to three times higher in rural women exposed to biomass smoke compared with urban women who are considerably less exposed.11 12

The worldwide prevalence of COPD is growing faster in women than in men.13 Over the past two decades, COPD-related mortality rates have also increased faster for women, and since the year 2000 more women than men have died from COPD.13 Many individual studies have been conducted worldwide to evaluate the relation between energy choice for cooking and COPD, and these studies lead to a wide variation in findings. Notwithstanding, controversies remain concerning the link between biomass fuel use and COPD in women. The meta-analyses previously conducted have either analysed the link between COPD and solid fuel14 or the association between the disease and biomass exposure in adults. The one meta-analysis that focuses on women only analysed studies in a rural environment and considered so many respiratory diseases that did not allow an in-depth assessment of the relation between COPD and indoor biomass smoke exposure in women. It remains a field that still requires considerable attention. Thus, we conducted a systematic review and meta-analysis to highlight the relationship between COPD and domestic biomass fuel use in women.

Methodology

Search strategy

Using MEDLINE and Scopus database, we did a systematic literature search according to Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines up to 31 December 2016, with keywords including “wood”, “charcoal”, “biomass”, “solid fuels”, “organic fuel”, “biofuel”, “female”, “women”, “COPD”, “chronic bronchitis”, “emphysema”, “chronic obstructive pulmonary disease”.

The same search terms were used for both databases.

The search was restricted to English and French languages. No limitations were set for participants’ ages. Studies were considered if they estimate the association between COPD and biomass smoke exposure.

To better fulfil our objectives, studies searching was based on the participants, interventions, comparators, outcomes and study design approach. Participants were women, cooking or cooking and heating with biomass fuels represented the exposure (intervention), cooking or cooking and heating with non-biomass fuel or clean fuel were the comparator and the outcomes were COPD phenotypes (COPD, chronic bronchitis (CB), emphysema). We considered case–control or cross-sectional studies.

So, studies were eligible if comparing exposure to biomass smoke to exposure to other fuels, conducted at any time and in any geographic location. They had to use case–control or cross-sectional designs. Papers had to provide calculable or reported ORs to estimate the association between COPD and biomass smoke with corresponding 95% CI in female; and they had to be based on an independent set of data from other studies. References in each of the identified papers were screened for any additional article that had not been identified in the original search. Articles were excluded when they could not distinguish statistical association between exposure to biomass smoke and respiratory diseases found in women from men. Studies without a proper comparator were also discarded.

Studies were included in the final analysis when they had considered COPD as an airflow limitation that is not fully reversible (assessed by postbronchodilator spirometry), either according to the American Thoracic Society/European Respiratory Society criteria15 (postbronchodilator forced expiratory volume in 1 s (FEV1)/forced vital capacity (FVC) ratio ≤ 0.70) or the Global Initiative for Obstructive Lung Disease criteria16 (presence of a postbronchodilator FEV1/FVC ratio <0.70) or the method of FEV1/FVC below the lower limit of normal value. Case–control studies based on medical confirmed case of COPD are also analysed as well as studies that had considered CB according to the British Medical Research Council criteria ‘daily productive cough for at least 3 consecutive months for more than 2 successive years’.17

Studies selection and data extraction were performed by two of the authors (AS and SMAS), following the established protocol and the consensual data extraction table. Disagreements were resolved by consensus, but sometimes the issue was discussed with CB.

Data management

We screened titles, abstracts and full texts according to study eligibility (inclusion criteria), and data were extracted using an internally validated data extraction form. When required, additional information was obtained from authors.

Data analysis

The ORs and their CI were extracted from the publications or calculated when the paper did not report ORs but provided sufficient information for its calculation. The ORs of COPD associated with biomass smoke were estimated using no exposure to biomass smoke as the reference. In this work, only a few studies provided adjusted estimates of OR. In addition, those which provided adjusted OR did not consider confounders in the same way. To avoid heterogeneity due to adjustment for confounding factors, pooling the mix of adjusted ORs and unadjusted ORs is not appropriate in this case.18 19 Then we preferred analysing using only the unadjusted ORs.19 Meta-analysis was performed using Stata software V.13. The ORs and the 95% CI were used to estimate the pooled effect size of all the studies. The homogeneity Q statistic and the I² index were computed. A random-effects model was used when the heterogeneity was high (I²>50%), given that the Cochran Q statistic is known to be anticonservative.20 When the heterogeneity was low (I²<50%), a fixed-effects model was used. The variance of the fixed-effects model was estimated with the Mantel and Haenszel method,21 and one of the random-effects model was obtained using the DerSimonian and Laird method.22

Subgroup analyses were performed with stratifications by study design (case–control and cross-sectional), geographic location (rural, including semirural; urban, including semiurban; and rural/urban location), phenotypes (COPD, emphysema or CB) and smoking status. The significance of pooled ORs was determined by z test. Two-tailed P values<0.05 were considered statistically significant. Publication bias was assessed through the Egger test and funnel plots.23

The study is reported in accordance with the  Meta-analysis Of Observational Studies in Epidemiology guidelines for meta-analysis and systematic reviews of Observational Studies in Epidemiology.

Results

Characteristics of the articles

Up to 31 December 2016, our search identified 641 citations. After title review and duplicates removed, 164 abstracts were reviewed and 49 full-text articles were kept. We excluded 28 full-text papers as they failed to meet inclusion criteria or had insufficient information for data extraction. Twenty-one papers were eligible, and we found three more papers in references of full-text articles screened. The detailed selection process is shown in figure 1.

Figure 1

Preferred Reporting Items for Systematic Reviews and Meta-Analyses study selection flow chart.

The meta-analysis includes 24 articles divides into 5 case–control studies and 19 cross-sectional studies.

Concerning location, 11 studies were based on rural populations, 3 on urban, 4 on urban hospital, 5 on mixed rural/urban populations and 1 on mixed rural/semiurban/urban populations. Concerning gender, six mixed-gender studies had sufficient information to allow women data extraction; the remaining 18 studied only women. About smoking status of the subjects, 12 studies were conducted on non-smoking women and 12 on both smoking and non-smoking women with only one study providing data for non-smokers group. Considering phenotypes, 13 articles studied COPD, 10 solely CB and 1 studied both COPD and CB.

In total, the selected papers accounted for 19 099 subjects, of whom 669 suffered of CB, 1594 of COPD and the remaining 16 836 subjects were healthy participants.

Several biomass fuels were studied, including various combinations of biomass kind, or wood only.24–27 Comparator fuels included liquefied petroleum gas (LPG), gas or gas/electricity and kerosene. Two studies were conducted in Africa.28 29 The others were distributed between Asia, South America, Middle East and Europe.

The details of the included studies in the meta-analysis are shown in table 111 12 24–27 29–46 and (online supplementary table S1) (spirometry lung function test results). No unpublished or ongoing studies were retrieved.

Table 1

Studies included in the meta-analysis investigating the relation between COPD and biomass fuel exposure

Authors	Location	Study design	Population	Sample size	Definition of exposure	Definition of non-exposure	Phenotypes	OR (95% CI)	Smoking status	PBD FEV1% mean±SD (case/exposed)
Van Gemert et al30	Uganda Rural	CS	≥30 years	297	Indoor biomass fuel exposure	Non-exposed to biomass fuel	COPD	1.7 (0.53 to 5.51)	B	91.8±20.3 (COPD)*
Örnek et al31	Turkey Urban	CS	≥10 years	351	Biomass for cooking or heating	No biomass use	COPD	0.97 (0.42 to 2.22)	B	N/A
Jaganath et al32	Peru Rural/urban	CS	≥35 years	1487	Biomass fuel for daily cooking	Non-biomass fuel	COPD	2.22 (1.02 to 4.81)	B	N/A
Mukherjee et al33	India Rural	CS	23–43 years	1119	Biomass (dung, wood, dried leaves, jute stick, hay)	LPG	COPD	5.18 (1.81 to 20.3)	N	71.8±6.82 (biomass)
Dutta et al34	India Rural	CS	22–41 years	480	Cooking with biomass	Cooking with LPG	COPD	4.07 (1.34 to 12.36)	N	69.8±26.6 (biomass)
Alim et al35	Bangladesh Rural/urban	CS	≥15 years	420	Biomass	Gas	CB	4.62 (1.32 to 16.2)	N	N/A
Sukhsohale et al36	India Rural	CS	≥15 years	760	Biomass	LPG	CB	1.40 (0.82 to 2.41)	N	N/A
Laniado-Laborin et al37	Mexico Urban	CS	≥40 years	1380	Exposure to biomass smoke	Not exposed to biomass smoke	COPD	1.65 (1.24 to 2.2)	B	60.6±21.8 (COPD)
Johnson et al38	India Rural	CS	≥30 years	900	Biomass	Clean fuel (kerosene, LPG)	COPD	1.24 (0.36 to 4.25)	N	N/A
Desalu et al29	Nigeria Rural	CS	≥35 years	269	Biomass fuel	Non-biomass fuel	CB	3.75 (1.07 to 13.16)	N	70.8±9.50 (biomass)
Akhtar et al39	Pakistan Rural	CS	≥10 years	2557	Solid biomass fuels for cooking	LPG for cooking	CB	2.51 (1.65 to 3.83)	N	N/A
Liu et al40	China Rural/urban	CS	≥40 years	1719	Biomass (wood, crop residues)	LPG	COPD	3.11 (1.63 to 5.94)	N	N/A
Xu et al41	China Rural/urban	CC	≥35 years	1396	Firewood/straw	Electricity/gas	COPD	0.98 (0.76 to 1.27)	B	N/A
Orozco-Levi et al12	Spain Urban	CC	>50 years	120	Exposure to wood, charcoal or both smoke	Not exposed to wood, charcoal	COPD	2.24 (0.84 to 6.28)	B	51 (33–75) (COPD)
Regalado et al42	Mexico Rural	CS	≥38 years	845	Using biomass as cooking fuel	Using gas as cooking fuel	COPD	1.5 (0.5 to 4.3)	N	97.8±13.7(biomass)
Sezer et al43	Turkey Rural/urban	CC	≥38 years	148	Biomass use (wood, dung)≥10 years	No exposure to biomass	COPD	1.32 (0.63 to 2.73)	N	N/A
Ekici et al11	Turkey Rural	CS	>40 years	596	Biomass (Wood, grass, crop, dung)	LPG	CB	1.4 (1.2 to 1.7)	N	Exposition index (biomass) 96.2±17.4 (<68.8 hour-years) 95.8±21.5 (68.8–152.4 hour-years) 87.8±26.2 (>152.4 hour-years)
Kiraz et al44	Turkey Rural/urban	CS	≥25 years	344	Biomass for cooking and heating (dung, wood, sticks)	LPG	COPD CB	3.47 (1.19 to 10.11) 2.15 (1.07 to 4.33)	B	Study group (rural/biomass) FEV1 %: 80.96±12.61 (smokers) FEV1 %: 82.36±11.48 (non-smokers)
Uzun et al27	Turkey Rural/urban	CS	17–75 years	177	Biomass fuel use (wood)	Non-biomass fuel use	CB	3.36 (1.8 to 6.26)	B	N/A
Golshan et al25	Iran Rural	CS	1 month to 81 years (27.62±16.6)	561	Using wood fuel in the past	Using gas fuel	CB	2.91 (2.08 to 4.4)	B	N/A
Pérez-Padilla et al26	Mexico Urban	CC	>40 years	438	Wood smoke exposure while cooking	No exposure to wood smoke	CB	3.9 (2.0 to 7.6) N 7.2 (2.8 to 20)	B N	N/A
Dennis et al24	Colombia Urban	CC	≥35 years	208	Using wood as cooking fuel	Not using wood as cooking fuel	COPD	3.43 (1.69 to 7.05)	B	FEV1 % 57.60±13 (COPD)
Dutt et al45	India Urban	CS	15–60 years	195	Biomass	LPG, kerosene	CB	4.17 (0.46 to 38.02)	B	N/A
Behera et al46	India Rural	CS	N/A	3608	Cooking with ‘chulla’ in which biomass fuels are used	Cooking with LPG stove	CB	1.75 (1.03 to 2.98)	N	FEV1 %: 83.21±1.89 (symptomatic)

*A ll the sample (including men and women).

B, both smokers and non-smokers; CB: chronic bronchitis; CC, case–control; COPD, chronic obstructive pulmonary disease; CS, cross-sectional; FEV1, forced expiratory volume for 1 s; LPG, liquefied petroleum gas; N, non-smokers only; N/A, not available; PBD, postbronchodilator.

Publication bias

Begg’s funnel plot visualisation indicated no publication bias in the studies included in the meta-analysis (see figure 2).

Figure 2

Funnel plot for studies included in the meta-analysis.

Exposure to biomass smoke and COPD

From the raw data, we observe that COPD, all phenotypes included, represents 11.85% of the total studied population (2263/19 099).

A fixed-effects model was used for the analysis because no heterogeneity had been found among the selected studies. Considering all COPD phenotypes OR, the pooled analysis shows that individuals exposed to biomass are 1.38 times (OR 1.38, 95% CI 1.28 to 1.57) more likely to be diagnosed with COPD than those not exposed.

The pooled analysis of only the COPD phenotype failed to show a significant association (OR 1.20, 95% CI 0.99 to 1.40). Nevertheless, we observe that ORs are significantly higher for CB (OR 2.11, 95% CI 1.70 to 2.52) than for COPD (figure 3).

Figure 3

Forest plot for chronic obstructive pulmonary disease (COPD) comparing biomass smoke exposed with non-exposed to biomass smoke separated by COPD phenotypes in women. CB, chronic bronchitis; ES, effect size.

The studies were then stratified between non-cigarette smokers’ participants (n=13) and both cigarette smokers and non-cigarette smokers’ participants (n=12). The analysis showed heterogeneity. However, a significant effect was highlighted as well in studies on both smokers and non-smokers (OR 1.90, 95% CI 1.35 to 2.45) and in studies on only non-smokers (OR 1.80, 95% CI 1.40 to 2.20) (see online supplementary figure S1).

The analyses were also stratified by location (urban/rural). The OR was very strong in studies conducted in rural area (pooled OR 1.95, 95% CI 1.54 to 2.37) compared with the pooled ORs of studies conducted in urban area (pooled OR 1.61, 95% CI 1.20 to 2.02) or both locations (rural/urban) (pooled OR 1.11, 95% CI 0.87 to 1.35) (see figure 4). However, studies using both locations did not present any significant difference.

Figure 4

Forest plot showing the effect size of chronic obstructive pulmonary disease due to exposure to biomass fuels compared with other fuels, separated by location.

When stratifying by study design, the pooled ORs for cross-sectional and case–control studies were respectively 1.82 (95% CI 1.54 to 2.10) and 1.05 (95% CI 0.81 to 1.30). Association is found between COPD and biomass smoke exposure in both case–control and cross-sectional groups even if this association is not statistically significant in case–control studies group.

Discussion

The present systematic literature review and meta-analysis re-examines existing research findings about the association between domestic biomass smoke exposure and COPD in women.

A total of 24 individual studies that evaluated COPD or CB as a health outcome in women, in the context of biomass fuel exposure compared with other fuels, were analysed.

Based on 19 cross-sectional and 5 case–control studies, the meta-analysis provides confirmation that exposure to indoor biomass fuel smoke is associated with an increased risk of COPD. Women exposed to biomass fuel smoke were more at risk of developing COPD or CB than those exposed to other fuels with a reported OR of 1.38.

In our analyses of location/geographical subgroups, we found a significant association between biomass smoke exposure and COPD for women living in rural as well as urban areas. In rural environments, women exposed to biomass smoke were more at risk of developing COPD than non-exposed women. The same pattern of association has been found in urban areas. Even if the crude ORs were different, one could not conclude that the risk was higher in the rural areas, as in rural areas OR was estimated to be 1.95 (95% CI 1.54 to 2.37) and in urban areas OR was estimated to be 1.61 (95% CI 1.20 to 2.02).

Although there was no significant heterogeneity among the studies included in the meta-analysis, we stratified their results by cigarette smoking status of participants included in each study and found a significant heterogeneity among studies involving both cigarette smokers and non-smokers and those involving only non-cigarette-smoking women. These subgroup analyses showed that associations did not differ between both groups of studies. The pooled OR increased to 1.90 (95% CI 1.35 to 2.45) for those including cigarette smokers and non-smokers and to 2.55 (95% CI 2.06 to 3.05) for those including only cigarette non-smokers. Some studies suggested that biomass smoke may interact with cigarette smoking in the pathogenesis of COPD.47

When considering COPD and CB separately, women were more at risk of developing COPD and CB if exposed to biomass fuel smoke compared with not-exposed women. The relationship between biomass smoke exposure and spirometry-defined and/or hospital-diagnosed COPD was not significant in studies, with OR values ranging from 0.97 to 5.18. Among the 14 publications evaluating the relationship between biomass exposure and the COPD phenotype, half of them did not show statistically significant association. This difference between studies may be due to difference in terms of design, difference in the way of conceiving exposed and non-exposed groups. We pooled only crude OR estimates, without adjustment for potential confounders. Consequently, potential confounding factors such as age, smoking status, body mass index (BMI), kitchen ventilation, socioeconomic status, history of tuberculosis and women educational attainment, which may have influenced the individual study’s result,40 48 49 are not being taken into account. To deal with that, we stratified studies by smoking status (studies involving both smokers and non-smokers and studies involving only non-smokers); the association remains not statistically significant in the two groups (see online supplementary figure S2). Concerning the link between BMI and lung function, the authors did not reach a consensus yet, for some there is an effect,50 leading to an eventual underestimation of COPD among those who are overweight and obese51; for others there is none.52 In addition, several studies have shown that low BMI is an important risk factor for COPD.49 53

The stratification of the publications by study design showed that exposure to biomass fuel smoke is associated with COPD (COPD and CB phenotypes) regardless of the study design with pooled OR of 1.82 (95% CI 1.54 to 2.10) for cross-sectional studies and 1.05 (95% CI 0.81 to 1.30) for case–control studies (see online supplementary figure S3). Although the association between COPD and biomass smoke exposure is not statistically significant in case–control studies group. Our finding can also be explained by the effect of hospital-diagnosed studies. In fact, Kurmi and colleagues14 have found in previous meta-analysis a non-significant association between solid fuel and hospital-diagnosed COPD (OR 2.29, 95% CI 0.70 to 7.52) comparatively with their pooled effect size for lung function-defined COPD which was largely significant (OR 2.96, 95% CI 2.01 to 4.37).

Among case–control studies, only one concerned CB diagnosis.26 Among the four others that studied COPD, two revealed non-significant association between biomass exposure and COPD,12 43 and one did not find any association.41 Case–control design is prone to bias, especially when hospital-based,24 when regarding selection bias. By selecting cases from the hospital population, selected individuals would not be representative of all possible cases happening within the population (severe cases might be more present). Less severe cases (that can be biomass smoke-exposed COPD), asymptomatic or never-smoking patients with COPD could have been overlooked, resulting in an underestimation of the association’s strength between COPD and biomass smoke. Xu et al41 reported an OR of 0.98 (95% CI 0.76 to 1.27), incoherent with most of the results reported in the literature on the relationship between exposure to biomass smoke and COPD. The author explains this difference by the fact that, in households using biomass or traditional fuels, kitchens were usually large and doors and windows were usually opened during cooking, and in urban areas kitchens were often properly ventilated.41 Consequently, air quality in these households is less affected by the use of biomass fuel, hence the absence of statistical difference between biomass users and electricity/gas users concerning COPD. Another explanation of the absence of difference could be the small sample size of some studies.12 43 45

No significant publication bias was found in the studies. Additionally, no significant heterogeneity was found for the group of studies included within the meta-analysis. However, heterogeneity was found after stratification due to the small number of studies corresponding to each subgroup. At this case, random-effect analysis was performed. Most of the studies were cross-sectional, decreasing the level of evidence of the meta-analysis. However, rigorous statistical methods were used to get accurate results.

This paper gathers 25 different results from 24 different publications. The publication period ranged from 1991 to 2015, representing almost 25 years. In comparison with the previous meta-analysis, additional studies were found. The large studies set of this systematic literature review enables us to better evaluate the association between biomass smoke exposure and COPD diagnosis in women compared with previous the meta-analysis based on nearly a dozen studies.14 54 The fact that our study focused especially on women allows us a more detailed analysis concerning this group of subjects than general studies involving all adult subjects.

Although our estimation of the OR was slightly lower than the OR found in the previous meta-analysis, the trend of association between domestic biomass exposure and COPD/CB was consistent with these studies.14 47 54 The findings confirm that use of clean fuels such as LPG, gas or electricity can reduce the risk of COPD in women. However, some studies highlight that clean cooking fuel choice was significantly associated with household socioeconomic status (such as income and education) and location (urban vs rural).55 To be effective, interventions aimed at reducing impact of biomass fuel on COPD must take this into account. Therefore, in poor or rural communities, improving the efficiency of current fuel stoves and energy user behaviours (fuel drying, avoiding smoke exposure as much as possible during cooking, improved kitchen ventilation, properly used and maintained stoves, promoting outdoor cooking) will be more effective in reducing smoke emission and exposure rather than by attempting to replace the solid fuel stoves with any clean fuel stoves.56 57 However, ‘for communities benefiting from a cheaper and more reliable access to clean fuels, then strategies to support a switch to LPG or other liquid or gaseous fuels have a higher chance of success’.56 A 9-year prospective cohort study performed in southern China revealed that improving kitchen ventilation and biomass stoves was associated with a reduced decline in FEV1 by 13 mL/year (95% CI 4 to 23 mL/year) compared with those who took up neither intervention. According to the same study, the use of clean fuels (biogas) instead of biomass for cooking reduced the FEV1 decline by 12 mL/year (95% CI 4 to 20 mL/year).58 ‘Compared with participants without improved ventilation for cooking, those with improvement for 5–9 y had a lower risk of COPD, with an adjusted OR of 0.39 (95% CI, 0.15 to 0.99)’, and there was no significant difference between the clean fuel and ventilation interventions.58 Aunan’s study highlights that a significant difference in COPD prevalence was observed between women who used stoves without chimney compared with those who used improved stoves (with a chimney), after adjustment for age, socioeconomic status and ventilation (OR 3.48, 95% CI 1.02 to 11.90).59 Chapman et al found that installation of a chimney was associated with a reduction in the incidence of COPD among women compared with people who did not have chimneys (relative risk 0.75 (0.62 to 0.92, P=0.005)).60

These results suggested that attention must be paid to the burden faced primarily by women in relation to traditional fuel like biomass fuels and traditional stoves use, particularly in rural areas. Cleaner energy for cooking is women’s and children’s respiratory health improvement, but not only! Improving traditional cooking stoves could be a more accepted and less cost-effective alternative, especially in the context of poverty. Healthier cooking means also environment protection and economic empowerment.

Conclusion

This study confirms that biomass smoke exposure is associated with COPD in women. An increased attention must be paid to cooking energy and cooking stoves improvement in view of the burden primarily faced by women in relation to traditional fuels like biomass and traditional stoves use, particularly in rural areas.

Additional studies with well-designed lung function measurement methods are needed to further highlight the causal link between lung function-diagnosed COPD and indoor exposure to biomass.

Many low-income, middle-income countries face an important energy transition geared towards energy substitution. Modern fuels such as LPG are more and more used but mostly used alongside traditional ones, mainly in urban areas. Additional data are needed to further explore the benefit of the usage of mixed fuel for cooking or heating on respiratory health, particularly on COPD reduction, in a context of energy transition, as seen more and more in many low-income, middle-income countries.