Relationship between pulmonary function and indoor air pollution from coal combustion among adult residents in an inner-city area of southwest China.
Study subjects
Nonsmoking adult residents of coal-using households, YA Road, Zunyi city, Guizhou, China
110 adults from the 52 coal-using households; 64.5% female, age range 20 to 74 years, mean height 163.7 +/- 6.6 cm, mean weight 61.1 +/- 7.1 kg. All were nonsmokers identified by questionnaire, aged 18 years and older, and had lived in YA Road for more than 3 years. Mean 4.4 +/- 1.2 persons and 3.9 +/- 1.2 rooms per home. 30.9% had a history of recurrent pulmonary tract infections and 32.7% had environmental tobacco smoke exposure (a smoker in the household). Age did not differ significantly from the control group (P = 0.08), nor did height (P = 0.23) or weight (P = 0.16); the number of rooms per home (P < 0.05) and the prevalence of recurrent pulmonary tract infections (P < 0.05) did, and household size was borderline - the Results text gives P = 0.053 for it while Table 1 prints P = 0.06, one of three places where the text and Table 1 disagree.
Cross-sectional study of adult residents of an inner-city area (YA Road) of Zunyi city, Guizhou province, southwest China - a traditional industrial city with abundant coal reserves and the heaviest acid-rain pollution in China - with all respiratory measurements made during the winter of October 2011 to March 2012
Nonsmoking adult residents of households using modern energy sources, YA Road, Zunyi city, Guizhou, China
121 adults from the 52 households using modern energy sources, selected by simple random sampling from the people using modern energy sources in the same two residence communities; 57.9% female, age range 19 to 71 years, mean height 164.8 +/- 6.8 cm, mean weight 62.6 +/- 7.5 kg. All were nonsmokers identified by questionnaire, aged 18 years and older, and had lived in YA Road for more than 3 years. Mean 4.1 +/- 1.3 persons and 4.4 +/- 0.9 rooms per home. 13.2% had a history of recurrent pulmonary tract infections and 25.6% had environmental tobacco smoke exposure.
Cross-sectional study of adult residents of an inner-city area (YA Road) of Zunyi city, Guizhou province, southwest China, with all respiratory measurements made during the winter of October 2011 to March 2012
Lung function assays
Spirometry - adult residents of coal-using households (coal-exposed group)
Ventilatory function in the 110 nonsmoking adults from coal-using households, measured during the winter of October 2011 to March 2012: FEV1 2.75 +/- 0.56 L, FVC 3.22 +/- 0.62 L, FEV1/FVC 83.13 +/- 3.95%, PEFR 6.00 +/- 0.90 L/s (Table 1). All four parameters were lower than in the 121 non-coal controls; Table 1 marks each difference P < 0.05 while the Results text calls the same comparisons highly significant at P < 0.0001. The PM2.5 relative concentration measured during cooking in these households was 569.5 +/- 50.1 cpm in the kitchen, 543.2 +/- 65.0 cpm in the living room and 416.2 +/- 5.9 cpm outdoors; the kitchen value was significantly higher than both the living room and outdoors (P < 0.0001). Cooking was mainly done indoors on an open-fire traditional stove in a small kitchen without effective ventilation, or in a kitchen adjacent to the living room, and most coal used by Zunyi households is a high-sulfur coal of approximately 3% sulfur content. This group also had a significantly higher prevalence of recurrent pulmonary tract infections (30.9% vs 13.2%, P < 0.05) and fewer rooms per home (3.9 +/- 1.2 vs 4.4 +/- 0.9, P < 0.05); environmental tobacco smoke exposure did not differ significantly (32.7% vs 25.6%, P = 0.23).
Exposure conditions
| Condition | Agent | Concentration | Duration |
|---|---|---|---|
| Indoor air pollution from domestic coal combustion for cooking and heating, quantified as the PM2.5 relative concentration measured in the kitchen during cooking (living room 543.2 +/- 65.0 cpm; outdoors 416.2 +/- 5.9 cpm); exposure duration not assessable, as the authors could not determine cooking times EXPOSURE:jie-coal-smoke-indoor |
fine respirable suspended particulate matter ENVO:01000415 |
569.5 counts per minute (mean; standard deviation ±50.1 counts per minute; n=52) | — |
Outputs
| Group | Output | Description |
|---|---|---|
| treated | Measured FEV1, FVC, FEV1/FVC and PEFR in adults from coal-using households LF:jie-coal-exposed-spirometry-output-coal · under EXPOSURE:jie-coal-smoke-indoor |
fev1: 2.75 liter (mean; standard deviation ±0.56 liter; n=110)
fvc: 3.22 liter (mean; standard deviation ±0.62 liter; n=110)
fev1 fvc ratio: 83.13 percent (mean; standard deviation ±3.95 percent; n=110)
peak expiratory flow: 6.00 liter per second (mean; standard deviation ±0.90 liter per second; n=110)
|
Evidence
FEV1 (L) 2.75 ± 0.56 3.49 ± 0.77 ,0.05 FVC (L) 3.22 ± 0.62 3.87 ± 0.70 ,0.05 FEV1/FVC (%) 83.13 ± 3.95 91.63 ± 5.46 ,0.05 PEFR (L/s) 6.00 ± 0.90 6.80 ± 1.17 ,0.05
Table 1 rows giving the mean +/- SD of FEV1, FVC, FEV1/FVC and PEFR for the coal-exposed group (first column, recorded on this assay's output) and the non-coal group (second column, recorded on the companion control assay), with Table 1's P < 0.05 marker.
Coal (++) (n=110) Coal (– –) (n=121) P
Table 1 column header establishing that the coal-exposed column has n = 110 and the non-coal column n = 121, the sample sizes recorded on the two assays' outputs.
In a cross-sectional study of 104 households, pulmonary function measurements were assessed and compared in 110 coal users and 121 non-coal users (≥18 years old) who were all nonsmokers.
States the study design, the two group sizes and the two eligibility facts carried on this assay's study_subject: adults 18 years and older, all nonsmokers.
(P ,0.0001) were observed between coal-exposed and unexposed individuals, with mean values lower in those exposed to coal (Table 1).
The Results text's reading of the crude group differences in all four pulmonary function parameters, at P < 0.0001, and the direction of the difference; Table 1 marks the same comparisons P < 0.05, and both readings are noted in the description.
The indoor and outdoor PM2.5 relative concentrations during the cooking period in the coal-exposed and unexposed groups were 569.5±50.1 vs 477.0±58.5 cpm (kitchen), 543.2±65.0 vs 452.7±51.2 cpm (living room), and 416.2±5.9 vs 417.0±5.1 cpm (outdoors), respectively.
Gives the kitchen PM2.5 relative concentration (569.5 +/- 50.1 cpm) recorded as this condition's exposure_concentration, together with the living room and outdoor values carried in the condition's name and the description.
During coal combustion, the relative concentration of PM 2.5 in the kitchen was also significantly higher than in the living room and outdoors (P ,0.0001).
Supports the within-home gradient stated in this assay's description: the kitchen is the highest-exposure room during coal combustion.
Significant differences were present for the number of rooms in the home (P ,0.05) and for individuals who had recurrent pulmonary tract infections (30.9% in exposed vs 13.2% in unexposed groups, Table 1).
Supports the two significant baseline imbalances noted in this assay's description and study_subject: fewer rooms per home and a higher prevalence of recurrent pulmonary tract infections in the coal-exposed group.
There was an indication of an imbalance in the two groups regarding the number of persons living in the home (P=0.053).
The Results text's p-value for the household-size imbalance recorded on this assay's study_subject, P = 0.053.
Persons living in a home 4.4 ± 1.2 4.1 ± 1.3 =0.06
Table 1's row for the same household-size comparison, which prints P = 0.06 where the Results text gives P = 0.053 - the third of the three text-vs-table p-value disagreements flagged in the file header. Also gives the mean household size of both cohorts.
Cooking is mainly done either indoors, with an open-fire traditional cooking stove in a small kitchen without effective ventilation, or in an open or closed kitchen adjacent to the living room within the house.
Describes the exposure setting recorded in this assay's exposure condition and description: unventilated indoor combustion adjacent to living space.
All respiratory measurements were performed during the winter (from October 2011 to March 2012).
Fixes the measurement season recorded in this assay's description and clinical_context, the period of heaviest domestic coal use.
Most of the coal used by Zunyi households is a high-sulfur coal with a sulfur content of approximately 3% (13).
Characterises the fuel behind the coal-smoke exposure recorded on this assay, noted in the description.
Third, we cannot comment on the duration of exposure to coal smoke because we could not adequately assess cooking times in households.
Justifies omitting exposure_duration from this assay's coal-smoke condition: the study could not assess how long participants were exposed.
Our main finding was impaired pulmonary function in both adult males and females who used coal compared to non-users.
The authors' own statement of the finding this assay and its companion record, and of the key event it informs (decreased lung function in coal users of both sexes).
Second, when interpreting the results of this study, it is important to note that the cross-sectional design limited causal inference.
Supports recording this paper as a cross-sectional association only, and the decision to record no key_event_relationships from it.
Multivariate linear regression analysis was used to assess the association between coal combustion and pulmonary function with other sociodemographic factors (e.g., age, gender, height, and exposure to tobacco smoke, etc.) as variable controls.
Identifies the statistical_test recorded on every response comparison in this file: multivariate linear regression of pulmonary function on coal combustion with sociodemographic covariates.
In multivariate analysis, current exposure to coal smoke was associated with a 31.7% decrease in FVC, a 42.0% decrease in FEV₁, a 7.46% decrease in the FEV₁/FVC ratio, and a 23.1% decrease in PEFR in adult residents.
The paper's prose framing of the four adjusted coal coefficients as percent decreases. Tables 2-5 print the same four numbers as absolute coefficients (-0.317 L, -0.420 L, -7.463 percentage points, -0.231 L/s), which is how the comparisons record them; this quote documents the discrepancy rather than resolving it.
Coal (yes vs no) – –0.317 – –0.396; ––0.238 ,0.0001
Table 2 row giving the adjusted FVC coefficient for coal use, -0.317 L (95%CI -0.396; -0.238), P < 0.0001, recorded on COMP:jie-fvc-coal-vs-noncoal.
Coal (yes vs no) – –0.420 – –0.525; ––0.316 ,0.0001
Table 3 row giving the adjusted FEV1 coefficient for coal use, -0.420 L (95%CI -0.525; -0.316), P < 0.0001, recorded on COMP:jie-fev1-coal-vs-noncoal.
Coal (yes vs no) – –7.463 – –9.797; ––5.129 ,0.0001
Table 4 row giving the adjusted FEV1/FVC coefficient for coal use, -7.463 percentage points (95%CI -9.797; -5.129), P < 0.0001, recorded on COMP:jie-fev1fvc-coal-vs-noncoal.
Coal (yes vs no) – –0.231 – –0.460; ––0.002 ,0.05
Table 5 row giving the adjusted PEFR coefficient for coal use, -0.231 L/s (95%CI -0.460; -0.002), P < 0.05, recorded on COMP:jie-pefr-coal-vs-noncoal; the P < 0.05 threshold is why that comparison carries no p_value.
In the multivariate models, we found that PM 2.5 relative concentrations in the kitchen showed a significant effect on FVC, FEV 1, and PEFR.
Establishes which three of the four pulmonary function parameters the kitchen PM2.5 relative concentration significantly affected, and by omission why the FEV1/FVC slope is recorded as no_change.
Our models indicate that there was a 2-L decrease in FVC and a 3-L decrease in FEV1, as well as an 8 L/s decrease in PEFR per 1 cpm PM2.5 exposure among adult participants.
The paper's prose framing of the three significant kitchen PM2.5 slopes. The tabulated coefficients are -0.002 L, -0.003 L and -0.008 L/s per cpm, a thousandfold smaller than the litre figures stated here; the comparisons record the tabulated values and this quote documents the discrepancy.
PM2.5 relative concentration in kitchen (cpm) – –0.002 – –0.004; 0.000 ,0.05
Table 2 row giving the FVC coefficient per 1 cpm of kitchen PM2.5, -0.002 L (95%CI -0.004; 0.000), P < 0.05, recorded on COMP:jie-fvc-per-cpm-kitchen-pm25.
PM2.5 relative concentration in kitchen (cpm) – –0.003 – –0.005; 0.000 ,0.05
Table 3 row giving the FEV1 coefficient per 1 cpm of kitchen PM2.5, -0.003 L (95%CI -0.005; 0.000), P < 0.05, recorded on COMP:jie-fev1-per-cpm-kitchen-pm25.
PM2.5 relative concentration in kitchen (cpm) – –0.008 – –0.013; ––0.003 ,0.05
Table 5 row giving the PEFR coefficient per 1 cpm of kitchen PM2.5, -0.008 L/s (95%CI -0.013; -0.003), P < 0.05, recorded on COMP:jie-pefr-per-cpm-kitchen-pm25.
PM2.5 relative concentration in kitchen (cpm) 0.008 – –0.044; 0.060 0.757
Table 4 row giving the null FEV1/FVC coefficient per 1 cpm of kitchen PM2.5, 0.008 percentage points (95%CI -0.044; 0.060), P = 0.757, recorded on COMP:jie-fev1fvc-per-cpm-kitchen-pm25.
Spirometry - adult residents of households using modern energy sources (control group)
Ventilatory function in the 121 nonsmoking adults from households using modern energy sources rather than coal: FEV1 3.49 +/- 0.77 L, FVC 3.87 +/- 0.70 L, FEV1/FVC 91.63 +/- 5.46%, PEFR 6.80 +/- 1.17 L/s (Table 1). These are the comparator values for the coal-exposed assay. The PM2.5 relative concentration measured during cooking in these households was 477.0 +/- 58.5 cpm in the kitchen, 452.7 +/- 51.2 cpm in the living room and 417.0 +/- 5.1 cpm outdoors; indoor values were significantly lower than in the coal-using homes - the Results text gives P < 0.0001 for the kitchen and living room contrast while Table 1's own PM2.5 rows are marked only P < 0.05, the same text-vs-table disagreement the header flags for the spirometry rows - whereas the outdoor concentrations of the two groups did not differ (417.0 +/- 5.1 vs 416.2 +/- 5.9 cpm, P = 0.480) - the contrast between the groups is therefore an indoor one, not a difference in ambient air. These households had been drawn from the same inner-city area, where biomass fuels have largely been replaced by kerosene, liquid petroleum, gas or electricity while coal remains a major cooking and heating fuel.
Exposure conditions
| Condition | Agent | Concentration | Duration |
|---|---|---|---|
| Households using modern energy sources (kerosene, liquid petroleum, gas or electricity) instead of coal; control condition, with the PM2.5 relative concentration measured in the kitchen during cooking (living room 452.7 +/- 51.2 cpm; outdoors 417.0 +/- 5.1 cpm) COND:jie-noncoal-modern-fuel |
fine respirable suspended particulate matter ENVO:01000415 |
477.0 counts per minute (mean; standard deviation ±58.5 counts per minute; n=52) | — |
Outputs
| Group | Output | Description |
|---|---|---|
| control | Measured FEV1, FVC, FEV1/FVC and PEFR in adults from non-coal households LF:jie-noncoal-control-spirometry-output-control · under COND:jie-noncoal-modern-fuel |
fev1: 3.49 liter (mean; standard deviation ±0.77 liter; n=121)
fvc: 3.87 liter (mean; standard deviation ±0.70 liter; n=121)
fev1 fvc ratio: 91.63 percent (mean; standard deviation ±5.46 percent; n=121)
peak expiratory flow: 6.80 liter per second (mean; standard deviation ±1.17 liter per second; n=121)
|
Evidence
FEV1 (L) 2.75 ± 0.56 3.49 ± 0.77 ,0.05 FVC (L) 3.22 ± 0.62 3.87 ± 0.70 ,0.05 FEV1/FVC (%) 83.13 ± 3.95 91.63 ± 5.46 ,0.05 PEFR (L/s) 6.00 ± 0.90 6.80 ± 1.17 ,0.05
Table 1 rows; the second numeric column gives the non-coal control group's mean +/- SD for FEV1 (3.49 +/- 0.77 L), FVC (3.87 +/- 0.70 L), FEV1/FVC (91.63 +/- 5.46%) and PEFR (6.80 +/- 1.17 L/s), the values recorded on this assay's output.
PM2.5 relative concentration in kitchen (cpm) 569.5 ± 50.1 477.0 ± 58.5 ,0.05 PM2.5 relative concentration in living room (cpm) 543.2 ± 65.0 452.7 ± 51.2 ,0.05 Outdoor PM 2.5 relative concentration (cpm) 416.2 ± 5.9 417.0 ± 5.1 =0.480
Table 1 rows giving the control group's kitchen PM2.5 relative concentration (477.0 +/- 58.5 cpm) recorded as this condition's exposure_concentration, and establishing that the outdoor concentrations of the two groups did not differ (P = 0.480).
In the inner-city areas of Zunyi, biomass fuels have been widely replaced at home by cleaner energy sources, such as kerosene, liquid petroleum, gas, or electricity, but coal is still a major source of fuel for cooking and heating throughout the year, particularly in the winter.
Identifies the modern energy sources that define this control condition and confirms that coal remains a major alternative fuel in the same inner-city area.
the relative concentrations of PM 2.5 in the kitchen and living room during coal combustion were significantly higher than during non-coal combustion (P ,0.0001), but this was not the case for outdoor concentrations.
The Results text's P < 0.0001 for the indoor between-group PM2.5 difference stated in this assay's description, and its confirmation that the outdoor concentrations did not differ; Table 1's PM2.5 rows mark the same indoor contrast only P < 0.05.
The individuals in the coal-exposed and unexposed groups were similar regarding demographic characteristics: 64.5% were female (exposed) vs 57.9% (unexposed), with age ranges of 20 to 74 (exposed) and 19 to 71 years (unexposed)
Gives the sex distribution and age range of both cohorts recorded on the two assays' study_subject records, and the authors' judgement that the groups were demographically similar.
Age (year) 43.7 ± 12.5 40.8 ± 12.4 =0.08
Table 1 row giving the mean age +/- SD of both cohorts (43.7 +/- 12.5 years coal-exposed, 40.8 +/- 12.4 years control) recorded on the two study_subject records, with the non-significant group difference.
PM 2.5 relative concentration during cooking was measured in 52 households in the coal-exposed group and 52 households in the unexposed group.
Gives the number of households behind each group's PM2.5 relative concentration, the sample_size recorded on both exposure conditions' concentrations.
Response comparisons
Adjusted (multivariate) difference in FVC for coal use vs no coal use
In the multivariate model, coal use was associated with an FVC coefficient of -0.317 L (95%CI -0.396; -0.238), P < 0.0001 (Table 2). The abstract and Results text describe this same number as 'a 31.7% decrease in FVC'; the tabulated coefficient is an absolute difference in litres, and it is recorded as such here. The paper reports only P < 0.0001, so the tabulated p_value 0.0001 is the stated upper bound. Crude group means were 3.22 +/- 0.62 L (coal) vs 3.87 +/- 0.70 L (non-coal); the control output lives on the non-coal assay.
Statistical test: multivariate linear regression adjusted for gender, age, weight, height, number of persons living in a home, number of rooms in a home, history of recurrent respiratory tract infections, environmental tobacco smoke exposure, and PM2.5 relative concentrations in the kitchen, living room and outdoors
Adjusted (multivariate) difference in FEV1 for coal use vs no coal use
In the multivariate model, coal use was associated with an FEV1 coefficient of -0.420 L (95%CI -0.525; -0.316), P < 0.0001 (Table 3). The abstract and Results text describe this same number as 'a 42.0% decrease in FEV1'; the tabulated coefficient is an absolute difference in litres, and it is recorded as such here. The paper reports only P < 0.0001, so the tabulated p_value 0.0001 is the stated upper bound. Crude group means were 2.75 +/- 0.56 L (coal) vs 3.49 +/- 0.77 L (non-coal); the control output lives on the non-coal assay.
Statistical test: multivariate linear regression adjusted for gender, age, weight, height, number of persons living in a home, number of rooms in a home, history of recurrent respiratory tract infections, environmental tobacco smoke exposure, and PM2.5 relative concentrations in the kitchen, living room and outdoors
Adjusted (multivariate) difference in FEV1/FVC ratio for coal use vs no coal use
In the multivariate model, coal use was associated with an FEV1/FVC coefficient of -7.463 percentage points (95%CI -9.797; -5.129), P < 0.0001 (Table 4). The abstract describes this as 'a 7.46% decrease in the FEV1/FVC ratio'; because the outcome is itself a percentage, the coefficient is a difference in percentage points rather than a relative decrease. The paper reports only P < 0.0001, so the tabulated p_value 0.0001 is the stated upper bound. Crude group means were 83.13 +/- 3.95% (coal) vs 91.63 +/- 5.46% (non-coal); the control output lives on the non-coal assay.
Statistical test: multivariate linear regression adjusted for gender, age, weight, height, number of persons living in a home, number of rooms in a home, history of recurrent respiratory tract infections, environmental tobacco smoke exposure, and PM2.5 relative concentrations in the kitchen, living room and outdoors
Adjusted (multivariate) difference in PEFR for coal use vs no coal use
In the multivariate model, coal use was associated with a PEFR coefficient of -0.231 L/s (95%CI -0.460; -0.002), P < 0.05 (Table 5) - the only one of the four coal coefficients whose confidence interval comes close to zero. The abstract describes this same number as 'a 23.1% decrease in PEFR'; the tabulated coefficient is an absolute difference in L/s, and it is recorded as such here. The paper gives only the P < 0.05 threshold, so p_value is omitted. Crude group means were 6.00 +/- 0.90 L/s (coal) vs 6.80 +/- 1.17 L/s (non-coal); the control output lives on the non-coal assay.
Statistical test: multivariate linear regression adjusted for gender, age, weight, height, number of persons living in a home, number of rooms in a home, history of recurrent respiratory tract infections, environmental tobacco smoke exposure, and PM2.5 relative concentrations in the kitchen, living room and outdoors
Adjusted (multivariate) change in FVC per 1 cpm increase in kitchen PM2.5 relative concentration
Each 1 cpm increase in the kitchen PM2.5 relative concentration was associated with an FVC coefficient of -0.002 L (95%CI -0.004; 0.000), P < 0.05 (Table 2). Note the discrepancy with the paper's own prose, which states 'a 2-L decrease in FVC ... per 1 cpm PM2.5 exposure': the tabulated coefficient is 0.002 L, i.e. about 2 mL per cpm, a thousandfold smaller than the stated litre figure. The tabulated value is recorded. The paper gives only the P < 0.05 threshold, so p_value is omitted. Continuous-exposure regression slope, so no control/treated output pair exists (see the section note).
Statistical test: multivariate linear regression, coefficient per 1 count per minute of kitchen PM2.5 relative concentration, adjusted for gender, age, weight, height, number of persons living in a home, number of rooms in a home, history of recurrent respiratory tract infections, environmental tobacco smoke exposure, and PM2.5 relative concentrations in the living room and outdoors
Adjusted (multivariate) change in FEV1 per 1 cpm increase in kitchen PM2.5 relative concentration
Each 1 cpm increase in the kitchen PM2.5 relative concentration was associated with an FEV1 coefficient of -0.003 L (95%CI -0.005; 0.000), P < 0.05 (Table 3). Note the discrepancy with the paper's own prose, which states 'a 3-L decrease in FEV1 ... per 1 cpm PM2.5 exposure': the tabulated coefficient is 0.003 L, i.e. about 3 mL per cpm. The tabulated value is recorded. The paper gives only the P < 0.05 threshold, so p_value is omitted. Continuous-exposure regression slope, so no control/treated output pair exists (see the section note).
Statistical test: multivariate linear regression, coefficient per 1 count per minute of kitchen PM2.5 relative concentration, adjusted for gender, age, weight, height, number of persons living in a home, number of rooms in a home, history of recurrent respiratory tract infections, environmental tobacco smoke exposure, and PM2.5 relative concentrations in the living room and outdoors
Adjusted (multivariate) change in PEFR per 1 cpm increase in kitchen PM2.5 relative concentration
Each 1 cpm increase in the kitchen PM2.5 relative concentration was associated with a PEFR coefficient of -0.008 L/s (95%CI -0.013; -0.003), P < 0.05 (Table 5) - the only one of the four kitchen PM2.5 coefficients whose confidence interval excludes zero on both sides. Note the discrepancy with the paper's own prose, which states 'an 8 L/s decrease in PEFR per 1 cpm PM2.5 exposure': the tabulated coefficient is 0.008 L/s, i.e. about 8 mL/s per cpm. The tabulated value is recorded. The paper gives only the P < 0.05 threshold, so p_value is omitted. Continuous-exposure regression slope, so no control/treated output pair exists (see the section note).
Statistical test: multivariate linear regression, coefficient per 1 count per minute of kitchen PM2.5 relative concentration, adjusted for gender, age, weight, height, number of persons living in a home, number of rooms in a home, history of recurrent respiratory tract infections, environmental tobacco smoke exposure, and PM2.5 relative concentrations in the living room and outdoors
Adjusted (multivariate) change in FEV1/FVC ratio per 1 cpm increase in kitchen PM2.5 relative concentration
Each 1 cpm increase in the kitchen PM2.5 relative concentration was associated with an FEV1/FVC coefficient of 0.008 percentage points (95%CI -0.044; 0.060), P = 0.757 - a null result, recorded because it is the one pulmonary function parameter on which kitchen PM2.5 showed no significant effect, consistent with the paper's statement that the significant effects were on FVC, FEV1 and PEFR. The point estimate is positive, so change_direction is no_change rather than decreased. Continuous-exposure regression slope, so no control/treated output pair exists (see the section note).
Statistical test: multivariate linear regression, coefficient per 1 count per minute of kitchen PM2.5 relative concentration, adjusted for gender, age, weight, height, number of persons living in a home, number of rooms in a home, history of recurrent respiratory tract infections, environmental tobacco smoke exposure, and PM2.5 relative concentrations in the living room and outdoors
Protocols
FGC-A++ portable spirometry against American Thoracic Society acceptance criteria
Lung function examined by measuring FVC, forced expiratory volume in 1 s (FEV1), FEV1/FVC and peak expiratory flow rate (PEFR) according to standard guidelines, using a portable electronic FGC-A++ spirometer (Anhui Institute of Electronic Science, China). Participants sat upright with both feet flat on the ground, were instructed to inhale completely, place the meter in the mouth and exhale with maximal force once the lips were sealed around the mouthpiece, with an investigator demonstrating the maneuver. Per American Thoracic Society guidelines, maneuvers were accepted only if back-extrapolated volume was low (<5% of the FVC and <0.15 L), both FVC and FEV1 were within 0.20 L of the best-effort FVC and FEV1, and little volume accumulated at the end of the effort. Each subject performed three expiratory maneuvers; analysis used the largest FVC, the largest FEV1, the ratio of the largest FEV1 to the largest FVC, and the largest PEFR.
Equipment: Portable electronic FGC-A++ spirometer (Anhui Institute of Electronic Science, China)
LD-3K light-scattering dust monitoring of PM2.5 relative concentration during cooking
PM2.5 concentrations inside and outside the households were measured during cooking with a digital dust monitor (LD-3K; Sibata Scientific Technology Inc., Japan), a portable light-scattering monitor with a laser diode as the light source that determines the relative concentration of PM2.5 from the intensity of the laser beam scattered by particles. The relative concentration is reported as counts per minute: cpm = count value / measuring time in minutes. All air sampling was done 1.2-1.5 m above the floor. Three measurements were taken 1 m from the centre of the cooking stove inside the kitchen, three from random sites inside the living room at least 3 m from the kitchen, and three outdoors at least 20 m from the household as environmental controls. Each measurement was maintained for over 1 min and three readings were averaged, with 1 min between the three readings in a room and 5 min between rooms.
Equipment: Digital dust monitor LD-3K (Sibata Scientific Technology Inc., Japan)
Sociodemographic and smoking questionnaire
Questionnaire administered at home after consent, used to assess sociodemographic factors and to identify smoking status (only nonsmokers were eligible), gender, history of recurrent respiratory tract infections, environmental tobacco smoke exposure (defined by the presence of a smoker in the household), number of persons living in the home and number of rooms in the home. Household cooking fuel source was identified for each household visited and used to assign the coal and non-coal categories.