Introduction

Acute lower respiratory infections (ALRIs), predominantly pneumonia, account for more deaths among children under-five than any other infectious disease [1]. India has almost three times the number of ALRI deaths of any other country [2]. Vaccination is an important tactic to reduce childhood bacterial pneumonia and the conjugate pneumococcal and Hemophilus influenzae B vaccinations are increasingly available to children in low- and lower-middle-income countries [3,4]. Both vaccines are recommended in India [5], but the conjugate pneumococcal vaccine was not available outside the private sector until May 2017 [6]. While the immune response to the vaccines are developing in the first six months of life [7], there is an important role for other strategies to prevent ALRIs.

Household air pollution (HAP) from cooking and heating the home with solid fuels such as wood, crop waste, coal and animal dung on open fires or inefficient cookstoves is responsible for more disease and illness than any other environmental risk factor worldwide [7], including India [8]. The inefficient burning of solid fuels releases harmful levels of pollutants, including fine particulate matter [9]. Fine particulate matter, or PM_2.5, can travel deep into the lungs and enter the blood stream when inhaled leading to increased risk for respiratory and cardiac diseases [10]. The World Health Organization set the 24-hour standard for mean concentration of fine particulate matter in the ambient air at 25 μg/m³[11]. Rural households burning solid fuels can reach mean daily PM_2.5 concentrations of 609 μg/m³ in the kitchen [12]. Women and children have the highest personal exposures to HAP because of the time they spend inside the house and in the proximity of cookstoves [12]. HAP exposure due to cooking from solid fuels increases the risk for ALRIs among children [13,14], although a recent meta-analysis reported no benefit of improved cookstoves on ALRIs either [15]. In India, there is evidence that exposure to HAP increases the incidence of life-threatening respiratory illnesses, such as ALRIs, in children under the age of six [16]. While no evidence exists that quantifies the risk between duration of exposure to HAP and ALRI incidence within the Indian context, this relationship has been observed in Sierra Leone [17].

The critical need to reduce HAP has been recognized in recent years and was reaffirmed by the 2018 Global Burden of Disease Study [18]. The Global Alliance for Clean Cookstoves is leading the effort in driving adoption of clean cookstoves and fuel. However, rollout and acceptance of clean cooking alternatives is progressing slowly [19]. The Indian Government’s initiative Pradhan Mantri Ujjwala Yojina, launched in May 2016, is already providing access to clean cooking gas for about 31 million of the 50 million households below the poverty line [20]. However, there are an additional 50 million households in India still using solid fuel and the children living there are at risk of death from ALRI. If there is a direct association between duration of exposure to pollutants from use of solid fuel and risk of developing ALRI, potential interim strategies need to be considered such as educating the mothers to limit the child’s exposure to burning solid fuel, reducing the number of hours of cooking time or encouraging other household members to care for the infant away from the stove while the mother is cooking.

The Eunice Kennedy Shriver National Institute of Child Health and Human Development’s Global Network for Women and Children’s Health Research supports a Maternal and Newborn Health (MNH) Registry of pregnant women and their babies living in rural communities in six low- and lower-middle-income countries [21]. In the Nagpur site in India, we have the Institutional Review Board approval to follow the infants born to women enrolled in the Registry for six months after birth. Therefore, the Nagpur site of the Registry provides an ideal population to address questions about risk factors for ALRI in rural communities where about two thirds of households use solid fuels [22]. We hypothesized that the number of hours per day a child’s mother spent in front of a lit or smoldering solid fuel cookstove would be an easily measurable proxy for duration of exposure of the child to HAP as the child is often in the house while the mother is cooking [17]. The aim of the HAP-ALRI add-on study to the MNH Registry is to determine whether there is an association between duration of time in front of burning solid fuel and ALRI in the rural villages near Nagpur, India.

Materials and methods

Study design and setting

The HAP-ALRI add-on study was a prospective observational study that enrolled children born to women participating in the Global Network for Women’s and Children’s Health Research MNH Registry Study (registered on clinicaltrials.gov under NCT01073475) in Nagpur, India and followed them from birth through 6 months of age. Nagpur is approximately 310 meters above sea-level and there is little variation in altitude between study villages. Nagpur has an ALRI season that runs from October through March during which temperatures are cooler, averaging 68–83°F (20–28°C), and there is less rainfall compared to the rest of the year.

The details of the MNH Registry Study can be found elsewhere [21]. Briefly, women are enrolled in the Registry as early as possible in their pregnancy. In the Nagpur site, the women are recruited from villages in the neighboring areas of 20 rural Primary Health Centers (PHCs) that served a population of approximately 536,000 in 2010 (Fig 1). Women are followed through labor and delivery and until 42 days postpartum to collect morbidity and mortality data on the mother and newborn using MNH data collection forms.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Four districts in Nagpur and locations of primary health centers.

https://doi.org/10.1371/journal.pone.0224374.g001

Results

Study population

The study population consisted of babies born between August 20, 2013 and March 2, 2014 to mothers enrolled in the MNH Registry in Nagpur, India. Of the 1,667 children enrolled in the study, 81 were excluded due to multiple or missing gestation or missing information on household primary stove, hours spent in front of solid fuel cookstove, ALRI status or covariates (Fig 2). Of the remaining 1,586 children, 302 (19%) experienced at least one ALRI episode during the study while 1,284 children experienced no episodes.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Study population flow diagram.

https://doi.org/10.1371/journal.pone.0224374.g002

Overall, 107 (6%) of the households used only clean cookstoves, 930 (56%) used only polluting cookstoves and 628 (38%) used a mixture of clean and polluting cookstoves (stove stacking). Of those using a mixture of clean and polluting cookstoves, 63% considered the clean stove their primary stove and 37% considered the polluting cookstove their primary stove.

Only 6% (n = 96) of our population moved to another residence for at least 30 days during their pregnancy. These numbers are not sufficient to conduct a secondary analysis. For the women who moved, the distribution of our exposure variable is similar between residences (S1 Table).

Table 1 displays the distribution of variables by outcome. Compared with children with no episodes of ALRI, children with ≥1 ALRI were less likely to live in homes with a clean cookstove (21.8% vs. 33.7%; p<0.001) and be in the rich wealth category (32.1% vs. 42.4%; p<0.01), and more likely to live in homes with smokers (18.9% vs. 14.1%; p = 0.04) or be born to mothers aged <25 (67.9% vs. 59.6%; p<0.01). Sex of the child, birth weight and parity of their mother are similarly distributed in the two groups.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Background characteristics of 1586 children followed from August 20, 2013 to March 2, 2014.

https://doi.org/10.1371/journal.pone.0224374.t001

Table 2 displays the results from the multivariable logistic regression models that included the main effects. Interaction terms were excluded from the final models since none of the tested interactions showed a substantial effect (exposure and smoking p = 0.06; exposure and birth weight 0 = 0.13; exposure and wealth p-0.44). The adjusted results indicated that the odds of ≥1 ALRI increased from 1.2 (95% CI: 0.7–2.2) for 1-≤2 hours to 2.1 (95% CI: 1.4–3.3) for >3 hours in front of a solid fuel cookstove (p<0.01) and with increasing poverty from 1.2 (95% CI: 0.9–1.6) among the middle group to 1.4 (95% CI: 1.2–1.7) among the poor group (p<0.001).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Adjusted ORs and 95% CIs to predict children with ≥1 ALRIs.

https://doi.org/10.1371/journal.pone.0224374.t002

Discussion

In our large observational study in rural Central India, we found that after adjusting for other covariates, the odds of a child having ≥1 ALRI increases from 1.2 (95% CI: 0.7–2.2) for ≤1 hour to 2.1 (95% CI: 1.4–3.3) for >3 hours that the mother spends in front of a solid fuel cookstove (p<0.01) compared with those with no exposure. Our study showed that infants are usually inside the house when their mother is cooking, and in poorly ventilated households. another study has shown that this increases the risk for ALRI [12].

Our results are consistent with detailed and costly exposure measurements made in the RESPIRE study where Smith et al [14] found a direct exposure-response relationship between HAP and ALRIs, with increasing exposure, measured by personal carbon monoxide exposure, being associated with increased risk of ALRIs. However, as noted above, even though our results are consistent with the RESPIRE study, the results need to be confirmed in a study focused on the relation between duration of time spent cooking and the measures of personal exposure and ALRI. In Nepal, Bates et al [27] found a similar relationship between exposure to biomass cooking fuel and ALRI by measuring PM_2.5 concentrations in household kitchens. Rylance et al [28] found that particulate exposure caused a dose-dependent increase in inflammatory cytokine production and decrease in phagocytosis of S. pneumoniae and oxidative burst, which would be consistent with HAP exposure interfering with an immune response to the pneumococcus. Our study suggests that a mother’s duration of cooking could be a simple proxy marker of exposure to household air pollutants in assessing risk of ALRI in young infants when a direct measure of the child’s exposure is not available. However, additional studies are needed to understand the complexities of measures of air quality, the actual exposure of the child to particulate matter and CO and other pollutants combined with improved assessment of the diagnosis of ALRI in relation to duration of cooking.

Our study has several strengths. First, we followed a large number of children using the Maternal and Newborn Health Registry study which has been enrolling and following women and their babies since 2009 and is run by highly trained study staff who receive on-going central and refresher trainings to ensure consistency of the data. Second, we adjusted for confounding, particularly wealth and tobacco smoking using detailed information on wealth as defined in DHS [26]. Third, we used a daily diary to improve mothers’ recall of number of days and timing of illness.

Our study also has some limitations. First, our definition of acute lower respiratory infection relied on maternal recall instead of physician diagnosis or chest x-ray. While some mis-classification of ALRI may have occurred, we expect it to be non-differential by the cookstove as ASHA workers were not trained to specifically look for ALRIs in houses for the type of stove and were also unaware of the study hypothesis. Second, use of solid fuels in the household is only a proxy of actual exposure to HAP and it would have been ideal to have actual exposure measurements to further evaluate the exposure-response relationship. Actual measurements would have also helped us to better classify households with clean cookstoves according to their true exposure level based on other sources of HAP. However, over 95% of households in our population have electricity so we feel confident that sources other than the cooking stove such as kerosene lamps are not a major contributor to HAP. Similarly, since we did not measure the child’s actual exposure to HAP, duration of mother’s cooking with a biomass stove is only a proxy of the child’s exposure and as noted above the actual exposure of the child to HAP and duration of cooking in the household needs to be studied further. Ambient air pollution data were not collected as part of this study. We did not specifically study lack of breastfeeding in our population because 99.4% of the children were exclusively breastfed during the study period. Similarly, we were not able to study maternal smoking status as over 99% of mothers reported not smoking. We also did not specifically evaluate immunization status because the conjugate pneumococcal vaccine was not widely available during the study period (only 15% of the children had received at least one of the three doses by age 6 months), but the H influenza vaccine status was up to date in 45% of the children during the study period. Kitchen characteristics were similar across our sample (75% cooked in a separate room used as a kitchen). Finally we were unable to study the effect of undernutrition because we did not conduct anthropometry during this study.

In summary, our study indicates an association between increased duration of exposure to HAP and development of ALRI in an infant under the age of 6 months. We believe that until HAP can be measured accurately, assessing a child’s exposure to HAP based on the mother’s duration of cooking using a biomass stove when the child is inside the household may be a simple way of evaluating the infant’s risk of ALRI. Until households can access clean cooking fuels, finding other ways to reduce an infant’s time indoors while a biomass stove is in use could decrease episodes of ALRI. The challenge will be finding logistically feasible ways for this to occur.

Conclusions

The findings of our study indicate that increasing the time that a mother spends cooking over a biomass stove while her child is in the house may be associated with the child developing ALRI. Finding ways to reduce the number of hours that infants are exposed to polluting fuels during cooking time, such as educating mothers to have other household members care for the infant away from the stove while the mother is cooking could reduce the risk of ALRI in children under six months of age. This approach could provide an interim solution until all Indian households can access clean cooking fuel. How feasible this solution is in the Indian context needs to be further studied.

Supporting information