Our results show that pregnant women living in Nairobi often face a variety of environmental exposures, regardless of educational attainment. In addition, women with lower educational attainment had greater exposure to certain pollutants. Overall, cooking emissions, vehicle exhaust, road dust, the Dandora dumpsite, secondhand smoke, pests, rubbish burning, and proximity to sewers were exposures most frequently reported by participants in residential, ambient, and work settings.

As expected, participants with higher educational attainment were more likely to earn a monthly salary and pay more rent, supporting the use of educational groups as a proxy for SES. Higher educational attainment may be related to some of the choices and behaviors that influence environmental exposures. For example, those with higher educational attainment had stays half as long as those with lower educational attainment, potentially reflecting a means to change their residence motivated by the pregnancy. Participants with higher educational attainment were more likely to report using LPG and less likely to report using kerosene in their own house. Similarly, participants with higher educational attainment were less likely to report work-related kerosene cooking exposures. Work-related exposure to rubbish burning also differed by educational attainment. However, higher educational attainment was less closely linked with exposures related to residential proximity (e.g., sewers and dumpsite), HAP (e.g., rubbish burning), AAP (e.g., cooking smoke and vehicle exhaust), and work (e.g., charcoal and wood cooking), as might be expected.

Studies in other African geographies offer important context for our observations. In a review of environmental health research on the continent, ambient air pollution, indoor air pollution, heavy metals, pesticides, and dietary mold were among the exposures most studied [36] and mirrored the exposures most often reported by participants in our study. In a study of fetal growth among healthy pregnant women at low risk of adverse pregnancy outcomes and environmental exposures, indicators of environmental health among Nairobi participants were somewhat better than our observations a decade later [37]. For example, housing was less dense (median 0.5 people per room versus 2 people per room), < 5% reporting mold/mildew odors (compared to 23% in this study), 18% reported peeling paint (compared to 26% observed in the current study), just over half reported pests such as mice rats and cockroaches (compared to 71% in this study), and 88.4% had gas and 9.4% had electric stoves [37], compared to the 65% using LPG and 2.8% using electricity most days to daily in our study.

Beyond Kenya, a systematic review in Ethiopia examined exposures during pregnancy and birth outcomes and reported that biomass fuels, environmental tobacco smoke, and pesticides were among those most commonly experienced and associated adverse birth outcomes [38]. Studies in Durban and Bloemfontein, South Africa examining birth outcomes, reported much higher use of electricity as a cooking energy source (> 90%) than we observed, though nearly all participants also reported cooking indoors exclusively [39,40]. However, smoking tobacco indoors was more prevalent in one of these studies compared to our results, where 49% of households permitted smoking indoors and 76% reported exposure to environmental tobacco smoke indoors [39], compared to our results where 95% reported exposure rarely to not at all and 13% reported a smoker in the home. Also more prevalent, was proximity to industrial sources (15%) [39]. Further study of the awareness, resources, and agency to reduce environmental exposures among pregnant women in sub-Saharan Africa is warranted.

Housing characteristics influence indoor air quality and HAP, as well as water and sanitation behaviors, which can contribute to poor environmental health conditions [41,42]. Nearly all participants had access to improved water sources (piped water) and sanitation (flush toilets rather than latrines) at or near their house – positive features for environmental health. However, one room houses were common and may not provide sufficient living space and increase risk of disease transmission and increased concentrations of pollutants. Furthermore, while many houses had windows and ventilation, emission-generating activities like cooking still occur near all other aspects of domestic life. Pests (such as cockroaches and rats) and mold were reported by 71% and 23% of participants, respectively, contributing to poor indoor air quality conditions. Twenty-six percent of participants reported chipping or peeling paint, which could pose a lead hazard despite the 2018 ban on lead in paint [43]. Half of participants also reported living in their residence less than fourteen months, potentially indicating precarity of tenure and housing quality.

Household fuel use is among the most important factors influencing HAP. In contrast to rural Kenyans, where over 90% rely mostly on wood [44,45], few participants reported wood use daily or most days (<1%). Despite this lower use of solid fuels, common urban fuels may also produce harmful emissions. One 2012–2014 study in Nairobi reported kerosene was the predominant fuel among 69.7% of residents in the informal settlements of Korogocho and Viwandani [46]. Ten years later in the current study, 32% reported kerosene use most days to daily, with higher usage (49%) among participants with lower educational attainment. Instead of solid fuels or kerosene, LPG was used by 65% of participants most days or daily, which is comparable to the 60% reported in the 2022 Kenya Demographic and Health Survey [47]. Such LPG use reflects the broader trend of LPG becoming a more common fuel in LMICs as international campaigns and governments promote cleaner fuels [48,49]. Furthermore, urbanization, socioeconomic status, educational attainment, and income have been associated with greater adoption of LPG [50,51], which may partly explain the high use of LPG among participants.

Cleaner stoves are not without risk, however; LPG users, particularly in low ventilation settings, may still be exposed to PM concentrations that exceed World Health Organization guidelines [52], as well as other contaminants including NO₂ [53], benzene [54], and a range of other hazardous air pollutants [55] which pose risks to human health. Additionally, users of cleaner cooking fuels may also use alternative fuels, such as kerosene, charcoal, or wood in fuel stacking [50,56]. In this study, 26 participants reported using both LPG and kerosene most days to daily. Kerosene cooking fuel use is associated with increased respiratory morbidity, adverse blood pressure effects, and mortality [57]. Beyond HAP, ambient air quality guidelines may not be achieved due to the practice of fuel stacking [7].

Despite the high proportion of women that reported using cleaner fuels in their own house, 73% reported being frequently exposed to outdoor cooking pollution, potentially from neighbors or businesses. Participants also frequently reported AAP sources including vehicle exhaust (66%), welding shop activity (32%), unpaved/dusty roads (32%), the Dandora dumpsite (32%), and rubbish burning (24%) within 1 km of their residence. Similar themes emerged from other recent reports of perceived air pollution, where rubbish burning, cooking fuels, dump sites, and industry were top concerns [58,59]. In contrast, in this study industry was one of the least reported sources of exposure (4%). In a 2013 study in Nairobi, researchers observed low awareness of air pollution, with 35–55% having never heard information about air pollution exposure, despite the common practice of waste burning from a nearby dumpsite [58]. Our results possibly indicate a growing awareness of environmental exposures over time.

Main work-related or occupational exposures reported by participants were vehicle exhaust and unpaved/dusty roads; this experience is supported by prior research in which women working alongside roads and potentially walking to work were exposed to PM_2.5 concentrations of 62.7 ± 20.5 µg/m³ from resuspended road dust and refuse burning in Nairobi [60]. Another study linked work as a street vendor with a reduction in birth weight [61]. In the present study, a common work-related exposure was smoke from tobacco and marijuana, with 52% of participants reporting exposure at least some days, despite there being few active smokers in this study (1.8%), and smoke from marijuana and cigarettes in the house not being frequently identified (3.8% reported exposure at least some days). Exposure to cigarette smoke only (56%) was higher in our results than other studies where 12–19% [58] and 3.2% [46] of Nairobi women reported exposure to cigarette smoke in the workplace. In the same study, 14.7% of participants reported one or more members of their household smoked, with 6.7% reporting smoking inside the house [46]. Our results highlight that despite being a well-recognized prenatal exposure of concern that should be eliminated, exposure to secondhand smoke is still prevalent in Nairobi among pregnant women, especially in work settings.

Exposure to rubbish burning is an exposure of particular concern, occurring within households, outdoors, and in work contexts [62,63]. Encouragingly, 70% of participants reported burning rubbish at home rarely to not at all. For the 26% of participants that reported this practice some days and 4.5% reporting it most days to daily, there is an opportunity for education to reduce this potentially hazardous exposure. A higher percentage of participants reported exposure to rubbish burning in the ambient environment and workplace (~40% some days and ~25% most days to daily). Reducing these exposures could be more challenging, potentially requiring community action, and is compounded by the proximity and operations of the Dandora dumpsite.

Seventy-four percent of participants reported their highest level of education was secondary school, compared with 50% of women in the 2022 Kenyan Demographic and Health Survey [47]. This difference, combined with the potential for participants’ spouses to be similarly educated [64,65], suggests relatively higher SES among women in this cohort and potentially limited generalizability of our results. Similarly, participants generally lived within a specific neighborhood and their experiences may not be shared by pregnant women in other parts of Nairobi or other cities in sub-Saharan Africa. Another limitation of this study was that women that opted to participate in the study may differ from those who did not, possibly causing selection bias. Finally, data were based on participant self-report and recall, which could be biased if participants were less aware of certain pollution sources, provided answers they felt were socially appropriate, or inaccurately recalled their practices and behaviors. Despite these potential limitations, our results provide useful insights into environmental health conditions experienced by the pregnant women in this study.

Our findings - that pregnant women in Nairobi face a variety of environmental exposures and that lower educational attainment is associated with greater exposure to certain pollutants - build upon the limited understanding of residential, ambient, and work-related environmental risk factors for this population. Furthermore, improving understanding of social and economic factors that are linked to housing conditions, household fuel use, HAP, AAP, and work-related exposures is critical for focusing interventions and control measures that are more likely to have sustained and effective reductions for individuals at highest risk. However, despite its importance, government entities and budgets to address environmental hazards in pregnancy, particularly for HAP [66], are lacking. Continued efforts throughout sub-Saharan Africa are needed to identify environmental health priorities for pregnant women for well-informed policy.