Abstract
Background
There is little evidence on the association between cooking with polluting fuels, particularly charcoal, and heart disease from low- and middle-income countries, which have the highest burdens of polluting fuels and cardiovascular disease (CVD). We sought to assess the relationship between self-reported primary home cooking with charcoal and prevalent cardiovascular disease in the low-income country Haiti.
Methods
Demographic and clinical data were collected from 3,005 adults in the Haiti Cardiovascular Disease Cohort, recruited from Port-au-Prince using multistage random sampling from 2019 to 2021. Primary cooking fuel was self-reported in the household: charcoal orliquified petroleum gas (LPG). Prevalent CVD (heart failure, stroke, myocardial infarction) was physician-adjudicated using epidemiological criteria similar to international cohorts. Multivariable generalized estimating equations with a Poisson distribution estimated prevalence ratios of prevalent CVD by polluting charcoal vs. LPG.
Results
Among 2,865 adults in the analytic sample, median age was 41 years (IQR 28–55), 57.9% female, and 88.4% reported using charcoal cooking fuel. Age-adjusted prevalence of any CVD was 13.7% (95% CI 12.2%, 15.4%) with the most common subtype being heart failure. Cooking with charcoal versus LPG was associated with higher prevalence of heart failure (1.63 prevalence ratio; 95% confidence interval 1.09, 2.44) after multivariable adjustment, lower prevalence of stroke, and not significantly associated with myocardial infarction or any CVD.
Conclusion
Most urban Haitian adults cook with charcoal, an uncommon primary cooking fuel in other countries, and charcoal is associated with higher prevalence of heart failure compared to LPG. Polluting charcoal cooking fuel is a potentially modifiable risk factor for heart failure in low-income settings that needs remediation and intervention.
Trial registration
Registry: clinicaltrials.gov.
Supplementary Information
The online version contains supplementary material available at 10.1186/s12940-026-01292-w.
Keywords: Haiti, Low- and middle-income countries, Heart failure, Cardiovascular disease, Household air pollution, Charcoal
Background
While ambient air pollution is an established risk factor for atherosclerotic cardiovascular disease (CVD) [1], there is less evidence on the relationship between household air pollution and heart disease, in particular heart failure. Furthermore, low-and middle- income countries (LMICs) have the highest burden of both air pollution and CVD globally [2, 3], but there is a paucity of data from LMICs such as Haiti [4]. One of the largest sources of household air pollution is cooking with polluting solid fuels, including coal, biomass (wood, agricultural debris), and charcoal [5]. Polluting solid fuels are used by 3 billion people for household energy in LMICs [5], resulting in the generation of particulate matter, black carbon, and gaseous co-pollutants such as carbon monoxide, which have been linked in high-income countries to higher rates of non-fatal and fatal CVD [1].
Haiti is the poorest country in the Western Hemisphere, and CVD is the leading cause of mortality, estimated to account for 26% of all adult deaths [2]. Ambient air pollution is reported to be high in Haiti based on modeled data with an age-standardized mortality rate of 210.5 per 100,000 population [6]. However accurate primary data on household air pollution levels are limited in this setting [7]. Our prior data have established that the most common type of heart disease in urban Haiti is early-onset heart failure, associated with hypertension among other factors, and indoor cooking was associated with higher hypertension prevalence and higher systolic blood pressure [8, 9]. The relationship between household air pollution from cooking with solid fuels and heart failure in LMICs has not been well established. Furthermore, the World Health Organization has called for more research investigating the cardiometabolic effects of air pollution, especially in LMICs where the burden is highest [10].
As a next step to address this knowledge gap, we evaluated the cross-sectional association between self-reported primary home cooking fuel use with physician-adjudicated prevalent CVD outcomes including heart failure, stroke, and myocardial infarction using the Haiti Cardiovascular Disease Cohort Study.
Methods
Study design and population
The Haiti Cardiovascular Disease Cohort Study (NCT03892265) is a longitudinal, population-based cohort of Port-au-Prince residents selected using multistage random sampling [11]. One hundred census blocks across Port-au-Prince were randomly selected, and GPS waypoints were randomly assigned within blocks to identify households for recruitment. The number of waypoints was proportional to the population size of each block, based on the Institute of Haitian Statistics. Adults were then recruited in person from each household for study enrollment. Between March 2019 and August 2021, 3,005 participants were enrolled for ongoing longitudinal follow-up to evaluate the prevalence and incidence of CVD risk factors and diseases, including hypertension, diabetes, obesity, dyslipidemia, kidney disease, poor diet, smoking, physical inactivity, and non-traditional risk factors like environmental pollutants and neighborhood variables. Inclusion criteria were: adults ≥ 18 years, primary residence in Port-au-Prince, and absence of any serious medical condition or cognitive impairment preventing participation as previously described [11].
The present analysis is based on cross-sectional data from enrollment. Given 97.7% of the 3,005 participants used either charcoal or liquified petroleum gas (LPG), we excluded participants who used other types cooking fuels (kerosene n = 49; electricity n = 14, wood n = 3, other n = 4). We also excluded pregnant females (n = 32), and participants who did not prepare meals at home (n = 38) for a final analytic sample of 2,865 (95.3%) participants (Figure S1).
The cohort is ongoing and is conducted at Groupe Haitien d’Etude du Sarcome de Kaposi et des Infections Opportunistes clinics (GHESKIO), a medical non-profit organization that has operated continuously over four decades in Haiti to provide clinical care and conduct research on HIV and related infectious and chronic diseases. This study was approved by institutional review boards at Weill Cornell Medicine and GHESKIO (1803019037), with written participant consent.
Measurements
Sociodemographic information was measured by self-report from participants, including age, sex, highest education level, income, and marital status. Education was categorized as “primary or lower” or “secondary or higher”. Income was measured in the local currency of Haitian gourdes, converted to US dollars using the 2019–2020 exchange rate, and categorized according to World Bank definitions of poverty including “<1 US dollar a day”, “1 to 10 US dollars a day”, and “>10 US dollars a day”. Marital status was categorized as “married/living together”, “single”, and “widowed/divorced/separated”.
Traditional CVD risk behaviors including tobacco smoking status, alcohol intake, physical activity, fruit/vegetable and salt intake were measured using the WHO STEPwise Approach to NCD Risk Factor Surveillance instruments [12]. Smoking status was categorized as “current/former” or “never”, alcohol intake was grouped into “≤1 drink daily” or “>1 drink daily”, physical activity was categorized as “moderate high” if participants reported yes to moderate activity > 150 min/week or vigorous activity > 75 min/week and “low” if they reported no to both. Fruit/vegetable intake was categorized into “<5 servings/day” or “≥5 servings/day” based on World Health Organization (WHO) serving sizes and limits. Salt intake was categorized as “high” or “moderate-low” also based on WHO STEPs.
Primary cooking fuel type and cooking location were reported for the entire household by the primary member of the household. Primary cooking fuel type included charcoal, wood, kerosene, liquified petroleum gas (LPG), electricity, or other (e.g. agricultural residue) in accordance with the WHO [10]. Cooking location was grouped as inside home, in a separate building, or outside. Participants were categorized as being a primary cook if they worked as a cook, food vendor, or were the main household cook.
Study physicians and nurses performed a clinical exam of all participants, collecting medical history, vital signs, physical exam, laboratory and imaging study data [8]. The physician screened for myocardial infarction (MI) symptoms using a translated version of the validated Rose angina questionnaire [13], for heart failure using a symptom checklist adapted from the AHA guidelines for CVD endpoint events [14], and for stroke using the Questionnaire for Verifying Stroke-Free Status (QVSFS) [15]. For vital signs, blood pressure (BP) measurements were taken by trained study staff using standardized protocols adapted from the AHA and WHO [12], with semi-automated electronic BP machine (OMRON HEM 907) with appropriate cuff sizes. Three BPs were measured, and the average of all three was used in this analysis. Height and weight were measured and used to calculate body mass index (BMI, kg/m2). A complete physical exam was performed. Laboratory measurements included serum hematology (Abbott CELLDYN 3200), creatinine, glucose, total cholesterol, high density lipoprotein (HDL), low density lipoprotein (LDL) (Vitros 250/350).
Electrocardiography (ECG) (Nasiff CardioCard PC) was performed on all participants, and echocardiography (Sonosite M-turbo ultrasound machine using a P21 × (5-1 MHz) probe) was performed for participants with either systolic BP ≥ 140 mmHg, diastolic BP ≥ 90 mmHg, abnormal ECG, or clinical concern for heart disease. Trained study staff were credentialed in echocardiography.
Definitions of cardiovascular risk factors
Hypertension was defined as systolic BP (SBP) ≥ 140 mmHg, or diastolic BP (DBP) ≥ 90 mmHg, or self-report of taking antihypertensive medication in the past two weeks, based on WHO guidelines [16]. Hypertension was further categorized into stage 1 (SBP 140–159 mmHg, or DBP 90–99 mmHg, or taking medications) or stage 2 (SBP ≥ 160 mmHg or DBP ≥ 100 mmHg) based on WHO definitions [16]. Prehypertension was defined as SBP 120–139 mmHg, or DBP 80–89 mmHg. Obesity was defined as BMI ≥ 30.0 kg/m2, and overweight as BMI 25.0 to 29.9 kg/m2. Diabetes was defined as any of the following criteria: enrollment fasting glucose value (FPG) ≥ 126 mg/dL or non-fasting glucose ≥ 200 mg/dL based on WHO definitions [17], patient self-report of taking diabetes medications in the last two weeks, or study physician diagnosis of diabetes based on clinical evaluation. Hypercholesterolemia was defined as either total cholesterol ≥ 240 mg/dL, low-density lipoprotein level ≥ 160 mg/dL or taking statins [18]. An estimated glomerular filtration rate (eGFR) was calculated using the 2021 Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) creatinine-based equation without ethnic factor. Renal dysfunction or chronic kidney disease (CKD) was defined as eGFR < 60 mL/min/1.73m2 or urine albumin to creatinine ratio ≥ 30 mg/g according to Kidney Disease: Improving Global Outcomes guidelines [19].
Outcomes: prevalent CVD
Prevalent CVD was defined as myocardial infarction (MI), stroke, or heart failure (HF), based on epidemiological criteria from the AHA, WHO, and European Society of Cardiology as previously reported in a separate study [8]. Prevalent CVD events were identified and adjudicated by physician review of patient reported symptoms, past medical history, physical exam signs, imaging data, and laboratory data, based on criteria similar to other CVD epidemiological cohorts [20]. Events were categorized into definite and probable, as detailed and described previously (see Supplement), and consistent with prior studies, both definite and probable cases were included as CVD events in this analysis [8]. The outcomes of prevalent CVD, MI, stroke, and HF were binary.
Statistical analysis
Participant demographic characteristics were summarized by cooking fuel type (charcoal or LPG). Skewed data was summarized using median and interquartile range (25th to 75th percentiles). To evaluate the association between cooking with charcoal versus LPG on prevalent CVD, we used a series of models: (1) adjusted for age and sex, (2) adjusted for socioeconomic (age, sex, income) and behavioral factors (physical activity, smoking), and (3) adjusted for socioeconomic, behavioral and CVD risk factors (BMI category (e.g. obesity), hypertension category, hypercholesterolemia, diabetes, chronic kidney disease). While diet and alcohol intake were selected for inclusion a priori, given extreme data imbalance with almost 100% of participants reporting poor diet with < 5 servings of fruits and vegetables daily or < 1 drink daily, these variables were not included in the final regressions. We used generalized estimating equations with a Poisson distribution to account for clustering of participants by household, with a log-link function, using the R package geepack [21]. Adjusted prevalence ratios and robust standard errors are reported.
Confounders as well as traditional CVD risk factors that may be potential mediators of the relationship between charcoal and CVD were chosen a priori. Model 2 is the main regression of interest, and Model 3 is a sensitivity analysis to explore the effect of polluting fuels when adjusting for traditional CVD risk factors. Results from Model 3 should not be interpreted as from a formal mediation analysis. In the primary analysis, we conducted complete case analysis to include only participants with complete data on variables used in the generalized estimating equations.
We conducted a series of sensitivity analyses to check the robustness of the main results to different assumptions. First, given concerns that income may greatly define the cooking fuel type such that it’s inclusion removes the variation in the exposure, we ran model 2 above without income. Second, we removed current and former smokers from models 1–3 as their personal exposure of air pollution was considered high. Due to this restriction, these associations should be interpreted for the population of non-smokers. Third, based on our prior demonstration of differential effects of cooking fuel type on blood pressure between sexes, we conducted an adjusted multivariable regression including an interaction term between sex and fuel type [9]. Fourth, we used multiple imputation to account for missing data (up to n = 88, 3.0%) as a sensitivity analysis, with 10 iterations using the R package mitml, for models 1–3 above. Imputations were based on all covariates, and assumes data were missing at random. The multivariable regression was performed on each imputed dataset, and estimates and standard errors were pooled (see Supplement). Analyses were conducted in R 4.5.0.
Results
Among 3005 participants in the Haiti Cardiovascular Disease cohort, 2935 (97.7%) used either polluting charcoal or cleaner LPG, and 2865 (95.3%) met inclusion criteria for the final analysis (Figure S1). Participants came from 1911 total households, with a median household size of 2 participants. Median age was 41 years (IQR 28–55), 57.9% female, with 70.3% earning < 1 US dollar a day (Table 1).
Table 1.
Enrollment demographic and clinical characteristics of haiti cardiovascular disease cohort by primary cooking fuel type, n = 2865
| Total N = 2,865 |
Charcoal N = 2,532 |
LPG N = 333 |
|
|---|---|---|---|
| Demographics, N (%) | |||
| Age: Median (25th to 75th percentiles) | 41 (28, 55) | 41 (28, 55) | 37 (28, 52) |
| Female sex | 1,658 (57.9%) | 1,447 (57.1%) | 211 (63.4%) |
| Education | |||
| Primary or Lower | 1,051 (36.8%) | 975 (38.6%) | 76 (22.8%) |
| Secondary or Higher | 1,807 (63.2%) | 1,550 (61.4%) | 257 (77.2%) |
| (Missing) | 7 | 7 | 0 |
| Income (daily) | |||
| < 1 USD / day | 2,010 (70.3%) | 1,784 (70.7%) | 226 (67.9%) |
| 1 to 10 USD / day | 349 (12.2%) | 311 (12.3%) | 38 (11.4%) |
| > 10 USD / day | 499 (17.5%) | 430 (17.0%) | 69 (20.7%) |
| (Missing) | 7 | 7 | 0 |
| Marital Status | |||
| Married/Living together | 1,115 (39.0%) | 977 (38.7%) | 138 (41.4%) |
| Single | 1,581 (55.3%) | 1,401 (55.5%) | 180 (54.1%) |
| Widowed/Divorced/Separated | 162 (5.7%) | 147 (5.8%) | 15 (4.5%) |
| (Missing) | 7 | 7 | 0 |
| Cooking | |||
| Cooking location | |||
| Inside home | 492 (17.2%) | 416 (16.4%) | 76 (22.8%) |
| Outside | 958 (33.4%) | 922 (36.4%) | 36 (10.8%) |
| In separate building | 1,411 (49.2%) | 1,190 (47.0%) | 221 (66.4%) |
| (Missing) | 4 | 4 | 0 |
| Cook status | |||
| Primary Cook | 1,069 (37.4%) | 935 (37.0%) | 134 (40.2%) |
| Not primary cook | 1,789 (62.6%) | 1,590 (63.0%) | 199 (59.8%) |
| (missing) | 7 | 7 | 0 |
| Health Behaviors | |||
| Smoking status | |||
| Never | 2,638 (92.7%) | 2,325 (92.4%) | 313 (94.8%) |
| Current/Former | 208 (7.3%) | 191 (7.6%) | 17 (5.2%) |
| (Missing) | 19 | 16 | 3 |
| Alcohol intake | |||
| < 1 drink a day (low) | 2,744 (96.2%) | 2,424 (96.2%) | 320 (96.4%) |
| ≥ 1 + drinks a day (moderate or higher) | 107 (3.8%) | 95 (3.8%) | 12 (3.6%) |
| (Missing) | 15 | 14 | 1 |
| Physical activity | |||
| ≤ 150 min/week (low) | 1,443 (50.6%) | 1,263 (50.1%) | 180 (54.1%) |
| > 150 min/week (moderate-high) | 1,410 (49.4%) | 1,257 (49.9%) | 153 (45.9%) |
| (Missing) | 12 | 12 | 0 |
| Fruit/vegetable intake | |||
| < 5 servings a day | 2,836 (99.3%) | 2,507 (99.4%) | 329 (98.8%) |
| ≥ 5 servings a day | 20 (0.7%) | 16 (0.6%) | 4 (1.2%) |
| (Missing) | 9 | 9 | 0 |
| Salt intake | |||
| Moderate-Low | 367 (12.8%) | 318 (12.6%) | 49 (14.7%) |
| High | 2,491 (87.2%) | 2,207 (87.4%) | 284 (85.3%) |
| (Missing) | 7 | 7 | 0 |
| CVD Risk Factors | |||
| Hypertension categories | |||
| Normotension (SBP < 120 and DBP < 80) | 1,391 (48.9%) | 1,230 (48.9%) | 161 (48.9%) |
| Prehypertension (SBP 120–139 or DBP 80–89) | 612 (21.5%) | 536 (21.3%) | 76 (23.1%) |
| Hypertension Stage 1 (SBP ≥ 140 or DBP ≥ 90) | 516 (18.2%) | 457 (18.2%) | 59 (17.9%) |
| Hypertension Stage 2 (SBP ≥ 160 or DBP ≥ 100) | 323 (11.4%) | 290 (11.5%) | 33 (10.0%) |
| SBP: Median (25th to 75th percentiles) | 119 (107, 138) | 119 (107, 139) | 119 (108, 136) |
| DBP: Median (25th to 75th percentiles) | 72 (63, 84) | 72 (63, 84) | 73 (63, 84) |
| BMI category | |||
| Underweight/Normal < 24.9 kg/m2 | 1,618 (56.6%) | 1,480 (58.5%) | 138 (41.6%) |
| Overweight 25.0–29.9 kg/m2 | 750 (26.2%) | 642 (25.4%) | 108 (32.5%) |
| Obese ≥ 30.0 kg/m2 | 492 (17.2%) | 406 (16.1%) | 86 (25.9%) |
| (Missing) | 5 | 4 | 1 |
| Diabetes Mellitus | 107 (3.8%) | 95 (3.9%) | 12 (3.7%) |
| (Missing) | 84 | 72 | 12 |
| Hypercholesterolemia | 357 (12.5%) | 321 (12.7%) | 36 (10.8%) |
| (Missing) | 7 | 6 | 1 |
| Chronic kidney disease | 241 (8.7%) | 215 (8.7%) | 26 (8.0%) |
| (Missing) | 79 | 69 | 10 |
Statistics presented are median (interquartile range); n (%), percentages are calculated with denominator of those who responded. Salt intake adapted from the WHO STEPS questionnaire
LPG Liquified petroleum gas, USD US dollar, CVD Cardiovascular disease, SBP SystoSlic blood pressure; DBP Diastolic blood pressure, BMI Body mass index
Charcoal vs. liquified petroleum gas use
For cooking fuel use, 2532 (88.4%) used charcoal and 333 (11.6%) used LPG. Almost half of participants cooked in a separate building (49.2%), and only 37.4% were primary cooks, of whom 93.2% were female. Participants who cooked with charcoal versus LPG had a lower proportion of females (57.1% vs. 63.4%), and were more likely to be slightly older, less educated, earn < 1 USD/day, and cook outside (Table 1). Participants who cooked with charcoal also had lower levels of obesity compared to those who cooked with LPG. The remaining variables were similar between the two groups.
Association between charcoal use and CVD
For cardiovascular disease prevalence at study enrollment, 339 (11.8%) had heart failure with 274 (80.8%) heart failure with preserved ejection fraction and 31 (9.1%) heart failure with reduced ejection fraction, 73 had stroke (2.5%), and 26 participants had MI (0.9%). In total, 394 participants had any CVD (13.7%). The age-adjusted prevalence of any CVD was 13.7% (95% CI 12.2%, 15.4%). Figure 1 displays the prevalence of CVD by fuel type. Table S1 describes the clinical characteristics of prevalent CVD.
Fig. 1.
Title: Prevalence of Cardiovascular Disease by Fuel Type
Caption: Bars represent 95% confidence intervals for prevalence
In Model 1 adjusting for age and sex, charcoal use vs. LPG was associated with higher prevalence ratio (PR) of heart failure (1.67 PR; 95% CI 1.12, 2.48), and lower prevalence of stroke (0.44 PR; 95% CI 0.25, 0.76), but not associated with MI or any CVD (Table 2). In Model 2 adjusting for socioeconomic and behavioral factors, charcoal remained associated with higher prevalence of heart failure (1.63 PR, 95% CI 1.09, 2.44) and lower prevalence of stroke (PR 0.44, 95% CI 0.25, 0.77) but with decreased magnitude of effect (Table 2). In sensitivity analyses, removing income did not meaningfully change the magnitude of this association, suggesting it was neither a confounder nor colinear with fuel type (Table S2). In Model 3 adjusting for socioeconomic and CVD risk factors, charcoal was associated with heart failure at 1.82 PR (95% CI 1.25, 2.69) and stroke at 0.49 PR (95% CI 0.28, 0.84) (Table 2, Table S3). In sensitivity analyses, removing current and former smokers (Table S4) and accounting for missing variables with multiple imputation (Table S5), the direction and magnitude of associations were similar, although point estimates for multiple imputation were slightly higher.
Table 2.
Association between charcoal vs. LPG and adjudicated CVD prevalence in regressions
| Model: Charcoal vs. LPG |
Any CVD | Heart Failure | Stroke | MI | ||||
|---|---|---|---|---|---|---|---|---|
| PR | 95% CI | PR | 95% CI | PR | 95% CI | PR | 95% CI | |
| Adjusted for age and sex | 1.17 | (0.87, 1.60) | 1.67 | (1.12, 2.48) | 0.44 | (0.25, 0.76) | 1.16 | (0.27, 5.00) |
| Adjusted for socioeconomic, behavioral factors* | 1.15 | (0.85, 1.57) | 1.63 | (1.09, 2.44) | 0.44 | (0.25, 0.77) | 1.08 | (0.25, 4.62) |
| Adjusted for socio-economic, behavioral and CVD risk factors† | 1.27 | (0.95, 1.70) | 1.82 | (1.25, 2.69) | 0.49 | (0.28, 0.84) | 0.99 | (0.24, 4.01) |
Complete case analysis was conducted on 2757 participants out of the 2865 participants in the final analytic sample (96.2%)
Diet and alcohol were not included given data imbalance as almost 100% of participants reported poor diet with < 5 servings of fruits and vegetables daily or < 1 drink daily
PR Prevalence ratio, CI Confidence interval
*Adjusted for: age, sex, income, physical activity, smoking
†Adjusted for: age, sex, income, physical activity, smoking, BMI category (e.g. obesity), hypertension category, hypercholesterolemia, diabetes, chronic kidney disease
After stratification by sex, the difference in CVD prevalence by fuel type was greater in females compared to males (Figure S2) for any CVD and heart failure. For example, the prevalence of heart failure was 15.1% in charcoal vs. 7.1% in LPG among females, but only 9.0% in charcoal vs. 6.6% in LPG among males.
In adjusted multivariable regressions with an interaction term for fuel type and sex, the interaction term was significant for heart failure (p value 0.005) but not for MI, stroke, or any CVD. Among females, cooking with charcoal was associated with a higher prevalence ratio of heart failure (1.99 PR; 95% CI 1.23, 3.22) (Table S6).
Discussion
In this cross-sectional study of the population-based Haiti Cardiovascular Disease Cohort, very high levels of charcoal use were reported. In adjusted regressions, polluting charcoal use was associated with higher prevalence of heart failure, and lower prevalence of stroke, robust to multiple sensitivity analyses. Sex modified the relationship only for heart failure, where females exposed to charcoal vs. LPG had a higher prevalence of heart failure, compared to males exposed to charcoal vs. LPG.
Our finding that 88.4% of participants used charcoal is notable for two reasons: this is higher than the usage of polluting fuels reported for other LMICs, and other LMICs largely use coal or wood as the primary form of polluting cooking fuel [22]. Haitians meet 80% of their national energy needs through firewood and charcoal production, with charcoal being the second-largest agricultural commodity in the country [23]. For other LMICs, the WHO estimates that 33% of people living in LMICs mainly use solid fuels, with many countries shifting to the cleaner fuel LPG and electric stoves in recent years, including India, Indonesia, and Peru [22]. Most other countries with similar levels of solid fuel use to Haiti are in sub-Saharan Africa, including Central African Republic, South Sudan, Rwanda, and Burundi (i.e., > 85% using solid fuels) [24].
Solid fuel use is the major contributor to household air pollution worldwide [10], especially elevating levels of particulate matter, black carbon, and gaseous co-pollutants, which have been associated with CVD in past studies [1, 25, 26]. Air pollution is thought to contribute to CVD by activating systemic oxidative stress responses, activating the lung autonomic nervous system reflex arcs, and directly affecting blood vessels [1]. Both short-term and long-term increases in air pollution increase CVD risk, although most prior studies focus on ambient air pollution rather than household air pollution [1]. In a meta-analysis on adverse health effects associated with household air pollution, data pooled from 16 studies found use of polluting fuels was associated with a pooled risk ratio of 1.10 (95%CI 1.09–1.11) for ischemic heart disease, 1.09 (95%CI 1.04–1.14) for cerebrovascular disease, and 1.07 (95%CI 1.04–1.11) for cardiovascular death [27].
We found polluting charcoal use was associated with higher risk of prevalent heart failure and lower risk of prevalent stroke but was not associated with prevalent MI in urban Haiti. Furthermore, we found the relationship with heart failure only existed for females. The inclusion of traditional CVD risk factors in the sensitivity analysis in Model 3 resulted in larger effect sizes for the association between charcoal and prevalent CVD. It is unclear whether this is due to confounding, a mediated effect or possible collider bias, limiting the interpretation of this model. Lastly, in sensitivity analyses after multiple imputation, effect sizes were also slightly larger than the main analysis, suggesting participants initially excluded for missing data may have had a stronger association between cooking fuel use and CVD outcome.
There are much less data on the association between air pollution and heart failure, compared to the other CVD outcomes in LMICs [1]. In the meta-analysis on adverse health outcomes associated with household air pollution, only one study (Prospective Urban and Rural Epidemiology Study) measured heart failure as an outcome, and did not find an association between polluting fuel use and heart failure [28]. The remaining literature focuses on ambient air pollution. Short-term increases in ambient air pollution increase the risk of heart failure hospitalizations, and in some studies with incident heart failure and death (2.1% higher risk) from pooled data in South Korea, Japan, Canada, and the USA [29], but the association is inconsistent across studies. Possible mechanisms linking air pollution to heart failure include increased systemic blood pressure and vasoconstriction [30–32], pulmonary vasoconstriction leading to increased pulmonary and right ventricular diastolic filling pressures [33], and adverse ventricular remodeling including myocardial fibrosis [34]. The fact that this association was observed only in females in this study could be due to a few factors. First, almost half of our participants cooked in a separate building from their home, suggesting exposure to charcoal may be limited to cooking times. Females are more likely to be the primary cooks in their families, and are more likely to have heart failure in Haiti [8, 9]. Our finding of charcoal associated with lower prevalence of stroke, and null findings for MI, are contradictory with prior research which showed increased risk for both conditions from solid fuel use. These findings may be related to the limitations of cross-sectional data, including unmeasured confounders, inability to assess prior polluting fuel use (including frequency, duration, and timing) that occurred before a CVD event possibly leading to survivor bias, potential misclassification of exposure due to self-report of fuel use, or limited power from the small number of prior non-fatal ischemic events in this young population.
Polluting cooking fuel use is a potentially modifiable exposure and risk factor for CVD, with a disproportionately high burden among poorer populations [22]. Targeting polluting fuel use and shifting to cleaner fuels could prevent CVD. The Household Air Pollution Intervention Network (HAPIN) tested the impact of a LPG cooking stove and fuel on health outcomes, but was not powered for CVD events and focused on pneumonia in children, growth in children, and blood pressure in older females [35]. HAPIN found the intervention compared to polluting fuels (largely coal) reduced personal air pollution exposures, and led to a smaller increase in gestational blood pressure among pregnant females versus controls that continued to use solid fuel cookstoves [36]. However, HAPIN did not decrease blood pressure perhaps due to the shorter time scale of intervention [36].
There are many barriers to the transition from polluting to cleaner fuels in LMICs. A market analysis of Port-au-Prince by the Clean Cooking Alliance on fuel usage revealed the most common barriers included upfront cost of stoves that use LPG, ongoing access issues with gas, and safety concerns including fear of explosions [37]. Future studies should explore multilevel strategies to address barriers and facilitators to increasing LPG and electric stove use in high burden contexts like Haiti, including at a societal level.
Strengths of the current study include prospective data collection from a population-based cohort in a data-scarce region of the world, physician-adjudicated CVD outcomes incorporating symptoms, physical exam findings, laboratory measures, and imaging including electrocardiograms and echocardiograms. Adjudicators were blinded to cooking fuel type. Limitations include cross-sectional data analysis that limits causal inference about the observed associations, especially with temporality of the charcoal exposure and CVD outcome. We could not ascertain if participants used “stove stacking”, or using different stoves and fuels for different purposes (e.g. charcoal with cooking beans, LPG with cooking vegetables). We did not ascertain cooking fuel type at work for participants who worked as cooks. In addition, there was a small number of events for myocardial infarction, perhaps related to low sensitivity of the assessment methods or survival bias. These limitations may have biased our results to the null for myocardial infarction, or explain the inverse association we found for stroke. Duration of exposures, ventilation around cookstoves, and other exposure metrics were not available at time of this analysis.
In conclusion, most urban Haitian adults cook with charcoal, an uncommon primary cooking fuel in other countries, and charcoal is significantly associated with higher prevalence of heart failure compared to LPG. Polluting charcoal cooking fuel is a potentially modifiable risk factor for heart failure in low-income settings that needs urgent remediation and intervention.
Supplementary Information
Acknowledgements
We thank the participants of the Haiti Cardiovascular Disease Cohort, the GHESKIO community health workers and participating communities in Port-au-Prince. We also thank the Haitian College of Cardiology and Haitian Ministry of Health for providing leadership in the design and interpretation of the study.
Abbreviations
- CVD
Cardiovascular disease
- LMIC
Low- and middle-income country
- GHESKIO
Haitian Group for the Study of Kaposi’s Sarcoma and Opportunistic Infections
- LPG
Liquified petroleum gas
- WHO
World Health Organization
- AHA
American Heart Association
- NCD
Non-communicable Disease
- SBP
Systolic blood pressure
- DBP
Diastolic blood pressure
- BMI
Body mass index
- CKD
Chronic kidney disease
- MI
Myocardial infarction
- HF
Heart failure
- PR
Prevalence ratio
- HDL
High density lipoprotein
- LDL
Low density lipoprotein
Authors’ contributions
Conceptualisation: LDY, RS, MLM. Data curation: RS, RSS, KC, VR, GF, SEM, AS, JP, EH, FJL, AO, CB, RR. Formal analysis including directly accessing and verifying the data: AO, LDY. Funding acquisition: MLM. Investigation: LDY, RS, MLM. Methodology: LDY, AO, RR. Project administration: CB, VR, MMD, JWP, MLM. Resources: VR, MMD, JWP, MLMSoftware: LDY, AO. Supervision: JLP, MDH, VGDR, MMD, JWP, MLM. Validation: LDY, AO. Visualisation: AO. Writing: original draft: LDY, RS. Writing: review & editing: all authors contributed equally. LY and MLM have the final responsibility for the decision to submit the study for publication.
Funding
This study was registered on clinicaltrials.gov NCT03892265 and funded by NIH grants R01HL143788, D43TW011972, K23HL177149, K24HL163393, and K24HL175228.
Data availability
Researchers who provide a methodologically sound proposal may have access to a subset of deidentified participant data, with specific variables based on the proposal. Proposals should be directed to the principal investigator Dr Margaret McNairy. To gain access, data requestors will need to sign a data access agreement. Data are available following publications through 3 years after publication and will be provided directly from the PI.
Declarations
Ethics approval and consent to participate
This study was approved by institutional review boards at Weill Cornell Medicine and GHESKIO (1803019037), with written participant consent.
Consent for publication
Not applicable.
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Lily D. Yan and Rodney Sufra contributed equally to this work.
References
- 1.Brook RD, Rajagopalan S, Pope CA, Brook JR, Bhatnagar A, Diez-Roux AV, et al. Particulate Matter Air Pollution and Cardiovascular Disease. Circulation. 2010;121(21):2331–78. [DOI] [PubMed]
- 2.GBD 2021 Causes of Death Collaborators. Global burden of 288 causes of death and life expectancy decomposition in 204 countries and territories and 811 subnational locations, 1990–2021: a systematic analysis for the Global Burden of Disease Study 2021. Lancet. 2024;403(10440):2100–32. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Fuller R, Landrigan PJ, Balakrishnan K, Bathan G, Bose-O’Reilly S, Brauer M, et al. Pollution and health: a progress update. Lancet Planet Health. 2022;6(6):e535–47. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.World Bank. World Development Indicators. Available from: https://data.worldbank.org/indicator. Cited 2024 May 6.
- 5.International Energy Agency (IEA). International Renewable Energy Agency (IRENA), United Nations Statistics Division (UNSD), World Bank, World Health Organization. Tracking SDG 7: The Energy Progress Report 2022. Available from: https://trackingsdg7.esmap.org/downloads. Cited 2023 Jan 30.
- 6.WHO. Indicator Metadata Registry Details. Available from: https://www.who.int/data/gho/indicator-metadata-registry/imr-details/2476. Cited 2023 Jan 12.
- 7.van Donkelaar A, Martin RV, Brauer M, Hsu NC, Kahn RA, Levy RC, et al. Global Estimates of Fine Particulate Matter using a Combined Geophysical-Statistical Method with Information from Satellites, Models, and Monitors. Environ Sci Technol. 2016;50(7):3762–72. [DOI] [PubMed] [Google Scholar]
- 8.Yan LD, Sufra R, St Sauveur R, Jean-Pierre MC, Apollon A, Malebranche R, et al. Spectrum of prevalent cardiovascular diseases in urban Port-au-Prince, Haiti: a population-based cross-sectional study. Lancet Reg Health Am. 2024;33:100729. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Rasul R, Sufra R, Pierre MCJ, Sauveur RS, Rouzier V, Inddy J et al. Prevalence of Cooking with Polluting Fuels and Association with Elevated Blood Pressure Among Adults in Port au Prince, Haiti: A Cross-Sectional Analysis. Glob Heart. 2025;20(1). Available from: https://globalheartjournal.com/articles/10.5334/gh.1405. Cited 2025 Feb 28. [DOI] [PMC free article] [PubMed]
- 10.WHO. WHO global air quality guidelines [Internet]. Geneva: World Health Organization. 2021. Available from: https://www.who.int/publications/i/item/9789240034228. Cited 2025 May 19.
- 11.Lookens J, Tymejczyk O, Rouzier V, Smith C, Preval F, Joseph I, et al. The Haiti cardiovascular disease cohort: study protocol for a population-based longitudinal cohort. BMC Public Health. 2020;20(1):1633. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.WHO. WHO STEPS surveillance manual. The WHO STEPwise approach to chronic disease risk factor surveillance. Geneva: WHO; 2017. Available from: https://www.who.int/teams/noncommunicable-diseases/surveillance/systems-tools/steps. Cited 2023 Sept 19.
- 13.Cook DG, Shaper AG, MacFarlane PW. Using the WHO (Rose) angina questionnaire in cardiovascular epidemiology. Int J Epidemiol. 1989;18(3):607–13. [DOI] [PubMed] [Google Scholar]
- 14.Hicks KA, Tcheng JE, Bozkurt B, Chaitman BR, Cutlip DE, Farb A, et al. 2014 ACC/AHA Key Data Elements and Definitions for Cardiovascular Endpoint Events in Clinical Trials: A Report of the American College of Cardiology/American Heart Association Task Force on Clinical Data Standards (Writing Committee to Develop Cardiovascular Endpoints Data Standards). J Nucl Cardiol Off Publ Am Soc Nucl Cardiol. 2015;22(5):1041–144. [DOI] [PubMed] [Google Scholar]
- 15.Jones WJ, Williams LS, Meschia JF. Validating the Questionnaire for Verifying Stroke-Free Status (QVSFS) by Neurological History and Examination. Stroke. 2001;32(10):2232–6. [DOI] [PubMed] [Google Scholar]
- 16.WHO. Guideline for the pharmacological treatment of hypertension in adults. Geneva: World Health Organization. 2021. (WHO Guidelines Approved by the Guidelines Review Committee). Available from: http://www.ncbi.nlm.nih.gov/books/NBK573631/. Cited 2023 Dec 15.
- 17.World Health Organization. Definition and diagnosis of diabetes mellitus and intermediate hyperglycaemia: report of a WHO/IDF consultation. 2006.
- 18.Lloyd-Jones Donald M, Leip Eric P, Larson Martin G, D’Agostino Ralph B, Alexa B, Peter W. Prediction of Lifetime Risk for Cardiovascular Disease by Risk Factor Burden at 50 Years of Age. Circulation. 2006;113(6):791–8. [DOI] [PubMed] [Google Scholar]
- 19.Levin A, Stevens PE, Bilous RW, Coresh J, Francisco ALMD, Jong PED, et al. Kidney disease: Improving global outcomes (KDIGO) CKD work group. KDIGO 2012 clinical practice guideline for the evaluation and management of chronic kidney disease. Kidney Int Suppl. 2013;3(1):1–150. [Google Scholar]
- 20.Safford MM, Brown TM, Muntner P, Durant RW, Glasser S, Halanych J, et al. Association of Race and Sex With Risk of Incident Acute Coronary Heart Disease Events. JAMA J Am Med Assoc. 2012;308(17):1768–74. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Højsgaard S, Halekoh U, Yan J. The R Package geepack for Generalized Estimating Equations. J Stat Softw. 2006;15:1–11. [Google Scholar]
- 22.IEA IRENA, World Bank UNSD. WHO. Tracking SDG 7: The Energy Progress Report. World Bank, Washington DC; 2023. Available from: https://trackingsdg7.esmap.org/data/files/download-documents/sdg7-report2023-full_report.pdf.
- 23.International Bank for Reconstruction and Development. Charcoal In Haiti: A National Assessment of Charcoal Production and Consumption Trends. World Bank, Washington DC. 2017. Available from: https://documents1.worldbank.org/curated/en/697221548446232632/pdf/134058-CharcoalHaitiWeb.pdf.
- 24.Local Burden of Disease Household Air Pollution Collaborators. Mapping development and health effects of cooking with solid fuels in low-income and middle-income countries, 2000-18: a geospatial modelling study. Lancet Glob Health. 2022;10(10):e1395–411. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Ayres J. Cardiovascular disease and air pollution: a report by the Committee on the Medical Effects of Air Pollutants. UK Dep Health. 2006.
- 26.Song X, Hu Y, Ma Y, Jiang L, Wang X, Shi A, et al. Is short-term and long-term exposure to black carbon associated with cardiovascular and respiratory diseases? A systematic review and meta-analysis based on evidence reliability. BMJ Open. 2022;12(5):e049516. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Lee KK, Bing R, Kiang J, Bashir S, Spath N, Stelzle D, et al. Adverse health effects associated with household air pollution: a systematic review, meta-analysis, and burden estimation study. Lancet Glob Health. 2020;8(11):e1427–34. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Hystad P, Duong M, Brauer M, Larkin A, Arku R, Kurmi OP, et al. Health Effects of Household Solid Fuel Use: Findings from 11 Countries within the Prospective Urban and Rural Epidemiology Study. Environ Health Perspect. 2019;127(5):057003. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Zhang D, Chen W, Cheng C, Huang H, Li X, Qin P, et al. Air pollution exposure and heart failure: A systematic review and meta-analysis. Sci Total Environ. 2023;872:162191. [DOI] [PubMed] [Google Scholar]
- 30.Pieters N, Plusquin M, Cox B, Kicinski M, Vangronsveld J, Nawrot TS. An epidemiological appraisal of the association between heart rate variability and particulate air pollution: a meta-analysis. Heart Br Card Soc. 2012;98(15):1127–35. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Brook RD, Brook JR, Urch B, Vincent R, Rajagopalan S, Silverman F. Inhalation of fine particulate air pollution and ozone causes acute arterial vasoconstriction in healthy adults. Circulation. 2002;105(13):1534–6. [DOI] [PubMed] [Google Scholar]
- 32.Tsai DH, Riediker M, Wuerzner G, Maillard M, Marques-Vidal P, Paccaud F, et al. Short-term increase in particulate matter blunts nocturnal blood pressure dipping and daytime urinary sodium excretion. Hypertens Dallas Tex 1979. 2012;60(4):1061–9. [DOI] [PubMed] [Google Scholar]
- 33.Rich DQ, Freudenberger RS, Ohman-Strickland P, Cho Y, Kipen HM. Right heart pressure increases after acute increases in ambient particulate concentration. Environ Health Perspect. 2008;116(9):1167–71. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Wold LE, Ying Z, Hutchinson KR, Velten M, Gorr MW, Velten C, et al. Cardiovascular remodeling in response to long-term exposure to fine particulate matter air pollution. Circ Heart Fail. 2012;5(4):452–61. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Liao J, Kirby MA, Pillarisetti A, Piedrahita R, Balakrishnan K, Sambandam S, et al. LPG stove and fuel intervention among pregnant women reduce fine particle air pollution exposures in three countries: Pilot results from the HAPIN trial. Environ Pollut Barking Essex 1987. 2021;291:118198. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Ye W, Steenland K, Quinn A, Liao J, Balakrishnan K, Rosa G, et al. Effects of a Liquefied Petroleum Gas Stove Intervention on Gestational Blood Pressure: Intention-to-Treat and Exposure-Response Findings From the HAPIN Trial. Hypertens Dallas Tex 1979. 2022;79(8):1887–98. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Clean Cooking Alliance. Clean cooking in Haiti: Urban and peri-urban consumer segmentation. 2020. Available from: https://cleancooking.org/reports-and-tools/cleancooking-in-haiti-urban-and-peri-urban-consumer-segmentation-n.
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
Researchers who provide a methodologically sound proposal may have access to a subset of deidentified participant data, with specific variables based on the proposal. Proposals should be directed to the principal investigator Dr Margaret McNairy. To gain access, data requestors will need to sign a data access agreement. Data are available following publications through 3 years after publication and will be provided directly from the PI.