Introduction
On February 6, 2023, at 04:17 a.m. local time, a 7.7 Mw magnitude earthquake struck Turkey, with its epicenter in Pazarcık, Kahramanmaraş, along the East Anatolian Fault line. Two subsequent major seismic events followed: a 7.6 Mw earthquake centered in Elbistan and a 6.4 Mw earthquake in Yayladağı, Hatay. Together, these disasters affected 11 provinces (Kahramanmaraş, Hatay, Adıyaman, Malatya, Kilis, Gaziantep, Adana, Osmaniye, Şanlıurfa, Diyarbakır, and Elazığ) and approximately 14 million people (see Figure 1). 1 These disasters resulted in 53,537 deaths, 107,213 injuries, and the collapse or severe damage of hundreds of thousands of buildings, causing an estimated economic loss of $148.8 billion (9% of Turkey’s 2023 GDP).2, Reference Kahraman3
National Air Quality Monitoring Network and distribution of air quality monitoring stations in earthquake-affected provinces.
Beyond the immediate mortality and morbidity, earthquakes create prolonged public health challenges, in part through environmental disruption, especially deteriorating air quality.Reference Onbası, Ogenler, Yapici, Okuyaz and Bolaç4 Post-earthquake air pollution arises from building collapses, debris removal, reconstruction activities, and increased indoor stove use for heating. Dust, smoke, asbestos, and other hazardous substances released from damaged structures, combined with heavy machinery emissions and improper debris management, contribute to both outdoor and indoor pollution.Reference Zanoletti and Bontempi5, Reference Hsu, Huang and Huang6
Air pollution is associated with increased risk of hospital admissions for chronic diseases (e.g., cardiovascular, respiratory, and metabolic conditions) and acute conditions including infectious diseases, CO poisoning, headaches, and bronchospasms.Reference Uprety, Ozaki and Higuchi7, Reference Simkhovich, Kleinman and Kloner8 In particular, PM2.5, PM10, and gaseous pollutants (CO, nitrogen dioxide [NO₂], sulfur dioxide [SO₂], ozone [O₃]) can exacerbate systemic inflammation, oxidative stress, and hypoxemia, thereby potentially worsening myocardial injury.Reference Pota, Suwannasom and Chattipakorn9 Previous studies have also suggested that outdoor air pollution may influence the prognosis of CO poisoning, with the AQI serving as an independent predictor of hospital admission and cardiac complications.Reference Sabak, Gulen and Satar10
In this context, we hypothesize that combined post-disaster indoor CO exposure and elevated outdoor air pollution contribute to increased myocardial injury in patients with CO poisoning. Investigating this relationship 1 year after the Kahramanmaraş earthquake provides novel insights into disaster-related environmental health risks and clarifies the potential clinical relevance of air-quality monitoring.
Materials and Methods
Study Design
This retrospective, multicenter study was conducted in 10 hospitals across 8 cities affected by the February 6, 2023, earthquake (Kahramanmaras, Hatay, Adana, Adiyaman, Kilis, Gaziantep, Malatya, and Osmaniye). Ethical approval was obtained from the relevant ethics committee, and permissions were secured from all participating centers. The Ethics Committee of Adana City Training and Research Hospital (Meeting date: March 28, 2024, Ethics committee decision no: 3254, Meeting number: 149) has approved this study. Due to the retrospective nature of the study, informed consent was not obtained from the patients included in the study. The study was performed in accordance with the Declaration of Helsinki and good clinical practice standards.
Study Population
Patients aged 18 years or older who presented to emergency departments in the earthquake-affected provinces for CO poisoning between December 1, 2023, and March 31, 2024, were included. Eligible patients had complete medical records and accessible hospital information system data. For non-smokers, CO poisoning was defined as a carboxyhemoglobin (COHb) level >5%, while for smokers, the threshold was >10%.
Exclusion criteria included:
-
• Pregnant women,
-
• Patients with concurrent poisoning from other substances (e.g., alcohol, medications, cyanide),
-
• Patients who were deceased upon arrival,
-
• Cases presenting >24 hours after exposure,
-
• Intentional poisonings,
-
• Patients with a history of neurological diseases,
-
• Individuals under 18 years, and
-
• Patients with inaccessible data.
Data Collection
Collected data included demographic characteristics, presenting symptoms, source and timing of CO exposure, vital signs, blood gas (COHb) and laboratory parameters, complications, emergency department (ED) length of stay, hospitalization duration, and clinical outcomes.
Common complications assessed included:
-
• Leukocytosis (>11.0 × 10Reference Pota, Suwannasom and Chattipakorn9/L),
-
• Hyperglycemia (>126 mg/dL),
-
• Acute renal failure (serum creatinine increase ≥0.3 mg/dL within 48 hours),
-
• Hypoxemia (<92%),
-
• Rhabdomyolysis,
-
• Central nervous system depression, and
-
• Myocardial injury, the primary outcome, was defined as cardiac troponin elevation above the 99th percentile upper reference limit (URL) for the locally used assay within 24 hours of ED presentation. To harmonize values across centers using different assays, troponin concentrations were expressed as multiples of the assay-specific URL, where assay type and URL were available (troponin/URL). ECG (new ST-segment/T-wave changes, new arrhythmias) and echocardiographic findings (new regional wall motion abnormalities or reduced left ventricular ejection fraction) were recorded when available and treated as secondary/supportive evidence of cardiac involvement, but were not mandatory for outcome definition. We acknowledge that echocardiography was performed at the discretion of clinicians and was not uniformly available across centers.
Laboratory reference ranges were standardized across centers, and venous blood gas COHb levels were measured using the Radiometer ABL90 Flex analyzer (Radiometer, Copenhagen, Denmark).
Air quality data
Hourly air quality data (AQI, PM2.5, PM10, CO, SO₂, NO₂, O₃) and meteorological data (temperature, humidity, wind speed) were obtained from the Turkish National Air Quality Monitoring Network (Ministry of Environment, Urbanization and Climate Change) (see Figures 1 and 2).11 This system utilizes the Air Quality Index (AQI), which categorizes regions hourly based on pollutant concentrations: good (0-50), moderate (51-100), sensitive (101-150), unhealthy (151-200), poor (201-300), and hazardous (301-500) (Table 1). For each patient, we assigned exposure from the geographically nearest monitoring station to the patient’s recorded residence or exposure location. In Kahramanmaraş, Adıyaman, and Malatya, only AQI, SO₂, and PM10 data were available. In provinces where pollutant-specific data (notably PM₂.₅ and NO₂) were not available, AQI or province-level averages were used for descriptive purposes; these provinces were handled separately in sensitivity analyses. For Hatay, air quality data were sourced from a station located 20 kilometers away. This may have introduced exposure misclassification; therefore, sensitivity analyses were performed to minimize potential bias. Sanlıurfa, Diyarbakır, and Elazığ were excluded due to technical and data access challenges (see Figure 1).
Data sample for the Kahramanmaras province, the epicenter of the earthquake, on any given day.
Air Quality Index (AQI) cut-off values
The National Air Quality Index (AQI) has been developed by adapting the Environmental Protection Agency (EPA) AQI to align with national legislation and threshold values. The AQI is calculated based on six primary pollutants, which include particulate matter (PM2.5 and PM10), carbon monoxide (CO), sulfur dioxide (SO₂), nitrogen dioxide (NO₂), and ozone (O₃).
Sample Size
Sample size was calculated using G*Power software, yielding a requirement of 91 patients with a 95% confidence interval, α = 0.05, and a power of 85% (1-β = 0.85).
Outcomes
The primary outcome is the relationship between AQI and myocardial injury in patients with CO poisoning. The secondary outcome was to compare air pollutant parameters in the affected provinces with EPA thresholds and to evaluate differences in AQI values between the year before and the year after the earthquake.
Statistical Analysis
Data were analyzed using SPSS v22 (SPSS Inc., Chicago, Illinois, USA). Categorical variables were expressed as frequencies and percentages, and analyzed using the Chi-square test. Continuous variables were tested for normality using the Kolmogorov-Smirnov test and reported as means ± standard deviations. The paired T-test was used for normally distributed data, and the Mann-Whitney U test for non-normally distributed data. A multivariable logistic regression analysis was conducted to evaluate the association between myocardial injury and factors such as age, sex, COHb, AQI, PM10, SO2, and comorbidities. Variables included in the multivariable logistic regression model were selected a priori based on biological plausibility and prior literature rather than on statistical significance in univariate analyses. Specifically, age, sex, comorbidities, COHb levels, AQI, PM10, and SO2 were considered potential confounders due to their known or suspected associations with cardiovascular outcomes. This approach was chosen to reduce the risk of residual confounding and to provide a more robust model specification. Correlation analyses between continuous variables (e.g., AQI and ED length of stay) were performed using Pearson or Spearman coefficients as appropriate. Statistical significance was defined as P <0.05, but effect sizes were interpreted in the context of clinical relevance.
Results
A total of 151 patients were included in the study, distributed across the following provinces: Kilis (n = 27), Adiyaman (n = 27), Gaziantep (n = 27), Hatay (n = 24), Adana (n = 16), Kahramanmaraş (n = 14), Malatya (n = 9), and Osmaniye (n = 7) (see Figure 3). Of the patients, 58.3% were female, with a mean age of 40 ± 17.7 years. The highest incidence of presentations occurred in December (34.4%) and during night shifts (41.1%). Heating stoves were the leading cause of CO poisoning, accounting for 79.5% of cases. The most common presenting symptoms were nausea and vomiting (51%), headache (48.3%), and dizziness (36.4%). Hypertension was the most frequently observed comorbidity, affecting 17.2% of patients. The average values of patients’ vital signs and laboratory parameters are shown in Table 2.
Flowchart of the patients included in the study.
Distribution of demographic and clinical data of Carbon monoxide poisoning cases
COHb: Carboxyhemoglobin; CRP: C-reactive protein; HCO3: Bicarbonate; WBC: White blood cell.
Meteorological and Pollutant Parameters
The average meteorological and pollutant parameters during patient admissions, categorized by the city of patient presentation, are summarized in Table 3. The lowest recorded average air temperature was in Malatya (7.6°C [IQR 7.1-9.7°C]), the highest humidity in Hatay (99% [IQR 84-100]), and the strongest wind speed in Kilis (2.3 ± 1.2 km/h). The AQI and other pollutant concentrations relative to EPA cut-off values are also shown in Table 3.
Distribution of meteorological parameters and pollutant parameters by province
Note: Data are presented as mean ± standard deviation or median (25–75 IQR), as appropriate. Values above the reference range are indicated in bold.
* Normal range according to EPA (Environmental Protection Agency) (https://www.epa.gov/criteria-air-pollutants/naaqs-table).
The mean AQI across all cities was 55 (IQR 44-56), and the average PM2.5 concentration was 17.5 μg/m33 (IQR 10-27), both exceeding EPA pollution limits. Technical challenges limited PM2.5 measurements to four cities (Kilis, Gaziantep, Adana, and Osmaniye), and all recorded values exceeded EPA thresholds. The highest AQI value was recorded in Hatay (56 [IQR 51-60.5]). Except for Adıyaman, all provinces had AQI values exceeding EPA recommendations. The highest CO concentration was recorded in Hatay (1621.5 ± 1111.4 μg/m3). PM10 values also surpassed EPA limits, with the highest values observed in Osmaniye (80 μg/m3 [IQR 72-82]), Kahramanmaraş (65 μg/m3 [IQR 52-77]), Hatay (56 μg/m3 [IQR 45.5-89), and Kilis (54.3 ± 15.7 μg/m3) (see Table 3).
Comparison of AQI Values in the Year Preceding (AQI1: 2022) and the Year Following (AQI2: 2023-2024) the Earthquake by Province
Using the admission dates, times, and locations of the study participants, AQI values were recorded for the year preceding the earthquake (2022, AQI1) and the year following the earthquake (2023-2024, AQI2). Analysis revealed that the mean AQI values in Kahramanmaraş, the earthquake epicenter (AQI1 = 48.5 [IQR 48-55], AQI2 = 55 [IQR 55-80]; P = 0.007), and Hatay, the city with the most significant destruction Hatay (AQI1 = 42 [IQR 33-53], AQI2 = 56 [IQR 51-60.5]; P <0.001), were significantly higher post-earthquake compared to the previous year (see Table 4). In another metropolitan city, Gaziantep, the AQI (AQI1 = 44 [IQR 41-56], AQI2 = 55 [IQR 54-55.5]; P = 0.014) was also found to be significantly higher after the earthquake (see Figure 4).
Comparison of AQI values in the year preceding (AQI1: 2022) and the year following (AQI2: 2023-2024) the earthquake by province
Note: Bold values indicate statistical significance.
AQI1: AQI values in the year preceding earthquake (2022).
AQI2: AQI values in the year following earthquake (2023-2024).
Geographic distribution of CO poisoning cases and average AQI levels 1 year after the earthquake.
Complications and Treatments
The most frequently encountered complications were hyperglycemia (29.8%), leukocytosis (19.9%), and myocardial injury (15.9%). The distribution of complications by province is presented in Table 5. Statistically significant differences were observed in the incidence of leukocytosis (P = 0.004) and myocardial injury (P <0.001) between provinces. The highest incidence of leukocytosis and myocardial injury was observed in Adana. This difference between provinces is attributed to disparities in access to health care services. Among the cities included in the study, Adana was the only city with a 24/7 hyperbaric oxygen therapy (HBOT) center, which resulted in the referral of more severe cases to this center. The average AQI of patients with myocardial injury was 55 (IQR 37-55), while the average AQI of patients without myocardial injury was 55.5 (IQR 54-55.9), with no statistically significant difference (P = 0.295).
Distribution of complications by province
Note: Bold values indicate statistical significance.
To further assess the factors associated with myocardial injury, a multivariable logistic regression analysis was performed, including age, sex, comorbidities, COHb, AQI, PM10, and SO2 levels as covariates. The results indicated that age (OR: 1.054, 95% CI: 1.011-1.099, P = 0.013) and COHb levels (OR: 1.059, 95% CI: 1.005-1.116, P = 0.032) were independently associated with myocardial injury. In contrast, AQI (P = 0.160), PM10 (P = 0.793), and SO2 (P = 0.993) were not significantly related to the outcome. These findings suggest that individual-level exposure markers may be more predictive of cardiac complications than ambient pollution indices alone (see Table 6).
Multivariate logistic regression analysis of independent predictors of myocardial injury
Note: Bold values indicate statistical significance.
The average ED length of stay was 5 hours (IQR 3-6). A statistically significant but weak correlation was found between AQI and ED length of stay (r = 0.218, P = 0.007), which is unlikely to be clinically meaningful.
High-flow nasal cannula oxygen therapy (HFNCO) was administered to 9.9% (n = 15) of patients, and 10.6% (n = 16) received HBOT. HBOT distribution by province was as follows: Kilis (n = 1), Adiyaman (n = 1), Adana (n = 10), Kahramanmaras (n = 1), Malatya (n = 2), and Osmaniye (n = 1). The average AQI of patients who received HBOT was 55 (IQR 37-55.25), while the average AQI of those who did not receive HBOT was 55.5 (IQR 54-55.9), with no statistically significant difference (P = 0.558). The hospitalization rate was 11.9% (n = 18), with a median hospital stay of 1.5 days (IQR: 1-2 days). No neuropsychiatric sequelae developed in any of the patients, and no mortality was observed in any of the cases.
Discussion
This study aimed to evaluate the relationship between carbon monoxide (CO) poisoning cases presenting to emergency departments in earthquake-affected cities and concurrent changes in air quality one year after the February 6, 2023, Kahramanmaraş earthquake. We observed a significant increase in overall air pollution indices after the disaster, with mean AQI values in several provinces exceeding the thresholds recommended by the EPA. Notably, both Kahramanmaraş, the earthquake’s epicenter, and Hatay, the most severely affected province, showed statistically significant increases in AQI compared to the preceding year. (Kahramanmaraş: AQI1 = 48.5 [IQR 48-55], AQI2 = 55 [IQR 55-80]; P = 0.007; Hatay: AQI1 = 42 [IQR 33-53], AQI2 = 56 [IQR 51-60.5]; P <0.001). Earthquakes are sudden-onset disasters, primarily due to the immediate physical destruction they cause. However, the subsequent processes of debris removal and reconstruction contribute to slow-onset health crises, introducing PM and further deteriorating air quality.Reference Bayram, Rastgeldi Dogan and Şahin12, Reference Ha13 Particulate matter (PM)—comprising solid particles and liquid droplets ranging in size from 0.001 to 100 μm—is a critical pollutant recognized for its adverse effects on human health, ecosystems, and infrastructure.Reference Farrow, Miller and Myllyvirta14 In 2021, PM-related air pollution accounted for 8% of global Disability-Adjusted Life Years (DALYs), underscoring its major public health burden.15 In this study, AQI and PM10 levels in Kahramanmaraş exceeded EPA thresholds, while Hatay recorded the highest CO levels alongside AQI and PM10 concentrations surpassing regulatory limits. These findings highlight the environmental consequences of post-earthquake debris removal and reconstruction activities, emphasizing that the resulting deterioration in air quality has the potential to exacerbate both respiratory and cardiovascular conditions, particularly among vulnerable populations. Therefore, improving post-disaster air pollution monitoring systems and implementing targeted interventions to mitigate exposure are of critical importance to safeguard public health.Reference Uprety, Ozaki and Higuchi7, Reference Bayram, Rastgeldi Dogan and Şahin12
This study hypothesized that increased air pollution could have led to heightened myocardial damage in CO poisoning cases in the earthquake-affected regions. Although ambient air pollution levels were elevated, our findings did not reveal a statistically significant independent association between AQI or PM10 and acute myocardial injury in patients with CO poisoning. In the multivariable logistic regression model, age and COHb concentration emerged as independent predictors, while AQI and PM10 were not significant. These findings suggest that individual-level exposure markers may be more informative for predicting acute cardiac complications in CO poisoning than ambient air quality indices. A statistically significant correlation between AQI and ED length of stay was observed (r = 0.218, P = 0.007); however, this effect size was small and is unlikely to be clinically meaningful. The highest frequency of myocardial injury was observed in Adana. This may partly reflect referral bias, as Adana hosts the only 24/7 HBOT center in the region and likely received more severe cases. However, because we lacked detailed referral pathway data, this explanation remains speculative and should be interpreted with caution. These results suggest that while environmental pollution remains a relevant concern in post-disaster settings, it may not directly influence acute myocardial injury in CO poisoning cases. Nevertheless, long-term exposure to air pollution is known to increase cardiovascular risks.Reference Alexeeff, Liao and Liu16 This is particularly concerning in post-disaster settings, where individuals may experience combined exposure to both indoor and outdoor pollutants. The potential synergistic effects of these exposures warrant further investigation. Regular air quality monitoring and long-term follow-up studies are essential to better understand both the immediate and chronic health outcomes in affected populations.
CO poisoning is a multifactorial process influenced by the concentration and duration of CO exposure, individual susceptibility, and underlying comorbidities. In our cohort, indoor exposure from stoves was the primary cause of poisoning. Indoor air pollution, particularly from smoke, may elevate baseline COHb levels and increase vulnerability to CO-related toxicity.Reference Ruan, Deng and Wang17, Reference Manisalidis, Stavropoulou and Stavropoulos18 Ambient pollutants such as PM2.5 and PM10 have been associated in prior literature with oxidative stress, endothelial dysfunction, and adverse cardiovascular outcomes.Reference Simkhovich, Kleinman and Kloner8 Prior literature have demonstrated that elevated air pollution levels, including PM10 and PM2.5, are associated with higher risks of acute coronary syndromes,Reference Peters, Dockery and Muller19 ischemic heart disease,Reference Host, Larrieu and Pascal20 and myocardial infarction readmissions.Reference Von Klot, Peters and Aalto21 Recent evidence has even linked short-term exposure to air pollutants—such as PM2.5, NO₂, and CO—to increased risk of acute coronary events within hours of exposure.Reference Chen, Jiang and Hu22 While our dataset did not confirm an acute synergistic effect between indoor and outdoor pollutants on myocardial injury, such interactions remain biologically plausible and merit investigation in future prospective studies with larger sample sizes and more comprehensive exposure data.
Limitations
This study has several limitations. First, its retrospective design and the inclusion of data from only 8 of the 11 affected provinces (Şanlıurfa, Diyarbakır, and Elazığ were excluded due to data limitations) may restrict generalizability. In addition, in provinces such as Malatya, Adıyaman, and Kahramanmaraş, comprehensive monitoring data for certain pollutants (e.g., PM2.5, NO₂) were unavailable, potentially reducing the accuracy of regional comparisons. In particular, the absence of PM₂.₅ and NO₂ measurements may have led to under- or overestimation of local air quality, thereby limiting our ability to fully capture geographic variability in exposure. These gaps may have introduced regional bias, especially considering the heterogeneous environmental and infrastructural conditions after the earthquake. Despite these limitations, the inclusion of 8 provinces and the application of standardized data collection methods offer meaningful insight into post-disaster public health risks associated with air quality. Second, the number of air quality monitoring stations in the earthquake-affected area was limited, and these stations measure only a restricted set of air pollution parameters. Additionally, for Hatay, air quality data were obtained from a monitoring station located approximately 20 kilometers from the patients’ residential areas, which may have introduced potential exposure misclassification. To ensure better public and environmental health protection following large-scale disasters, expanding the network of monitoring stations and broadening the range of parameters measured would be essential. Despite these challenges, the available monitoring data remain useful for the overall assessment of air quality trends. Third, myocardial injury was primarily defined based on elevated troponin levels, which provided a standardized and widely accepted criterion. However, ECG and echocardiographic findings were not uniformly available across centers and were used as supportive rather than mandatory criteria. This heterogeneity in outcome assessment may have limited the validity of comparisons across sites. Fourth, socioeconomic status, access to care, or smoking history was not systematically recorded, and data on other comorbidities were limited. Although we performed sensitivity analyses including diabetes mellitus, asthma, and coronary artery disease, unmeasured confounding cannot be excluded. Fifth, detailed referral pathway data were not available, which limits our ability to fully explain the higher complication rates observed in Adana. Finally, the relatively small number of patients presenting from Malatya, a province that experienced severe destruction, may reflect the relocation of a large proportion of the population to container housing units heated by electric stoves or air conditioners. This factor may have influenced both exposure levels and observed health outcomes. In light of these limitations, our findings should be interpreted with caution. Nevertheless, this study provides novel insights into the environmental and clinical aspects of CO poisoning in a post-disaster context and highlights the importance of implementing robust air-quality monitoring and conducting prospective, multicenter studies to better clarify the acute and long-term cardiovascular risks associated with disaster-related pollution.
Conclusion
In conclusion, earthquakes not only cause physical destruction but also accelerate environmental health crises. One year after the February 6 earthquake, AQI and PM2.5 levels in the affected region were found to exceed EPA thresholds. Notably, AQI values in Kahramanmaraş, Hatay, and Gaziantep were significantly higher compared to the previous year, indicating a substantial increase in air pollution. Although our analysis did not identify a statistically significant association between AQI or PM10 levels and myocardial injury among CO poisoning cases, individual-level factors—particularly COHb concentration and age—appeared to be more predictive of myocardial injury. These findings highlight the need to enhance air quality monitoring networks and to conduct regular environmental and clinical assessments following disasters. Future large-scale and prospective studies are warranted to better understand the cumulative and potentially synergistic effects of indoor and outdoor air pollutants on cardiovascular health in post-disaster settings.
Data availability statement
Data will be made available on reasonable request.
Acknowledgments
None.
Author contribution
All authors contributed to the study conception, material preparation, data collection, and design. Analyses were performed by MG, SS, MS, and SA. The first draft of the manuscript was written by MG, SS, MS, and SA, and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.
Funding statement
None.
Competing interests
The authors have no conflicts of interest to declare.
