Based on our results, it can clearly be seen that CO poisoning is a substantial public health problem in Turkey which kills hundreds of people annually. The stagnating pattern of CO related deaths in the 10-year period may show the underestimated importance of CO poisoning which may be caused by low death rates in previous reports. To our knowledge, there were only two nationwide reports showing 84 and 39 CO related deaths in the years 2008 and 2010 respectively, which does not seem realistic in a developing country [6, 13]. In our opinion, we may also have undercalculated the exact risk due to the fact that our methodology was only based on media reports. However, our study is the first report that shows the trend of CO related deaths in the last decade with more reasonable death rates in a developing country.
In a study conducted by the World Health Organization (WHO) in all its 53 Member States of European Region [6], despite heterogeneity in reporting between countries, the annual death rate of CO poisoning was found to be 2.24 per 100,000 population. In most countries, the trends of CO-related death rates were either steady or slowly decreasing similar to our findings. In that report, for instance, in Belarus and the Czech Republic the average annual death rates by CO poisoning were 11.99 and 2.62 per 100,000 population respectively. Considering the similar socioeconomic status of both of these developing countries to Turkey, more precise methods to show the real burden of the present health problem need to be applied. Local media in Turkey is a continuously working area that may have the potential to reach the farthest parts of Turkey. By evaluating their reports, we reached an average of 267 deaths per year that helps to emphasize the importance of the current problem to take preventive measures. Winter, unsurprisingly, was the season in which the death incidents peaked similar to other reports [14,15,16]. Also, we calculated the risk of CO-related death according to the provinces to discover the most vulnerable places to poisoning. The Middle Anatolian region had the highest risk of CO-related deaths. The Middle Anatolian region is a rural area with strong winters that could explain the widespread and relatively long utilization of stoves for heating purposes. Additionally, socioeconomic status is a known risk factor of CO poisoning [17]. The socioeconomic status in Turkey improves from east to west regions gradually. The socioeconomic status of Middle Anatolian region is relatively low compared to other parts of Turkey, other than the Eastern Anatolian region, which may have added to the increased risk of stove related deaths. Interestingly, in the Eastern Anatolian region, which has the lowest socioeconomic status in Turkey, relatively low CO-related death rates were observed. In our opinion, the lack of local media penetration into the Eastern parts of Turkey lead to the underreporting of CO poisoning related death incidents. Therefore, the risk of death may actually be higher in the Eastern Anatolian region than reported. Istanbul which is the most developed province in Turkey, have lower Co-related death rate. Major but relatively underdeveloped eastern province (Gaziantep) have 8 times more death rate compared to Istanbul. Ankara which is capital and second major developed city and izmir which is third major and developed city have similar death rate. But even here, there are twofold death rate compared to Istanbul. These differences can be thought to be due to the weather conditions and the density of natural gas use.
A study conducted by Metin et al., that investigated CO poisoning cases and related deaths in 2010 in Turkey using ICD-10 (International Statistical Classification of Diseases and Related Health Problems) coding system, showed only 39 CO related death incidents [13]. The majority of the poisoning cases, whose sources were known, were caused by stoves and show similar seasonal pattern with our study. However, in a developing country where stoves are still widely used, that number seemed extremely low and could indicate underreporting of CO poisoning incidents. We therefore evaluated media reports, where the incidents were recorded regularly, to obtain more logical findings. Certainly, underreporting still existed in our data, since incident reports from the Eastern Anatolia region, the coldest region of Turkey with a relatively low socioeconomic status, were extremely rare. However, 0.35 deaths per 100,000 population is the highest rate that has been reported so far. Another study executed by Akköse et al. from Bursa [7], evaluated CO related emergency department and intensive care unit admissions in 10 years’ period. All the poisoning cases were reported as accidents except one case which was reported as a suicide attempt. The majority of the poisoning cases were due to coal heaters which is in accordance with our study. Of the 305 admissions, 10 patients died in the study period. Considering Bursa is one of the provinces with high CO related death incidents in our study, those findings reveal that deaths could generally occur prior to hospital admission in CO poisoning. And also, local hospital admissions may show the real burden of the poisoning rather than single institution records. Other reports including CO-related deaths in Turkey are from medico-legal autopsies. In a review which was performed by Karapirli et al. [16], total 47,523 autopsies from 7 cities were evaluated within 27-year period. Although study period differed between provinces, 980 CO-related deaths were found. Although not report a comparable rate, the study is one of few studies and showing about 2% CO deaths of autopsies in Turkey. Taken into account significantly higher number of deaths from those reports, to obtain the nationwide burden of CO-related death incidents collaboration between forensic medicine, public health departments and emergency departments might be needed.
The results of our study showed that stoves are still frequently being used and are the cause of death especially in rural areas with lower socioeconomic status. The majority of the deaths occurred in domestic conditions with the use of old fashion stoves for heating purposes in accordance with the literature [18]. Currently, there are no proper measures being taken to prevent CO-related deaths in Turkey. This may have caused underestimation of the problem due to lack of data regarding to real burden of the poisoning. Certain measures can be taken to decrease the burden of the CO-related accidents. Increased use and regular maintenance of CO detectors have been shown to be an effective way to diminish the both morbidity and mortality related to the poisoning with additional cost saving benefits [19]. To decrease stove related poisoning incidents, encouragement to use natural gas in the both heating and cooking systems should be deliberated. If there is no alternative heating system without stoves, the use of quality coal should be encouraged. The content of coal (calorie, sulfur, etc.) and properties (humidity, ash etc.) affect the quality of coal and poor quality coal are reported to produce CO gas due to incomplete combustion [20]. Public education campaigns and prevention programs organized by the authorities’ can be reduced mortality and morbidity in winter months. Schools, governmental and non-governmental organizations can educate the community about the ventilation of places where coal stoves are used as heating system. Warnings on CO poisoning can be ensured through media and social media on risky days. Routine evaluation of the chimneys by professionals to prevent any blockage that may lead to backflow of the smoke is another crucial step to hinder CO to accumulate within closed places. Even though stove is the main source of poisoning, regular maintanence of water heater, gas heater and other engine should have done to prevent deaths. Although not common, using a motor vehicle, generator or any engine using gasoline within an improperly ventilated area should also be avoided [21].
Limitations
There are several limitations of our study. First and foremost, we evaluated the death incidents from media reports without any forensic medicine-based diagnosis. By doing this, we may have included some deaths not related with CO poisoning. In addition, lack of media reports from some parts of Turkey might have failed to show the real risk in those areas. (to ignore this issue, not being able to make news due to more intensive agenda at that time, not worth publishing the death news which occurred after the incident, etc.) In our opinion, we may also have undercalculated the exact risk due to the fact that our methodology was only based on media reports. It can be possible that over-reporting due to population or interest in some regions. Also, there might be some duplications left due to misspelling or absence of victim names, despite this we tried to exclude the duplicated cases in many ways. However, this is a report which showed logical CO-related death rates in a developing country in a 10-year period.