2022 Volume 63 Issue 6 Pages 267-273
The frequent occurrence of mushroom poisoning cases in the wet season in Thailand has long been recognized but has never been quantitatively analyzed. This study aims to analyze mushroom poisoning cases in Thailand between 2003 and 2017 and focused on their association with the rainfall. The results revealed 22,571 cases and 106 deaths in this period. Cases were higher for females than males, adults than children, and agriculturists than people in other occupations. Cases were higher in the northeastern and northern regions than in the central, south, and east regions. There are strong effects of seasonality on mushroom poisoning cases. Over the 15-y period, 17,337 cases occurred in the wet season (May-Sep) compared with 5,234 cases in the dry season (Oct-Apr). The number of monthly poisoning cases were high in men, mature adults, agriculturists, people living in the rural areas and people living in the Northeastern and Northern provinces. Moreover, a strong positive correlation was shown between the number of monthly poisoning cases and the amount of monthly rainfall.
Mushroom poisoning is an important food-borne disease causing a problem for public health in many countries. Reports from various countries have been published, such as Japan (Ishihara & Yamaura, 1992), China (Chen, Zhang, & Zhang, 2014; Jiang, Luo, & Hao, 2018; Li et al., 2020), Malaysia (Supramaniam & Mohanadas, 1980), Nepal (Das, Parajuli, & Jayakumar, 2007), Turkey (Cevik & Unluoglu, 2014; Vişneci, Acar, Özdamar, Güven, & Patat, 2019), and the USA (Brandenburg & Ward, 2018). In Thailand, annual summaries of mushroom poisoning have been published consistently in the Annual Epidemiological Surveillance Reports since 2003 and are published frequently in the Weekly Epidemiological Surveillance Reports (Wiengpitak, 2004; Niramitsantipong, 2008; Kanjanasombut, Ratanarat, & Uppapong, 2012). Chaiear, Limpaiboon, Meechai, and Poovorawan (1999) reported five clinical cases of fatal poisoning from Amanita virosa. Boonthavikoon (2002) provided a description of the morphology and habitats of six edible and 17 poisonous mushrooms in Thailand, with notes on their toxins. Boonpratuang, Choeyklin, Promkium-on, and Teeyapan (2012) examined 113 samples of mushrooms suspected as poisonous species, revealing nine distinct species: Amanita bisporigera, A. verna, Boletus patouillardii, Chlorophyllum molybdites, Entoloma talisporum, Geastrum saccatum, Pisolithus tinctorius, Psilocybe cubensis, and Scleroderma verrucosum. Parnmen et al. (2016) identified mushroom samples from nine cases between 2008 and 2014 and reported Amanita exitialis to be a poisonous species for the first time. Tawatsin, Parnmen, Thavara, Siriyasatien, and Kongtip (2018) provided information on the mushroom poisoning incidence in Thailand between 2008 and 2017 and provided details on four groups of toxins from poisonous mushrooms.
This study aims to analyze the occurrence of mushroom poisoning in Thailand from the primary available data in a rather long period record (2003-2017), comparing with those reported in other countries. We also attempt to elucidate the factors that are correlated with an increase of the mushroom poisoning. The information acquired from this study will be beneficial to forecast and reduce the mushroom poisoning incidence in the future.
Data on mushroom poisoning cases were collected by the Center of Epidemiological Information, Bureau of Epidemiology, Department of Disease Control, Ministry of Public Health and are available online through the Annual Epidemiology Surveillance Report (https://apps.doe.moph.go.th/boeeng/annual.php). The datasets used in the analysis included monthly cases and deaths throughout the whole country and in each province between 2003 and 2017; yearly cases and deaths categorized by gender, age class, and occupation in the same period; the total mid-year populations; and the mid-year populations of each gender and each age class. Although case and case rate datasets on mushroom poisoning in Thailand reported by the Annual Epidemiology Surveillance Report date back to 1971, the categorical datasets such as cases per gender, age classes or occupation are available only since 2003. The datasets in 2018 are not included in the analysis because the number of cases is too high (4,084 cases, 3.3 times higher than in 2017), which resulted in an outlier in the statistical analysis. The dramatic increase in the number of patient cases in 2018 is worth further analysis if sufficient data are available. The dataset was verified by using the online Annual Epidemiology Surveillance Report (http://www.boe.moph.go.th/boedb/surdata/). The verification from the two sources revealed few discrepancies in 2012, 2015, and 2016, and the data from the Annual Epidemiology Surveillance Report were selected in this study.
The average monthly rainfall data (2003-2017) were collected by the Meteorological Department, Ministry of Information and Communication Technology and is available online through the website of the National Statistical Office (http://webhost.nso.go.th/nso/project/search_cen/index.jsp). Additional data from the Provincial Agriculture and Cooperatives Database (http://www.moac-info.net/modules/reports/I802.php) were used to complete missing or incomplete records.
2.2. Data analysesThe datasets were analyzed for descriptive statistics. The mushroom poisoning cases or deaths were converted into rates (per 100,000 people) by dividing the number of cases or deaths by the mid-year population. Variance distributions were visualized by histograms and tested for normality by the Shapiro-Wilk test. The preliminary test results revealed that the monthly cases were nonnormally distributed, so nonparametric statistics were further employed, and the median was used as a representative for the central tendency. The Friedman test was used for multiple comparisons of categories across the time periods. The significance level before adjustment was set such that alpha was 0.05. The Wilcoxon signed-rank post hoc test with the Bonferroni significance adjustment was used for paired category comparisons across the time periods. Charts showing the case data against the months were constructed to visualize the changes over time and possible association patterns. The additive decomposition model of time series analysis was used to help clarify the seasonality and trend or cycle. Spearman's rank correlation analysis was employed to determine the correlation between monthly cases and monthly rainfall of the whole country, and of each province. The provinces in which fewer than 10 cases occurred between Jan 2003 and Dec 2017 were not analyzed for the correlation because of relatively low samples for correlation confidence. The calculations were performed using SPSS 11.5 for Windows. A provincial vector map of Thailand was obtained from City Hubs Corporation Co., Ltd. (https://www.cityhubs.net/news/43-thailand-powerpoint-map.html). SPSS or Microsoft Excel originally generated charts including time series and boxplots before being included in illustration plates.
During the 15-y period from Jan 2003 to Dec 2017, a total of 22,571 mushroom poisoning cases and 106 deaths were reported in Thailand. The yearly cases ranged from 1,232 cases in 2014 (1.9 per 100,000 people) to 2,148 cases in 2012 (3.3 per 100,000 people). The lowest yearly death was 1 in both 2005 and 2006, and the highest was 24 in 2012 (0.04 per 100,000 people). The monthly cases ranged from 9 cases in Apr 2016 and Feb 2017 to 438 cases in May 2012, which included the greatest number of deaths-19.
A total of 9,245 cases (41%) in men and 13,326 cases (59%) in women (0.69 man by 1 woman) were reported during this 15-y period. The yearly cases in women ranged from 727-1,289 (2.3-3.9 per 100,000 people), while those in men were between 457 and 859 (1.4-2.7 per 100,000 people). The yearly cases and case rates in women were significantly higher than those in men (p = 0.001) (Fig. 1A, B). However, there were 53 deaths of both men and women (Fig. 1C), and there was no significant difference between males and females in yearly deaths (p = 0.69) (female yearly deaths ranged from 0-8, with a median of 3, while those of males ranged from 0-7, with a median of 2) (Fig. 1C). Cases were higher in groups of mature adults (35-44 y) (case median of 261) than children (14 y or below) (case medians of 157 or 70) or elderly people (65 y or older) (case median of 185). However, cases of mature adults (34-44 y) and elderly people (65 y or older) were not significantly different. (Fig. 1D). The people at risk by mushroom poisoning in descending order include agriculturists, laborers, students, children, housewives, government officers, and traders. Cases in agriculturists (case median of 762) were higher than in other people (Fig. 1E). Cases in patients living in rural areas (case median of 1,301) were significantly higher than those in patients living in urban areas (case median of 189) (p = 0.003) (Fig. 1F). Regarding annual case distributions, mushroom poisoning cases have been reported every month (Fig. 1G). The numbers of accumulated cases between 2003 and 2017 in each calendar month were 513; 517; 569; 896; 3,730; 3,623; 3,860; 3,487; 2,637; 1,443; 762, and 534, from Jan to Dec respectively. The number of cases increased drastically in May and remained high until Aug, after which they declined. The number of monthly cases from May to Aug (case medians between 224 and 255) was non-significantly different (p≥0.00076) but was significantly higher than the number from Oct to Apr (case medians between 29.5 and 94) (Fig. 1G). Seasonality and correlations between cases and rainfall are analyzed and discussed below.
Regarding spatial distribution, 76 provinces were reported to have cases of mushroom poisoning, with the cases in each province ranging from 1 to 4,005, while zero cases were reported in only one province (Supplementary Table S1). Cases were high in many northeastern and northern provinces than in the south, central, and east provinces (Fig. 2A). According to the provincial distribution by quartile, provinces whose case rates were in the top quartile were the border provinces in the northern and northeastern regions and one province in the southern region (Fig. 2B). The case rates were higher than 2.75 but ranged mostly from 4-6; however, some were greater than 10 per 100,000 people in a few provinces. The third quartile included some provinces in the northeastern region, many provinces in the lower northern region and provinces mostly in the upper southern and eastern regions; the case rates were 1.2-2.7 per 100,000 people. Provinces with case rates in the lower quartiles (quartiles 1-2) included central provinces and some provinces scattered in the lower southern and middle northeastern regions, with case rates lower than 1.1 per 100,000 people. The case and case rate showed a somewhat similar pattern (Fig. 2A, B).
In Thailand, May to Sep constitutes the wet season, while Oct to Apr constitutes the dry season (the cool dry period last from Oct to Feb, and the hot dry period occurs during Mar and Apr). The mushroom poisoning cases were highly affected by seasonality. During the 15-y period, the number of cases in the wet season (17,337 cases, or 77% of the total cases) was approximately three times higher than that in the dry season (5,234 cases, or 23%). Seasonality is also clearly shown as having an effect in the time series chart according to the additive decomposition model analysis (Fig. 3). The seasonal indexes of each month from May to Sep were 133.2, 109.4, 131.5, 116.8, and 54.1, respectively, while the indexes of the remaining months ranged between -32 and -95. The chart also shows an increasing trend of mushroom poisoning between 2003 and 2012 and a decreasing trend between 2013 and 2017 (Fig. 3). Regarding spatial distribution, 66 provinces had a higher number of cases in the wet season than in the dry season, whereas eight provinces had the opposite pattern, and two provinces had equal cases between the wet season and the dry season (Supplementary Table S1).
The distribution of the monthly mushroom poisoning cases was positively skewed (1.07 skewness, mean of 152.3 and median of 79). Approximately one-third of the months (63 mo) had between 25 and 75 cases. Additionally, the monthly rainfall data showed a bimodal distribution. The first group represented months in the dry season, in which the monthly rainfall was less than 150 mm, with the majority between 25 and 75 mm (the lowest was 3.6 mm in Feb 2005 and 2014). The latter group represented months in the wet season, in which the monthly rainfall was more than 150 mm, with the majority between 200 and 220 mm (the highest was 319.7 mm in Sep 2011).
In Thailand between 2003 and 2017, the annual period with high mushroom poisoning patients corresponded with the period of the high rainfall (Fig. 4A). The analysis showed a strong positive Spearman's rho correlation coefficient and a significant correlation (rs=0.801, p<0.001) between the average monthly rainfall and monthly mushroom poisoning cases in the country. The scatter diagram between their ranks clearly revealed a monotonic relationship; namely, an increase in mushroom poisoning cases occurred with increasing rainfall (Fig. 4B).
In summary, the mushroom poisoning cases between 1971 and 2002 (32 y) were 14, 682, and the patient deaths were 209. The fewer number of patients despite the longer period between 1971 and 2002 compared with the patients in the period between 2003 and 2017 may have resulted from the capability of patients to go to hospitals. Compared with Japan and other Asian countries, Thailand has a relatively high annual average number of patients and patient deaths (Supplementary Table S2). Moreover, while the number of patients in Japan has tended to decline (Ishihara & Yamaura, 1992; Yamaura, 2013), the number of patients in Thailand still fluctuates.
People living in rural areas of Thailand are more vulnerable and exposed to poisonous mushrooms than those living in densely populated cities. Villagers in rural areas have an opportunity to collect mushrooms from forests nearby. This rationale has resulted in higher epidemiology rates in the sub-district administrative organization (SAO) than in municipalities (Fig. 1F). Many rural people are farmers resulting in high cases of mushroom poisoning among agriculturists than in other occupations (Fig. 1E). Mushroom is a pricey forest product usually harvested by villagers for household consumption or sale in their local markets (Puangmalee & Suksard, 2013; Noinarai, Sunthornhao, & Pothitan, 2020). Meanwhile, many local epidemiological newsletters issued warnings of mushroom collection from the wild because most mushroom poisoning cases have resulted from such harvest (Kongyu, Saritapirak, & Iamsirithawon, 2005). In contrast, people living in the municipalities or in cities usually buy cultivated edible mushrooms from supermarkets. Chan, Lam, Chiu, Tse, & Lau (2016) mentioned that in an urban territory such as Hong Kong, the commercially sold cultivated mushrooms are easily available and consumed by most citizens everyday but mushroom poisoning occurs from people collecting and consuming wild mushrooms in the belief of their nutritional and medicinal values. Poisoning cases from commercial or cultivated mushrooms were also reported in Switzerland (Schenk-Jaeger et al., 2012). Vişneci et al. (2019) reported that 34% of the patients in Konya province of Turkey were caused by cultivated mushrooms whereas 66% were caused by wild mushroom consumption. However, food poisoning cases from eating commercial or cultivated mushrooms are very rare in Thailand.
Epidemiological inequality in gender (between men and women) or age (between children and adults or elders) is observed for mushroom poisoning in Thailand. In Thailand females have a higher patient record (59%) than males (41%), in agreement with those published by Chan et al. (2016) (57% of female) in Hong Kong, Schenk-Jaeger et al. (2012) (51.4% of female) and Keller et al. (2018) (56.9% of female) in Switzerland, Eren et al. (2010) (59% of female), Yardan et al. (2010) (67.5% of female), Cevik and Unluoglu (2014) (53.3% of female), and Vişneci et al. (2019) (53.0% of female) in Turkey. In contrast, in Florida, USA, during 2003-2007 and in Yunnan, China, during 2004-2016, most poisoning cases occurred to males than females, namely, 65% of male in Florida and 59% of male in Yunnan (Kintziger et al., 2011; Jiang et al., 2018). We think these variabilities resulted from the socioeconomic or socio-culture of each locality. In Thailand, women have more opportunities to be exposed to poisonous mushrooms in the gathering and food preparation process.
The distribution of patients in different age classes in Thailand corresponds to many foreign records that the number of adult patients was higher than children, for example, in Turkey (Eren et al., 2010) and Switzerland (Schenk-Jaeger et al., 2012). In Thailand, children mostly spend their days in kindergartens (3-6 y) or schools (7-18 y) where food with good nutrition is prepared under the national policy on the school lunch program (Kongnoo, Loysongkroa, Chotivichien, Viriyautsahakul, & Saiwongse, 2014) resulting in less risk to poisonous mushroom consumption.
In Thailand, rainfall is considered as one of the most important factors that account for the high productivity of wild mushrooms and high mushroom poisoning cases. Indeed, high productivity of wild mushroom in the rainy season in Thailand was indicated in previous study, in which the biodiversity of wild mushroom in the Western province in Thailand was shown to be correlated positively with the amount of rainfall (Sutjaritvorakul, Permpoonsinsup, Srigobue, and Koomsubsiri, 2017). Moreover, in Thailand, sale prices of wild mushrooms are generally higher than those of other forest products (Puangmalee & Suksard, 2013; McLellan & Brown, 2017). Thus, during the rainy season, local people in the North and Northeastern regions rush into forests to search and harvest wild mushrooms for household consumption or sale along roadsides and in local markets (Jones, Whalley, & Hywel-Jones, 1994; Pegler & Vanhaecke, 2001; Sangvichien & Taylor- Hawksworth, 2001). A large amount of wild mushroom harvest during the heavily rainy often results in accidental mixing of poisonous species with edible ones because several poisonous and edible mushrooms are morphologically similar, such as the poisonous Amanita exitialis and the edible Amanita princeps, frequently misidentified species (Rungsriwong & Siriarayaporn, 2016; Parnmen et al., 2016; Tawatsin et al., 2018). The Department of Disease Control of Thailand, therefore, often warns to be careful with the risk of wild mushroom consumption during the rainy season (Sakaraserane et al., 2015; Sompong et al., 2020). Such information supports the positive correlation between rainfall and mushroom poisoning cases in this study.
There is a limitation to this study. The unreported cases from people with minor or no harm or people who have no access to public health facilities lead to an underestimated number of cases (Brandenburg & Ward, 2018). However, data from the long period record used for the analysis in this study provide more confidence for epidemiological interpretation. The benefit of this study is to increase awareness about poisonous mushrooms, especially during the early wet season as rainfall and mushroom poisoning cases are strongly correlated.
This study does not involve conversations, interviews or diagnoses of any patient directly. The data used in this study are from public reports, and the data sources are clearly declared. No individual identification or personal information of a patient is disclosed. None of the authors have any conflicts of interest. All the authors have made significant contributions to this work.
The whole study is a part of basic information analysis of the mushroom importance in the project “Shaping modern fungal taxonomy - an integrative approach using morpho-molecular characterization, chemotaxonomy, and Next Generation Sequencing of invertebrate-pathogenic fungi and edible mushrooms” from BIOTEC, Grants P-19-50231. Center of Epidemiological Information, Bureau of Epidemiology, Department of Disease Control, Ministry of Public Health; and National Statistical Office are acknowledged for the primary data accession. City hubs corporation Co. Ltd. is thankful for a free download vector map.
