Nationally, hanging is now the most common method of suicide in Sri Lanka [25] [26]. However, our data reveal that in rural agricultural parts of Sri Lanka where pesticides are widely used and available, they remain more than twice as important as hanging for suicide. Notably, this data set identifies the insecticides carbosulfan and profenofos to be the key pesticides involved in suicides, likely responsible for more than half of pesticide suicides and almost 1 in 5 of all means of suicides. Many of the unknown fatal OP or carbamate insecticide poisoning are likely to be due to these commonly used insecticides as well.
The pattern of pesticides responsible for suicides changed markedly between 2002–08 [18] and 2011–16 (figure 4). In 2002–08 and in 2011–16, the proportion of suicides due to herbicides deceased from 45% to 25% whereas suicides due to insecticides increased from 54% to 75% respectively. In 2002–08, carbosulfan and profenofos were responsible for 7% of pesticide suicides in a large prospective secondary hospital cohort in the North Central Province of Sri Lanka [18]. However, ingestion of carbosulfan or profenofos was responsible for 59 (56.2%) of the suicides for the period of 2011–16.
A recent hospital-based study from Peradeniya, in central Sri Lanka, identified profenofos as a key cause of death but also multiple long intensive care admissions due to the OP insecticide quinalphos [27]. However, although cases of quinalphos self-poisoning have been noted in North Central Province (ref), few deaths were reported in this study - perhaps because quinalphos is not a commonly used insecticide among local farmers.
It is important to know why carbosulfan and profenofos are such commonly used in suicides in recent years. A study conducted by Rathish and colleagues reported that carbosulfan and profenofos were the two highest selling pesticides in local rural shops after 2015 [28]. One explanation is that the withdrawal from agriculture of dimethoate and fenthion over three years (2008–2011) followed by chlorpyrifos in 2014 has resulted in a rapid increase in use of carbosulfan and profenofos in agriculture and then suicide. In addition, the importance of carbosulfan and profenofos for suicide is probably because of their relatively higher case fatality [29] [30] compared to other remaining pesticides [18].
Of note, although the suicide rate in Sri Lanka is continuing a downward trend with these recent bans [10], the reduction would be faster still if chlorpyrifos had not been banned and this relatively less toxic OP compound had remained the most common insecticide used in agriculture, and therefore in self-harm. The herbicide that appears to have replaced glyphosate is diuron (Weerasinghe, unpublished observations) which is of low acute toxicity (WHO toxicity class III slightly hazardous) [15] and was not involved in the suicides reported here.
We noted several suicides due to paraquat, chlorpyrifos, glyphosate or propanil during the study period, with a few occurring after their bans in 2011 and 2014. This is likely due to patients ingesting pesticides stored in the farming households or bought after the ban, as stocks ran down in local shops. However, previous studies suggest that bans will result in pesticides being no longer used for suicide within a few years [31]. Our data also supported this—no deaths involving dimethoate or fenthion occurred in this study period (figure 4).
Previous studies in Sri Lanka [19] [32], Bangladesh [33] and South Korea [34] provide no evidence that pesticide bans result in reduced agricultural yield and/or increased input costs to the farmer. However, pesticide regulations need to be carefully planned to reduce potential adverse effects on agricultural output and costs, with support of integrated pest management and replacement of highly hazardous pesticides with less hazardous compounds. According to Sri Lanka’s manual of pesticide management recommendation, profenofos is the only recommended insecticide for gall fly (Neolasioptera falcata) in Cucurbits [35]. Therefore, studies are required to identify substitutes and alternative strategies for pest control before regulation of such insecticide.
A common concern about limiting access to one lethal method used for suicide (means restriction) is that individuals will simply switch to other methods of suicide [36]. For example, regulation of selected pesticides might result in method substitution, possibly to methods of suicide with higher case fatality (such as hanging). However, the data from Sri Lanka and other countries [6] indicates that individuals simply switch to other, less toxic pesticides, resulting in fewer deaths from pesticides poisoning and any forms of suicide. In addition, withdrawing two pesticides from agricultural practice does not impact for pesticide availability for self-harm as many other pesticides are available in the market. The evidence for switching to more lethal method is not strong in Asia [37].
Suicide prevention efforts will need to be multifaceted [38]. In addition to regulation efforts, community-based prevention through public health interventions is especially important to reduce acts of self-poisoning. Over the last decade, a considerable attention has been given to public health approaches in Sri Lanka, such as improved household storage and vendor-based sales restrictions. Unfortunately, our recent cluster randomized controlled trial showed no reduction in pesticide self-poisoning as a result of improved pesticide storage [14]. Vendor-based sales restrictions are now being piloted, but no evidence currently exists on the effectiveness of this approach [39]. Multiple structural and social factors are associated with suicide in South Asia (for example: debt, poverty, cash cropping, intimate partner violence, and alcohol) [40]. However, these risk factors are more diffuse and would require a large policy consensus for action. The best evidence currently indicates that national pesticide regulation is likely to be more effective than community-level approaches [9].
Limitations
The main limitation is the lack of data on the pesticide ingested in 46% of cases due to the lack of forensic toxicology results with which to confirm exposure. However, we hypothesize that many of these deaths will have been due to carbosulfan poisoning since this insecticide has a relatively fast onset of poisoning [30], causing death before presentation to hospital where data on the pesticide ingested was best collected. However, we have generally found the history to be very accurate for identifying the pesticide where blood analyses have been done retrospectively [29] [30].
Our data are from one part of the island; there is no parallel sentinel monitoring of pesticide suicide in other parts of Sri Lanka. However, the area is typical for other agricultural regions of Sri Lanka and it is likely that the findings are generalizable to rural Sri Lanka generally.