Case fatality of agricultural pesticides after self-poisoning in Sri Lanka: a prospective cohort study

Abstract

Background: Pesticide poisoning is among the most common means of suicide globally, but can be prevented with regulation of the most hazardous agents. We aimed to compare the lethality of pesticides ingested by our cohort, seek evidence on variation between human and regulatory animal toxicity, and establish change over time in the case fatality of individual pesticides in Sri Lanka.

Methods: We examined the case fatality of agricultural pesticides in a prospective cohort in nine hospitals serving rural populations in Sri Lanka. We included all patients (>11 years) who had presented to a South Asian Clinical Toxicology Research Collaboration study hospital during the study period. Patients were enrolled by clinical research assistants and were regularly reviewed. Identification of the ingested pesticide was generally on the basis of history or positive identification of the container, supported by nested blood analysis.

Findings: From March 31, 2002, to Dec 31, 2019, 34 902 patients (median age 29 years [IQR 21-40]; 23 060 [66·1%] male) presented with a possible or known pesticide self-poisoning. We identified 23 139 specific pesticides that were ingested. Poisoning was fatal in 2299 (6·6%) patients. Case fatality varied greatly from 0·0% (several substances) to 41·8% (paraquat). The three most toxic agents (ie, paraquat, dimethoate, and fenthion) were banned between 2008 and 2011. Since 2013, the five agents causing the most deaths (ie, profenofos, propanil, fenobucarb, carbosulfan, and quinalphos) had a case fatality of 7·2-8·6%. A steady decline was seen in overall case fatality of pesticide poisoning (10·5% for 2002-06 to 3·7% for 2013-19), largely attributable to pesticide bans. A modest fall in case fatality for non-banned pesticides was also seen.

Interpretation: Declines seen in case fatalities of poisonings with non-banned pesticides suggest that medical management improved over time. The human data for acute toxicity of pesticides should drive hazard classifications and regulation. We believe that a global benchmark for registration of pesticides should include a less than 5% case fatality after self-poisoning, which could prevent many deaths and have a substantial effect on global suicide rates.

Funding: The Wellcome Trust and the National Health and Medical Research Council of Australia.

Translations: For the Sinhala and Tamil translations of the abstract see Supplementary Materials section.

Figure 1

Study profile Death-substance pairs are higher than the total number of deaths as deaths were counted against each pesticide ingested. In the lower half of the figure, people and deaths add up to more than 34 902 and 2299 as there were 98 people (nine deaths) who ingested known pesticides and other unknown pesticides or substances and are counted twice. SACTRC=South Asian Clinical Toxicology Research Collaboration.

Figure 2

Changes over time in case fatality for the 18 most common pesticides and three largest unknown categories for 2002–06 versus 2007–12 versus 2013–19 Data are case fatality (95% CI). *Pesticides banned between 2010 and 2015.

Figure 3

Pesticide bans and changes in case fatalities, 2002–19 Timing of seven bans and changes in overall pesticide case fatality (A), percentage of poisoning deaths due to banned pesticides (B), and the number of poisoning deaths due to banned pesticides (C). In A, the grey shaded area shows years when there were progressive restrictions on imports. The four later bans shown were for environmental reasons. Although the number of poisonings from these agents was reduced, the bans did not contribute to reduced case fatality. Propanil is listed as it was subject to a regional but not national ban from 2015.