Main findings
We conservatively estimate that approximately one third (n = 258,234) of the world's 873,000 suicides in 2002 were caused by pesticide ingestion. Our estimate may well be too low, since it is influenced by the reliability of official suicide statistics for India. Recent detailed studies of rural India suggest that the true number of suicides may be 2–3 fold higher than official estimates and concentrated in rural areas where pesticide poisoning is common[
67,
64] a pattern that is also found in China [
74].
The estimated proportion of suicides attributable to pesticide self-poisoning varies considerably across the WHO's six regions: in Europe we estimate 3.7% of suicides employ pesticides, the Americas: 4.9%, Eastern Mediterranean: 16.5%, Africa: 22.9%, South East Asia: 20.7% and Western Pacific: 55.8%.
Whilst our estimate of pesticide deaths is similar to that of 300,000 given in recent reviews [
6,
7], this is the first time an attempt has been made to estimate the number of pesticide self-poisoning using data from all six WHO regions. Our review indicates that previous crude extrapolations based on data from a limited number of countries have provided a relatively robust estimate of the extent of the problem. Pesticide poisoning and hanging are the two most commonly used methods of suicide worldwide, though precise data on their relative frequency are lacking. An important difference between suicides using these two methods is that deaths from pesticide self-poisoning are considerably more amenable than those from hanging to prevention by both restricting access to pesticides and improving the medical management of pesticide poisoning.
The number of deaths from pesticide self-poisoning under-estimates the true burden on health services of pesticide related self-harm[
75]. Acts of pesticide self-poisoning are associated with a case fatality of between 1% and 70%, depending on the particular pesticide taken. Paraquat and aluminium phosphide self-poisoning have case-fatality in excess of 70%[
11,
12] whereas case fatality for the organophosphorus (OP) insecticides dimethoate and chlorpyrifos are 23% and 8% respectively [
76]. Based on the crude assumption that overall case fatality following pesticide self-poisoning worldwide is between 10% (based on estimated case fatality in China, personal communication Michael Phillips) and 20% (based on data from India [30% case fatality, Table
2] and Malaysia 16%[
72]), our conservative estimate of 258,234 pesticide deaths arise from between 1,291,170 and 2,582,340 episodes of pesticide self-poisoning annually.
OP insecticides appear to be the most commonly ingested pesticides in rural Asia, accounting for around two thirds of cases[
77]. Depending on the particular OP ingested, around 20–30% of patients require intubation [
78,
79]. Two-thirds of intubated patients will be intubated for a median of 45 hours post admission. One third of intubated patients will develop late respiratory failure (intermediate syndrome) requiring intubation for a median of 284 hours [
79]. As a result, across rural Asia, intensive care units are often filled with OP poisoned patients on ventilators, preventing the admission of other acutely ill patients [
80,
75]. Relatively few (<10%) of those poisoned with other pesticides require intubation. Patients ill with paraquat or aluminium phosphide poisoning are usually not intubated or admitted to intensive care because of their poor prognosis. These figures allow us to generate a crude estimate of person days of ventilation required each year of 1,147,000 to 2,294,000 days based on the high and low estimates of the number of episodes of pesticide self-poisoning (see Additional file
1 for details). This would require the constant use of 3,140 to 6,280 ventilators worldwide solely for managing self-poisoning with pesticides. But ventilators are not available in many parts of the developing world so lack of access to ventilators makes a substantial contribution to deaths from pesticide ingestion. Furthermore, the use of ventilators, where these are available, for cases of pesticide poisoning will compromise the capacity of hospitals to manage other life threatening conditions.
The global distribution of fatal pesticide self-poisoning does not mirror pesticide sales. The largest sales (29% of the world market) are in Europe [
81] which accounted for only 2% (6,080/258,234) of pesticide suicides (Table
1). In contrast, Asia (comprising the Western Pacific and SE Asia regions in Table
1) accounts for approximately 25% of the world market [
81] but 91% (235,620/258,234) of deaths [
81]. This is probably due to the different agricultural practices in these regions – the large number of small holders practising agriculture in Asia allows easy access to pesticides, while the very restricted number of people working the land in Western countries means that few people are able to obtain pesticides in the quantities and strength that farmers use. Of note, the greatest market growth in pesticide sales is in Africa and the Middle East with a 9.1% rise in sales between 2004 and 2005[
81] – so there may be a corresponding growth in pesticide self-harm in these regions in years to come.
Limitations
The main limitation of this review is the absence of good quality cause-specific suicide mortality data for a number of the world's largest countries – most notably Indonesia, Iran, Russia, Germany and Ethiopia. Missing data from Europe are less problematic than data from Africa and the Eastern Mediterranean, as estimates for these latter regions are based on particularly limited data. Nevertheless, less than 10% of the world's suicides come from these areas. The availability of better quality data from these regions is therefore unlikely to greatly affect our estimates, unless the overall number of suicides in these regions has been grossly under-estimated. Our figures might be further refined through requests for pesticide mortality data from individual governments and hand searches of the grey literature.
The second limitation is in relation to the likely poor quality of India's suicide mortality data. India is the second largest country in the world, contributing almost 20% to the world's population. In a sensitivity analysis we used plausible estimates for suicide rates and the contribution of pesticides to suicide in India. This analysis resulted in an increase of our estimate of pesticide deaths to 371,594 – an increase of around 110,000 deaths. Obtaining nationally representative data on suicide rates, and the contribution of pesticide suicides to suicides in India, should be a priority for health funding agencies over the next decade. Poor quality data from other parts of the world are also likely to result in under-estimations. This is particularly the case as most pesticide suicides occur in rural (farming) districts where the quality of suicide data collection is likely to be poorer than in urban locations. Thirdly, we have no strong evidence that the proportion of pesticide suicides in countries within WHO mortality strata for each region is the same; mortality patterns are a crude marker of agricultural practices and the availability of toxic pesticides. Fourthly, in some instances the data obtained from the literature predated WHO's 2002 estimate of world suicides by up to 20 years e.g. figures for Nigeria were based on data for 1979–1988[
22] and for Ireland for 1982–1994 [
47]. Changes in the use of pesticides for self-harm may have resulted in us under- or over-estimating pesticide suicides within particular strata. Lastly, some of our estimates for pesticide suicides are based on combined accidental and suicide deaths; but where such deaths are separately categorised, the great majority of pesticide deaths are suicides so it is unlikely that conflating intentional and unintentional pesticide deaths in these countries would substantial alter the overall results [
82,
38].
Policy options for reducing pesticide suicides
Pleas for national and international action to restrict the sale of pesticides to reduce their impact on human health date back over 40 years. In 1966 Ganapathi and colleagues wrote in the Journal of the Indian Medical Association "The present authors plead for a restriction in the sale of such lethal agents. It is hoped that a considerable number of young lives could then be saved from such a measure.... [
83]."
In 2005 WHO launched a global initiative to tackle the burden of pesticide suicides[
84]. Inactivity prior to this is likely to reflect the tensions between the perceived benefits of pesticides in increasing crop yields in low-income countries, where under-nutrition is an important contributor to poor health, and concern about the health effects of excessive exposure. Over the years the pesticide industry has responded to safety concerns with a number of initiatives to reduce the global death toll from pesticide poisoning. However, the pesticide industry is highly profitable – global sales of pesticides amounted to $31 billion in 2005[
81] - and there are clear conflicts of interest where such initiatives may compromise profits. Industry led initiatives are to be welcomed, but national and international health policy makers should recognise that they may not focus on those aspects of the problem most likely to reduce mortality [
85].
Restricting access to the most toxic pesticides is of paramount importance to reduce the number of severe self-poisoning episodes[
86]. National policies and systems for enforcement need to be put in place immediately to phase out WHO Class I and the most toxic Class II pesticides. Likewise, in the longer-term, cost-effective alternatives have to be made available to farmers to reduce the overall use of pesticides. Further investments are needed to promote the safe storage of pesticides at community level. To further reduce the case-fatality from pesticide self-harm investments are needed to improve quality, affordability and accessibility of health care close to the affected communities [
87]. A feature of pesticide suicides is that many patients reach hospital alive, providing opportunities for resuscitation. Data from China indicate that almost two thirds of pesticide deaths received some resuscitation that failed – improved training and resources for medical care may have prevented a number of these deaths[
88]. Likewise, research and investments are needed to support community and facility based interventions aimed at reducing the incidence of self-harm overall. Such interventions need to be based upon policies and strategies that give due consideration to local culture, patterns of self-harm and institutional frameworks and not just a transfer of strategies from Europe or North America.