Motor vehicle exhaust gassing suicides in Australia: epidemiology and prevention

Monash University Accident Research Centre - Report #139 - 1998

Author: V. Routley

Full report in pdf format [852KB]


Suicide is a major public health problem in Australia. In 1996 carbon monoxide poisoning from motor vehicle exhaust gas was the second major method, accounting for almost 22% of suicides. For some survivors there were lasting effects on the heart and brain.

The aim of the study was to reduce the overall suicide rate by making motor vehicle exhaust gas suicide substantially more difficult to undertake and complete. Since the method is common and relatively lethal this has high potential to reduce total suicides. The study involved a literature review and obtaining of background information on suicide and methods of suicide, particularly exhaust gas suicide, data analysis, personal communication with technical experts, being informed of other current or unpublished studies, an examination of a sample of exhaust gas suicide Victorian State Coroner’s case files and an awareness of activities to reduce the method.

Motor vehicle exhaust gas suicides have increased in rates, as a proportion of suicides and in numbers since at least 1968, despite the introduction of catalytic converters in 1986. The method is most commonly used by middle-aged males. Its usage varies between states and it is a relatively favoured method in Australia compared with other countries. Hospital admissions have increased considerably in recent years, at a faster rate than deaths. In a sample of Coroner’s records 36% of victims’ vehicles had catalytic converters. Almost all used a hose or pipe leading into the interior of the vehicle with ventilation sealed. They were most frequently undertaken at home or at an open-air location away from home.

While recognising that the causes of suicide are complex there are design changes which potentially make the method more difficult to affect. These are multi-gas (CO, O2 and possibly CO2) sensing devices installed in the vehicle cabin which emit a warning light followed by an alarm and then shut down the engine when the levels become life threatening, exhaust modifications which make it difficult to fit a hose or pipe and further improvements in engine design and catalytic conversion techniques.

Recommendations are for the introduction of mandatory regulations for new vehicles to ensure that life threatening gas levels cannot be reached by passing a hose from the exhaust into the vehicle over the lifetime of the vehicle. In-service vehicles should be required to replace existing exhaust pipes with new safety designs to make it substantially more difficult to attach a hose. A study of exhaust gassing suicide attempters and the Suicide Module of the National Coronial Information System should collect information which further clarifies potential countermeasures and risk factors.

The Australian Medical Association has convened a multi-sectoral committee, on which MUARC is represented. It is progressing research to reduce exhaust gassing suicides, relevant to the above design changes and recommendations, with the assistance of $30,000 from the Department of Health & Family Services.

Executive Summary

Suicide is a major problem in Australia. In recent years there have been approximately 2000 deaths each year from suicide. Since 1990 suicide has been more common than motor vehicles crashes as a cause of death. In 1996 motor vehicle exhaust gas was the second major method of suicide, accounting for almost 22% of suicides. It was the method most commonly used by middle-aged males.

There is evidence that reduction in access to a lethal means potentially reduces the overall suicide rate. The potential for reduction is greatest where the restricted method is commonly used. Motor vehicle exhaust gassing fits many of these criteria.

The study

The study involved a literature review and obtaining background information, data analysis, personal communication with technical experts, being informed of other studies which were unpublished or in process and an examination of a sample of exhaust gas suicide Victorian State Coroner’s case files.


To reduce the overall suicide rate in Australia by making motor vehicle exhaust gas suicide substantially more difficult to undertake and complete.

Objectives of this study

  • To review published and unpublished literature on the aetiology of suicide focussing on access to the means, particularly motor vehicle exhaust gas suicides.
  • To describe the epidemiology of motor vehicle exhaust gas suicides in Australia.
  • To investigate :-
    • the transference between means of suicide
    • the relationship of exhaust gas suicide to vehicle numbers
    • the Victorian Coroner’s files for the means and circumstances by which this method of suicide is usually attempted
    • possible solutions to exhaust gas suicide
    • whether the legislated environmental limits on carbon monoxide (CO) emissions have impacted on the numbers of suicide attempters selecting and succeeding by this method.
    • the international situation in regard to patterns, trends and possible solutions to exhaust gas suicide.
  • To make recommendations to relevant authorities and to disseminate the findings of this study.


Social factors which are considered risk factors for suicide are those which relate to social bonding: unemployment; never married, widowed or divorced marital status and non-church attendance. Age, sex, immigration and country of birth are also found to influence suicide. Physical and mental health states are also known to be associated with the above risk factors eg. terminal illness, depression, schizophrenia.

Methods of suicide

Physical availability and socio-cultural acceptability are considered necessary preconditions for the choice of suicide methods. Methods chosen vary considerably in lethality. Males have a propensity to use more immediate and violent methods than females (eg firearms, hanging) and this partially accounts for their higher overall suicide rates. Lethality rates have been reported at 85% for firearms, 80% for hanging, 77% for exhaust poisoning, 75% for drowning and 23% for drug overdose.

In Australia, in 1995, four methods accounted for 84% of suicides - hanging, motor vehicle exhaust gas, firearms and poisoning. There have been reductions over time in firearms and poisoning. Exhaust gassings have constantly increased since at least 1968, despite the introduction of catalytic converters in 1986. Hanging has seen the most dramatic rise. The less frequent methods - jumping from high places, drowning and cutting and piercing vital structures have varied little in the proportion of suicides they represent (approximately 3% each). Poisoning is clearly the preference for females, although not to the previous extent. Impulsiveness appears to play an important role, especially in youth suicide.

Cantor et al (1996) concluded that restricting the availability of a particular method of suicide often, but not invariably, reduces overall suicide rates. A complex interaction of factors will determine this outcome. If the method were made more difficult, then it could take longer to affect, thus enhancing the possibility of the potential victim changing their mind or being intercepted. Additionally, another method may not be acceptable.

Carbon monoxide

Carbon monoxide (CO) is colourless, odourless and tasteless and is produced from the incomplete combustion of organic fuels. It attaches itself to the body’s red blood cells, making the cells unable to carry oxygen. The brain and heart are the most susceptible to toxicity because they depend most heavily on oxygen to function. Lasting effects on these organs can include myocardial infarction, deterioration of personality and impaired memory. Symptoms of carbon monoxide poisoning are normally a headache, drowsiness, then loss of consciousness (LOC) and finally death. If the poisoning is not fatal hypoxic brain injury can occur with possible symptoms of confusion, disorientation, incontinence, amnesia, short-term memory loss and/or muteness.

Environmental CO requirements

The required maximum levels of CO in motor vehicle exhaust gases have been 24.3g/km from July 1976 (ADR27A), 9.3 g/km from 1986 for new passenger vehicles (ADR37-00) and 2.1 g/km for new models (ADR37-01) from 1997 and for all new passenger vehicles from 1998. In order to cope with the unleaded petrol legally required since 1986, vehicles usually require catalytic converters (ADR37-00, AS2877).

Data analysis

  • The motor vehicle exhaust gas suicide (MVEGS) rates for Australia in 1995 were calculated to be 4.69/100,000 males, 1.02/100,000 females and 2.85/100,000 total. In 1995, there were 509 cases of motor vehicle exhaust gas suicides in Australia. These represented 21.5% of suicide cases.
  • The preference for this method of suicide varies between states. It appears to be a relatively favoured method in Australia, particularly in the ACT and WA, compared with other countries
  • MVEG has been more favoured in middle age and by males. In the 30-50 year age group MVEG has been the leading means of suicide and 82% of all motor vehicle exhaust gas suicides were male. Numbers were substantial in the 20-24 age group, this being the age group which has had the highest suicide frequency overall.
  • Hospital admissions for Australian motor vehicle exhaust gas suicide attempts have increased steadily in recent years (1994/95 data are not available), with the data showing an exponential trend (R2=0.99). Deaths also increased, but show no clear trend.
  • MVEG related self-harm represents a smaller proportion of hospitalised attempted suicides (2%) than successful suicides (21.6%). This is likely to be the result of the lethality of MVEG as a means of suicide compared to other means.
  • The age and sex distribution of MVEG related admissions is generally similar to that seen for completed suicides. It appears however that younger males are relatively more common among hospitalisations. This suggests that attempts among younger males are more likely to be detected before death occurs presumably due to higher involvement of their family.

Coroners’ records

  • Thirty-six percent of suicide vehicles had catalytic converters. This proportion was not significantly different from the 39% of vehicles with catalytic converters in the Victorian fleet.
  • In the 93 cases where the method was specified, all but 4 used a hose or pipe leading into the interior of the vehicle with ventilation sealed.
  • The location of the vehicle in the 85 cases, where it was specified, was most frequently at a home (51% cases), at an enclosed workplace (7%) or at an open-air location away from home (including farm paddocks) (43%).
  • The median carboxyhaemoglobin level (COHb), for the 88 cases for which it was noted, was 79%. The median COHb level of victims in the groups with and without catalytic converters was identical (77.1%).
  • Alcohol was detected in 24 cases, benzodiazepines in 10, paracetamol in 8, cannabis in 7 and other pharmaceuticals, usually in a ‘cocktail’ in 7.

Several overseas studies which have examined the relationship between suicides and MVEG have found that the imposition of emission controls reduces MVEGS. This observation may in part relate to the comparatively newer vehicle fleet compared with Australia (and to lower allowable emission levels.)

Explanations for suicides in later model vehicles

It appears that MVEGS have not reduced as much as expected with reductions in emission levels and it is of note that the pattern between Australia, Japan and the USA may have similarities ie a lagged levelling off after the introduction of CO exhaust limits. Considerably less reductions are anticipated than would be expected if exhaust emission CO levels of 2.1g/km made MVEGS impossible.

In explanation, it appears that there are several situations where the testing for these legislated environmental CO limits may not be particularly relevant to the suicide lethality situation eg 1) a hose or pipe is led into the interior of the vehicle and ventilation is sealed. In this situation other gases eg. oxygen and CO2 and other factors such as heat and humidity may have a synergistic and additive effect. It appears inappropriate that CO limits are the same for each vehicle regardless of the vehicle’s cabin volume 2) the engine idles. In environmental CO testing, there are 3 phases to the vehicle testing and none of these involves idle only.

There are several situations where CO emissions may exceed the legislated limit such as 1) where the engine idles from a cold start, with a delay of between 1.5 and 3 minutes before the catalytic converter has warmed up and is operating efficiently. Since carboxyhaemoglobin concentration rises most rapidly when first exposed to CO the initial absorption rate would be particularly high 2) modern vehicles may be better sealed ie. less ventilation and 3) the vehicle does not have a well-functioning catalytic converter.


The cost per year attributable to 509 Australian motor vehicle exhaust poisoning deaths was approximately $386.6 million in 1996, based on conservative figures. If a device were developed which prevented 50% of these suicides and it cost $36, the full costs would be recovered in just 2 years. Alternatively if a device cost $72, or if it prevented only 25% of these suicides, the break even period would be 4 years. Furthermore there are additional benefits of preventing several hundred hospital admissions resulting from failed MVEGS attempts.

While it is considered that ultimately the solution should be in terms of performance requirements for the vehicle there are several design solutions possible for the prevention (or minimisation) of MVEGS:

Design Solutions

1. The mandatory incorporation into new vehicles of a multi-gas sensing device which monitors carbon monoxide, oxygen and possibly carbon dioxide levels and when these become life threatening displays a warning light, then emits an alarm and finally shuts down the engine. The operation of window winding devices could also be incorporated.

Devices for the household which meet the Underwriters-Laboratory Inc. (UL) standard 2034 are available in the U.S for between AU$43 and AU$100 and a similar concept could possibly be fitted to the engine management system of a motor vehicle. Currently these alarms are activated at CO levels which produce COHb levels of 10% or above.

A CO detector attached to the engine management system of a motor vehicle would have the added advantage of identifying and preventing unintentional poisonings and driver fatigue due to leaks of carbon monoxide from unsealed boots, rusted holes and access through open windows.

2. The exhaust pipe on new vehicles could be modified to incorporate a device inside the pipe, so a hose cannot be inserted, and to make the end of the pipe irregular, to make it difficult to fit a hose. New designs would need to overcome any problems of backpressure, vibration or noise and would need to meet current exhaust regulations (eg exhaust gases should be emitted beyond the vehicle).

Both approaches showed sufficient promise of a potential cost-effective contribution to the prevention of MVEG and warrant further expenditure on research. Research would better define what is involved and provide more precise estimated values for the assumptions made above.

3. Further improvements could be made to engine design and in catalytic conversion techniques to complete the combustion process and thereby virtually eliminate carbon monoxide emissions.

It is recognised that design changes will not eliminate all suicides. Complex social, economic and psychological reasons underlie the causes of suicide and the solutions for these require a multi-faceted longer term approach.


1.  Mandatory regulations should be introduced for new vehicles which will make exhaust gas suicide virtually impossible, by ensuring that life threatening levels of CO, O2 and possibly CO2 cannot be reached by passing a hose from the exhaust into the vehicle with sealed ventilation. A sensor is preferred because it would cater for deterioration in catalytic converter and engine performance over time.

2. Regulations should be introduced for in-service vehicles requiring replacement exhaust pipes to be of new safety designs to make it substantially more difficult to attach a hose.

3. A study of MVEGS attempters (ie those admitted to hospital) should collect information which further clarifies potential countermeasures and risk factors. Variables on which it would be important to collect data are: how long the vehicle ran, if they were interrupted, if they reneged, the reason for selection of the MVEG method, the practical details of how they made the attempt eg. equipment used, the blood alcohol content and if possible the presence of other drugs, if they had previously attempted suicide by exhaust gas or otherwise and information on the make, model and year of manufacture of the vehicle. Such data would be most useful in gaining further insight into how MVEGS are undertaken and the extent to which catalytic converters are making an impact. Data collected so far has concentrated on deaths.

4. The Suicide Module of the National Coronial Information System should collect information encoded and/or in text including the make, model and year of manufacture of the vehicle; COHb level; drugs, medications and other gases detected at pathology; practical details of the attempt including equipment used; details of previous attempts. Additionally there should a move towards consistency in the medical and legal definitions of suicide.

These actions should be undertaken quietly, without media attention. Adding to the risk group’s knowledge of how to undertake this method, or risking ‘copycat’ suicides, should be avoided.


The Mental Health Branch of the Department of Health & Family Services was given responsibility for the allocation of $30,000 to alleviate exhaust gassing suicides.

Since November 1996 there have been four meetings convened by the Australian Medical Association and initially the Federal Office of Road Safety to reduce vehicle exhaust gas suicides with representatives from various interest groups.

The priority recommendations for the allocation of the H&FS $30,000 have been revised. These are now:

  • $6,000 to bring Prof David Penney, Director of Surgical Research and Professor of Physiology, Wayne State University, US, an expert on the physiological effects of CO to the April 1998 meeting in Melbourne to determine an appropriate CO level for a CO detector. He will produce a report in September, 1998.
  • $15,000 to research MVEGS attempters who have been treated for CO poisoning at the Prince of Wales Hyperbaric unit. The study is in progress and the reasons for their method choice, demographic profile, details of psychiatric records and any previous suicide attempts by MVEGS or other methods and the details of their latest MVEGS attempt, including the presence or otherwise of catalytic converters are being studied.
  • $5,000 to an industrial design student at Flinders University to research the available exhaust pipe designs and to design a tailpipe which is low cost (including fitting), which does not cause problems of backpressure or reduce performance etc and which allows a vacuum or garden hose or other pipe to be fitted only with great difficulty. A design has been produced which meets these criteria.
  • $4,000 to develop a computer model which simulates the exhaust gassing suicide scenario. Mr Jerry Moller has now submitted a paper which includes a model of CO and O2 exhaust gas concentrations to the Department of Health and Family Services Mental Health Branch.

In addition, the Australian Automobile Association has allocated $15,000 to RMIT in Melbourne for the development of a CO sensor for the vehicle cabin. Currently the emphasis is on developing CO, O2 and CO2 detectors. The focus will then be to link these into a multi-sensor system. Additional funds are likely to be required for development to the next stage.

International enquiries have identified very little activity to reduce suicides by this method and it appears that this committee is leading the way internationally.