An index of fatal toxicity for drugs of misuse

Abstract

To determine the lethal toxicity of five commonly-used illicit substances by relating the number of associated deaths to their availability.
An index of toxicity was calculated for each of five drugs [heroin, cocaine/crack, ecstasy (MDMA), amphetamine and cannabis] as the ratio of the number of deaths associated with that substance to its availability in the period 2003-2007. Three separate proxy measures of availability were used (number of users as determined by household surveys, number of seizures by law enforcement agencies and estimates of the market size). All data are related to England and Wales only.
There was a broad correlation between all three denominators of availability. Not unexpectedly, heroin and cannabis showed, respectively, the highest and lowest toxicities. The index of fatal toxicity of MDMA was close to that of amphetamine and cocaine/crack. There was a rank correlation between this index and other measures of lethal toxicity based on safety ratios.
These results are contrary to widely-held public views of the relative fatal toxicity of MDMA.

Full text

An index of fatal toxicity for drugs of misuse
Leslie A. King*and John M. Corkery
27 Ivar Gardens, Basingstoke, RG24 8YD, UK
Objective To determine the lethal toxicity of five commonly-used illicit substances by relating the number of associated deaths to their
availability.
Methods An index of toxicity was calculated for each of five drugs [heroin, cocaine/crack, ecstasy (MDMA), amphetamine and cannabis] as
the ratio of the number of deaths associated with that substance to its availability in the period 2003–2007. Three separate proxy measures of
availability were used (number of users as determined by household surveys, number of seizures by law enforcement agencies and estimates
of the market size). All data are related to England and Wales only.
Results There was a broad correlation between all three denominators of availability. Not unexpectedly, heroin and cannabis showed,
respectively, the highest and lowest toxicities. The index of fatal toxicity of MDMA was close to that of amphetamine and cocaine/crack.
There was a rank correlation between this index and other measures of lethal toxicity based on safety ratios.
Conclusions These results are contrary to widely-held public views of the relative fatal toxicity of MDMA. Copyright #2010 John Wiley
& Sons, Ltd.
key words — lethality; mortality statistics; drug users; law enforcement seizures; market size; illicit substances
INTRODUCTION
The lethal toxicity of a drug in humans cannot be
assessed by a simple inspection of the number of
deaths associated with that substance; the incidence
of poisoning is also determined by availability. For
prescription medicines, it has been shown that an index
of fatal toxicity (T) can be constructed by relating the
number of fatalities in a given period to the number of
prescriptions for that drug (King and Moffat, 1981,
1983). Values of Twere broadly consistent with other
measures of drug toxicity as well as physico-chemical
parameters such as octanol–water partition coefficients
(P). Values of the coefficient Prepresent the ratio of
concentrations of a compound in the two phases of a
mixture of two immiscible solvents (i.e. water and
octanol), and are a measure of the differential solubility
of the compound between these two solvents. Partition
coefficients are useful, for example, in estimating the
distribution of drugs within the body. Hydrophobic
drugs with high partition coefficients are preferentially
distributed to hydrophobic compartments of cells while
hydrophilic drugs (low partition coefficients) are found
mostly in hydrophilic compartments such as the blood.
For illicit substances, it is clearly not possible to use
prescription data. The objective of this paper was to
determine the lethal toxicity of five commonly-
used illicit substances by relating the number of
associated deaths to three proxy measures of avail-
ability: (i) national user-surveys; (ii) seizures by law
enforcement agencies; and (iii) estimates of market
size.
METHODS
Information on the total number of deaths in England
and Wales, where selected substances (whether or not
other substances were also indicated) were mentioned
on death certificates, was obtained from the Office for
National Statistics for each year in the period 2003–
2007 (Anon, 2008). These are summarised in Table 1.
In most drug-related deaths in England and Wales, an
inquest is held by a coroner. In the overwhelming
majority of such cases an autopsy will be conducted by
an approved pathologist. The latter will usually also
ask for toxicological investigations to assist in the
interpretation of the physical examination. Such
investigations will enable the pathologist to determine
if drugs were present in the body at the time of death,
and what role, if any, they played in the death. If the
pathologist concludes on the evidence that a drug or
human psychopharmacology
Hum. Psychopharmacol Clin Exp 2010; 25: 162–166.
Published online 26 January 2010 in Wiley InterScience
(www.interscience.wiley.com) DOI: 10.1002/hup.1090
* Correspondence to: L. A. King, 27 Ivar Gardens, Basingstoke, RG24 8YD,
UK. Tel: 01256-363084. E-mail: les@king.myzen.co.uk
Copyright #2010 John Wiley & Sons, Ltd.
Received 7 September 2009
Accepted 18 November 2009

combination of drugs caused or contributed to the
death these details will be recorded as the medical
cause of death in their report to the coroner. Assuming
that the coroner agrees with this conclusion, the details
will then be written on the death certificate.
In some instances the deceased will have a history of
drug use, abuse or dependence. This may have given
rise to a particular disease or condition which then
contributes to death, for example injecting heroin over
a long period use may lead to hepatitis C or the
acquisition of HIV/AIDS, while the prolonged use of
stimulants may induce cardiac infarction, etc. This
causal connection may also be described on the death
certificate.
The mention of a drug on a death certificate does not
necessarily indicate that substance was a direct cause
of death; it may have been an entirely inconsequential
factor, or it may have been the combination of
substances that led to death. This is particularly true for
cannabis, which has only rarely been the direct cause of
any death. Even with MDMA/MDA, a study of deaths
in the period 1996–2002 by Schifano et al. (2003)
showed that many involved other drugs, mostly
alcohol, cocaine, amphetamines and opiates. These
random effects occur with all substances, and may
partly cancel each other, but to overcome this problem
it was considered appropriate to consider only those
deaths where there was mention of a single substance
on a death certificate; these are listed in Table 2. Deaths
associated with amphetamines were calculated as ‘all
amphetamines’ less ‘MDMA/Ecstasy’. Population use
of those substances in England and Wales was derived
from ‘Drug Misuse Declared’ (England and Wales),
part of the British Crime Survey (BCS) for the same
period (Chivite-Matthews et al., 2005; Roe, 2005; Roe
and Man, 2006; Murphy and Roe, 2007; Hoare and
Flatley, 2008). Table 3 shows the number of individuals
aged 16–59 who used those substances in the last year.
Drug seizures made by law enforcement agencies,
which include Police and HM Revenue and Customs,
for the same period in England and Wales (Table 4)
were compiled by Smith (2008) and Smith and Dodd
(2009). The size of the illicit drug market was
estimated by Pudney et al. (2006) using a variety of
demand-side surveys and supply-side indicators for the
period 2003–2004. Table 5 shows the mean ‘baseline
estimates’ of market size in tonnes for England and
Wales, where quantities had been corrected by those
authors (Pudney et al., 2006) to ‘pure quantities’ using
information on average drug purities. The original data
for Ecstasy were given as millions of tablets, corrected
for purity (drug content). In Table 5, this has been
converted to tonnes of Ecstasy by assuming a tablet
weighs 333 mg.
The various data sets use slightly different nomen-
clature. The terms ‘Ecstasy’ and ‘Ecstasy-type’ should
be understood in almost every case, and are used
hereafter, to mean MDMA (3,4-methylenedioxy-
methylamphetamine). Although once common in the
mid-1990s, other ‘Ecstasy-like’ substances such as
MDA (3,4-methylenedioxyamphetamine) and MDEA
(3,4-methylenedioxy-ethylamphetamine) were rarely
seen in the period covered. Similarly ‘amphetamines’
effectively means amphetamine itself since methy-
lamphetamine and related phenethylamines were
uncommon. Cocaine and crack usage has been
combined in both the BCS data and the market size
data; it is used here in preference to the BCS category
Table 1. Deaths in England and Wales, where selected substances were
mentioned on death certificates (all mentions)
Substance 2003 2004 2005 2006 2007 Mean
Heroin and morphine 696 751 842 713 829 766.2
Cocaine 129 154 176 190 196 169.0
Amphetamines
(not MDMA/ecstasy)
31 37 45 44 50 41.4
MDMA/ecstasy 50 43 58 48 47 49.2
Cannabis 11 19 19 17 12 15.6
Table 2. Deaths in England and Wales, where selected substances were
mentioned on death certificates (sole mention)
Substance 2003 2004 2005 2006 2007 Mean
Heroin and morphine 481 491 558 496 587 522.6
Cocaine 43 48 53 68 84 59.2
Amphetamines
(not MDMA/ecstasy)
16 21 26 20 28 22.2
MDMA/ecstasy 29 24 33 27 28 28.2
Cannabis 1 1 2 2 1 1.4
Table 3. Number (thousands) of individuals aged 16–59 who used selected substances in the last year in England and Wales
Substance 2003–2004 2004–2005 2005–2006 2006–2007 2007–2008 Mean
Heroin 43 38 39 41 34 39.0
Cocaine þcrack 810 667 822 886 778 792.6
Amphetamines 483 430 426 421 329 417.8
Ecstasy 614 556 502 567 470 541.8
Cannabis 3364 3040 2775 2616 2382 2835.4
Copyright #2010 John Wiley & Sons, Ltd. Hum. Psychopharmacol Clin Exp 2010; 25: 162–166.
DOI: 10.1002/hup
index of fatal toxicity 163

of ‘any cocaine’. Similarly, data in the BCS for
‘Heroin’ is used rather than ‘Opiates’. The mortality
statistics refer to ‘Heroin and morphine’ as a category,
but in the present paper this is listed as just ‘Heroin’.
Almost all of those deaths will have arisen from heroin,
but toxicological analysis will have usually only
recorded the presence of its major metabolites, namely
morphine and morphine glucuronide. Crack is only
rarely distinguished from powder cocaine on death
certificates, so figures for deaths comprise both
substances.
For amphetamine, cannabis, cocaine, heroin and
MDMA, three indices of fatal toxicity were calculated
as the mean number of deaths per annum divided by the
availability of that drug, where availability is firstly the
mean number of users, secondly the mean number of
seizures per annum, and thirdly the market size. Each
index was normalised such that for heroin the value
was 1000. The use of the mean death rate for the period
2003–2007 was justified since, although absolute
numbers rose slightly during the 5 years, the relative
numbers were more stable. A similar argument can be
applied to the usage and seizure data, but in any event,
almost all poisonings are acute events that reflect
current availability. Schifano et al. (2006) have
demonstrated that fatalities where ecstasy-type drugs
were mentioned have a strong positive (Spearman)
correlation with last year use (r¼0.854, p<0.01) and
the number of seizures (r¼0.805, p<0.01). Even
stronger correlations have been found (Schifano and
Corkery, 2008) for cocaine (including crack) fatalities:
last year use (r¼0.901, p<0.001); number of seizures
(r¼0.946, p<0.001). Both studies also found
negative correlations between fatalities and price,
albeit at lower levels: ecstasy-type drugs (r¼0.710,
p<0.05); cocaine (r¼0.882, p<0.001). These
findings support the choice of indicators used in the
present study.
RESULTS
The three fatal toxicity scales are plotted in Figure 1
(any mentions) and in Figure 2 (sole mentions).
Logarithmic transforms were used because, on all
measures, the index spanned three orders of magnitude.
Apart from the high toxicity of heroin and low toxicity
of cannabis, Figures 1 and 2 also demonstrate that there
is a broad rank correlation between the three measures
of availability, and that the lethal toxicity of MDMA is
close to that of amphetamine and cocaine/crack.
Amongst the four data sets considered here, only the
estimates of market size, as calculated by Pudney et al.
(2006), included confidence intervals. The upper and
lower bounds were typically þ/30–50% of mean
values, but are not included in Figures 1 and 2 for the
sake of clarity. An overall mean index of fatal toxicity
for the five drugs is shown in Table 6.
DISCUSSION
The mechanism of acute fatal poisoning will not be the
same for all of the substances considered here. Heroin
will often cause death through respiratory depression,
but poisonings from both amphetamine and MDMA
are more likely to result from pyrexia and cardiac
failure, and, for MDMA, hyponatraemia. The various
data sets are all subject to distorting factors. It is not
known to what extent deaths from amphetamine are
Table 4. Number (thousands) of drug seizures in England and Wales
Substance 2003 2004 2005 2006–2007 2007–2008 Mean
Heroin 10.569 11.074 13.331 13.205 13.463 12.33
Cocaine þcrack 11.662 12.812 18.262 22.665 27.509 18.58
Amphetamines 5.862 6.174 7.425 8.030 8.412 7.18
Ecstasy-type 6.114 5.938 6.336 7.752 6.807 6.59
Cannabis (total) 82.752 77.482 114.20 137.13 164.89 115.3
Table 5. Baseline estimates of market size (tonnes) of pure drugs in
England and Wales
Substance Mean (2003–2004)
Heroin 7.04
Cocaine þcrack 16.81
Amphetamines 3.60
Ecstasy 4.57
Cannabis 360.33
Figure 1. Correlation between three measures of the index of fatal toxicity
based on all mentions on death certificates. The abscissa shows deaths/users
(D/U) and the ordinate shows both deaths/seizures (D/S) and deaths/market
quantity (D/Q), all as logarithmic transforms
Copyright #2010 John Wiley & Sons, Ltd. Hum. Psychopharmacol Clin Exp 2010; 25: 162–166.
DOI: 10.1002/hup
164 l. a. king and j. m. corkery

increased because of the occasional practice of
injecting this drug. In other words, for the majority
of (ingesting) amphetamine users, the index of toxicity
might be lower. Since routine toxicological analysis
would not shed light on whether the deceased used
cocaine powder or smoked crack cocaine, the index is
an aggregate where it should be recognised that the two
forms of cocaine might have quite different toxicities.
In the study period, the ratio of crack cocaine users to
all cocaine users was stable at 5–6%. While other
measures of prevalence in the BCS, such as ‘ever used’
or ‘used in the last month’, could have been chosen, it is
unlikely that they would have changed the overall
conclusions.
It could be argued that the number of seizures is as
much a reflection of law enforcement activity as it is of
availability. But by focussing on the number of seizures
instead of the quantities seized, which are known to be
much more volatile, and recognising that many
seizures arise indirectly from routine Police and
Customs activity rather than from targeted operations,
the relative numbers of seizures of different substances
are likely to be more stable than their absolute number.
Finally, the usage data, as collected in the BCS, does
not capture those who do not live in normal domestic
premises. It is believed, for example, to underestimate
the number of individuals, such as some problematic
heroin and crack users, with more chaotic lifestyles.
When based on seizures, the apparent toxicity of
amphetamine, cocaine, MDMA and cannabis is higher
than when based on users. This could, therefore, be
interpreted as support for the suggestion that the BCS
does indeed underestimate heroin users. An equally
plausible deduction from Figures 1 and 2 is that there is
an excess of cannabis seizures relative to the user base
compared to the other drugs.
Results from the National Programme on Substance
Abuse Deaths (np-SAD), as reported by Schifano
(2008), showed that in the period 1997–2007, when
compared with amphetamine, ecstasy was more
frequently identified on its own at post mortem
(23.15 and 17.15%, respectively). Although not large,
this difference is significant at the 0.01 level (x
2
).
Secondly, ecstasy deaths were less frequently associ-
ated with concomitant/contributory factors (10.7%)
compared to 27.7% for amphetamine (x
2
,p<0.001).
However, the personal circumstances of the individuals
concerned were broadly similar, for example, sex
(mostly male), ethnicity (mostly white), cause of
death (mostly accidental), age at death (mostly below
34 years), and most had a previous history of drug
addiction.
The fatality index treats all users of a drug as an
equivalent data point. In other words, the index relates
to an ‘average’ user. Clearly, a one-time user has a
Table 6. Mean index of fatal toxicity
Substance
Index
(any mention)
Index
(sole mention) Overall index
Heroin 1000 1000 1000
Cocaine þcrack 56 28 42
Amphetamine 36 28 32
MDMA 45 37 41
Cannabis <1<1<1
Figure 2. Correlation between three measures of the index of fatal toxicity based on sole mentions on death certificates. The abscissa shows deaths/users (D/U)
and the ordinate shows both deaths/seizures (D/S) and deaths/market quantity (D/Q), all as logarithmic transforms
Copyright #2010 John Wiley & Sons, Ltd. Hum. Psychopharmacol Clin Exp 2010; 25: 162–166.
DOI: 10.1002/hup
index of fatal toxicity 165

much lower risk of dying from a drug than a frequent
user. Similarly, a user who regularly consumes more of
a drug on each occasion than another user is also at
greater risk of a fatal outcome. Unfortunately, reliable
information on what constitutes occasional/regular use
or heavy/light use of illicit substances is not available.
Figures 1 and 2 show that the relative fatal toxicity
of the five substances considered here is largely
independent of whether all deaths are used or just those
where there is sole mention of the substance on the
death certificate. This suggests that, in any future
studies where there is a need to calculate the fatal
toxicity of a further substance, it would be appropriate
to consider all deaths involving that substance and not
the sub-set of ‘sole mentions’. Although multiple
mentions of substances on death certificates might
provide information on polydrug use, this has not been
pursued in the present work. Regardless of whether one
or several drugs are indicated on death certificates, a
further limitation of this study is that some deaths could
have been caused by other undetected substances.
A different index of toxicity was constructed by
Gable (2004). A safety ratio for a number of substances
was defined as the acute lethal dose divided by the dose
most commonly used for non-medicinal purposes,
where the higher the ratio the safer the drug. The safety
ratios for the five substances examined by the present
authors were: heroin 6, cocaine 10, methylampheta-
mine 10, MDMA 16 and cannabis >1000. Assuming
that methylamphetamine can be equated to amphet-
amine, then the inverse of the rank order is similar to
that found in the present study.
It is not clear why deaths from MDMA should attract
so much publicity whereas other poisonings often go
unrecorded in the media (Forsyth, 2001). It has often
been stated that ‘Ecstasy can and does kill unpredic-
tably; there is no such thing as a safe dose’ (House of
Commons Home Affairs Committee, 2002). The
analysis presented here gives reason to believe that,
while an individual death may be unpredictable
(Henry, 1992), at the population level the number of
poisonings caused by MDMA is entirely predictable.
Despite the limitations inherent in the use of the
proxy measures for the availability of illicit drugs, this
exercise has demonstrated that the technique outlined
here can be usefully employed, as one of a number of
methods, to assess the intrinsic lethality of substances.
CONCLUSIONS
These results are contrary to widely-held public views,
where there is a perception that MDMA is much more
dangerous than many other substances; they show that
MDMA, amphetamine and cocaine have a similar fatal
toxicity.
ACKNOWLEDGEMENTS
The authors assert that they have no conflicts of interest. This
work has not been published elsewhere and is not currently
under review elsewhere. No ethical issues arise in this work.
REFERENCES
Anon. 2008. Report: Deaths related to drug poisoning in England and
Wales, 2003-07, Health Statistics Quarterly 39, Office for National
Statistics, Autumn 2008.
Chivite-Matthews N, Richardson A, O’Shea J, et al. 2005. Drug misuse
declared: Findings from the 2003/04 British Crime Survey, England and
Wales, Home Office Statistical Bulletin 04/05, Home Office: London.
Forsyth AJM. 2001. Distorted? A quantitative exploration of drug fatality
reports in the popular press. Int J Drug Policy 12: 435–453.
Gable RS. 2004. Comparison of acute lethal toxicity of commonly abused
psychoactive substances. Addiction 99(6): 686–696.
Henry JA. 1992. Ecstasy and the dance of death: severe reactions are
unpredictable. Br Med J 305: 5–6.
House of Commons Home Affairs Committee. 2002. The government’s
drugs policy: is it working? Third report of session 2001-02, Volume 1:
Report and Proceedings of the Committee, The Stationery Office:
London.
Hoare J, Flatley J. 2008. Drug misuse declared: Findings from the 2007/08
British Crime Survey, England and Wales, Home Office Statistical
Bulletin 13/08, Home Office: London.
King LA, Moffat AC. 1981. Hypnotics and sedatives: an index of fatal
toxicity. Lancet (i): 387–388.
King LA, Moffat AC. 1983. A possible index of fatal toxicity in humans.
Med Sci Law 23(3): 193–197.
Murphy R, Roe S. 2007. Drug misuse declared: Findings from the 2006/07
British Crime Survey, England and Wales, Home Office Statistical
Bulletin 18/07, Home Office: London.
Pudney S, Badillo C, Bryan M, Burton J, Conti G, Iacovou M. 2006.
Estimating the size of the UK illicit drug market. In Singleton N, Murray
R, Tinsley L (eds). Measuring Different Aspects of Problem Drug Use:
Methodological Developments,Home Office Online Report 16/06, Home
Office: London.
Roe S. 2005. Drug misuse declared: Findings from the 2004/05 British
Crime Survey, England and Wales, Home Office Statistical Bulletin
16/05, Home Office: London.
Roe S, Man L. 2006. Drug Misuse Declared: Findings from the 2005/06
British Crime Survey, England and Wales, Home Office Statistical
Bulletin 15/06, Home Office: London.
Schifano F. 2008. Ecstasy and amphetamines; the National Programme on
Substance Abuse Deaths (np-SAD) mortality data, Presentation to
the Advisory Council on the Misuse of Drugs, London, 26 September 2008.
Schifano F, Corkery J. 2008. Cocaine/crack cocaine consumption, treatment
demand, seizures, related offences, prices, average purity levels
and deaths in the UK (1990–2004). J Psychopharmacol 22(1):
71–79.
Schifano F, Corkery J, Deluca P, Oyefeso A, Ghodse AH. 2006. Ecstasy
(MDMA, MDA, MDEA, MBDB) consumption, seizures, dosage levels
and deaths in the UK (1994–2003). J Psychopharmacol 20(3): 456–
463.
Schifano F, OyefesoA, Corkery J, Cobain K, Jambert-Gray R, Martinotti G,
Ghodse AH. 2003. Death rates from ecstasy (MDMA, MDA) and
polydrug use in England and Wales 1996–2002. Hum Psychopharmacol
18: 519–524.
Smith K. 2008. Seizures of Drugs in England and Wales, 2006/07, Home
Office Statistical Bulletin 12/08, Home Office London.
Smith K, Dodd L. 2009. Seizures of Drugs in England and Wales, 2007/08,
Home Office Statistical Bulletin 08/09, Home Office London.
Copyright #2010 John Wiley & Sons, Ltd. Hum. Psychopharmacol Clin Exp 2010; 25: 162–166.
DOI: 10.1002/hup
166 l. a. king and j. m. corkery