Series

Addiction 1
Extent of illicit drug use and dependence, and their
contribution to the global burden of disease
Louisa Degenhardt, Wayne Hall

This paper summarises data for the prevalence, correlates, and probable adverse health consequences of problem use Lancet 2012; 379: 55–70
of amphetamines, cannabis, cocaine, and opioids. We discuss findings from systematic reviews of the prevalence of See Editorial page 2
illicit drug use and dependence, remission from dependence, and mortality in illicit drug users, and evidence for See Perspectives pages 20, 21,
acute and chronic effects of illicit drug use. We outline the regional and global distribution of use and estimated and 22
health burden from illicit drugs. These distributions are likely to be underestimates because they have not included This is the first in a Series of
all adverse outcomes of drug use and exclude those of cannabis—the mostly widely used illicit drug. In high-income three papers about addiction

countries, illicit drug use contributes less to the burden of disease than does tobacco but a substantial proportion of National Drug and Alcohol
Research Centre, Faculty of
that due to alcohol. The major adverse health effects of cannabis use are dependence and probably psychotic disorders Medicine, University of
and other mental disorders. The health-related harms of cannabis use differ from those of amphetamine, cocaine, New South Wales, Sydney,
and opioid use, in that cannabis contributes little to mortality. Intelligent policy responses to drug problems need NSW, and Burnet Institute and
better data for the prevalence of different types of illicit drug use and the harms that their use causes globally. This Centre for Health Policy,
Programs and Economics,
need is especially urgent in high-income countries with substantial rates of illicit drug use and in low-income and School of Population Health,
middle-income countries close to illicit drug production areas. University of Melbourne,
Melbourne, VIC, Australia
Introduction is more scarce than it is for the drugs included in this (Prof L Degenhardt PhD); and
University of Queensland
Illicit drugs are drugs for which non-medical use has paper (webappendix p 1).42–46 Their exclusion is because of Centre for Clinical Research
been prohibited by international drug control treaties for the scarcity of evidence rather than any judgment about and Centre for Youth
half a century because they are believed to present the contribution of these drugs to disease burden. We Substance Abuse Research,
unacceptable risks of addiction to users.1,2 International were also unable to separately discuss the magnitude of University of Queensland,
Queensland, Australia
control has since been extended from plant-based adverse outcomes attributable to prescribed pharma- (Prof W Hall PhD)
drugs—heroin, cocaine, and cannabis—to synthetic ceutical opioids. Although increased prescription of these
drugs, such as amphetamines and methylenedioxy-
metamfetamine (MDMA), and pharmaceutical drugs
such as buprenorphine, methadone, and benzodi- Key messages
azepines (panel 1). • The illegality of opioids, amphetamines, cocaine, and cannabis precludes the accurate
In this paper, we summarise data for the prevalence, estimation of how many people use these drugs, how many people are problem users,
correlates, and probable consequences of use of the and what harms their use causes.
amphetamines, cannabis, cocaine, and opioids—the • An estimated 149–271 million people used an illicit drug worldwide in 2009:
most commonly used and studied illicit drugs. We 125–203 million cannabis users; 15–39 million problem users of opioids,
discuss findings from systematic reviews of data for the amphetamines, or cocaine; and 11–21 million who injected drugs.
prevalence of illicit drug use and dependence,3–8 • Levels of illicit drug use seem to be highest in high-income countries and in
remission from dependence,9 and mortality in illicit countries near major drug production areas, but data for their use in low-income
drug users (panel 2).10–13 We attribute adverse health countries are poor.
effects to these drugs using findings from reviews of • Cannabis use is associated with dependence and mental disorders, including
published studies of the evidence on a range of acute psychoses, but does not seem to substantially increase mortality.
and chronic harms of illicit drug use.8,14,19,35–41 We provide • Illicit opioid use is a major cause of mortality from fatal overdose and dependence;
a brief summary of adverse health effects for different HIV, hepatitis C, and hepatitis B infections from unsafe injection practices are
drug types referencing other reviews (webappendix important consequences in people who inject opioids, cocaine, or amphetamines.
pp 3–5 for more details). We also summarise earlier • Adverse health outcomes such as mental disorders, road-traffic accidents, suicides,
global burden of disease studies that estimated the and violence seem to be increased in opioid, cocaine, and amphetamine users. To
regional and global distribution of health burden from what extent these associations are causal is unclear, because confounding variables
illicit drug use and compared this with the burden are not always controlled and quantification of risk is poor.
attributable to alcohol and tobacco use.29–33 • Global burden of disease estimates suggest that in high-income countries, the
We do not discuss the prevalence of or disease burden contribution of illicit drug use is a substantial proportion of that attributable to alcohol.
related to MDMA (ecstasy), hallucinogenic drugs, • These estimates probably underestimate the true burden because only a few effects of
inhalants, or the non-medical use of benzodiazepines problem use of opioids, cocaine, and amphetamines are included. The global burden
and anabolic steroids because information about the of disease 2010 study will address these limitations.
prevalence of their use and quantification of their harms

www.thelancet.com Vol 379 January 7, 2012 55

 Series

Correspondence to: The availability and quality of data for estimation varies
Prof Louisa Degenhardt, Panel 1: Major types of illicit drugs globally. Evidence3–7 shows that the four drug classes
National Drug and Alcohol
Research Centre, Faculty of • Amphetamine-type stimulants are a class of synthetic, (opioids, amphetamines, cocaine, and cannabis) are used
Medicine, University of sympathetomimetic amines with powerful stimulant in most countries, but quantitative estimates of such use
New South Wales, Sydney, NSW effects on the CNS. are more scarce. This is especially so for estimates of drug
2052, Australia dependence.3–7 The best data come from developed
l.degenhardt@unsw.edu.au
• Cannabis is a generic term for preparations
(eg, marijuana, hashish, and hash oil) derived from the countries in Europe, North America, and Australasia.
Cannabis sativa plant that produce euphoria and Consequently, much uncertainty exists in the determination
See Online for webappendix
relaxation, heighten the senses, and increase sociability. of the global number of people who use illicit drugs.
• Cocaine is an alkaloid that is a powerful CNS stimulant The 2011 world drug report by the UN Office on Drugs
derived from the coca plant (Erythroxylum coca). and Crime (UNODC)28 shows this uncertainty by
• Opioids include derivatives from the opium poppy providing a range of prevalence estimates for countries
(Papaver somniferum), such as heroin and morphine, and and regions. UNODC estimated that 149–271 million
their synthetic analogues (eg, methadone, fentanyl). people aged 15–64 years (3·3–6·1%) had used an illicit
Opioids relieve pain, produce euphoria, and can cause coma drug at least once in 2009. The drugs used varied
and respiratory depression in high doses. substantially across regions (table 1), and these numbers
are not mutually exclusive, because some people used
more than one drug type. The greater uncertainty
Panel 2: Search strategy and selection criteria surrounding estimates for cannabis and amphetamines
than for cocaine and opioid use is attributable to the
We discuss the findings from reviews of published studies of scarcity of credible estimates of their prevalence of use in
prevalence, natural history, and mortality related to illicit drug many countries, and the varying prevalence seen within
use.3–14 We did searches of peer-reviewed studies (from 1990 to countries that have made estimates.3
2008) with methods recommended by the Meta-analysis of The global number of cannabis users was estimated
Observational Studies in Epidemiology (MOOSE) group,15 at 125–203 million people (2·8–4·5% of the global
systematic searches of online databases,16,17 internet searches population aged 15–64 years in 2009).28 The highest levels
for other evidence of drug use, and consultation with experts in of recorded use were in the established market economies
HIV and illicit drug use around the world. Data extraction of North America, western Europe, and Oceania. Between
followed written protocols in line with STROBE guidelines18 14 million and 56 million people aged 15–64 years were
(with cross-checking and tests of internal consistency) and data estimated to have used an amphetamine-type stimulant
graded according to predefined variables. (0·3–1·3%). The highest levels of use were near
We also draw on systematic reviews of illicit drug use as a amphetamine-manufacturing countries in southeast
cause of adverse health outcomes,19–23 cross-national studies Asia. For cocaine, the number of users worldwide ranged
of illicit drug use and dependence,24,25 reviews of illicit drug from 14 million to 21 million (0·3–0·5% of the population
markets,26–28 and estimates of the contribution of illicit drugs aged 15–64 years). The largest market was North America,
to the global burden of disease.29–34 then western and central Europe and South America.
The global number of opioid users was estimated at
12–21 million people.28 More than half these users lived
drugs has been accompanied by increases in morbidity in Asia, and the highest levels of use were along the main
and mortality in some countries,47 data for the magnitude drug trafficking routes out of Afghanistan.
of risks of iatrogenic dependence and mortality in users The health risks of illicit drug use increase with the
are not available.47,48 In countries where use of these drugs frequency and quantity of use. People who use these
has been studied, a substantial proportion of problem drugs only once or twice have, at most, a very small
users had pre-existing problems with opioids. In these increase in mortality, which is difficult to detect in
countries, estimates of opioid-dependent people include epidemiological studies. Problematic drug use, however,
both heroin and pharmaceutical opioid users.47 most clearly harms the health of users. It is defined by
the International Classification of Diseases (10th revision)
The prevalence of drug use and dependence as “harmful use” and “dependence”.54 A classification of
Major challenges exist in the accurate estimation of harmful drug use needs evidence that substance use is
the prevalence of an illegal, and often stigmatised, causing physical (eg, organ damage) or psychological
behaviour like illicit drug use. This is especially so in harm (eg, drug-induced psychosis). A classification of
cultural settings where illicit drug use can lead to drug dependence needs the presence of three or more
imprisonment, and where research participants cannot indicators of dependence for at least a month within the
be assured of confidentiality or freedom from reprisals previous year.54 A similar classification is used by the
for disclosing their drug use behaviours. By necessity, a American Psychiatric Association.55 Such indicators of
range of imperfect methods have to be used to estimate dependence include the following: a strong desire to take
the prevalence of use in such areas (panel 3). the substance; an impaired control over use; a withdrawal

56 www.thelancet.com Vol 379 January 7, 2012

 Series

syndrome on cessation or reduction of use; tolerance to

the effects of the drug; the need for larger doses to achieve Panel 3: How do we estimate the number of people who use illicit drugs?
the desired psychological effect; a disproportionate No gold-standard method exists for the estimation of the true size of the population
amount of time spent by the user obtaining, using, and of illicit drug users.50 No method is ideal for all drugs or all countries. This absence
recovering from drug use; and continuing to take the of consistency in measurement and potential biases poses major challenges for
drug despite the problems that occur. Dependence can cross-national comparisons.26,51 The best strategy is to look for convergence of results
occur with all four drug types discussed in this paper. from different indirect methods of estimation.50 Information adapted from reference 27.
Findings from a systematic review of data for the
prevalence of injecting drug use suggested that, Direct methods
worldwide, 11–21 million people injected drugs in General population or household surveys
2007 (figure; table 2).8 Injecting drug use has been In these surveys, participants are asked if they have used various drugs in the past month,
documented in 151 countries;56 its prevalence has been the past year, or in their lifetime (monthly, past year, and lifetime prevalence).
estimated in 61.8 Prevalence varies substantially across The main strength of this approach is that it accurately estimates prevalence if
regions, and between and within countries. National representative population samples are obtained, if people honestly disclose their drug
estimates varied from 0·02% of people aged 15–64 years use, and if drug users are equally distributed around the country. The major limitations
in India and Cambodia, to typically 1–2% in Azerbaijan, are that drug users are probably less likely to be available or to agree to an interview if
Georgia, Mauritius, Russia, Estonia, Malaysia, Canada, contacted; they might be reluctant to admit drug use (especially if they fear adverse
Ukraine, Puerto Rico, and Australia. Russia, China, and consequences from doing so); illicit drug use is often concentrated in large cities
the USA accounted for more than 40% of the estimated (information that national surveys might not be able to capture); and marginalised
population of injecting drug users (IDUs) worldwide groups with high rates of drug use are often missed (eg, homeless people and prisoners).
(table 2).8 The expense of these surveys limits their use in developing countries. Surveys
Global and regional estimates have been made of the underestimate the prevalence of the most harmful and stigmatised forms of illicit drug
number of problematic drug users. In 2009, UNODC use, such as opioid and injecting drug use,26 in ways that probably vary between
estimated that there were 15–39 million problem drug countries and cultures.
users globally (ie, IDUs, or problem users of opioids, School surveys
cocaine, or amphetamines).28 In these surveys, school-attending children or young people (typically in secondary
No global estimates of the prevalence of specific forms schools) are asked whether they have used various licit and illicit drugs ever or in the past
of drug dependence exist. Drug dependence (combined year, and, if so, how often.
for all illicit drugs and illicit pharmaceutical use) was
assessed in WHO’s World Mental Health household The strength and limitations of this approach are shared with general population surveys.
surveys in 27 countries in five WHO regions.57 These An additional limitation is that they exclude young people who have left school, who are
surveys recorded substantial geographical variation in most likely to have used illicit drugs. This population might be a large proportion of
rates of illicit drug use24 and dependence.57 Generally, young people in some countries.
rates of drug dependence were higher in more developed Indirect methods
countries.57 The variation is probably attributable to a These methods use different sources of data to indirectly estimate the total number of
combination of differences in prevalence and to cultural drug users.52 A simple approach is the multiplier method, in which, for example, the
differences in preparedness to report illicit drug use and number of people who receive drug treatment in a year (an indicator) is multiplied by an
related problems. In the past 20 years, nine countries estimate of the proportion of drug users who receive treatment in a year (the multiplier)
have estimated the prevalence of amphetamine depen- to estimate the total size of the drug-using population. Other indirect methods include
dence; seven of cannabis; five of cocaine; and 25 of heroin capture-recapture and back-projection estimates.52
and other opioids (table 3).3 The scarcity of estimates of
These methods are less expensive than surveys because they use existing data. Their major
specific forms of drug dependence severely limits the
limitations are uncertainty about the quality of indicator data and the validity of the
ability to make evidence-based statements about the
multipliers. These problems are usually addressed by making multiple indirect estimates
global scale of illicit drug problems.
with different indicators of illicit drug use (eg, deaths, number in drug treatment, arrests,
The estimation of global trends in illicit drug use over
treatment for complications of drug use), different multipliers, and different methods of
time is even more difficult. Drug use is routinely assessed
estimation. Often, a combined estimate is produced from these different sources.53
in few high-income countries; assessment of trends in
other countries often relies on indirect indicators of drug
supply, drug use, and problems related to drug use.27 routes have changed. African countries are now used for
Some indicators suggest that global illicit drug con- transhipment of illicit drugs to European markets and
sumption (and its related burden) has increased since illicit drug use has reportedly increased in these countries.
1990. Injecting drug use, for example, is now reported in In the established market economies of western Europe,
more countries,8,58 and HIV in IDUs is more prevalent in USA, and Australia, cannabis use has stabilised or
eastern Europe, and Asia.8 The global manufacture of decreased, whereas the use of MDMA and metamfetamine
amphetamines has increased, as has the number of has increased.26,59
problem metamfetamine users in southeast Asia and the Analysts differ in their interpretation of global trends.
Middle East.59 Opium production and heroin trafficking A UNODC report concluded that illicit drug use had

www.thelancet.com Vol 379 January 7, 2012 57

 Series

Cannabis users (N [%]) Opioid users (N [%]) Amphetamines-group users (N [%]) Cocaine users (N [%])
Africa 21 630 000–59 140 000 (3·8–10·4%) 890 000–3 210 000 (0·2–0·6%) 1 180 000–8 150 000 (0·2–1·4%) 940 000–4 420 000 (0·2–0·8%)
North Africa 4 780 000–10 620 000 (3·6–8·0%) 130 000–550 000 (0·1–0·4%) ND 30 000–50 000 (<0·1–<0·1%)
West and central Africa 11 380 000–31 840 000 (5·2–14·6%) 410 000–1 070 000 (0·2–0·5%) ND 550 000–2 300 000 (0·3–1·1%)
East Africa 2 340 000–8 870 000 (1·7–6·5%) 140 000–1 310 000 (0·1–1·0%) ND ND
Southern Africa 3 130 000–7 810 000 (3·9–9·8%) 210 000–280 000 (0·3–0·3%) 280 000–780 000 (0·4–1·0%) 270 000–730 000 (0·3–0·9%)
The Americas 40 950 000–42 860 000 (6·7–7·0%) 1 180 000–1 910 000 (0·2–0·3%) 5 170 000–6 210 000 (0·8–1·0%) 8 280 000–8 650 000 (1·4–1·4%)
North America 32 520 000–32 520 000 (10·7–10·7%) 1 000 000–1 630 000 (0·3–0·5%) 3 460 000–3 460 000 (1·1–1·1%) 5 690 000–5 690 000 (1·9–1·9%)
Central America 550 000–610 000 (2·2–2·5%) 20 000–20 000 (0·1–0·1%) 320 000–320 000 (1·3–1·3%) 120 000–140 000 (0·5–0·6%)
The Caribbean 440 000–2 060 000 (1·6–7·6%) 50 000–80 000 (0·2–0·3%) 30 000–530 000 (0·1–1·9%) 110 000–330 000 (0·4–1·2%)
South America 7 410 000–7 630 000 (2·9–3·0%) 110 000–170 000 (0·0–0·1%) 1 340 000–1 890 000 (0·5–0·7%) 2 360 000–2 480 000 (0·9–1·0%)
Asia 31 340 000–67 970 000 (1·2–2·5%) 6 440 000–12 020 000 (0·2–0·4%) 4 330 000–38 230 000 (0·2–1·4%) 400 000–2 300 000 (<0·1–0·2%)
East and southeast Asia 5 440 000–24 160 000 (0·4–1·6%) 2 800 000–4 990 000 (0·2–0·3%) 3 480 000–20 870 000 (0·2–1·4%) 400 000–1 070 000 (<0·1–0·2%)
South Asia 16 830 000–28 110 000 (1·9–3·1%) 1 380 000–3 170 000 (0·3–0·4%) ND ND
Central Asia 1 950 000–2 260 000 (3·8–4·4%) 320 000–320 000 (0·6–0·6%) ND ND
Near and Middle East 6 060 000–12 360 000 (2·4–4·8%) 1 940 000–3 170 000 (0·8–1·4%) 460 000–4 330 000 (0·2–1·7%) 40 000–650 000 (<0·1–0·3%)
Europe 28 730 000–29 250 000 (5·2–5·3%) 3 110 000–3 470 000 (0·6–0·6%) 2 540 000–3 180 000 (0·5–0·6%) 4 300 000–4 750 000 (0·8–0·9%)
West and central 22 750 000–22 860 000 (7·1–7·1%) 1 010 000–1 170 000 (0·3–0·4%) 2 030 000–2 120 000 (0·7–0·7%) 3 990 000–4 090 000 (1·2–1·3%)
Europe
East and southeast 5 980 000–6 380 000 (2·6–2·6%) 2 100 000–2 300 000 (0·9–1·0%) 510 000–1 050 000 (0·2–0·5%) 310 000–660 000 (0·1–0·3%)
Europe
Oceania 2 160 000–3 460 000 (9·3–14·8%) 40 000–50 000 (0·2–0·3%) 470 000–640 000 (2·0–2·8%) 330 000–400 000 (1·4–1·7%)
Global estimates 124 810 000–202 680 000 (2·8–4·5%) 11 66 000–20 660 000 (0·3–0·5%) 13 690 000–56 410 000 (0·3–1·3%) 14 250 000–20 520 000 (0·3–0·5%)

Data are N and % of population aged 15–64 years in each region. ND=insufficient data for subregional-specific estimate. Data comprised published country-level estimates from the 2010 World Drug Report.28 Estimates
were made only when direct estimates were published for at least two countries with at least 20% of the region’s population aged 15–64 years. Regions with fewer data and less certainty have greater ranges. The scarcity
of robust data for levels of drug use, particularly in countries such as India and China, preclude an accurate estimate of the global population of illicit drug users. Regions with estimates from more countries have more
precise estimates. Adapted from reference 28 by use of regions defined by the UN Office on Drugs and Crime.

Table 1: Estimated number of people aged 15–64 years who used illicit drugs at least once in the past year, 2009

stabilised and perhaps decreased between 1998 and peak in young adulthood and decrease as young people
2007.60 A RAND report undertaken for the EU, by enter relationships, marry, have children, engage in
contrast, concluded that globally illicit drug use had, at further education, and enter the workforce. People who
best, remained stable, and possibly increased during the do not make these transitions are more likely to persist
same period.26 in their drug use.63
The natural history of dependence on illicit drugs has
The natural history and risk factors for use been poorly studied in prospective cohort studies. Most
and dependence of these studies have recruited cohorts of users seeking
Studies in high-income countries, with high levels of treatment or entering the criminal justice system, groups
cannabis use, have reported a common temporal whose trajectory of use can differ from users who do not
ordering of drug initiation—alcohol and tobacco, enter these systems. The restricted evidence suggests
followed by cannabis use, and then other illicit drugs. that a minority of individuals will no longer meet criteria
This pattern persists after control for possible con- for dependence a year after diagnosis.9 This proportion is
founders.19,61,62 This pattern is not consistent across higher for cannabis and amphetamines than it is for
countries.25 Use of other illicit drugs is more prevalent heroin and cocaine.9
than is use of cannabis in some countries (eg, Japan), Most of what we know about risk factors for problem
and the association between initiation of alcohol, tobacco, use of opioids comes from retrospective studies of
and cannabis, and other illicit drug use is stronger in treatment populations rather than prospective studies
some countries (eg, the USA) than in others (eg, the of representative cohorts of young people.64 We know
Netherlands).25 Variations in patterns of drug initiation most from cohort studies about risk factors and
between countries and cultures suggest that entry into pathways for regular cannabis use in developed
illicit drug use is dependent on social factors and drug countries.65 Similar risk factors seem to predict early
availability, as well as characteristics of users and social cannabis use in developing countries.65 These risk
settings that facilitate or deter use. factors can be divided into social and contextual factors,
Drug use is consistently more common in boys and family factors, individual factors, and peer affiliations
men than in girls and women.24,63 Rates of cannabis use during adolescence.

58 www.thelancet.com Vol 379 January 7, 2012

 Series

Eastern Europe
3 476 500 IDUs
Canada and USA
2 270 500 IDUs
Western Europe Central Asia
1 044 000 IDUs 247 500 IDUs Southeast
and east Asia
South Asia 3 957 500 IDUs
Caribbean 569 500 IDUs
186 000 IDUs Middle East and north Africa
1 778 500 IDUs

Sub-Saharan
Africa
1 778 500 IDUs
Latin America
2 018 000 IDUs

Oceania
Prevalence of HIV in IDUs 193 000 IDUs
<5%
5% to <10%
10% to <15%
15% to <20%
≥20%

Figure: Estimated number of injecting-drug users (IDUs) and regional prevalence of HIV in people who inject drugs
Adapted from reference 8.

Estimated number of people who inject Estimated Estimated number of people with HIV Estimated
drugs regional who inject drugs regional
mid-point IDU mid-point HIV
prevalence (%) prevalence in
IDUs (%)
Low Middle High Low Middle High
Eastern Europe 2 540 000 3 476 500 4 543 500 1·50% 18 500 940 000 2 422 000 27·0%
Western Europe 816 000 1 044 000 1 299 000 0·37% 39 000 114 000 210 500 10·9%
East and southeast Asia 3 043 500 3 957 500 4 913 000 0·27% 313 000 661 000 1 251 500 16·7%
South Asia 434 000 569 500 726 500 0·06% 34 500 74 500 135 500 13·1%
Central Asia 182 500 247 500 321 000 0·64% 16 500 29 000 47 000 11·8%
Caribbean 137 500 186 000 241 500 0·73% 6000 24 000 52 500 12·9%
Latin America 1 508 000 2 018 000 2 597 500 0·59% 181 500 580 500 1 175 500 28·8%
Canada and USA 1 604 500 2 270 500 3 140 000 0·99% 127 000 347 000 709 000 15·3%
Pacific Islands 14 500 19 500 25 000 0·36% <250 500 500 1·4%
Australia and New Zealand 105 000 173 500 236 500 1·03% 500 2500 6000 1·5%
Middle East and north Africa 89 000 121 000 156 500 0·05% 1500 3500 6500 2·9%
Sub-Saharan Africa* 534 500 1 778 500 3 022 500 0·43% 26 000 221 000 572 000 12·4%
Extrapolated global estimates 11 008 500 15 861 500 21 222 000 0·37% 764 000 2 997 500 6 589 000 18·9%

Adapted from reference 8 by use of regional groupings defined by the the Joint UN Programme on HIV/AIDS. IDU=injecting drug user. *The estimates for sub-Saharan Africa
should be viewed with caution because the prevalence estimates were derived from three of 47 countries in the region (South Africa, Mauritius, and Kenya). Furthermore, the
estimated range of IDU for this region was derived by applying the regional observed error; this large error band shows the substantial uncertainty around these estimates.

Table 2: Injecting drug use and HIV in people aged 15–64 years who inject drugs, 2008

The major social and contextual factors that increase backgrounds more likely to use illicit drugs.68 Less well
the likelihood of use are drug availability, use of tobacco studied structural risk factors include poverty and social
and alcohol at an early age (ie, early adolescence),66 and and cultural factors. Family factors that increase risk
social norms for the toleration of alcohol and other drug during adolescence include poor quality of parent–child
use.67 Socioeconomic background is also an important relationships,69 parental conflict,25,70,71 and parental and
correlate of use, with people from more disadvantaged sibling drug use.69,72,73

www.thelancet.com Vol 379 January 7, 2012 59

 Series

Studies identified Regions Range of adult past-year

prevalence estimates
Amphetamines Nine countries: four indirect prevalence estimates and Western Europe, eastern Europe, 0·10–0·73%
five representative household survey-based estimates North America, Australia,
southeast Asia
Cannabis Seven countries: all representative household Western Europe, North America, 0·10–1·50%
survey-based estimates Australia, southeast Asia
Cocaine Five countries: one indirect prevalence estimate and Western Europe, southeast Asia, 0·07–0·52%
four representative household survey-based estimates North America
Heroin and other opioids 25 countries: 17 indirect prevalence estimates and Western Europe, eastern Europe, 0·11–0·82%
eight representative household survey-based estimates North America, southeast Asia,
Australia, south Asia

Age ranges varied across studies so not all estimates were directly comparable. Only national estimates are included in these ranges; additional studies had sub-national estimates.3–7

Table 3: Systematic reviews of prevalence studies of dependence on illicit drugs

Individual risk factors include being male,71 the will use more than one of the four drug types discussed
personality traits of novelty74 and sensation seeking,75 in this paper. Therefore, the acute and long-term health
early oppositional behaviour and conduct disorders in effects of their drug use might be even greater than it is
childhood,73,76 and poor school performance, low commit- for people using only one drug type. Little work has
ment to education, and early school leaving.77,78 Affiliating quantified these potential interactions, but they are likely
with antisocial and drug-using peers is one of the to be important.
strongest predictors of adolescent alcohol and other drug Many studies have recorded associations between
use79,80 that operates independently of individual and illicit drug use and various health-related harms, but
family risk factors.78,81 determination of whether such associations are causal is
These risk factors often co-occur. Young people who more difficult. To make a causal inference it is necessary
initiate substance use at an early age are often exposed to to document an association between drug use and the
many social and family disadvantages, come from families adverse outcome, confirm that drug use preceded the
with problems and a history of parental substance use, are outcome, and exclude alternative explanations of the
impulsive, have performed poorly at school, and are association, such as reverse causation and confounding.90
affiliated with delinquent peers. Young people with many Cohort studies of problem amphetamine, cocaine, and
of these risk factors start alcohol, tobacco, and illicit drug heroin users suggest that these drugs increase the risk
use at an early age, and often develop problem drug use.79 of premature death, morbidity, and disability. These
Risk factors for drug dependence can differ between studies have rarely controlled for social disadvantage,
countries, although few studies have directly examined but the mortality excess is too large to be wholly
this.24 A study24 of initiation and progression to dependence accounted for by this confounding;91 the major causes of
in 17 countries showed that the following variables increased mortality are plausibly and directly related to
predicted the development of illicit drug dependence in illicit drug use.36
users: earlier onset of drug use, use of multiple types of Tables 4 and 5 compare the availability of evidence, the
illicit drugs, and development of externalising (eg, conduct quality of evidence, and the strength of associations seen
disorder) and internalising (eg, depression) disorders for each drug type for a range of putative acute and
before the age of 15 years. These findings are lent support chronic outcomes. Several things are apparent. First, the
by those from cohort studies in high-income countries, risks of cannabis use are much smaller than those of
which have recorded that early onset drug use, and mental other illicit drugs, largely because cannabis does not
health problems, are risk factors for dependent drug use,82 produce fatal overdoses and it cannot easily be injected.
and that mental health problems increase the risk of Second, the quality of evidence varies widely across drug
problem drug use. and health outcomes—data for cannabis are largely from
prospective population-based cohorts, whereas data for
Health consequences of illicit drug use the other drug types are from selected cohorts of treated
Four broad types of adverse health effects of illicit drug opioid, cocaine, and amphetamine users.
use exist:1 the acute toxic effects, including overdose; the Third, the magnitude of the effect is often poorly
acute effects of intoxication, such as accidental injury quantified. Especially in view of the known potential for
and violence; development of dependence; and adverse serious adverse health and social consequences from
health effects of sustained chronic, regular use, such as opioids and psychostimulants, a clear need exists for
chronic disease (eg, cardiovascular disease and cirrhosis), more prospective, quantitative, longitudinal studies of
blood-borne bacterial and viral infections, and mental specific patterns of drug use (or common combinations)
disorders (tables 4, 5). Many people who use illicit drugs and specific outcomes of such use, to produce better

60 www.thelancet.com Vol 379 January 7, 2012

 Series

Cannabis Opioids Amphetamines Cocaine

Effect Level of Size of Reference Effect Level of Size of Reference Effect Level of Size of Reference Effect Level of Size of Reference
evidence effect evidence effect evidence effect evidence effect
Acute toxic effects × ·· 0 19, 38 √ C CMR 0·7 11 √ C ? 13, 39, √ C ? 12, 39
(fatal overdose) 109
Acute intoxication effects
Accidental ? ·· ·· 19, 38 √ C CMR 0·16 11 ? ·· ·· 13, 39 ? ·· ·· 12, 39
injury
Motor vehicle √ D ? 19, 38, 41 ? ·· ·· 41 ? ·· ·· 13, 41 ? ·· ·· 12, 41
accidents
Drug-induced √ A OR 2–3 19, 38, 21 × ·· 0 39 √ A ? 37, 39 √ E ? 39, 84
psychotic
symptoms
Myocardial ? E ·· 19, 38 × ·· 0 39 √ E ? 37, 39 √ E ? 39
infarction
Dependence √ A 9% 61 √ A 23% 101 √ A 11% 101 √ A 16% 101
(lifetime risk %)
Adverse health effects of chronic use
Cardiovascular ? ·· ·· 19, 38 √ E ? 39 √ C ? 39 √ E ? 39
pathology
Liver disease × ·· 0 19, 38 √ C ? 39 ? C ? 37, 39 ? ·· ·· 39
Pulmonary ? ·· ·· 19, 38 √ E ? 39 ? C ? 39 ? ·· ·· 39
disease
Cancers ? ·· ·· 10, 19, 38 ? C ? 11, 85 ? ·· ·· 39 ? ·· ·· 39
Neurotoxic ? C ·· 19, 38 × ·· ·· 39 √ ·· ? 37 √ ·· ? 39
effects
Psychotic √ B OR 2–3 19, 20 ,23 × ·· 0 39 √ D ? 86 √ D ? 84
disorders
Common ? B ·· 19, 23 √ D ? 87 √ D ? 86 √ D ? 87
mental
disorders
Suicide × B 0 10 √ C CMR 0·12 11 ? ·· ? 37, 86, 87 ? ·· ? 37
Increased × B 1 10 √ C 14·7 11 √ C 6·2† 13 √ C 4·7–7·6‡ 12
mortality (95% CI
(standardised 12·8–16·5) *

mortality ratios)

A=experimental or controlled evidence supports this finding. B=findings across cohorts, representative population-based. C=findings across cohorts of drug users. CMR=crude mortality rate per 100 person-years.
D=findings across cross-sectional studies, representative population-based, or case-control studies. E=cross-sectional associations in non-representative samples of drug users, case series suggesting outcome.
n/a=not applicable. OR=odds ratio. SMR=standardised mortality ratio. ×=this drug does not seem to have an effect on the outcome. √=the outcome might be increased by the use of this drug. ?=Insufficient data
exists for this drug and this outcome to allow conclusions about the association between the two. *Pooled SMR estimated from random effects meta-analysis (very high heterogeneity existed across studies;
stratified analyses investigated this heterogeneity in further analyses and demographic and regional differences were clearly evident). †Only one study from the Czech Republic reported SMRs (this should be
interpreted with caution). ‡Range from several studies only—interpret with caution.

Table 4: Major potential acute and chronic consequences of illicit drug use

estimates of expressed risk. Nonetheless, the major Cannabis Opioids, amphetamines, and cocaine
causes of increased mortality are plausibly and directly Effect Size of effect
related to illicit drug use.36 Tables 4 and 5 indicate that
HIV × 0 Risk of HIV infection via injection with an HIV-infected needle: about 1 in
although evidence links opioid, cocaine, and amphet- 125 injections88 (figure, table 2).
amine use with more adverse outcomes than cannabis, Hepatitis C × 0 The prevalence of hepatitis C antibodies varies widely in IDUs, from 1% to
gaps remain in knowledge about the causal nature and greater than 90% prevalence.14,22
magnitude of the risks. Hepatitis B × 0 Prevalence substantially increased.14
Drugs differ in their most direct effects. To overdose
Infective × 0 Most infections due to Staphyloccus aureus.39 Risk rarely quantified;
fatally on cannabis is difficult, if not impossible, whereas endocarditis two US cohorts of cocaine IDUs noted 3–10% of deaths due to sepsis or
fatal overdose is a well-known risk for the other major illicit endocarditis.12
drugs. The risk of overdose is increased when opioids are Tuberculosis × 0 Has been noted in some countries as especially prevalent as an HIV co-infection.89
used in combination with other CNS depressants, such as
alcohol and benzodiazepines,92,93 and when an individual ×=this drug does not seem to have an effect on the outcome.
resumes opioid use after periods of abstinence during
Table 5: Consequences of unsafe drug injection
drug treatment or imprisonment.36 Stimulant-related

www.thelancet.com Vol 379 January 7, 2012 61

 Series

overdoses can trigger fatal cardiac arrhythmias and A consistent association exists in longitudinal studies
strokes,39,94,95 which are otherwise very rare causes of death between early onset of cannabis use, regular cannabis use,
in healthy young adults.39,96 Few cohort studies have been and a later diagnosis of schizophrenia, which increasing
done to examine the magnitude of risk in stimulant users, evidence suggests is not caused by confounding.19,20,104–106
making the estimation of the magnitude of overdose risk Meta-analyses of prospective population-based studies
difficult, although the evidence that does exist suggests have noted a doubling of the risk of psychotic outcomes in
that crude mortality rates for drug overdose do not differ regular cannabis users, after controlling for confounders,20,23
much from those seen across cohorts of opioid users.12,13 and that the age of onset of schizophrenia is about 2·7 years
More thorough study of the rates and causes of death in earlier for cannabis users who develop the disorder.107
psychostimulant users is needed. Cannabis use is a biologically plausible contributory cause
Cannabis use impairs cognitive and behavioural of schizophrenia in vulnerable individuals.21
functions,97 especially for sustained-attention tasks, so A less consistent association exists between cannabis
the risk of road-traffic accidents can increase if users use and depression, and the evidence for a causal role
drive while intoxicated. Controlled studies have recorded between cannabis use and depression is less convincing
statistically significant deficits in driving performance, than it is for psychotic symptoms and disorders.19,23
but studies under more realistic road conditions report Anxiety, depression, and other illicit drug use are very
more impairment to a lesser extent.19,38,41 Case-control strongly associated,99 but to ascertain whether these
studies have recorded weak associations between disorders precede and contribute to the development of
cannabis use and culpability for road-traffic accidents, problem drug use, or are exacerbated by such use, is
with higher risks in individuals who use more difficult. For example, conduct disorders, depression, and
cannabis.19,38 These risks are less than those for alcohol, anxiety disorders, which develop in adolescence and early
and fewer drivers use cannabis—the estimated adulthood, predispose young adults to use illicit drugs at
proportion of road-traffic accidents attributable to an early age, thereby increasing the risk of their developing
cannabis in France between 2001 and 2003 was 3% dependence. Longitudinal studies provide strong evidence
(vs 30% for alcohol).98 The relative contribution of that heavy alcohol use is a causal factor in depressive
cannabis use to road-traffic accidents will vary between disorders.108 Similar longitudinal analyses are needed to
countries according to the prevalence of cannabis use understand the relation between different types of illicit
and access to motor vehicles. drug use and depression and other mental disorders.
Other illicit drugs can adversely affect an individual’s Reviews have concluded that insufficient evidence
ability to drive,41 although data for the effect of opioids is available to decide whether a causal relation exists
and stimulants on driving is equivocal.41 Nonetheless, between cannabis use and suicide.10,19 By contrast, rates of
road-traffic accidents, falls, drowning, and related injuries self-reported suicide attempts in problem opioid, cocaine,
are a more common cause of death in opioid and and amphetamine users109 are much higher than they are
stimulant users than in their non-using peers. The in non-drug-using peers of the same age, sex, and
contribution of these causes to drug-related disease socioeconomic status.110 The association is probably
burden might have been underestimated, because few mediated by depression, rates of which are high in
cohort studies report deaths from trauma, and such problem drug users.36 The intoxicating effects of these
deaths in drug users might not have been recorded as drugs, and the stresses of an illicit-drug-dependent
drug-related.36 A pooled estimate from cohort studies of lifestyle, probably increase suicide risk in depressed drug
opioid users suggested that the trauma-related crude users. Meta-analyses have produced a pooled crude
mortality rate was 0·16 per 100 person-years (95% CI mortality rate for suicide in opioid-dependent individuals
0·12–0·21).11 of 0·12 per 100 person-years (95% CI 0·08–0·16).11
In the USA, an estimated 20% of people who use an Because cannabis cannot be readily injected, the risks
illicit drug will meet the criteria for dependence;99 the of unsafe injecting arise from only opioid, cocaine, and
proportion reported in Australia is much the same.101 amphetamine use. HIV infection risk after injection with
Illicit drugs differ in their dependence risk,100,101 ranging an HIV-contaminated syringe has been estimated at
from 9% of lifetime cannabis users to 23% of lifetime 0·67%.88 The sharing of other contaminated drug-use
heroin users in one study (webappendix p 2).101 Such paraphernalia presents an unquantified but probably
variance is attributable to differences in pharmacological lower risk. The risk of sexual transmission of HIV
effects (drugs with a rapid onset and shorter duration of between HIV-positive IDUs and their sexual partners is
effect have a higher dependence risk) and route of admin- much lower at 0·02–0·05% per heterosexual sex act;111–113
istration (drugs that are smoked or injected have a higher risk during receptive anal intercourse between men can
dependence risk than do those that are swallowed or used be 0·82% (95% CI 0·24–2·76%).114
intranasally). More heroin injectors meet dependence Pronounced geographical variations exist in the
criteria than do cannabis smokers.101 Amphetamine and prevalence of injecting drug use and HIV infection in
cocaine users who smoke or inject have a higher risk of IDUs (figure; table 2). Injecting drug use has been
dependence than do those who use intranasally.102,103 reported in 151 countries,36 with 0·8–6·6 million (of an

62 www.thelancet.com Vol 379 January 7, 2012

 Series

estimated 11–21 million injectors in 2007) estimated to Burden of disease attributable to illicit drug use
be living with HIV.8 Existing estimates of drug-related Since 1993, estimates of the causes of global disease
HIV have been insufficient in their account of its burden have used disability-adjusted life years
geographical variation; new estimates being made for (DALY)1243to combine disease burden from premature
the 2010 global burden of disease analysis will be more mortality with that from disability. This metric allows a
accurate, in view of the increase in the amount of data comparison of the contribution across diseases, injuries,
for the extent of both injecting drug use and HIV in and risk factors. In 2002, the comparative risk assessment
people who inject drugs. exercise124 estimated the proportion of disease burden
The viruses that cause hepatitis B and hepatitis C attributable to alcohol, tobacco, and injecting drug use.
infections are also spread by sharing contaminated These estimates explicitly accounted for variations in
injection equipment.115,116 Large proportions of IDUs are prevalence of different diseases or injuries, considered
infected with hepatitis C, with an estimated 10·0 million age and sex differences, and included mortality as well
(range 6·0–15·2 million) injectors thought to be positive as morbidity.
for hepatitis C antibodies in 2010:14 75–85% of these Global mortality attributable to illicit opioid use was
develop chronic hepatitis C infections117–119 that can estimated at 100 000 deaths in 1990, 62% of which were
potentially lead to cirrhosis, liver failure, and in high-income countries.125 An estimate for 2000 (which
hepatocellular carcinoma.120 The proportion of people defined illicit drug use as injecting or problem use of
with hepatitis C who develop cirrhosis is estimated at 7% amphetamines, cocaine, or opioids) estimated all-cause
after 20 years of infection, and 20% after 40 years.121 mortality, and mortality attributable to AIDS, overdose,
Additional stresses on the liver from heavy alcohol suicide, and trauma from cohort studies of problem
intake, liver fibrosis, and HIV or hepatitis B co-infection, illicit drug users (table 6).29 Major regional differences
can increase rates and speed of the development of were recorded in the quality of data for the prevalence of
complications.121 Many individuals living with hepatitis C use, and estimates relied heavily on studies of mortality
report fatigue, poor sleep, and abdominal pain, which in problem drug users in high-income countries.126 Use
impair quality of life as much as diabetes does.122 In of findings from such studies is a major limitation of
countries with low rates of HIV infection in IDUs, these estimations—reviews have since shown that
the burden of hepatitis C in IDUs might be compara- mortality in drug users varies geographically and
tively higher. according to country income.11
Findings from reviews show no evidence that cannabis The 2000 study estimated that the median number
use increases overall mortality,10,19 which contrasts with of deaths attributed to illicit drugs was about
mortality from other types of illicit drug use (webappendix 200 000 (241 000 from summing all four causes, and
p 6). A meta-analysis of mortality in opioid users calculated 197 000 with an estimate of all-cause mortality).29
a pooled standardised mortality ratio of 14·7 (95% CI Uncertainty intervals around each estimate were wide
12·8–16·5).11 These risks varied geographically, with, for (102 000–322 000 and 82 000–408 000, respectively);
example, lower increases in mortality in Australia, and nonetheless, the 2000 estimate29 was double the
higher increases in Italy.11 Fewer cohort studies of cocaine 1990 estimate.125 WHO estimates of global DALYs
and amphetamine users12,13 report increased premature attributable to amphetamine, cocaine, or opioid use in
mortality; mortality increases in these cohorts seem less 2004 suggested that use of these drugs accounted for
pronounced than they are for opioid users. 0·9% of global DALYs, varying widely across regions

AIDS mortality Opioid overdose mortality Suicide mortality Trauma mortality Illicit drug all-cause Alcohol Tobacco
mortality all-cause all-cause
(middle) (middle)
Low Middle High Low Middle High Low Middle High Low Middle High Low Middle High
Africa 2000 5000 9000 1000 2000 3000 <1000 1000 2000 1000 4000 7000 13 000 28 000 42 000 213 000 158 000
Americas 12 000 17 000 30 000 7000 8000 11 000 806 1000 5261 2000 7000 12 000 37 000 61 000 83 000 279 000 802 000
Europe 1000 2000 10 000 7507 14 000 20 000 2000 7000 15 000 2000 5000 7000 17 000 33 000 47 000 538 000 1 605 000
Eastern 2000 5000 10 000 7000 17 000 26 000 1000 3000 4000 1000 4000 6000 9000 15 000 22 000 16 000 186 000
Mediterranean
Southeast Asia 6000 59 000 111 000 3000 23 000 45 000 2000 15 000 28 000 2000 4000 6000 8000 17 000 26 000 229 000 1 035 000
Western 5000 11 000 18 000 3000 4000 4000 <1000 1000 3000 8000 10 000 12 000 17 000 44 000 104 000 526 000 978 000
Pacific
Total 26 000 105 000 191 000 29 000 69 000 111 000 8000 32 000 57 000 18 000 34 000 50 000 102 000 197 000 322 000 1 804 000 4 800 000

Adapted from references 29,30,32,131. WHO regional definitions used. 31–33

Table 6: Estimated mortality attributable to injecting or problematic drug use according to several major causes, compared with alcohol and tobacco—2000 Global Burden of Disease
comparative risk assessment

www.thelancet.com Vol 379 January 7, 2012 63

 Series

HIV/AIDS DALYs Drug use Poisoning Suicide/ Trauma† DALYs Total illicit drugs Total alcohol Total tobacco
disorders* DALYs self-inflicted DALYs DALYs DALYs
DALYs injuries DALYs
Number % Number % Number % Number % Number % Number % Number % Number %
(000s) (000s) (000s) (000s) (000s) (000s) (000s) (000s)
Africa 0 0 939 000 100 9000 0·8 46 000 3·7 136 000 0·7 1 131 000 0·3 7 759 000 2·1 1 930 000 0·5
Americas 231 000 10·7 2 446 000 100 55 000 9·3 81 000 5·0 297 000 2·8 3 110 000 2·2 13 102 000 9·1 8 837 000 6·1
Europe 620 000 52·5 1 369 000 100 23 000 1·1 170 000 5·5 213 000 1·7 2 395 000 1·6 17 342 000 11·4 17 725 000 11·7
Eastern Mediterranean 199 000 21·6 1 675 000 100 7000 1·7 68 000 6·2 168 000 1·1 2 117 000 1·5 763 000 0·5 2 793 000 2·0
Southeast Asia 588 000 9·6 1 252 000 100 17 000 0·9 445 000 6·2 283 000 0·6 2 585 000 0·6 12 066 000 2·7 12 764 000 2·8
Western Pacific 788 000 54·1 674 000 100 22 000 1·7 39 000 0·7 363 000 1·4 1 886 000 0·7 18 393 000 6·9 12 848 000 4·8
Global DALYs 2 426 000 4·1 8 355 000 100 133 000 1·8 849 000 4·3 1 460 000 1·1 13 223 000 0·9 69 424 000 4·5 56 897 000 3·7

Extracted from reference 31. WHO regional definitions used. *Cannabis was not included in these estimates. †Included road-traffic accidents, falls, fires, drownings, and other unintentional injuries—these
31–33

estimates specifically excluded violence as a potential consequence of illicit drug use.

Table 7: Estimated disability-adjusted life years (DALYs) attributable to illicit drug use according to several major causes, compared with alcohol and tobacco, 2004

(table 7). Drug dependence (excluding cannabis) was Harms of illicit drug use not captured in burden of
the largest of the four causes of global illicit drug burden disease estimates
assessed (68%), followed by HIV/AIDS (18%). Burden of disease estimates do not include the adverse
These estimates indicate that illicit drug use is a social effects on drug users, such as stigma and
substantial global cause of premature mortality and discrimination, or the adverse effects that drug users’
morbidity. They were acknowledged to be underestimates behaviours have on public amenity (eg, public drug use,
because they did not include cannabis and MDMA, or drug dealing, and discarded injection equipment) and
the burden attributable to hepatitis B, hepatitis C, or public safety (eg, violence between drug dealers, and
drug-related violence.29 The Australian burden of disease property crime to finance illicit drug use).
study included a greater number of drug-related Interactions also exist between illicit drug policy and
outcomes, and its findings suggest that existing global drug-related harm. Both internationally and nationally,
figures substantially underestimate illicit-drug-related policies focus on the reduction of supply and use by
burden (panel 4; table 8). criminalisation of drug use and supply. Criminalisation
increases the price of illicit drugs,132 and probably
Comparison of illicit drugs with tobacco and alcohol discourages some people from using these drugs. The
Although far from perfect, the existing global burden of prevalence of illicit drug use is therefore probably lower
disease estimates provide a common metric to compare than it might be if their sale and use was as legal as
the harms caused by illicit drugs with those of alcohol alcohol and tobacco. This is not true for solely removing
and tobacco—regionally and globally—while taking criminal penalties for use.133
account of differences in prevalence and harms. Conversely, the higher price of illicit drugs probably
Comparison of existing estimates of use and burden of makes it more likely that some who use illicit drugs
disease for illicit drugs, alcohol, and tobacco (table 9) will engage in criminal activities to finance their use
draws attention to four main points. Globally, many (eg, by drug dealing, property offences, and fraud).134
fewer people use illicit drugs than use alcohol (roughly Furthermore, violence is often associated with illicit drug
one-tenth). Nonetheless, estimated levels of problem markets, presenting a risk to the wellbeing of drug
use of opioids, cocaine, or amphetamines are an users.135 Cohort studies of opioid users suggest a pooled
appreciable proportion of those for alcohol use disorders homicide crude mortality rate of 0·10 per 100 person-
(0·3–0·9% vs 1·2%). Tobacco use is far more widespread, years (95% CI 0·07–0·13),11 and findings from a meta-
and so its contribution to disease burden was greater analysis of toxicological studies of homicide victims
than that for alcohol or illicit drugs. Finally, the show that about 6% of victims tested positive for
estimated number of attributable deaths and DALYs cannabis, 11% for cocaine, and 5% for opioids.136 A
were much higher for alcohol use disorders than for review35 concluded that “the distal factors surrounding
problem illicit-drug use (3·8% and 4·5% for alcohol illicit drug markets appear to play a larger role in illicit
and 0·4% and 0·9% for illicit drugs, respectively). The drug-related homicide than the proximal effects of
higher number of years of life lost from illicit drug use [these] substances”. Drug-related law enforcement often
(2·1 million vs 1·5 million for alcohol) shows the comprises a large proportion of the social costs of illicit
concentration of illicit drug deaths in younger people, drug use.137,138 Countries that are sites of illicit drug
whereas alcohol and tobacco deaths occur in middle- production or trafficking might have substantial social,
aged and older adults. political, and health disruption from the activities of the

64 www.thelancet.com Vol 379 January 7, 2012

 Series

Panel 4: The burden of disease attributable to illicit drugs, Deaths Disability-adjusted

life years (DALYs)
tobacco, and alcohol in Australia
Number % total Number % total
The disease burden attributable to opioids, amphetamines,
Heroin and polydrug use 263 0·2 16 758 0·6
cocaine, and cannabis in Australia (table 8) is of
Hepatitis C 759 0·6 11 709 0·4
international relevance, because Australia has good data for
Cannabis use 0 0·0 5206 0·2
mortality, and rates of most types of illicit drug use (cocaine
Suicide and self-inflicted 204 0·2 4458 0·2
excepted) are much the same as those in other high-income
injuries
countries. A 2003 study examined more consequences of
Hepatitis B 329 0·2 3637 0·1
drug use than had previous global studies,127 allowing some
Other 150 0·1 9696 0·4
assessment of the extent to which global estimates
Total
underestimate disease burden.
Illicit drugs 1705 1·3 51 463 2·0
In 2003, illicit drug use in Australia caused similar levels of Alcohol 1084 0·8 61 091 2·3
disease burden as did alcohol (2·0% vs 2·3%).127 More deaths Tobacco 15 511 11·7 204 788 7·8
were attributed to illicit drugs than to alcohol, on the
Adapted from reference 127.
assumption that moderate alcohol use reduced cardiovascular
heart disease mortality in middle-aged adults. Injecting drug Table 8: Burden of disease attributable to illicit drugs, alcohol, and
use and opioid and polydrug use accounted for more than half tobacco in Australia, 2003
the contribution of illicit drugs to disease burden. Cannabis
dependence, psychosis, suicide, and road-traffic crashes
accounted for 0·2% of the total disease burden and 10% of the Illicit drugs Alcohol Tobacco
burden for all illicit drugs.127 The inclusion of cannabis use, N Global % N Global % N Global %
hepatitis C, and hepatitis B produced an estimate that was Past-year users 149–272 million* 3·3–6·1% 2830 million† 44% ·· ··
1·6 times greater than one based on the fewer outcomes Problem or 15–39 million‡ 0·3–0·9% 76·3 million§ 1·2%§ 1670 million¶ 26%¶
included in the 2000 global burden of disease estimates. dependent users

The estimates for the 2010 global burden of disease Attributable 0·25 million 0·4% 2·25 million 3·8% 5·11 million 8·7%
deaths†
project128,129 will include estimates of disease burden
Years of life lost 2·1 million 0·23% 1·5 million 0·17% NR NR
attributable to cannabis use, and a greater number of illicit due to use
drug use disorders and adverse consequences of illicit drug disorders†
use.128 The Australian data suggest that the new global Attributable 13·22 million 0·9% 69·42 million 4·5% 56·90 million 3·7%
estimates will be substantially higher than previous ones. DALYs†

*Any illicit drug (including cannabis); estimates refer to individuals aged 15–64 years made by the UN Office on Drugs and
Crime for 2009.27,28 †These data refer to 2004, and were extracted from WHO Global Burden of Disease spreadsheets31—
large-scale criminal networks involved, as is the case in cannabis is not included in the estimates for illicit drugs. ‡Estimates made by the UN Office on Drugs and Crime for 2009
in individuals aged 15–64 years, and refer to problem drug users, not necessarily dependent drug users—problem or
Afghanistan and Mexico.139 dependent cannabis users are excluded in this definition.27,28 §These data are reported in the 2004 WHO Global Status of
The dominant policy focus on supply reduction and Alcohol Report,130 and refer to individuals aged 15 years or older with alcohol use disorders—prevalence estimate
criminalisation of drug use can also adversely affect the approximated from population data reported in references 31 and 132. ¶These data refer to current smokers aged
15 years or older in 2004 and are from references 31 and 32. NR=not specifically reported.
health and wellbeing of illicit drug users in the following
ways: by increasing the health risks of illicit drug use Table 9: Comparison of existing estimates of use and burden of disease for illicit drugs, alcohol, and tobacco
(eg, if users engage in risky injecting to avoid arrest by
police);140,141 by increasing risks of engaging in sex work coverage globally is also low, with structural factors
or other illegal activities to finance drug use, exposing (including policy and legal bans—eg, on use of agonist
users to violence and sexual risk; by discouraging opioids) a major impediment to improved coverage.142,143
treatment-seeking (for fear of negative consequences);141,142
by reducing access to interventions that reduce risk, Discussion
through creating legal obstacles to, or policy limits on, A substantial proportion of young adults in developed
service provision;140–143 and by increasing the risks of countries have used an illicit drug at some time in
imprisonment and its attendant health risks.140,142,144 their lives. Worldwide, around one in 20 people aged
Some countries have been successful in ensuring that 15–64 years might have done so in the past year.
services are accessible to, and accessed by, people who Cannabis is the drug most often used and the most
use drugs. Some have achieved high coverage of HIV widely available because of widespread domestic
prevention services for IDUs,56 and others provide good production in many countries. A minority of individuals
access to drug treatment and other services for who use illicit drugs become dependent on or inject
dependent drug users.56,145 However, globally, a very low them. The prevalence of dependence on these drugs has
proportion of the population who inject drugs has access rarely been directly assessed, but it seems to be more
to interventions to reduce HIV infection.56 Treatment common in high-income countries.

www.thelancet.com Vol 379 January 7, 2012 65

 Series

Many important questions cannot be answered. How drug-induced psychosis, violence, and HIV and hepatitis C
many people who use a drug will go on to become infections when injected with non-sterile equipment.
dependent? How long do people use drugs for? And for Much of the burden attributable to injecting drug
how long do they remain dependent on them? Does use can be prevented or reduced by needle and syringe
the risk of dependence vary over time and between programmes, opioid substitution treatment, and anti-
individuals? Do users move in and out of harmful and retroviral therapy.143 Burden is also probably worsened
dependent use, and if so, when and why? How large is by the criminal status and stigmatisation of injecting
the risk for adverse outcomes, including early death? Do drug use, high rates of imprisonment, and little political
these risks vary between countries, and demographic interest in funding interventions to reduce these
subgroups? How much does criminalisation of drugs risks.140–143
reduce their prevalence of use? How much of the harm The major adverse health effect of cannabis use is
related to illicit drugs derives from their illegal status? dependence, which in young adults is correlated with,
Until we have better answers to these questions, and probably a contributory cause of, psychosis and other
statements about the exact magnitude of the health, mental disorders. The health-related harms of cannabis
social, and financial burden of illicit drug use cannot be use have never been quantified on a global scale, but they
made with accuracy. This makes the formulation of are qualitatively different from the other major drug
evidence-informed drug policies and programmes types, in that cannabis contributes more to morbidity
difficult.1 Without knowing the size of the population at than mortality because it cannot be injected and does not
risk, identification of appropriate interventions and the cause fatal overdose.
size of target populations is difficult. A major unintended consequence of the criminalisation
On the basis of available evidence, most of the disease of drug use is the inability to collect high quality data for
burden attributable to illicit drugs is concentrated in patterns of use and harms. High-income countries often
problem or dependent drug users, especially people use general population and school surveys to monitor
who inject drugs. Existing estimates underestimate the trends in drug use, but these probably underestimate the
contribution of illicit drugs to the global burden of use of more highly stigmatised drugs that account for
disease because they do not include all adverse outcomes most of the harms (panel 3). Routinely collected mortality
of illicit drug use. Even so, these estimates suggest that and morbidity data can be used to monitor trends in
drug dependence, HIV infection, and drug overdose are those that are more directly related to drug use (such as
important causes of drug-related disease burden. Causes overdose deaths and numbers seeking treatment).
of burden might also be changing in high-income However, even in high-income countries with good
countries—as mortality from HIV decreases, the burden research infrastructure, illicit drug use might not always
attributable to chronic hepatitis C infection in IDUs be recognised (or recorded) as a contributory cause of
might increase. As yet, we have no estimates of the death or hospitalisation.
global burden attributable to cannabis use. Data for patterns of use and harm are very scarce for
In high-income countries, the contribution of illicit synthetic drugs that have emerged within the past two
drugs to burden of disease is less than that of tobacco, decades. Policies towards newly emerging drugs (eg,
but may be similar to alcohol (if moderate alcohol use mephedrone) are often made in response to media stories
truly has protective effects on cardiovascular mortality). and in ignorance of the scale of their use and the
This outcome is the product of the following: the lower problems arising from it.146 Decisions are often made on
prevalence of problem illicit drug use than of alcohol an implicit precautionary principle: when in doubt,
and tobacco use (reducing the number of individuals prohibit the use of a new substance. Once use of a drug
exposed), the occurrence of adverse outcomes of illicit has been prohibited, the decision is rarely revisited.
drug use at much younger ages than those for alcohol Intelligent policy responses to drug problems need
and tobacco (increasing the years of life lost or lived much better data for the prevalence of different types of
with disability due to illicit drug use), and the illicit drug use and the harms that their use causes to
consequences of injection of opioids and stimulants users and society, especially in high-income countries
(with injecting-related blood-borne viral infections being with substantial rates of illicit drug use. It is equally
major contributors to burden that are not experienced important in developing countries that are close to
by cannabis, alcohol, or tobacco users). Estimates of source countries, or whose citizens have ready access to
disease burden are much less certain in low-income and precursor chemicals to illicitly manufacture synthetic
middle-income countries. drugs. A need exists for the global community, including
In many high-income countries, illicit opioid use seems UN agencies, to address the technical and political
to be the most hazardous type of illicit drug use in terms challenges that many countries face in developing this
of mortality. The risks of amphetamine and cocaine use capacity. The second paper in this Series133 examines
have not been as well studied as those of opioids, but evidence for the effectiveness of a range of interventions
are probably less hazardous than opioids in terms of that aim to reduce the extent of drug use and harms
fatal overdose. They nonetheless cause dependence, related to such use.

66 www.thelancet.com Vol 379 January 7, 2012

 Series

Contributors 14 Nelson P, Mathers B, Cowie B, et al. The epidemiology of viral

LD and WH both contributed to the overall structure of and idea for the hepatitis among people who inject drugs: results of global
paper. LD retrieved data presented in the paper and led the writing with systematic reviews. Lancet 2011; 378: 571–83.
substantial input from WH. Both authors approved the final draft. 15 Stroup DF, Berlin JA, Morton SC, et al. Meta-analysis of
observational studies in epidemiology: a proposal for reporting.
Conflicts of interest JAMA 2000; 283: 2008–12.
WH and LD have been consultants to WHO for the purposes of 16 Calabria B, Phillips B, Singleton J, et al. Searching the grey
providing scientific advice on illicit drug epidemiology (LD, WH), literature to access information on drug and alcohol research:
burden of disease (LD, WH), and reviewing evidence for the comparative NDARC technical report (No 293). Sydney: National Drug and
harms of cannabis, tobacco, and alcohol (WH). LD has consulted to the Alcohol Research Centre, University of NSW; 2008.
UNODC, providing advice on presentation, analysis, and interpretation 17 Ali H, Calabria B, Phillips B, et al. Searching the grey literature
of data for illicit drugs. LD has also consulted to UNAIDS on injecting to access information on drug, alcohol and HIV/AIDS research:
drug use, HIV, and effective responses, including reviewing data for an update: NDARC technical report (No 314). Sydney: National
monitoring and assessment including data from the UN General Drug and Alcohol Research Centre, University of NSW; 2010.
Assembly 26th Special Session. 18 Vandenbroucke JP, von Elm E, Altman DG, et al. Strengthinging
the reporting of observational studies in epidemiology
Acknowledgments (STROBE): explanation and elaboration. PLoS Med 2007;
We thank Shane Darke and Stuart Kinner for useful advice on earlier 4: 1628–54.
drafts of this paper. Thanks also to the many researchers who have 19 Hall W, Degenhardt L. The adverse health and psychological
worked tirelessly on the systematic reviews that are mentioned in this effects of non-medical cannabis use. Lancet 2009; 374: 1383–91.
paper, particularly Bianca Calabria, Chiara Bucello, Paul Nelson, 20 Arseneault L, Cannon M, Witton J, Murray RM. Causal association
Bradley Mathers, Anna Roberts, and Jessica Singleton. We also wish to between cannabis and psychosis: examination of the evidence.
thank Jen McLaren, who recently passed away, for her work with us on Br J Psychiatry 2004; 184: 110–17.
the Global Burden of Disease project. Her contribution and collaboration 21 Degenhardt L, Hall W, Lynskey M, et al. Should we make burden
was greatly appreciated. of disease estimates for cannabis use as a risk factor for psychosis?
PLoS Med 2009; 6: e1000133.
References
1 Babor TF, Caulkins J, Edwards G, et al. Drug policy and the public 22 Hagan H, Des Jarlais D, Stern R, et al. HCV synthesis project:
good. Oxford: Oxford University Press, 2010. preliminary analyses of HCV prevalence in relation to age and
duration of injection. Int J Drug Policy 2007; 18: 341–51.
2 McAllister WB. Drug Diplomacy in the twentieth century.
London: Routledge, 2000. 23 Moore TH, Zammit S, Lingford-Hughes A, et al. Cannabis use and
risk of psychotic or affective mental health outcomes: a systematic
3 Degenhardt L, Bucello C, Calabria B, et al. What data are available
review. Lancet 2007; 370: 319–28.
on the extent of illicit drug use and dependence globally? Results of
four systematic reviews. Drug Alcohol Depend 2011; 117: 85–101. 24 Degenhardt L, Chiu W-T, Sampson N, et al. Toward a global view
of alcohol, tobacco, cannabis and cocaine use: findings from the
4 Bucello C, Degenhardt L, Calabria B, et al. What do we know
WHO World Mental Health Surveys. PLoS Med 2008; 5: e141.
about the extent of cocaine use and dependence? Results of
a global systematic review—NDARC technical report No 308. 25 Degenhardt L, Dierker L, Chiu W, et al. Evaluating the drug
Sydney: National Drug and Alcohol Research Centre, University use “gateway” theory using cross-national data: consistency
of NSW, 2010. and associations of the order of initiation of drug use among
participants in the WHO World Mental Health Surveys.
5 Calabria B, Degenhardt L, Nelson P, et al. What do we know
Drug Alcohol Depend 2010; 108: 84–97.
about the extent of cannabis use and dependence? Results of
a global systematic review—NDARC technical report No 307. 26 Reuter P, Trautmann F. A report on global illicit drugs markets
Sydney: National Drug and Alcohol Research Centre, 1998–2007. Utrecht: Trimbos Intstitute, 2009.
University of NSW, 2010. 27 UN Office on Drugs and Crime. World drug report 2009.
6 Degenhardt L, Calabria B, Nelson P, et al. What do we know about Vienna: United Nations, 2009.
the extent of meth/amphetamine use and dependence? Results of 28 UN Office on Drugs and Crime. World drug report 2011. http://
a global systematic review—NDARC technical report No 310. www.humansecuritygateway.com/showRecord.php?RecordId=35492
Sydney: National Drug and Alcohol Research Centre, University (accessed on July 5, 2011).
of NSW, 2010. 29 Degenhardt L, Hall W, Lynskey M, Warner-Smith M. Illicit
7 Nelson P, McLaren J, Degenhardt L, et al. What do we know about drug use. In: Ezzati M, Lopez AD, Rodgers A, Murray R, eds.
the extent of heroin and other opioid use and dependence? Results Comparative quantification of health risks: global and regional
of a global systematic review—NDARC technical report No. 309. burden of disease attributable to selected major risk factors,
Sydney: National Drug and Alcohol Research Centre, 2nd edn. Geneva: World Health Organization, 2004: 1109–76.
University of NSW, 2010. 30 Ezzati M, Lopez AD, Rodgers A, Vander Hoorn S, Murray CJL,
8 Mathers B, Degenhardt L, Phillips B, et al. Global epidemiology Comparative Risk Assessment Collaborating Group. Selected major
of injecting drug use and HIV among people who inject drugs: risk factors and global and regional burden of disease. Lancet 2002;
a systematic review. Lancet 2008; 372: 1733–45. 360: 1347–60.
9 Calabria B, Degenhardt L, Briegleb C, et al. Systematic reviews of 31 WHO. The global burden of disease: 2004 update. Geneva: World
prospective studies investigating “remission” from amphetamine, Health Organization, 2008.
cannabis, cocaine and opioid dependence. Addict Behav 2010; 32 Rehm J, Room R, Monteiro M, et al. Alcohol use. In: Ezzati M,
35: 741–49. Lopez AD, Rodgers A, Murray R, eds. Comparative quantification of
10 Calabria B, Degenhardt L, Hall W, Lynskey M. Does cannabis use health risks: global and regional burden of disease attributable
increase the risk of death? Systematic review of epidemiological to selected major risk factors, 2nd edn. Geneva: World Health
evidence on adverse effects of cannabis use. Drug Alcohol Rev 2010; Organization, 2004: 959–1107.
29: 318–30. 33 Ezzati M, Lopez A. Smoking and oral tobacco use. In: Ezzati M,
11 Degenhardt L, Bucello C, Mathers B, Ali H, Hickman M, Lopez AD, Rodgers A, Murray R, eds. Comparative quantification of
McLaren J. Mortality among dependent users of heroin health risks: global and regional burden of disease attributable
and other opioids: a systematic review and meta-analysis to selected major risk factors, 2nd edn. Geneva: World Health
of cohort studies. Addiction 2011; 106: 32–51. Organization, 2004: 983–57.
12 Degenhardt L, Singleton J, Calabria B, et al. Mortality among 34 Murray CJ, Lopez AD. The global burden of disease:
cocaine users: a systematic review of cohort studies. a comprehensive assessment of mortality and disability from
Drug Alcohol Depend 2011; 113: 88–95. diseases, injuries and risk factors in 1990 and projected to 2020.
13 Singleton J, Degenhardt L, Hall W, Zabransky T. Mortality among Cambridge, MA: Harvard University Press, 1996.
people who use amphetamines: a systematic review of cohort 35 Darke S. The toxicology of homicide offenders and victims:
studies. Drug Alcohol Depend 2009; 105: 1–8. a review. Drug Alcohol Rev 2010; 29: 202–15.

www.thelancet.com Vol 379 January 7, 2012 67

 Series

36 Darke S, Degenhardt L, Mattick RP. Mortality amongst illicit drug 60 UNODC. World drug report 2008. Vienna: United Nations Office
users. Cambridge: Cambridge University Press, 2006. on Drugs and Crime, 2008.
37 Darke S, Kaye S, McKetin R, Duflou J. Major physical 61 Kandel D, Yamaguchi K, Klein L. Testing the gateway hypothesis.
and psychological harms of methamphetamine use. Addiction 2006; 101: 470–72.
Drug Alcohol Rev 2008; 27: 253–62. 62 Hall W, Lynskey M. Is cannabis a gateway drug? Testing hypotheses
38 Hall W, Degenhardt L, Lynskey M. The health and psychological about the relationship between cannabis use and the use of other
consequences of cannabis use. Canberra: Australian Publishing illicit drugs. Drug Alcohol Rev 2005; 24: 39–48.
Service, 2001. 63 Bachman JG, Wadsworth KN, O’Malley P, Johnston L,
39 Karch SB. Karch’s pathology of drug abuse, 3rd edn. Schulenberg J. Smoking, drinking and drug use in young
Boca Raton, Florida: CRC Press, 2002. adulthood: the impacts of new freedoms and responsibilities.
40 Kaye S, McKetin R, Duflou J, Darke S. Methamphetamine and Mahwah, NJ: Lawrence Erlbaum, 1997.
cardiovascular pathology: a review of the evidence. Addiction 2007; 64 Darke S. The life of the heroin user: typical beginnings,
102: 1204–11. trajectories and outcomes. Cambridge: Cambridge University
41 Kelly E, Darke S, Ross J. A review of drug use and driving: Press, 2011.
epidemiology, impairment, risk factors and risk perceptions. 65 Hall W, Degenhardt L. Prevalence and correlates of cannabis use
Drug Alcohol Rev 2004; 23: 319–44. in developed and developing countries (invited review).
42 Boot BP, McGregor LS, Hall W. MDMA (Ecstasy) neurotoxicity: Curr Opin Psychiatry 2007; 20: 393–97.
assessing and communicating the risks. Lancet 2000; 66 Maddahian E, Newcomb MD, Bentler PM. Adolescent drug use and
355: 1818–21. intention to use drugs: concurrent and longitudinal analyses of four
43 Degenhardt L, Bruno R, Topp L. Is ecstasy a drug of dependence? ethnic groups. Addict Behav 1988; 13: 191–95.
Drug Alcohol Depend 2010; 107: 1–10. 67 Lascala E, Friesthler B, Gruenwald PJ. Population ecologies of drug
44 Degenhardt L, Hall W. The health and psychological consequences use, drinking and related problems. In: Stockwell T, Gruenwald P,
of “ecstasy” (MDMA) use—NDARC Monograph No 62. of NSW. Toumbourou J, Loxley W, eds. Preventing harmful substance use:
Sydney: National Drug and Alcohol Research Centre, University of the evidence base for policy and practice. Chichester: John Wiley
NSW; 2010. and Sons, 2005.
45 Larance B, Degenhardt L, Dillon P, Copeland J. Rapid assessment 68 Daniel JZ, Hickman M, Macleod J, et al. Is socioeconomic status
of performance and image enhancing drugs (PIEDs) in New South in early life associated with drug use? A systematic review of the
Wales: feasibility study 2005. Sydney: National Drug and Alcohol evidence. Drug Alcohol Rev 2009; 28: 142–53.
Research Centre, University of New South Wales, 2005. 69 Cohen DA, Richardson J, LaBree L. Parenting behaviors and the
46 Charlson F, Degenhardt L, McLaren J, Hall W, Lynskey M. onset of smoking and alcohol use: a longitudinal study. Pediatrics
Benzodiazepine related mortality. Pharmacoepidemiol Drug Saf 2009; 1994; 94: 368–75.
18: 93–103. 70 Flewelling RL, Bauman KE. Family structure as a predictor of initial
47 Compton WM, Volkow ND. Major increases in opioid analgesic substance use and sexual intercourse during early adolescence.
abuse in the United States: concerns and strategies. J Marriage Fam 1990; 52: 171–81.
Drug Alcohol Depend 2006; 81: 103–07. 71 Fergusson DM, Horwood LJ, Lynskey MT. Parental separation,
48 Turk D, Swanson K, Gatchel R. Predicting opioid misuse by chronic adolescent psychopathology, and problem behaviors.
pain patients: a systematic review and literature synthesis. J Am Acad Child Adolesc Psychiatry 1994; 33: 1122–33.
Clin J Pain 2008; 24: 497. 72 Needle RH, Su S, Doherty WJ. Divorce, remarriage, and adolescent
49 Martell B, O’Connor P, Kerns R, et al. Systematic review: drug involvement: a longitudinal study. J Marriage Fam 1990;
opioid treatment for chronic back pain: prevalence, efficacy, 52: 157–69.
and association with addiction. Ann intern med 2007; 146: 116. 73 Lynskey MT, Fergusson DM, Horwood LJ. The effect of parental
50 Hartnoll R. Cross-validating at local level. Hartnoll R, ed. alcohol problems on rates of adolescent psychiatric disorders.
Estimating the prevalence of problem drug use in Europe. Addiction 1994; 89: 1277–86.
Luxembourg: Office for Official Publications of the European 74 Cannon DS, Clark LA, Leeka JK, Keefe CK. A reanalysis of the
Communities, 1997: 247–61. tridimensional personality questionnaire (TPQ) and its relation
51 Degenhardt L, Hallam C, Bewley-Taylor D. Comparing the drug to Cloninger’s Type 2 alcoholism. Psychol Assess 1993; 5: 62–66.
situation across countries: problems, pitfalls and possibilities. 75 Lipkus IM, Barefoot JC, Williams RB, Siegler IC. Personality
http://www.idpc.net/sites/default/files/library/Beckley%20 measures as predictors of smoking initiation and cessation in the
Briefing%2019.pdf (accessed April 23, 2010). UNC Alumni Heart Study. Health Psychol 1994; 13: 149–55.
52 Hickman M, Taylor C, Chatterjee A, et al. Estimating the prevalence 76 Lynskey MT, Fergusson DM. Childhood conduct problems
of problematic drug use: a review of methods and their application. and attention deficit behaviors and adolescent alcohol, tobacco
UN Bulletin on Narcotics 2002; LIV (1 and 2): 15–32. and illicit drug use. J Abnorm Child Psychol 1995; 23: 281–302.
53 Hall W, Ross J, Lynskey M, Law M, Degenhardt L. How many 77 Townsend L, Flisher A, King G. A systematic review of the
dependent heroin users are there in Australia? Med J Aust 2000; relationship between high school dropout and substance use.
173: 528–31. Clin Child Fam Psychol Rev 2007; 10: 295–317.
54 WHO. The ICD-10 classification of mental and behavioural 78 Lynskey M, Hall W. The effects of adolescent cannabis use on
disorders—diagnostic criteria for research. Geneva: educational attainment: a review. Addiction 2000; 95: 1621–30.
World Health Organization, 1993. 79 Fergusson DM, Boden JM, Horwood LJ. The developmental
55 APA. Diagnostic and statistical manual of mental disorders: antecedents of illicit drug use: evidence from a 25 year longitudinal
DSM-IV-TR, 4th edn. Washington, DC: American Psychiatric study. Drug Alcohol Depend 2008; 96: 167–77.
Association, 2000. 80 Kandel DB, Andrews K. Processes of adolescent socialisation by
56 Mathers B, Degenhardt L, Ali H, et al. HIV prevention, treatment parents and peers. Int J Addict 1987; 22: 319–42.
and care for people who inject drugs: a systematic review of 81 Hawkins J, Catalano R, Miller J. Risk and protective factors
global, regional, and country level coverage. Lancet 2010; for alcohol and other drug problems in adolescence and early
375: 1014–28. adulthood: implications for substance abuse prevention.
57 Kessler R, Üstün T. The WHO World Mental Health Surveys: global Psychol Bull 1992; 112: 64–105.
perspectives on the epidemiology of mental disorders. Cambridge: 82 Toumbourou J, Stockwell T, Neighbors C, Marlatt G, Sturge J,
Cambridge University Press, 2008. Rehm J. Interventions to reduce harm associated with adolescent
58 Ball AL, Rana S, Dehne KL. HIV prevention among injecting substance use. Lancet 2007; 369: 1391–401.
drug users: responses in developing and transitional countries. 83 Beautrais AL, Joyce PR, Mulder RT. Cannabis abuse
Public Health Rep 1998; 113: 170–81. and serious suicide attempts. Addiction 1999; 94: 1155–64.
59 UNODC. Amphetamines and ecstasy: 2008 Global ATS 84 Satel S, Southwick S, Gawin F. Clinical features of
assessment. Vienna: United Nations Office on Drugs cocaine-induced paranoia. Am J Psychiatry 1991; 148: 495.
and Crime, 2008.

68 www.thelancet.com Vol 379 January 7, 2012

 Series

85 Randall D, Degenhardt L, Vajdic C, et al. Increasing cancer 110 Maloney E, Degenhardt L, Darke S, Mattick RP, Nelson E. Suicidal
mortality among opioid dependent persons in Australia—a new behaviour and associated risk factors among opioid-dependent
public health challenge for a disadvantaged population. persons: a case-control study. Addiction 2007; 102: 1933–41.
Aust N Z J Public Health 2011; 35: 220–25. 111 Marincovich B, Castilla J, Del Romero J, et al. Absence of hepatitis
86 Marshall BDL, Werb D. Health outcomes associated with C virus transmission in a prospective cohort of heterosexual
methamphetamine use among young people: a systematic review. serodiscordant couples. BMJ 2003; 79: 160.
Addiction 2010; 105: 991–1002. 112 Pedraza Ma-A, del Romero J, Roldán F, et al. Heterosexual
87 Hall W, Degenhardt L, Teesson M. Understanding comorbidity transmission of HIV-1 is associated with high plasma viral
between substance use, anxiety and affective disorders: broadening load levels and a positive viral isolation in the infected partner.
the research base. Addict Behav 2009; 34: 526–30. J Acquir Immune Defic Syndr 1999; 21: 120–25.
88 Baggaley RF, Boily MC, White RG, Alary M. Risk of HIV-1 113 Boily M, Baggaley R, Wang L, et al. Heterosexual risk of HIV-1
transmission for parenteral exposure and blood transfusion: infection per sexual act: systematic review and meta-analysis of
a systematic review and meta-analysis. AIDS 2006; 20: 805–12. observational studies. Lancet Infect Dis 2009; 9: 118–29.
89 Perlman D, Salomon N, Perkins M, Yancovitz S, Paone D, 114 Vitinghoff E, Douglas J, Judon F, McKiman D, MacQueen K,
Des Jarlais D. Tuberculosis in drug users. Clin Infect Dis Buchinder S. Per-contact risk of human immunodificiency virus
1995; 21: 1253–64. tramnsmision between male sexual partners. Am J Epidemiol 1999;
90 Hill S. The environment and disease: association or causation? 150: 306.
Proc R Soc Med 1965; 58: 295–300. 115 Donoghoe M, Wodak A. Health and social consequences of
91 English D, Holman C, Milne E, et al. The quantification of drug injecting drug use. In: Stimson G, Des Jarlais D, Ball A, eds.
caused morbidity and mortality in Australia, 1995. Canberra: Drug injecting and HIV infection: global dimensions and local
Commonwealth Department of Human Services and Health, 1995. responses. London: UCL Press, 1998: 42–57.
92 Darke S, Zador D. Fatal heroin “overdose”: a review. Addiction 116 MacDonald M, Crofts N, Kaldor J. Transmission of hepatitis C virus:
1996; 91: 1765–72. rates, routes, and cofactors. Epidemiol Rev 1996; 18: 137–48.
93 Warner-Smith M, Darke S, Lynskey M, Hall W. Heroin overdose: 117 Hallinan R, Byrne A, Amin J, Dore GJ. Hepatitis C virus
causes and consequences. Addiction 2001; 96: 1113–25. prevalence and outcomes among injecting drug users on opioid
94 Hall W, Degenhardt L. The adverse health and psychological replacement therapy. J Gastroenterol Hepatol 2005; 20: 1082–86.
effects of non-medical cannabis use. Lancet 2009; 374: 1383–91. 118 Ministerial Advisory Committee on AIDS Sexual Health
95 Hall WD, Pacula RL. Cannabis use and dependence: public health and Hepatitis: Hepatitis C Sub-Committee. Hepatitis C virus
and public policy. Cambridge, UK: Cambridge University Press, projections working group: estimates and projections of the
2003. hepatitis C virus epidemic in Australia 2006. Sydney: National
Centre in HIV Epidemiology and Clinical Research, University
96 Darke S, Kaye S, Duflou J. Comparative cardiac pathology among
of New South Wales, 2006.
deaths due to cocaine toxicity, opioid toxicity and non-drug-related
causes. Addiction 2006; 101: 1771–77. 119 Coutinho RA. HIV and hepatitis C among injecting drug users:
success in preventing HIV had not been mirrored for hepatitis C.
97 Borgwardt S, Allen P, Bhattacharyya S, et al. Neural basis of
BMJ 1998; 317: 424–25.
delta-9-tetrahydrocannabinol and cannabidiol: effects during
response intervention. Biol Psychiatry 2008; 64: 966–73. 120 Limburg W. Natural history, treatment and prevention of hepatitis C
in injecting drug users: an overview. In: Jager J, Limburg W,
98 Laumon B, Gadegbeku B, Martin JL, Biecheler MB, the SAM
Kretzschmar M, Postma M, Wiessing L, eds. Hepatitis C and
Group. Cannabis intoxication and fatal road crashes in France:
injecting drug use: impact, costs and policy options—EMCDDA
population based case-control study. BMJ 2005; 331: 1371.
monograph. Lisbon: European Monitoring Centre for Drugs and
99 Glantz M, Anthony J, Berglund P, et al. Mental disorders as Drug Addiction, 2004: 21–38.
risk factors for later substance dependence: estimates of optimal
121 Dore GJ, Freeman AJ, Law M, Kaldor JM. Is severe liver disease
prevention and treatment benefits. Psychol Med 2008; 39: 1365–77.
a common outcome for people with chronic hepatitis C?
100 Hall W, Teesson M, Lynskey M, Degenhardt L. The 12-month J Gastroenterol Hepatol 2002; 17: 423–30.
prevalence of substance use and ICD-10 substance use disorders
122 Foster G. Hepatitis C and quality of life. In: Jager J, Limburg W,
in Australian adults: findings from the national survey of mental
Kretzschmar M, Postma M, Wiessing L, eds. Hepatitis C and
health and well-being. Addiction 1999; 94: 1541–50.
injecting drug use: impact, costs and policy options—EMCDDA
101 Anthony JC, Warner L, Kessler R. Comparative epidemiology monograph. Lisbon: European Monitoring Centre for Drugs and
of dependence on tobacco, alcohol, controlled substances, and Drug Addiction, 2004: 79–86.
inhalants: basic findings from the national comorbidity survey.
123 World Bank. World development report 1993: investing in health.
Exp Clin Psychopharmacol 1994; 2: 244–68.
New York: Oxford University Press, 1993.
102 Volkow N, Fowler J, Wang G-J, Swanson J. Dopamine in drug abuse
124 Lopez AD, Mathers C, Ezzati M, Jamison D, Murray C. Global
and addiction: results from imaging studies and treatment
and regional burden of disease and risk factors, 2001: systematic
implications. Mol Psychiatry 2004; 9: 557–69.
analysis of population health data. Lancet 2006; 367: 1747–57.
103 McKetin R, Kelly E, McLaren J. The relationship between crystalline
125 Donoghoe M. Illicit drugs. In: Murray CJL, Lopez AD, eds.
methamphetamine use and methamphetamine dependence.
Quantifying global health risks: the burden of disease attributable to
Drug Alcohol Depend 2006; 85: 198–204.
selected risk factors. Cambridge, MA: Harvard University Press, 1996.
104 Macleod J, Oakes R, Copello A, et al. Psychological and social
126 Degenhardt L, Hall W, Warner-Smith M. Using cohort studies
sequelae of cannabis and other illicit drug use by young people:
to estimate mortality among injecting drug users that is not
a systematic review of longitudinal, general population studies.
attributable to AIDS. Sex Transm Infect 2006; 82: 56–63.
Lancet 2004; 363: 1579–88.
127 Begg S, Vos T, Barker B, Stevenson C, Stanley L, Lopez A.
105 Degenhardt L, Hall WD. Is cannabis a contributory cause
The burden of disease and injury in Australia 2003. Canberra:
of psychosis? Can J Psychiatry 2006; 51: 556–65.
Australian Institute of Health and Welfare, 2007.
106 Moore T, Zammit S, Lingford-Hughes A, et al. Cannabis use and
128 Degenhardt L, Whiteford H, Hall W, Vos T. Estimating the burden
risk of psychotic or affective mental health outcomes: a systematic
of disease attributable to illicit drug use and mental disorders: what
review. Lancet 2007; 370: 319–28.
is “global burden of disease 2005” and why does it matter? (Invited
107 Large M, Sharma S, Compton MT, Slade T, Nielssen O. Cannabis review). Addiction 2009; 104: 1466–71.
use and earlier onset of psychosis: a systematic meta-analysis.
129 Murray C, Lopez A, Black R, et al. Global burden of disease 2005:
Arch Gen Psychiatry 2011; 68: 555–61.
call for collaborators. Lancet 2007; 370: 109–10.
108 Gossop M, Griffiths P, Powis B, Williamson S, Strang J.
130 WHO. Global status report on alcohol. Geneva: World Health
Frequency of non-fatal heroin overdose: survey of heroin users
Organization, 2004.
recruited in non-clinical settings. BMJ 1996; 313: 402.
131 Hall W, Degenhardt L, Sindicich N. The contribution of illicit drug
109 Rossow I, Lauritzen G. Balancing on the edge of death:
use to the burden of disease. In: Heggenhougen H, ed. International
suicide attempts and life-threatening overdoses among
encyclopedia of public health. London: Elsevier Press, 2008: 523–30.
drug addicts. Addiction 1999; 94: 209–19.

www.thelancet.com Vol 379 January 7, 2012 69

 Series

132 Moore MH. Supply reduction and drug law enforcement. 140 Jürgens R, Csete J, Amon JJ, Baral S, Beyrer C. People who
Crime and Justice 1990; 13: 109–57. use drugs, HIV, and human rights. Lancet 2010; 376: 475–85.
133 Strang J, Babor T, Caulkins J, Fischer B, Foxcroft D, Humphreys K. 141 Strathdee SA, Hallett TB, Bobrova N, et al. HIV and risk
Drug policy and the public good: evidence for effective environment for injecting drug users: the past, present,
interventions. Lancet 2012; 379: 71–83. and future. Lancet 2010; 376: 268–84.
134 MacCoun R, Kilmer B, Reuter P. Research on drugs-crime 142 Wolfe D, Carrieri MP, Shepard D. Treatment and care for
linkages: the next generation. Toward a drugs and crime research injecting drug users with HIV infection: a review of barriers
agenda for the 21st century. Washington, DC: National Institute of and ways forward. Lancet 2010; 376: 355–66.
Justice Special Report, 2003; 65–95. 143 Degenhardt L, Mathers B, Vickerman P, Rhodes T, Latkin C,
135 WHO. The global burden of disease: 2004 update. Geneva: World Hickman M. Prevention of HIV infection for people who inject
Health Organization, 2008. drugs: why individual, structural, and combination approaches
136 Kuhns J, Wilson D, Maguire E, Ainsworth S, Clodfelter T. are needed. Lancet 2010; 376: 285–301.
A meta-analysis of marijuana, cocaine and opiate toxicology study 144 Merrall E, Kariminia A, Binswanger I, et al. Meta analysis of
findings among homicide victims. Addiction 2009; 104: 1122–31. drug related deaths soon after release from prison. Addiction
137 Collins D, Lapsley H. The costs of tobacco, alcohol and illicit 2010; 105: 1545–54.
drug use to Australian society in 2004/05. National Drug Strategy 145 WHO. ATLAS on substance use (2010): resources for the
Monograph. Canberra: Commonwealth Department of Health and preventions and treatment of substance use disorders. Geneva:
Ageing, 2007. World Health Organization, 2010.
138 Rehm J, Gnam W, Popova S, et al. The costs of alcohol, illegal 146 The Lancet. A collapse in integrity of scientific advice in the UK.
drugs, and tobacco in Canada, 2002. J Stud Alcohol Drugs 2007; Lancet 2010; 375: 1319.
68: 886–95.
139 Reuter P. Disorganized crime: the economics of the visible hand.
Cambridge, MA: MIT Press, 1983.

70 www.thelancet.com Vol 379 January 7, 2012