Drug-related deaths: what now needs to be done

We have quantified drug-specific mortality as deaths per 100,000 person-years/past-year users but for four only of the 20 drugs whose 16 dimensions of harm were rated by Professor Nutt and colleagues in the Lancet this month. The four are: heroin, methadone, cocaine and ecstasy.

Somewhat surprisingly, the recent Lancet paper did not set out the evidence-base on drug-specific mortality for each of its 20 drugs (including alcohol and tobacco) that was the backdrop for its raters’ judgments on drug-specific mortality. Their approach did not, of course, allow for drug-related mortality being importantly different between male versus female users; or according to user-age. But is it!
As we have seen, both demographic factors matter when appraising drug-specific mortality. For ease of access, a summary of the coinclusions reached in this series of articles, and the drug-related deaths per 100,000 users are appended at the end of this article.

What now needs urgent investigation, and how should we go about it?

First, the Office for National Statistics (ONS) should consider publishing regularly, eg every 2 years, the demography of drug-specific DRDs, as we have defined them; together with an annual account of the demography of DRDs with any mention of specific drugs such as: a) heroin/morphine, b) benzodiazepines, c) cocaine, d) methadone and e) ecstasy. In doing so, they’d set a new reporting standard for other European statistical offices.

Statistics on drugs-related deaths are only as good as the toxicology on which they rely, and so we’d like an agreed UK protocol for the conduct of forensic autopsies on drugs-related deaths.
Third, it is unacceptable that, in England and Wales (unlike in Scotland), the registration of drugs-related deaths which are referred to coroners (and almost all are) can be delayed by many months until the coroner has determined the cause of death. Until determination, the registration system in E&W can be oblivious even to fact of death. In Scotland, however, there is a legal requirement for all deaths to be registered within 8 days of their ascertainment. And so should there be in England! No wonder we can’t count migrant populations when we don’t even count the dead properly. Parliament should act quickly to plug this shameful gap in the registration system.
It matters because, when there’s a new epidemic – whether pandemic influenza deaths in children or deaths in 16-24 year olds from some ‘legal high’ or from a contaminant – England’s registration system is incapable of telling us even how many deaths in the relevant age-range have been referred to coroners’ inquiries until, often many months later, those inquiries reach their verdict on cause of death. Acute epidemics, however, usually have doubling times that are much faster than the doubling time for coroners’ verdicts!
Fourth, astonishingly, in England and Wales, there is no confidential look-back into whether those who die of drug-related deaths have a history of drug treatment or have been in prison. Scotland does this for ex-prisoners because there is compelling international evidence about the very high risk of opiate-related death in the first fortnight (elevated still in second fortnight) after release from prison, notably those with a history of heroin-injection.
Looking-back at previous drug treatments is now needed, for example, to unlock the mystery of (alcohol-aided) benzodiazepine-specific deaths in those aged over 35. Moreover, earlier this month in the BMJ, Cornish et al. reported that an all-causes mortality risk as high as 4,500 per 100,000 person-years in the 4 weeks after expiry of a client’s last opiate-substitution script (95 per cent CI: 3,145 to 5,920), a death-rate that is nine times the rate used to define major combat in Afghanistan.The case for looking-back to the expiry date of the deceased’s last opiate-substitution prescription is compelling. Of course, until England’s death registration problems are sorted and Scotland properly records the end of a treatment episode, this type of confidential inquiry is severely hampered.
Fifth, although we recognise that some of the deaths reported by Cornish et al. (2010) will have been the reason that the deceased did not return for his/her next script – rather than  that all deaths were caused by dropping out from, or being discharged from opiate substitution therapy – DRD rates of around 500 per 100,000 opiate-client/user years {as reported earlier this year by one of us (JRR) for the Edinburgh/Muirhouse Addiction Cohort}mean that National Drug Treatment Agency and regional treatment providers should surely seek clients’ permission to perform periodic database linkage of client lists against the deaths register so that providers are not ignorant about current and ex-clients’ mortality risks. 
Sixth, we’d like the British Crime Survey (BCS) to publish pooled estimates with standard errors, for example: based on 2-years’ or 4-years’ worth of British Crime Survey, we’d like BCS to publish the per-annum estimated number of past-year users of i) cocaine and ii) ecstasy and to do so by sex and age-group (16-24 years, 25-34 years, 35-59 years) simultaneously. Pooled information on whether the intensity of cocaine-use by male versus female past-year users or according to user-age, if available, might shed light on some of the sex and age effects on cocaine-related mortality that we have revealed.
Finally, why did the apparent lethality of ecstasy and cocaine for older users increase this century?

Summary of conclusions reached
At most 40% of older females’ drugs-related deaths were opiate-related: In the decade 1998-2007, only 667/1,818 female DRDs aged 35+ years had either heroin or methadone present (95% CI: 34% to 39%). What killed the other 60%?
Benzodiazepine-specific deaths occurred predominantly at 35+ years, with alcohol co-present at over a third: Of 1,058 benzodiazepine-specific DRDs, three-quarters (808) were aged 35+ years, and 307 of them had alcohol co-present (95% CI: 34% to 41%).
Lowest male: female ratio was 1.2:1 for benzodiazepine-specific deaths: Why? We urgently need denominators.
Increased lethality of Ecstasy in the 21st century: in 2004-2007, per-annum E-related DRD risk was estimated at 10 per 100,000 past-year users of Ecstasy (95% CI: 8 to 12).
Cocaine-related DRD risk at least double that for Ecstasy: in 2004-2007, all-ages per-annum C-related DRD risk was estimated at 20 to 25 per 100,000 past-year users of cocaine aged 16-59 years but . . .
Cocaine-related death-rate increases sharply with user-age: in 2004-2007, per-annum C-related DRD risk was 7 (95% CI: 5.8 to 8.5) per 100,000 past-year users of cocaine aged 16-24 years but estimated at 53 (95% CI: 47 to 59) per 100,000 past year cocaine-users aged 35+ years.
Apparently 40% increased lethality of cocaine for older users (35+ years) between 2001-2003 and 2004-2007: per-annum C-related DRD risk was 37 (95% CI: 31 to 43) per 100,000 past-year cocaine-users in 2001-2003 but 53 (95% CI: 47 to 59) per 100,000 past year cocaine-users. Why?
Male past-year cocaine users’ C-specific DRD risk was roughly twice that of their female counterparts: risk multiplier in 2004-2007 was 2.4 (95% CI: 1.7 to 3.4).
Male injectors’ heroin-specific DRD risk was twice that of female injectors: risk multiplier for males was 2.0 (95% CI: 1.7 to 2.3) which, for England and Wales, corroborates a previous finding in Scotland.
Methadone-specific DRD risk multiplier for males was only 1.3 (95% CI: 1.1 to 1.5): enhanced risk-reduction for males recruited into opiate-substitution therapy needs further investigation because the National Programme on Substance Abuse Deaths (np-SAD) reports that over 60% of methadone-specific DRDs reported to them by coroners do not occur in those to whom methadone was prescribed . . . we need to know more.

Drug-related deaths per 100,000 person-years/users
Heroin-specific DRDs in 2004-2007 {this series} ~ ’as for’ because not all heroin-specific DRDs occur in injectors, and IDUs also experience DRDs that are not heroin-specific:   

As for male current IDUs:                                    290 (95% CI: 260 to 330) at 15-24 years

                                                                       vs    340 (95% CI: 320 to 350) otherwise;
As for female current IDUs:                                 120 (95% CI:   90 to 150) at 15-24 years
                                                                       vs    190 (95% CI: 160 to 210) otherwise.
Methadone-specific DRDs in 2004-2007 {this series}:   
For male opiate-substitution clients:                    136 (95% CI: 120 to 150)
For female opiate-substitution clients:                 106 (95% CI:   89 to 124)
Methadone-only DRDs in 2004-2008 per 100,000 implied methadone-client years {Strang et al., BMJ 2010 and Straight Statistics}
Scotland, by implied methadone-client-years:     136 (95% CI:   86 to 129)
England, by implied methadone-client-years:      212 (95% CI: 194 to 230)
Any-methadone DRDs in 2004-2008 per 100,000 implied methadone-client years {Strang et al., BMJ 2010 and Straight Statistics}
Scotland, by implied methadone-client-years:      511 (95% CI: 465 to   558)
England, by implied methadone-client-years:      420 (95% CI:   395 to   444)
ALL-CAUSE mortality in opiate-substitute clients in UK’s General Practice Research Database, 1990-2005, followed up until 1 YEAR BEYOND expiry of last prescription {Cornish et al., BMJ 2010}
GPRD clients, all-cause mortality on treatment:   694 (95% CI:   521 to   866)
GPRD clients, all-cause mortality off treatment: 1320 (95% CI: 1080 to 1560)
Cocaine-specific DRDs at 16-59 years in 2004-2007 {this series & British Crime Survey}:
Per 100,000 past-year cocaine-users:                         9 (95% CI, Poisson only: 7 to 10)
Per 100,000 past-year male cocaine-users:              14 (95% CI, Poisson only: 12 to 16)
Per 100,000 past-year female cocaine-users:             6 (95% CI, Poisson only:   4 to   8)
Ecstasy-specific & Ecstasy- related DRDs at 16-59 years in 2004-2007 {this series & British Crime Survey}:
E-specific, per 100,000 past-year Ecstasy-users:      5 (95% CI, Poisson only: 4 to   6)
E-related,   per 100,000 past-year Ecstasy-users:      9 (95% CI, Poisson only: 7 to 11)
BY AGE, Cocaine-related DRDs  in 2004-2007 {this series & British Crime Survey}:
Per 100,000 past-year cocaine-users aged 16-59:    24 (95% CI:      22 to     26)
Per 100,000 past-year cocaine-users aged 16-24:      7 (95% CI:        5 to       9)
Per 100,000 past-year cocaine-users aged 25-34:    28 (95% CI:      25 to     31)
Per 100,000 past-year cocaine-users aged 35-59:    47 (95% CI:      47 to     59)

Our definitions: Each of 15,795 DRDs was cross-classified by era of registration (1998-2000, 2001-2003, 2004-2007), sex, age-group at death (15-24, 25-34, 35+ years), presence/absence of heroin/morphine (hereafter, H), presence/absence of methadone (M), presence/absence of cocaine (C), presence/absence of benzodiazepines (B), and presence/absence of alcohol (A). We defined cause-specific DRDs as DRDs with “one specified illegal drug present and three other specified illegal drugs absent” such as “H, not (M, C, B)” [H-specific] or “M, not (H, C, B)” [M-specific] or “C, not (H, M, B)” [C-specific].

Conflicts of interest: SMB serves on, and was inaugural chair of, the Surveys, Design and Statistics Subcommittee (SDSSC) of Home Office’s Scientific Advisory Committee. In 2008, SDSSC reported on 21st Century Drugs and Statistical Science. Authors have research or clinical interests in the epidemiology and prevention of drugs-related deaths. SMB holds GSK shares and MRC-funding as part of the NIQUAD cluster on quantifying drugs harms