Deaths, HIV infection, abstinence, and other outcomes in a cohort of injecting drug users followed up for 10 years

BMJ 1994; 309 doi: (Published 06 August 1994) Cite this as: BMJ 1994;309:369
  1. J R Robertson,
  2. P J M Ronald,
  3. G M Raab,
  4. A J Ross,
  5. T Parpia
  1. Edinburgh Drug Addiction Study, Muirhouse Medical Group, Edinburgh EH4 4PL
  2. MRC Biostatistics Initiative for AIDS and HIV Research in Scotland, Centre for HIV Research, Edinburgh EH9 3JN-PD- 940806.
  • Accepted 8 June 1994


Objective : To trace, follow up, and interview a group of patients known to be injecting drug users in order to establish current drug taking and other features related to drug use over 10 years.

Design : Descriptive follow up study of a cohort of injecting drug users established between 1982 and 1985.

Setting : General practice based patient population study initially, with later tracing of subjects throughout the United Kingdom through NHS Central Registries and current general practitioners.

Subjects : 203 injecting drug users recruited up to December 1985.

Main outcome measures : Survival, cause of death, abstinence, or continued drug use; HIV status; and demographic variables. Results - Of the 203 injecting drug users recruited into the study, over half were known to be HIV positive and 42 died of various causes (increasingly AIDS). From the start of follow up in 1990, 163 (91%) of the 180 survivors were traced, of whom 116 (71%) were interviewed. Dramatic changes had occurred in drug taking, with a move away from injecting towards oral drug use. A few patients, however, continued to inject. 90 (78%) of those interviewed had been in prison, of whom 37 (41%) had injected drugs while in prison.

Conclusions : The pattern of deaths had changed from being largely due to overdose in the early 1980s to predominantly AIDS related in later years. The reduction in deaths due to overdose may have been connected with but was not always causally related to a new approach by legal, social, and medical services. Drug use continues in a modified form in a large proportion of patients followed up. As a group, drug users require long term support for a multiplicity of problems.

public health implications

  • public health implications

  • HIV and AIDS now account for more deaths than overdose among Edinburgh drug injectors

  • Continued drug injecting is common, and use of oral substitute drugs or illegally acquired drugs is also problematical

  • The foreseeable future holds the prospect of continuing problems

  • Current policy does not adequately address the difficulties experienced by drug users, either with respect to their immediate needs or in providing for the continuing problems of an aging cohort greater risk of acquiring HIV by sexual activity than by injecting drugs in recent years, even in known drug users, are reported elsewhere.16


Since HIV infection was first transmitted by injecting drug use interest has focused on known drug users. Previously mortality and morbidity outcomes such as hepatitis B and later hepatitis C and death due to overdose, violence, and infections were studied.*RF 1-3* Abstinence has also been the subject of much research.*RF 4-6* Guidelines for clinical practice have reflected a new concern about the whole issue of drug use and HIV*RF 6-9* and show a change from the previous non-medical approach.*RF 10-11* Drug use is notoriously difficult to study because of changing patterns of drug use, evolving policy, and its illegality.

The Edinburgh drug addiction study recruited a cohort of known injecting drug users between 1980 and 1985 and has followed them up continually since then. Earlier reports on this cohort showed a high seroprevalence for HIV antibody12 and later a rapid decline in injecting drug use.*RF 13-15* A recent report showed a change to heterosexual transmission of HIV in this group.16 This paper reports a formal follow up evaluation of drug taking morbidity and mortality in the cohort up to mid-1993.

Subjects and methods

Individual drug injectors were recruited into the study after consultation with one of the doctors of the Muirhouse Medical Group. The cohort was closed at the end of 1985 and numbered 203 subjects. Follow up was carried out between January 1990 and June 1993. At the start of this study 23 of the original 203 subjects had died and attempts were therefore made to contact and interview 180 drug users (60 women, 120 men). Details of deaths were obtained through the General Register Office with which all patients were flagged at recruitment. When contact was made with the patient a structured interview was carried out either at the Muirhouse surgery, at the patient's home, in prison, or elsewhere. This provided demographic data, information on prison experiences, details of sexual behaviour, and drug use from 1982 to the time of interview. The drug use history included information about drugs taken throughout the user's career, route of administration, and sharing of injecting equipment. Subjects were asked whether they had injected during that year and if so the exact number of episodes of sharing.

The main phase of data collection was conducted between January 1990 and December 1992, 96% of all interviews (111/116) being completed between these dates. Additional information from current general practitioners was then added.


The main phase of recruitment into the study was during 1983-5, only 14 subjects being recruited before 1983, of whom two died before 1983. Thus mortality is described as a life table of survival from January 1983 with cohort members contributing to the population at risk on their date of recruitment, truncated at the end of 1992 to allow for possible delays in the reporting of deaths from the General Register Office. There were 40 deaths in the cohort from January 1983 to December 1992. Fifteen deaths were attributed to overdose, concentrated in the early years of follow up, and there were 16 AIDS related deaths, 10 of which occurred in 1991 (fig 1).

Of the 180 injecting drug users (120 men) alive in December 1989, 163 were traced by the end of June 1993. Sixty one (34%) were registered at the Muirhouse Medical Group and 107 were registered elsewhere at follow up (fig 2). Comparison of those interviewed and those traced but not interviewed showed no differences in sex distribution between the groups. Subjects who were interviewed were slightly younger than those who were not (mean age 32.3 years v 35.7).

Table I compares the rate of follow up by HIV status as measured during the study. The few interviews in the untested group reflects the fact that people who lost contact with the study had less opportunity to be tested. Thirty general practitioners supplied additional information. Six patients were HIV positive, 10 HIV negative.


Follow up of subjects alive in December 1989 by HIV status

View this table:

The median age of the interviewed group was 29.3 years (interquartile range 26-34). Seventy six patients in the group (66%) had one or more children. Of these patients, 29 of the 51 men (57%) and four of the 25 women (16%) did not live with any of their children. Of the children who lived away, most lived with their other parent; five lived with grandparents, three were adopted, one was with a foster family, and one was in care. Of the 37 women interviewed, 26 had had at least one full term pregnancy, seven at least one spontaneous abortion, and 13 at least one therapeutic abortion. The median number of pregnancies was two, maximum seven.


Survival of 201 injecting drug users 1983-92


Follow up status of all 203 cohort members

Interview data disclosed that 85 (73%) injecting drug users had been hospital inpatients at some time, 66 for drug related problems and 16 for psychological problems. Ninety (78%) had been in prison. Of these, 37 (41%) reported having injected drugs while in prison and 28 (76%) having shared injecting equipment in prison. A slightly higher proportion of HIV positive injecting drug users had both injected and shared needles in prison (table II).


Drug use among prisoners by HIV status

View this table:

Of the 116 injecting drug users who were interviewed, 91 reported a sexual relationship with at least one person in the previous six months: 71 had had one partner, 13 two partners, five three partners, one four partners, and one (who worked in the sex industry) 470 partners. Forty four (38%) of the injecting drug users had had sex with a partner known to be infected with HIV at some point. This proportion was higher (37/83; 45%) among HIV positive injecting drug users, though six (24%) of the HIV negative subjects also recalled having sex with an HIV positive person. Of all 91 injecting drug users reporting a sexual relationship in the six months before interview, 39 (64%) of 61 HIV positive subjects and five (22%) of 23 HIV negative subjects reported using condoms. Twenty six of the HIV positive subjects and one HIV negative subject reported always using a condom.

The proportion of the 116 injecting drug users using heroin had fallen steadily. However, the use of other drugs had risen. Figure 3 shows the numbers of patients interviewed using drugs in each year from 1982 to 1990. Pharmaceutical drugs included the benzodiazepines, methadone, and dihydrocodeine, and non-prescribable drugs were heroin, cocaine, cyclizine-dipipanone (Diconal), amphetamines, barbiturates, and buprenorphine.


Number of 116 injecting drug users interviewed who reported using drugs in each year 1982-90

Injecting drug use had continued to decline since 1983, the sharpest fall occurring between 1984 and 1986 (fig 3). Illegal drug use and sharing of equipment showed a similar trend (fig 4). However, there had been very little fall in the overall level of drug use in the cohort, with a shift towards pharmaceutical drugs. In the last years of the study benzodiazepines were being used in considerable quantities and accounted for a large part of the clinical workload, intravenous use increasing from 2% to 8% (fig 5).


Percentage of injecting drug users who reported sharing needles in each year 1982-90.


Benzodiazepine use by drug users 1982-90

Of the 116 injecting drug users interviewed, 71 (61%) reported using at least one of the above drugs in every year. Twenty one (18%) users stopped using drugs and did not restart in the period 1982 to 1990, and 24 (21%) reported episodes of abstinence and relapse. Of the 114 who reported injecting, 22 (19%) injected in each year, 54 (47%) stopped injecting, and 38 (33%) showed episodic intravenous drug use.


Possibly the most important contribution of this study was the observation on patterns of drug use - that is, the categorisations of drug users into the 18% (21/116) who stopped, and 21% (24/116) who stopped and restarted, and the 61%(71/116) who used continuously. Equally revealing were the three groups of the 114 subjects who reported injecting - 19% (22) injecting in every year, 47% (54) stopping injecting, and 33% (38) showing an episodic pattern of injecting. These findings reflect the various responses to aging, treatments, and changing times and the difficulty of assessing success and failure of treatments. The many subjects who reported continuous or episodic injecting must be a cause of concern and is of considerable interest given the length of follow up. Clearly, a large proportion of drug users sustain drug use over many years, either continuously or episodically.

The change in drug use is dramatically portrayed in figures 3 and 4. The slightly depressing replacement of illegal heroin by prescribed drugs was associated with an encouraging decrease in overall injecting. Once again, however, it is difficult to attribute the decrease in injecting to the prescribing as other factors may influence these changes and the decrease in injecting predated the main prescribing initiatives. An important factor influencing behaviour change was the change in law enforcement which began in 1984. Increased sentences for people convicted of dealing in heroin and even for those found in possession disrupted an active drug using community dramatically. Prescribing services, present in inadequate form throughout this time, were increased only after 1987.17 Among other influences were the aging of the cohort, their knowledge of HIV, and the new availability of information and other support services. Recent testing for hepatitis C antibody yielded extensively positive results.

This follow up study was successful in tracing a large proportion (163/180;91%) of the surviving members of the cohort, though only 71% (116) of these were finally interviewed. Those most difficult to locate tended to be the HIV negative or untested drug users who had moved from the practice and who had frequent changes of address. This may indicate a more chaotic lifestyle or alternatively be an indication of an improved situation. There was evidence of both from general practitioner contacts.

The change from overdose to AIDS as a predominant cause of death among injecting drug users is interesting. AIDS deaths will clearly continue but the fall in deaths from overdose is important and may be due more to the lack of heroin and decreased injecting than directly to less overall drug use. The overall mortality is likely to exceed that in other, pre-AIDS studies of survival.*RF 18-20*

Drug use in prison is reported in many studies, including a recent study in Edinburgh.21 From our data we cannot assess whether there might be a causal relation between being in prison and being HIV antibody positive.

The fertility of the group is important, especially in a cohort so heavily infected with HIV, though transmission from mother to child was infrequent in this series.22 The high rate of sexual partners is worrying, though subjects with HIV adopt a greater use of condoms. A high rate of HIV positive partners and a Cherubin and Sapira's vast review of the (mainly American) published work concludes that the “natural history” of intravenous drug use includes early damaging behaviour followed in the subjects' late 20s and early 30s by attempts to quit drug use and in their late 30s and early 40s by some success in stopping. They also suggested, however, that drug use often continues beyond these decades and that “fifty and sixty year old intravenous drug users” are not rare.23

Our data fit in with Cherubin and Sapira's suggested pattern of behaviour and fill in some of the details of behaviour during those long decades of drug taking. They should help those interested in drug use behaviour to understand the complexity of the problem and the consequent problems in treatment and prevention. They should also prepare policy makers for a long haul ahead.24

We are grateful to Dr Ken Roberts and Dr Kirsty Foster and the partners of the Muirhouse Medical Group for their support and especially Lesley Devogelaere for all her work on the project.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
View Abstract