PERSONAL PERSPECTIVE

THE SWISS HEROIN TRIALS: SCIENTIFICALLY SOUND?

INTRODUCTION

Switzerland's heroin maintenance project has made headlines around the
world. The dramatic announcement of its success in the summer of 1997 seemed
to promise an unconventional but effective way to treat heroin addiction. In
this country, the Associated Press reported, Switzerland declared its novel
experiment with statedistributed heroin a success, saying the drug giveaway
slashed crime, misery and disease associated with hardcore drug addiction
(Associated Press, 1997). Favorable commentary in the prestigious oped pages
of the New York Times (Lewis, 1998) and the Washington Post (Shenk, 1997),
and on the national evening news (ABC World News Tonight, 1997) urged policy
makers to consider heroin maintenance here.

Frustrated by prior efforts to curb use, the most visible failure being the
squalid deterioration of Zurich's Platzspitz Park and subsequently the
Letten railway station areas designated by the government for open air drug
usethe Swiss Federal Office of Public Health instituted the Swiss Scientific
Studies of Medically Prescribed Narcotics to Drug Addicts (hereafter the
Swiss Heroin Trials). The Federal Office of Public Health conducted the
trials from January 1, 1994 to December 31, 1996 and published the results
in July 1997, proclaiming success (Uchtenhagen, Gutzwiller, & DoblerMikola,
1997). The trials attracted enthusiastic attention from the international
health authorities and media, which hailed heroin prescription based on the
government's report of high retention rates, improved social integration,
and reduced crime among the enrolled heroin abusers (Nadelmann, 1998a;
Olson, 1997; Report of the International Narcotics Control Board for 1997,
1998; "Swiss Heroin Experiment Yields Positive Results," 1997). One Swiss
newspaper described a growing interest in adding cocaine distribution to the
heroin project: "Kokainder nachste Streitpunkt" ["Cocainethe Next Point to
Be Disputed"] (1997) the headline read.

The Federal Office of Public Health issued its detailed Germanlanguage
outcome report last summer, yet as early as 1996, the trials' principal
investigator and project directors traveled throughout Switzerland and other
countries, such as Australia, Austria, Germany, the Netherlands, and the
United States, promoting what they interpreted to be its positive results.
In Switzerland, the Federal Office of Public Health plans to triple
enrollment next year, to about 3,000, and in the year 2004, the Swiss
Parliament is scheduled to vote to decriminalize consumption, possession,
and sale of narcotics for personal use (Federal Council of Switzerland Asks
for Opinion, 1997).

Several weeks ago, the International Narcotics Control Board of the United
Nationsthe quasijudicial organization that monitors international drug
treaties expressed concern over this trend. The Board worried that, "before
(completion of) the evaluation by the World Health Organization (WHO) of the
Swiss heroin experiment, pressure groups and some politicians are already
promoting the expansion of such programmes in Switzerland and their
proliferation in other countries" (Report of the International Narcotics
Control Board for 1997, 1998). Earlier this year, the Netherlands, mindful
of the Swiss experiment, initiated a trial of heroin administration. Drug
policy experts and advocates in Australia, Belgium, Germany, Luxemburg, and
Norway, all citing the Swiss experience, have urged their governments to
establish heroin distribution trials or projects (Bammer, 1994; Farrell &
Hall, 1998; Central Committee on the Treatment of Heroin Addicts, 1997). In
the United States, prominent constituents of the harm reduction and
legalization movements championed the Swiss project (Nadelmann, 1998b).
Thus, numerous policy debates are being influenced by the outcome of the
Swiss study.

A careful review of the study, however, reveals critical design weaknesses
that challenge enthusiastic claims about the superiority of heroin
maintenance over conventional opiate treatment for heroin abusers, such as
oral methadone maintenance and residential, or abstinenceoriented, care.
Awareness of the Swiss trials' methodological limitations is critically
important for medical researchers, political leaders, and other parties who
may be contemplating major changes in drug policy based upon the Swiss
experience.

SWISS HEROIN TRIALS

The Swiss study sought to answer several questions. What are the effects of
prescribed opiates on health, drug use, and social functioning? Is heroin
administration suitable for heroin abusers whose previous treatment had been
unsuccessful? Is heroin effective compared with other therapies?

Initially, achieving abstinence from drugs was identified as the primary
clinical aim of heroin prescription programs. As the study progressed,
however, other priorities took its place, including recruitment into the
program, retention in the program, and improvement in medical and social
functioning (Uchtenhagen, 1993, 1996). Beginning in January 1994, subjects
were recruited into three study arms: heroin, morphine, or methadone
maintenance. These medications were selfadministered by injection on site at
the clinic. Eligibility was restricted to individuals at least 20 years old
who were dependent on opioids for a minimum of 2 years, had failed at least
two previous treatment attempts (residential care and/or oral methadone
maintenance), or who were deemed unlikely to respond to available treatment,
and who displayed evidence of social, physical, and psychological
disintegration.

The original plan of randomly assigning 250 patients to the heroin
administration group, 250 to morphine, and 200 to methadone had to be
abandoned during the first year, due to subject's strong preference for
heroin as well as problems with side effects (histamineinduced hives from
morphine and vascular irritation from intravenous methadone). Target
recruitment was subsequently revised twice so that the heroin arm would
contain 800 patients and the others 100 each (Uchtenhagen, 1996).

Overall, 1,146 patients were enrolled in the 18 participating research
clinics, but most of the data that appeared in the final Swiss report were
drawn from a smaller sample of 385 patientsthe first cohort recruited to the
study prior to sample size revision. Patients were interviewed at intake and
6month intervals up to 18 months. Outcome measures included selfreport of
drug use, health status, psychological symptoms, social functioning, and
crime. Urine toxicology screening, human immunodeficiency virus (HIV)
testing, and examination of police records were allowed only with permission
of patients; thus, efforts to verify selfreports were sporadic at best.

The average daily intravenous heroin dose was 490 mg. No attempt was made to
correlate magnitude of dose with outcomes. Patients selfadministered heroin
under sterile conditions at the clinic three times per day, 7 days per week,
and received oral methadone overnight to reduce any breakthrough withdrawal
symptoms. Patients were also permitted heroin cigarettes, longacting heroin
pills and suppositories, intravenous methadone, and intravenous or oral
morphine to complement the study medication, intravenous heroin. Extensive
counseling, social assistance (welfare benefits, public housing, medical
care), and psychiatric services, none mandatory, were offered as needed.

Results based on the initial cohort of 385, all of whom received heroin,
showed that retention was 89% at 6 months, 76% at 12 months, and 69% at 18
months. Anyone who kept attending the program, even intermittently, was
reported as a retained participant. Of the 128 patients who did not complete
the 18month study period, 50 sought standard oral methadone maintenance, 20
sought abstinenceoriented treatment, 15 dropped out, and 7 died. Other than
disposition, outcome data on these 128 patients were not reported. Three
features were directly associated with drop out: duration of addiction to
heroin; severity of cocaine consumption; and HIVpositive status. Thus,
patients who left the trial tended to be those with the most serious
addictionrelated problems, the very group for whom heroin administration was
intended (Uchtenhagen, 1996).

The Swiss Federal Office of Public Health report presents outcomes as a pre
and posttreatment analysis of the 266 patients who completed at least 18
months. According to their selfreport, 81% (not 100%) said they used illicit
heroin at admission, while 52% admitted to illicit heroin consumption at 18
months. One in five claimed to be unemployed at 18 months, down from almost
half (44%) at admission. Employment was defined by the project managers as
having a contract with an employer, regardless of whether this meant working
a few hours a week or maintaining a fulltime job. Homelessness declined from
12% to 1% and stable housing increased from 49% to 69% (as noted earlier,
housing was provided by the government).

The percentage of individuals claiming that they supported themselves with
illegal income went from 70% to 10%, permanent employment increased from 14%
to 32%, and the number supported by welfare increased from 18% to 27%. At
admission, 82% of completers reported cocaine use (intensity undefined) at
admission, compared to 52% at 18 months. Patients suffering from psychiatric
symptoms declined from 36% to 18% during the 18month period while “somatic
state” (undefined) improved from 79% to 86%.

SCIENTIFIC SOUNDNESS OF THE SWISS TRIALS

The above results are intriguing, but unfortunately they are not the product
of an experimental outcome study. Failure to meet scientific standards for a
controlled clinical trial was also noted by an independent evaluator from
the World Health Organization, who assessed the experimental plan of the
Swiss project (Hall, 1997). First, there were no comparison groups. Although
the study contained two other groups, methadone and morphine, plus a newly
added cohort of 350 who were to receive oral methadone, data from those
subjects were reportedly unavailable for comparison with the heroin group at
the time of the final Swiss report. This is key, since the study has been
promoted by the Swiss government and some foreign observers as evidence that
heroin prescription is better than methadone maintenance or
abstinenceoriented treatment; the latter was not even included in the
original design.

A second issue is the validity of outcome measures. Optimally, selfreports
of illicit drug use, crime, and employment should be corroborated,
especially if the results are to be used to undergird major policy
decisions. However, the urine collection was unreliable; timing was not
random, but rather was mutually agreed upon by clinic and patient. Even then
it was not routinely collected under observation, nor were police records
and employment records systematically examined.

A third methodological problem was the failure to control for the influence
of social services on outcome. According to the director of a Zurich clinic
(Locher, 1996), the heroin trials spent almost five times more per patient
for social services than standard methadone treatment spends on its
patients. Thus, even if outcome data are valid, it is unclear whether they
are attributable to a stabilizing effect of heroin, to the social services
provided, or to an interaction between the two.

Fourth, only individuals who completed 18 months were included in the
outcome data analysis. Data on those who left the trial prematurely were not
collected, yet, with over half of them going into other treatment programs,
it is possible they improved as well, perhaps as much or more than the
completers.

It should also be noted that the Swiss Heroin Study was intended for
“severely addicted” individuals, yet it appears that the baseline condition
of many abusers was not so dire. For example, 74% of patients were rated by
interviewers as being in “good or very good health” and 80% as having “good
or very good nutrition.” A large percentage, 76%, experienced between zero
and five withdrawal episodes in a lifetime; a relatively small number of
total withdrawals, given the sample's 10year mean duration of dependence. In
addition, 14% were only occasional users of heroin and 4% did not use heroin
at all, even though eligibility required 2 continuous years of dependence on
heroin.

There is also clear evidence that heroin administration diverted patients
from enrolling in abstinenceoriented treatment. Swiss residential
programswhich generally operate at full capacity with waiting listswere able
to fill only half their slots after the project began operating, according
to a number of treatment providers in Zurich (Belengungsrueckgang bei
Drogentherapiehausern [Attrition in Residential Drug Treatment], 1997). In
addition, 61% of the participants came to the trial directly from oral
methadone maintenance programs. It would have been useful to track this
subgroup of patients to note differential improvement, if any, after
switching to heroin prescription.

At the same time, however, these instances of patient diversion and
migration, coupled with the study's recruitment problems described earlier,
show why it is difficult to conduct a randomized trial with heroin: when
given the option of being assigned to heroin prescription, subjects will
often reject other assignments. This raises the question of whether heroin
treatment will give some abusers an incentive to fail conventional
treatments.

Finally, one of the major justifications for the Swiss heroin
programaverting HIV transmissioncould not be fully evaluated, as there was
no requirement that patients submit to HIV testing. The final report does
not mention how many were tested. Although it notes that three new cases of
HIV were identified, it failed to compare pre and posttreatment
seroconversion rates among heroin recipients to atrisk individuals in the
general population or to samples in methadone treatment, thus making the HIV
data extremely difficult to interpret.

A major original intent of the study was to move patients to abstinence
directly or through referral to abstinenceoriented treatment. According to
this criterion, the project had a success rate of only 5.2% (20 of the 385
patients transferred into abstinence treatment). But, according to the
criterion of retention, which was 69% at 18 months, the study was heralded
as an achievement. At first glance, 69% appears high, compared to retention
in conventional treatments, but it is not especially surprising that heroin
abusers would continue to come to a site where they could get
pharmaceuticalgrade heroin at no or low cost. In addition, the definition of
“retention” was extremely loose. Standard threshold definitions of retention
for example, 75% attendance in a 30day period, or showing up for two out of
three injection times per daywere not employed.

Furthermore, many individuals continued to commit crime and use illicit
drugs while receiving heroin. When their social situations improvedand it is
unclear whether the abusers' situations improved as much as was claimed,
since their selfreports were not consistently verifiedthe study design made
it impossible to determine how much improvement was a function of heroin
prescription per se, and how much could be attributed to the social services
provided.

DISCUSSION

The Swiss Heroin Trials cannot be considered a valid experiment. They failed
to randomly assign patients to comparison groups and to account for the
effect of social services on outcome. Heavy reliance on selfreport renders
even the outcome data questionable. Instead of an experiment, the trials
resembled more a demonstration project and, in this regard, they were deemed
“feasible” by an independent WHO evaluator who cited minimal problems with
diversion, heroin overdose, and neighborhood disruption (Hall, 1997).

Surprisingly, there has been little critical appraisal of the project in the
popular press abroad and virtually none in this country, although the United
Nations' International Narcotics Control Board “is concerned that
announcement of the (heroin project) results . . . have led to hasty
conclusions by some politicians and the media in several European countries”
(Report of the International Narcotics Control Board for 1997, 1998). One
would presume that such a controversial undertaking would be submitted to
rigorous scientific evaluation before being used as the basis for a policy
change. Yet the Swiss government has already expressed its intention to
regard heroin administration as a recognized therapy (“anerkannte Therapie”)
for heroin addiction (Federal Council of Switzerland Asks for Opinion,
1997).

An historical look at the administration of morphine and heroin shows that
it has been largely ineffective in helping opiate abusers achieve
abstinence. In the United States, “narcotics clinics” were established in
the early part of the century after the U.S. Treasury Department outlawed
maintenance prescribing as part of a vigorous effort to crack down on “dope”
doctors. By 1920, approximately 40 clinics were scattered across the country
to dispense morphine; only the New York City clinic offered heroin. However,
the federal government closed most of these clinics after only a year or two
of operation because of a Supreme Court decision that allowed the
prosecution of any physician who prescribed narcotics for addiction
maintenance (Musto, 1998).

Some clinics, like the one in New Haven, CT, were wellrun, but others were
not (Musto & Ramos, 1981). The New York City clinic was shut down by the
city's health department because of massive diversion of distributed heroin
by many of the nearly 10,000 enrolled heroin abusers. Treating such a large
number of heroin abusers was costly and, in the end, almost all of them
reverted to illicit narcotics after they had been detoxified (Musto, 1998).
In 1920, the American Medical Association issued a resolution condemning the
narcotics clinics (“Report of the Committee on the Narcotic Drug Situation
in the United States,” 1920), though in the early 1970s, the American Bar
Association rekindled the debate by suggesting that heroin maintenance was
feasible as a way to reduce crime and disease associated with heroin
addiction (American Bar Association Special Committee on Crime Prevention
and Control, 1972).

In the 1920s, the decade in which the United States discontinued its
clinics, Great Britain officially endorsed the ongoing activity of treating
addicted individuals with opiate drugs if previous efforts at withdrawal had
failed. In the 1960s, when the number of addicted individuals on
prescription heroin and cocaine abruptly increased, special treatment
clinics were established in London and elsewhere. By 1970, about one fifth
of England's 2,600 opiate abusers were receiving prescription heroin or
morphine (Lewis, 1973). Today, roughly 340, or less than 1% of England's
150,000 heroin abusers, are maintained on heroin, despite the fact that some
100 physicians are permitted to prescribe it (Strang & Sheridan, 1997). By
comparison, about 16,500 (11%) are enrolled in oral methadone maintenance.
This strongly suggests that British physicians do not find heroin useful.

Only one random design outcome study of the socalled British system has been
published. Hartnoll et al. (1980) randomly assigned 96 treatmentseeking
heroin abusers to treatment with injectable heroin or oral methadone. At 12
months, 88 subjects remained available to interviewers. Despite the greater
retention of the heroin sample (74%, compared to only 29% of methadone
recipients), followup revealed that the minimal improvements in employment
status, health, and consumption of nonopiate drugs were comparable between
the two groups.

Clearly, enormous frustration with drug problems, both here and abroad,
fuels much of the interest in innovative remedies. If heroin maintenance
could be shown, through rigorous comparison with conventional treatment, to
eliminate crime, illness, and drug use in intractable opiate abusers, it
might justify consideration as a form of therapy to be weighed against the
political and economic costs. However, the recent Swiss trials of supervised
heroin prescription trials do not withstand scientific scrutiny.

REFERENCES

ABC World News Tonight. (1997). March 26. New York: American Broadcasting
Company.

American Bar Association Special Committee on Crime Prevention and Control.
(1972). New Perspectives on Urban Crime (p. 62). Washington, DC: Author.

Associated Press. (1997). July 10. Bern, Switzerland: Author.

Bammer, G. (1994). The feasibility of the controlled supply of heroin to
opiate addicts. Canberra, Australia: National Centre for Epidemiology and
Population Health.

Belengungsrueckgang bei Drogentherapiehausern [Attrition in Residential Drug
Treatment]. (1997). Leben und Glauben, Idea, May 28.

Central Committee on the Treatment of Heroin Addicts. (1997). Investigating
the medical prescription of heroin: A randomized trial to evaluate the
effectiveness of medically coprescribed heroin and oral methadone, compared
to oral methadone alone in chronic, treatment refractory heroin addicts.
Utrecht, The Netherlands: Author.

Farrell, M., & Hall, W. (1998). The Swiss heroin trials: Testing alternative
approaches. British Medical Journal, 316, 639.

Federal Council of Switzerland asks for opinion. (1997). Press release,
December 19, Bern.

Hall, W. (1997). The Swiss scientific studies of medically prescribed
narcotics (Technical Report No. 43). University of New South Wales: National
Drug and Alcohol Research Centre (NDARC).

Hartnoll, R.L., Mitcheson, M.C., Battersby, A., Brown, G., Ellis, M.,
Fleming, P., & Hedley, N. (1980). Evaluation of heroin maintenance in
controlled trial. Archives of General Psychiatry, 37, 87796 884.

Kokainder nachste Streitpunkt [CocaineThe next point to be disputed].
(1997). TagesAnzeiger, July 11.

Lewis, E. (1973). A heroin maintenance program in the United States? Journal
of the American Medical Association, 223, 53996546.

Lewis, A. (1998). The noble experiment. The New York Times, January 5, p.
A21.

Locher, U. (1996). Coordinator of Zurich's Lifeline Project and Drug
Coordinator for the Zurich Social Welfare Department. Presentation to the
Lindesmith Center, New York City, January 4.

Musto, D.F. (1998). The American disease: Origins of narcotics control.
London: Oxford University Press.

Musto, D.F., & Ramos, M.R. (1981). Notes on medical history: A followup of
the New Haven morphine maintenance clinic of 1920.

New England Journal of Medicine, 304, 1071961077.

Nadelmann, E. (1998a). Commonsense drug policy. Foreign Affairs, 77,
11196126.

Nadelmann, E.A. (1998b). Europe could teach America much about drug policy.
International Herald Tribune, January 10, p. 6.

Olson, E. (1997). Swiss weigh fate of clinics offering legal heroin. The New
York Times, September 28, p. A3.

Report of the International Narcotics Control Board for 1997. (1998).
February 24. Vienna: United Nations.

Report of the Committee on the Narcotic Drug Situation in the United States.
(1920). Journal of the American Medical Association, 74, 1318.

Shenk, J.W. (1997). Hooked on dogma: U.S. drug warriors ignore Switzerland's
success with heroin addicts. The Washington Post, December 12, p. C1.

Strang, J., & Sheridan, J. (1997). Heroin prescribing in the British system
of the mid 1990s: Data from the 1995 national survey of community pharmacies
in England and Wales. Drug and Alcohol Review, 16, 79616.

Swiss heroin experiment yields positive results. (1997). Agence France
Presse, July 10

Uchtenhagen, A. (1993). Programme for a medical prescription of Narcotics:
Experimentation plan. Bern: Federal Office of Public Health.

Uchtenhagen, A. (1996). Programme for a medical prescription of narcotics:
2nd intermediate report. Bern: Federal Office of Public Health.

Uchtenhagen, A., Gutzwiller, F., & DoblerMikola, A. (1997). Programme for a
medical prescription of narcotics: Final report of the research
representatives. Paper presented at meeting of the Institute for Social and
Preventive Medicine, University of Zurich.

Received December 4, 1998; Revised February 3, 1999; Accepted March 3, 1999.
The authors would like to acknowledge the Program on Medical Science and
Society of the Ethics and Public Policy Center, Washington, DC. Requests for
reprints should be addressed to Sally L. Satel, MD, Ethics and Public Policy
Center, 1015 15th Street NW, Suite 900, Washington DC, 20005. Email:
slsatel@aol.com