Volume 172 Issue 6

Abstract Objective: To describe the epidemiological pattern of newly diagnosed HIV infection and AIDS among Indigenous Australians. Design and setting: National surveillance for newly diagnosed HIV infection and AIDS in Australia. Information on Indigenous status was sought at HIV/AIDS notification in all State/Territory health jurisdictions, except the Australian Capital Territory, and Victoria before June 1998. Main outcome measures: Number of people with newly diagnosed HIV per year and population rate of HIV diagnosis; demographic characteristics of people with HIV and AIDS diagnoses by Indigenous status. Results: From 1992 to 1998, 127 Indigenous Australians were newly diagnosed with HIV infection and 55 were diagnosed with AIDS. The population rate of HIV diagnosis among Indigenous Australians (5.23/100 000 per year) was similar to that among non-Indigenous Australians (5.51/100 000 per year). The annual number of HIV diagnoses among Indigenous people was relatively stable, but among non-Indigenous people it declined steadily over time. A higher proportion of Indigenous people diagnosed with HIV were women (26.8% v 8.9%; P < 0.001). Although male homosexual contact was the predominant source of exposure for both Indigenous (46.7%) and non-Indigenous (75.0%) people with HIV infection, exposure by heterosexual contact (36.7% v 15.3%; P < 0.001) was reported more frequently among Indigenous people. Conclusion: Although HIV incidence was similar among Indigenous and non-Indigenous Australians, the lack of a recent decline in incidence and the higher proportion of Indigenous people exposed to HIV by heterosexual contact indicate the need to intensify interventions to prevent HIV transmission among Indigenous people. Introduction The epidemic of HIV transmission peaked in Australia in the mid 1980s, and there was a subsequent peak in AIDS incidence of nearly 1000 cases in 1994.1 The estimated number of people diagnosed with HIV infection in Australia to the end of 1998 was 16 714, with an estimated 11 800 living with HIV infection. Although the peaks of both the HIV and AIDS epidemics in Australia have passed, HIV infection continues to be transmitted, predominantly through male homosexual contact, at an estimated level of 450 cases per year.1 Despite evidence of a relatively well-controlled HIV epidemic in Australia, evaluation of the Third National HIV/AIDS Strategy noted an increase in the reported number of Indigenous Australians diagnosed with HIV infection in the early 1990s.1 Furthermore, high rates of other sexually transmissible infections in some Indigenous communities indicate the potential for HIV transmission.1 To define the pattern of HIV infection among Indigenous Australians, and to assess time trends in new diagnoses of HIV infection and AIDS, we examined national HIV and AIDS notification data by Indigenous status for the years 1992-1998. National Health and Medical Research Council guidelines on ethical matters in Aboriginal and Torres Strait Islander health research were followed.2 Methods National surveillance procedures Surveillance procedures for newly diagnosed HIV infection and AIDS have been described previously.3,4 Briefly, newly diagnosed HIV infection and AIDS are notifiable conditions in each State or Territory health jurisdiction in Australia. Information sought at national notification of newly diagnosed HIV infection includes the State or Territory of diagnosis, postcode of residence, namecode (based on the first two letters of the family name and the first two letters of the first given name), sex, date of birth, Indigenous status, date of HIV diagnosis, CD4 cell count at HIV diagnosis, evidence of newly acquired HIV infection, and patient-reported source of exposure to HIV. Information sought at AIDS notifications also includes the date of AIDS diagnosis, AIDS-defining illnesses, and use of antiretroviral therapy before AIDS diagnosis. People with newly diagnosed HIV infection with evidence of newly acquired HIV infection (ie, a negative or indeterminate HIV antibody test result or a diagnosis of HIV seroconversion illness within 12 months of HIV diagnosis) were defined as having "newly acquired HIV infection". People with AIDS were classified as having "late HIV diagnosis" if HIV infection was newly diagnosed within three months of AIDS diagnosis. Indigenous status From 1985, information on Indigenous status, obtained through self-identification as Aboriginal or Torres Strait Islander, was routinely sought at notification of HIV infection and AIDS for people newly diagnosed in the Northern Territory, Queensland, South Australia, Tasmania and Western Australia. In New South Wales, Indigenous status has been sought for newly diagnosed cases of HIV infection and AIDS since 1992. Indigenous status was not available for people with HIV infection or AIDS diagnosed in the Australian Capital Territory, or from Victoria before June 1998. Information on Indigenous status has been sought nationally from 1995; available information on Indigenous status for cases diagnosed before 1995 was obtained retrospectively through State or Territory health authorities. Exposure category HIV exposure was classified as male homosexual contact, male homosexual contact plus injecting drug use, injecting drug use, heterosexual contact only, haemophilia/coagulation disorder, receipt of blood or tissue, mother with or at risk for HIV infection, and other or undetermined exposure. Statistical analysis A χ2 or Fisher's exact test and odds ratios were used to test for differences between Indigenous and non-Indigenous cases with respect to demographic characteristics (sex, residence), newly acquired HIV 1infection, late HIV diagnosis, HIV exposure category, and individual AIDS-defining illnesses. Residence was divided into "metropolitan" and "non-metropolitan" on the basis of postcode. "Metropolitan" was defined as capital city (including Canberra), and "non-metropolitan" was defined as other than capital city. In the analyses, cases without information on Indigenous status were grouped with non-Indigenous cases. The population-based rate of HIV diagnosis was calculated by Indigenous status and year (for States and Territories other than Victoria and the ACT) using Australian Bureau of Statistics (ABS) census data for 1996.5 Results Information on Indigenous status was available for 91% of people with newly diagnosed HIV infection. For the period 1992-1998, 5313 cases of newly diagnosed HIV infection were notified to the national HIV surveillance centre, of which 127 (2.4%) were Indigenous cases. For the same period, 3638 AIDS cases were notified, of which 55 (1.5%) were Indigenous cases. The annual number of HIV diagnoses among Indigenous people was relatively stable over this period (Box 1). In contrast, the annual number of HIV diagnoses among non-Indigenous people gradually declined over the years 1992-1998. During this period, the annual HIV diagnosis rate per 100 000 population among Indigenous people (diagnosed in States and Territories other than Victoria and the ACT) (5.23) was similar to that among non-Indigenous people (5.51) (Box 1). A higher proportion of Indigenous people with HIV were female (26.8% v 8.9%; P < 0.001) (Box 2). The median age at HIV diagnosis (30 years v 33 years; P < 0.001) and AIDS diagnosis (32 v 37 years; P < 0.001) was lower among Indigenous cases. The pattern of exposure to HIV reported by Indigenous people was different from that reported by non-Indigenous people both for newly diagnosed HIV infection and AIDS (Box 2). Although male homosexual contact was the predominant source of exposure to HIV for both Indigenous (46.7%) and non-Indigenous (75.0%) people, a history of heterosexual contact only was reported more frequently by Indigenous people (36.7% v 15.3%; P < 0.001). The proportion of Indigenous and non-Indigenous people with AIDS with "late HIV diagnosis" was similar (23.6% and 18.3%; P = 0.42), as was the proportion reporting antiretroviral therapy before AIDS diagnosis (56.4% and 62.2%; P = 0.5). No difference between Indigenous and non-Indigenous cases was observed in the median CD4 cell count at diagnosis of HIV and of AIDS. The spectrum of AIDS-defining illnesses for Indigenous and non-Indigenous people with AIDS is shown in Box 3. Cryptococcal disease (odds ratio [OR], 3.3; 95% CI, 1.4-7.6; P = 0.004), oesophageal candidiasis (OR, 1.8; 95% CI, 0.95-3.38; P = 0.05), and atypical mycobacterium (OR 8.3; 95% CI, 2.4- 25.42; P = 0.002) were more frequent among Indigenous AIDS cases, whereas Kaposi's sarcoma was less frequent (OR, 0.12; 95% CI, 0.01-0.80; P = 0.01). Among people with HIV, there were more Indigenous than non-Indigenous cases in non-metropolitan locations (36% v 16%) (P < 0.01). Similarly, among people with AIDS, there were more Indigenous than non-Indigenous cases in non-metropolitan locations (37% v 19%) (P = 0.002). Discussion The HIV epidemic among Indigenous Australians has been relatively limited to date, with an overall rate of HIV diagnosis comparable with that for non-Indigenous Australians over the years 1992-1998. However, there have been contrasting trends in these rates, with a declining rate of HIV diagnosis among the non-Indigenous population, but a relatively stable rate among Indigenous people. Features that distinguish the Indigenous from the non-Indigenous HIV epidemic are a higher proportion of women affected, a higher proportion with heterosexual exposure to HIV, a younger age at HIV and AIDS diagnosis, and a higher proportion of people with HIV in rural areas. The low proportion of people with "late HIV diagnosis" among both Indigenous and non-Indigenous AIDS cases would suggest that a large pool of undiagnosed HIV infection is not present in Australia. The very low HIV prevalence among prison entrants in all States and Territories, including those where Indigenous Australians constitute a large proportion of prison inmates, is further confirmation that HIV prevalence among Indigenous Australians remains low.6 Our findings also extend those of an earlier study that showed comparable rates of HIV infection in both the Indigenous and the non-Indigenous population in Queensland.7 In interpreting our findings, several limitations to the study methods need to be considered. Firstly, the lack of a uniform reporting system for Indigenous status in all States and Territories may result in under-reporting in some jurisdictions. However, there is evidence that in recent years Indigenous status has been more completely reported, with 91% of HIV notifications in those States/Territories other than the ACT and Victoria currently reporting Indigenous status.1 Secondly, reporting of Indigenous status was based on "self-identification", which may either not be reported correctly by the patient, or not requested by the clinician. If identifying as Indigenous is more likely in a census setting than in clinical practice, our rates of Indigenous HIV diagnosis may be underestimates. Thirdly, reported rates of HIV and AIDS diagnoses are dependent on the level and extent of HIV testing. Poor access to and uptake of confidential testing by some Indigenous people, and fear of possible stigmatisation arising from positive test results, may influence the extent of HIV testing among Indigenous people. The explanation for the apparently limited HIV epidemic among Indigenous Australians is almost certainly multifaceted. The drop in HIV transmission from the mid 1980s has meant that the extent of the Australian HIV epidemic has been limited compared with many other countries.1 Behaviour change among homosexual men was largely responsible for the initial reduction in HIV transmission from the mid 1980s,1 with other measures such as the widespread introduction of harm minimisation programs for injecting drug users,8 and high condom use and low rates of sexually transmissible infections among most sex workers9 contributing to the ongoing relatively low level of HIV transmission. The absence of substantial levels of HIV infection among injecting drug users and female sex workers1 may have limited the spread of HIV into the heterosexual population. Despite the fact that the proportion of HIV diagnoses attributed to heterosexual contact has increased in recent years, homosexual contact remains the exposure category for about 85% of new HIV diagnoses.1 Australia's Indigenous people are not a homogeneous group. There are many hundreds of language groups and a wide diversity of cultural, social, economic and geographical settings within and between Indigenous Australian communities. Most Indigenous Australians suffer a higher burden of illness and die at a younger age than non-Indigenous Australians for almost every type of disease or condition for which information is available.10 Indigenous Australians are more likely to have lower annual incomes, are less likely to have qualifications beyond secondary school,11 and are 15 times more likely to be imprisoned than non-Indigenous Australians.11 These factors, combined with the remote locations in which many Indigenous Australians live and the resulting poor access to health services, contribute to their vulnerability to sexually transmissible infections.12 Associations in other industrialised countries between socioeconomic disadvantage and HIV transmission from heterosexual exposure and injecting drug use13 highlight the need to provide HIV prevention services which reach all sectors of society. The higher proportion of Indigenous people with HIV in rural areas should alert policymakers to the need for access to culturally appropriate health services in these locations. Likewise, the higher proportion of Indigenous people with HIV infection who are women, who report heterosexual exposure only and who inject drugs shows a need for broadly focused HIV prevention programs. This demographic pattern, the relatively stable level of HIV diagnoses in Indigenous people, and the continuing high rates of other sexually transmissible infections among some Indigenous communities,1 highlight the need to strengthen both sexual health and harm-minimisation strategies for Indigenous Australians. Following the recommendations of the Evaluation of the Third National HIV/AIDS Strategy, several measures have been implemented in an attempt to reduce the higher rates of sexually transmissible infections among Indigenous Australians and the associated risk of HIV infection. These include the establishment of an Indigenous Australians' Sexual Health Working Party and the subsequent implementation of the National Indigenous Australians' Sexual Health Strategy 1996-97 to 1998-99, which proposed a comprehensive approach to HIV prevention through a range of strategies considering treatment and care, partnership agreements and a properly resourced workforce.14 In particular, the Strategy emphasises the need for access to primary care services for communities without adequate facilities for diagnosing and treating sexually transmissible infections and the provision of information on reducing the risk of acquisition. Strategies aimed at the underlying causes of low socioeconomic status, low levels of education and low levels of employment must also be employed in order to reduce the risk of transmission of HIV and other sexually transmissible infections in Indigenous Australians. Acknowledgements The National Centre in HIV Epidemiology and Clinical Research (NCHECR) is funded by the Commonwealth Department of Health and Aged Care. We would like to acknowledge the valuable input and feedback received from the National Australian Indigenous Sexual Health Working Party during the drafting of this article. We also thank Ms Yueming Li for statistical analyses, Ms Patty Correll (NCHECR) for her assistance in extracting data, and Ms Suzanne Blogg (National Centre for Epidemiology and Population Health [NCEPH]) for her guidance and assistance. We thank the doctors who reported cases of newly diagnosed HIV infection and AIDS under national surveillance procedures, and the National HIV Surveillance Committee for their collaboration. The National HIV Surveillance Committee comprises Ms Irene Passaris (ACT), Mr Robert Menzies (NSW), Dr Jan Savage (NT), Dr Hugo Ree (QLD), Ms Therese Davey (SA), Mr Neil Cremasco (TAS), Ms Cathy Keenan (VIC), Dr Gary Dowse (WA), Professor John Kaldor (NCHECR), and Ms Ann McDonald (NCHECR). References Commonwealth Department of Human Services and Health. Valuing the past -- investing in the future. Evaluation of the National HIV/AIDS Strategy 1993-94 to 1995-96. Canberra: AIDS/Communicable Diseases Branch, CDHSH, 1995. National Health and medical Research Council. Guidelines on ethical matters in Aboriginal and Torres Strait Islander health research. Canberra: NHMRC, 1991. McDonald AM, Crofts N, Blumer CE, et al. The pattern of diagnosed HIV infection in Australia, 1984-1992. AIDS 1994; 8: 513-519. Kaldor J, McDonald AM, Blumer CE, et al. The acquired immunodeficiency syndrome in Australia: incidence 1982-1992. Med J Aust 1993; 158: 10-17. Australian Bureau of Statistics. Population distribution, Indigenous Australians. Canberra: ABS 1997. (Catalogue No. 4705.0.) McDonald AM, Ryan J, Brown PR, et al. HIV prevalence at reception into Australian prisons, 1991-1997. Med J Aust 1999; 171: 18-21. Neilson G, Hill PS. Human immunodeficiency virus notifications for Aborigines and Torres Strait Islanders in Queensland. Med J Aust 1993; 158: 155-157. MacDonald M, Wodak A, Ali R, et al. HIV prevalence and risk behaviour in needle exchange attenders: a national study. Med J Aust 1997; 166: 237-240. O'Connor CC, Berry G, Rohrsheim R, et al. Sexual health and use of condoms among local and international sex workers in Sydney. Genitourin Med 1996; 72(1): 4-51. Australian Bureau of Statistics. The health and welfare of Australia's Aboriginal and Torres Strait Islander peoples, 1997. Canberra: ABS, 1997. (Catalogue No. 4704.0.) Office of the Aboriginal and Torres Strait Islander Social Justice Commissioner. Indigenous deaths in custody 1989 to 1996. Sydney: Human Rights and Equal Opportunity Commission, October 1996. Fairley CK, Bowden FJ, Gay NJ, et al. Sexually transmitted diseases in disadvantaged Australian communities. JAMA 1997; 278: 117-118. Centers for Disease Control and Prevention. HIV/AIDS Surveillance Report 1998; 10 (No. 2): 1-43. ANCARD Working Party on Indigenous Australians' Sexual Health, Commonwealth Department of Health and Family Services. The National Indigenous Australians' Sexual Health Strategy, 1996-1997 to 1998-1999. Canberra: CDHFS, 1997. Authors' details National Centre in HIV Epidemiology and Clinical Research, Sydney, NSW. Jillian A Guthrie, BA, MAE (Indigenous Health) also at National Centre for Epidemiology and Population Health, Australian National University, Canberra, ACT. Gregory J Dore, FRACP, MPH, Lecturer. Ann M McDonald, MPH, Coordinator, National HIV/AIDS Surveillance. John M Kaldor, PhD, Professor; and Head, Epidemiology Unit. Reprints will not be available from the authors. Correspondence: Professor J M Kaldor, National Centre in HIV Epidemiology and Clinical Research, Level 2, 376 Victoria Street, Darlinghurst, NSW 2010. jkaldorATnchecr.unsw.edu.au 2: Newly diagnosed HIV infection and AIDS, 1992-1998, by Indigenous status and selected characteristicsHIV diagnoses IndigenousNon-IndigenousP Odds ratio (95% CI)Total casesn=127n=5186Males93 (73.2%)4726 (91.1%)<0.0010.27 (0.17-0.41)Median age (years)3033Median CD4 cell count4844000.10Newly acquired HIV*24 (18.9%) 930 (17.9%)0.79Late HIV diagnosis?---HIV exposure categoryn=120?n=4507?Male homosexual contact56 (46.7%)3382 (75.0%) <0.0010.29 (0.20-0.43)Male homosexual contact and injecting drug use 12 (10.0%)191 (4.2%)0.0022.51 (1.29-4.78)Injecting drug use6 (5.0%)176 (3.9%)0.3Heterosexual contact only44 (36.7%)689 (15.3%)<0.0013.21 (2.16-4.77)Receipt of blood/tissue0 (0.0)34 (0.8%)0.4Mother-to-child transmission2 (1.7%)35 (0.8%)0.2Other/Undetermined7679 AIDS diagnoses IndigenousNon-IndigenousPOdds ratio (95% CI)Total casesn=55n=3583Males43 (78.2%)3411 (95.2%)<0.0010.18 (0.09-0.37)Median age (years)3237<0.001Median CD4 cell count90600.71Newly acquired HIV*---Late HIV diagnosis?13 (23.6%)675 (18.8%)0.42HIV exposure catergoryn=52?n=3405?Male homosexual contact26 (50.0%)2783 (81.7%)<0.0010.22 (0.12-0.38)Male homosexual contact and injecting drug use7 (13.5%)167 (4.9%)0.0162.95 (1.20-6.93)Injecting drug use1 (1.9%)127 (3.7%)0.4Heterosexual contact only17 (32.7%)238 (7.0%)<0.0016.28 (3.33-11.75)Receipt of blood/tissues0 (0.0)76 (2.2)0.3Mother-to-child transmission1 (1.9%)15 (0.4%)0.2Other/Undetermined3178 *A negative or indeterminate HIV antibody test result or a diagnosis of HIV seroconversion illness within 12 months of HIV diagnosis. ?HIV infection newly diagnosed within three months of AIDS diagnosis. ?The "other/undetermined" category was excluded from the calculation of the percentage of cases attributed to each HIV exposure category.

Jillian A Guthrie · Gregory J Dore · Ann M McDonald · John M Kaldor

Abstract Cardiovascular risk factors are commonly associated with erectile dysfunction and should be identified and treated. Patients with cardiovascular diseases should be assessed and counselled regarding their fitness for sexual activity. The danger of concurrent use of sildenafil and nitrates under any circumstances, regardless of age and sex, must be highlighted at all levels of the community. Sildenafil is absolutely contraindicated in patients receiving treatment with long-acting nitrates for ischaemic heart disease. Patients who need sublingual short-acting nitrates infrequently should not be precluded from taking sildenafil, provided they are aware that sildenafil is not to be taken within 24 h of taking the nitrate. There has been concern about the use of sildenafil (Viagra; Pfizer) for the treatment of erectile dysfunction (ED), particularly with regard to its possible role in the reported deaths and other serious cardiovascular events. Although sildenafil attracted considerable free media publicity in its debut in Australia and ranks as the most publicised new product this decade,1,2 consumer interest has been subdued and partly overshadowed by reports of 130 deaths involving sildenafil users in the United States between late March and mid-November 1998.3 It is therefore important that the association between sildenafil and these deaths be examined critically, so that the nature and the degree of risk may be identified and proper guidelines may evolve for the use of sildenafil. Cardiovascular disease and erectile dysfunction Cardiovascular disease and ED are known to be associated. In the Massachusetts Male Ageing Study,4 moderate or complete ED was 31% more prevalent among people with heart disease than in an age-matched cohort without heart disease. In a study in Perth, WA, the prevalence of complete ED among patients with hypertension, ischaemic heart disease and peripheral vascular disease was 26%, 38% and 57%, respectively, compared with 18.6% for the whole study.5 Reported ED in patients hospitalised for myocardial infarct or coronary artery surgery is of the order of 57%-64%.6,7 Conversely, in patients with severe ED, there is a 16% risk of severe, clinically occult ischaemic heart disease.8 Indeed, a statistically significant correlation has been shown between ED and the number of occluded coronary vessels.9 A significant number of patients requesting treatment for ED will have known or undiagnosed ischaemic heart disease, leading to considerable potential for adverse cardiovascular events. Moreover, many medications used for the treatment of cardiovascular disease may aggravate ED or complicate its treatment.10 Sildenafil and erectile dysfunction Sexual stimulation leads to the release of nitric oxide in the corpus cavernosum and results in an increase of cyclic guanosine monophosphate (cGMP), which produces smooth muscle relaxation and increased blood flow. Sildenafil is a selective inhibitor of cGMP-specific type 5 phosphodiesterase (PDE), the enzyme responsible for the degradation of cGMP in the corpus cavernosum. Thus, it enhances the effects of cGMP and permits an erectile response to be achieved or sustained (Box 1). The relevant pharmacodynamic and pharmacokinetic characteristics of sildenafil are summarised in Box 2.11-13 The efficacy of sildenafil in the treatment of ED has been demonstrated in 21 randomised, double-blind, placebo-controlled trials involving more than 3000 patients aged 19-87 years with ED of various aetiology.11 Sildenafil has been studied in men with ischaemic heart disease and with a wide range of other risk factors.14 A low incidence of serious or clinically significant adverse events, including cardiovascular events, was reported, comparable to that in patients with no known history of cardiovascular disorders. In Phase II-III studies involving 349 placebo patient-years and 693 sildenafil patient-years in randomised studies and 4220 patient-years in open-label studies, the incidence of myocardial infarction was lower in the sildenafil group than in the placebo group, although the difference was not statistically significant. The incidence of adverse events attributable to lowering of blood pressure in patients taking sildenafil was also low and no higher than in those receiving placebo. It was similar in patients taking concomitant antihypertensives and in those not taking these medications.14 There were reports of 26 deaths in about 5000 sildenafil patient-years, including 14 people with myocardial infarction and sudden death. None of the deaths was considered to be treatment related.14 Sildenafil and adverse cardiovascular events The unprecedented hype generated by the launch of sildenafil in the United States was dampened by reports of cardiovascular events, including deaths, allegedly associated with its use. A summary of reports of death among sildenafil users was posted by the Food and Drug Administration (FDA), with the pertinent remark that, in interpreting these reports, consideration should be given to the limitations of spontaneous reporting, such as under-reporting, duplication, marketing and medicolegal factors, incomplete or inaccurate clinical information, and the assumption of a cause-effect relationship.3 An overview of the FDA's updated summary of 130 reports of death between late March and mid-November 1998 is shown in Box 3. Deaths have also been reported in the Netherlands15 and Australia.16 Did sildenafil cause or contribute to the reported deaths? Sexual activity remains a potential trigger for myocardial infarction and sudden death may result from ischaemia or arrhythmia, although the relative and absolute risks are apparently low (Box 4). In the US general population with an age distribution similar to that of sildenafil users, there are about 400 deaths per million per week, and about 150 of these have a cardiovascular cause.29 In Australia, there are about 250 myocardial infarctions (50 fatal) per week affecting men aged 35 to 69 years.30,31 More than 70% of the deceased subjects in the FDA summary had overt or occult cardiovascular disease. Considering the high prevalence of risk factors for sudden cardiac death in users of sildenafil, the reported 130 deaths need to be viewed in the context of patient exposure to about 50 million sildenafil tablets, or more than 6 million prescriptions, during the same period. Pharmacologically, the action of sildenafil as a type 5 PDE inhibitor is highly specific. Potential for disaster seems to lie in the concurrent use of sildenafil and organic nitrates, as sildenafil potentiates the effect of nitrates and may lead to life-threatening hypotension (Box 1). Among the deaths reported in the FDA summary, there were 16-19 sildenafil users who had allegedly used or received glyceryl trinitrate or a nitrate-containing medication. In this subset of individuals, the concurrent use of nitrates would provide the possible causal link between sildenafil and death. Therefore, it is not unlikely that sexual activity in a vulnerable person and adverse drug interaction caused or contributed to the reported deaths. Recommendations Box 5 shows the recommended strategy for treating ED. To minimise adverse consequences, it is important that a patient's fitness for sexual and physical activity be assessed when treatment of ED is considered, and that the patient be appropriately counselled if sexual activity is inadvisable. In general, sexual intercourse should be safe if a patient can perform an activity equal to 5-6 metabolic equivalents (METS), such as climbing 20 stairs in 10-15 seconds without distress.32 Postinfarction patients who reach 5-6 METS on stress testing without ischaemia or arrhythmia can resume their normal sexual activity without risk.33 In one study, patients with a negative exercise test result did not demonstrate ischaemia on Holter monitoring during sexual intercourse.26 The frequent use of short-acting nitrates and ongoing therapy with long-acting nitrates are absolute contraindications to the use of sildenafil. The only option for patients in this situation is to avoid the use of sildenafil. A policy of refraining totally from the use of sildenafil in all patients receiving nitrates in whatever form and regardless of frequency would, of course, quarantine patients from the risk of drug interaction. Nevertheless, patients who have only an infrequent need for short-acting nitrates, such as sublingual glyceryl trinitrate, should not be precluded from the use of sildenafil. However, doctors need to ensure that patients fully understand the implications of the potential interaction of these therapies. Patients whose only exposure to nitrate therapy is infrequent use of sublingual glyceryl trinitrate tablets or spray should be advised that at least 24 hours from the last use of the short-acting nitrates should be allowed to elapse before the use of sildenafil. It is not definitely known when nitrates can be safely administered after a dose of sildenafil. Certainly, patients should be cautioned against the use of any form of nitrates for at least 24 hours after sildenafil. In elderly patients, and in those with hepatic and renal impairment or receiving medications which may inhibit the cytochrome P450 3A4 isoenzyme (eg, erythromycin, fluconazole, fluoxetine, cimetidine), consideration must be given to decreased sildenafil clearance. If a patient should develop angina within 24 hours of taking sildenafil, nitrates should be totally avoided. Doctors should ensure that their patients follow this advice carefully. If necessary, an increase in the dose of alternative anti-anginal agents should be considered. If a patient requires hospital admission, non-nitrate anti-anginal preparations such as ß-blockers or calcium-channel blockers can be used with due regard to the risk of hypotension. If complications should develop from the inadvertent concurrent use of a nitrate, the recommendations of the American College of Cardiology and the American Heart Association34 should be followed. These include resuscitative measures such as fluid infusion and judicious use of intravenous vasopressors to maintain blood pressure, as well as appropriate non-nitrate anti-anginal agents. If adherence to such guidelines is difficult, the alternative is to avoid sildenafil in favour of other therapeutic options for ED. Clinical judgement and discretion must prevail over generalisation, bearing in mind the risk and benefit and the priority of the therapeutic interventions involved. There are no established data on the safety and efficacy of sildenafil in patients with myocardial infarction or life-threatening arrhythmia within the preceding six months; patients with systolic blood pressure < 90 mmHg; or patients with cardiac failure or coronary artery disease causing unstable angina. Caution must be exercised in prescribing sildenafil for these patients and for patients receiving complicated multidrug antihypertensive therapy.34 Where appropriate, monitoring of blood pressure at the initiation of sildenafil therapy would identify patients with a hypotensive response to sildenafil. An appropriate educational program will reduce the risk of drug interaction by alerting pharmacists and doctors to the potential danger. Pharmaceutical companies marketing nitrate-containing medications should specify sildenafil as a contraindication in their product information. Paramedics, nursing staff, patients and the community at large should be instructed on the danger of the concurrent use of sildenafil and nitrates, and of the undesirable practice of sharing medication with relatives and friends (Box 6). It is important that a warning be given to every patient, regardless of sex or age. Women have been known to use sildenafil to heighten sexual arousal and young people may use amyl nitrite inhalation for recreational pursuit. Cardiovascular diseases affect an estimated 2.3 million Australians.35 The risk of angina increases with age, affecting 12.4% of men and 11.7% of women aged 65-69 years.36 These proportions are likely to be greater among patients with ED. A request for treatment of ED provides a window of opportunity to assess the patient for cardiovascular and other risk factors, including diabetes and abnormal lipid profile. Cardiovascular risk factors, if identified, should be vigorously treated according to established guidelines.37,38 The sildenafil controversy continues.39 It has been reported that the FDA continues to believe that sildenafil remains safe.39 However, continuing postmarketing surveillance is essential for sildenafil, as for any recently marketed drug. Disclosure B G A Stuckey has served as a principal investigator in clinical studies of sildenafil. K K Chew and B G A Stuckey have received sponsorship to attend conferences from Pfizer Aust Pty Ltd. P L Thompson serves on the international steering committee of a Pfizer-sponsored clinical trial of lipid lowering in the elderly. K K Chew, B G A Stuckey and P L Thompson have been invited to present papers at Pfizer-sponsored meetings. References Kiely M. Viagra saturates media. Marketing Globe. Marketing 1998; Dec: 58. Jones A. When the thrill has gone. Business Rev Weekly 1999; June 25: 90-95. US Department of Health, Food and Drug Administration. Postmarketing safety of sildenafil citrate (Viagra) and summary of reports of death in Viagra users received from marketing (late March through mid-November 1988). 24 November 1988. Feldman HA, McKinlay JB, Goldstein I, Longcope C. Erectile dysfunction, cardiovascular disease and cardiovascular risk factors: prospective results in a large random sample of Massachusetts men. J Urol 1998; 159 Suppl 5: 91 abstract 347. Chew KK, Earle CM, Stuckey BGA, et al. Erectile dysfunction in general medical practice: prevalence and clinical correlates. Int J Impot Res 2000; 12: 1-5. Wabrek AJ, Burchell RC. Male sexual dysfunction associated with coronary heart disease. Arch Sex Behav 1980; 9: 69-75. Gundle MJ, Reeves BR, Tate S, et al. Psychosocial outcome after aortocoronary artery surgery. Am J Psychiatry 1980; 137: 1591-1594. Anderson M, Nicholson B, Louie E, Mulhall JP. An analysis of vasculogenic erectile dysfunction as a potential predictor of occult cardiac disease. J Urol 1998; 159 Suppl 5: 30 abstract 118. Greenstein A, Chen J, Miller H, et al. Does severity of ischaemic coronary disease correlate with erectile function? Int J Impot Res 1997; 9: 123-126. Slag MF, Morley JE, Elson MK, et al. Impotence in medical clinic outpatients. JAMA 1983; 249: 1736-1740. Morales A, Gingell G, Collins M, et al. Clinical safety of sildenafil citrate (Viagra) in the treatment of erectile dysfunction. Int J Impot Res 1998; 10: 69-74. Viagra -- approved product information. Pfizer, 1998. Boolell M, Allen MJ, Ballard SA, et al. Sildenafil: an orally active type 5 cyclic GMP-specific phosphodiesterase inhibitor for the treatment of penile erectile dysfunction. Int J Impot Res 1996; 8: 47-52. Zusman RM, editor. Cardiovascular data on sildenafil citrate. Am J Cardiol 1999; 83(5A). Feenstra J, van Drie-Pierik RJHM, Lacle CF, Stricker BHC. Acute myocardial infarction associated with sildenafil. Lancet 1998; 352: 957-958. Viagra is here! Australian Adverse Drug Reactions Bulletin 1998; 17(4). Hellerstein HK, Friedman EH. Sexual activity and the postcoronary patient. Arch Intern Med 1970; 125: 987-999. Bohlen Y, Held JP, Sanderson MO, Paterson RP. Heart rate, rate-pressure product, and oxygen uptake during four sexual activities. Arch Intern Med 1974; 144: 1745-1748. Willich SN, Klatt S, Arntz HR. Circadian variation and triggers of acute coronary syndromes. Eur Heart J 1998; 19 Suppl C: C12-23. Nalbangtil I, Yigitbasi O, Kiliccioglu B. Sudden death in sexual activity. Am Heart J 1976; 91: 405-406. Ueno M. The so-called coital death. Jpn J Legal Med 1963; 17: 330-340. Renshaw DC, Karstaedt A. Is there (sex) life after coronary bypass? Comp Ther 1988; 14: 61-66. Lecomte D, Fornes P, Nicolas G. Stressful events as a trigger of sudden death: a study of 43 medico-legal autopsy cases. Forensic Sci Int 1996; 79: 1-10. Muller JE, Mittleman MA, Maclure M, et al. Triggering myocardial infarction by sexual activity. JAMA 1996; 275: 1405-1409. Johnston BL, Fletcher GF. Dynamic electrocardiographic recording during sexual activity in recent post-myocardial infarction and revascularization patients. Am Heart J 1979; 98: 736-741. Drory Y, Shapira I, Fisman EZ, Pines A. Myocardial ischaemia during sexual activity in patients with coronary artery disease. Am J Cardiol 1995; 75: 835-837. Kavanagh T, Shephard RJ. Sexual activity after myocardial infarction CMAJ 1977; 116: 1250-1253. Paolillo V, Marra S, Spadaccini F, Angelino PF. Dynamic electrocardiographic recording during sexual activity in recent post-myocardial infarction and revascularization patients [letter]. Am Heart J 1980; 100: 763. Health United States. 1998. Hyattsville, Maryland: US National Center for Health Statistics, 1998. Australian Institute of Health and Welfare. Heart, stroke and vascular diseases, Australian facts. Canberra: AIHW and the Heart Foundation of Australia, 1999. (Cardiovascular Disease Series No. 10. AIHW Cat. No. CVD 7.) Australian Bureau of Statistics. Causes of death, Australia. Canberra: ABS, 1997. (Cat No. 3303.0.) Cardiac rehabilitation: sex after a heart attack. In: Zaret BL, Moser M, Cohen LS, editors. Yale University School of Medicine Heart Book. New York: Hearst Books, 1992; 351. Tardif GS. Sexual activity after a myocardial infarction. Arch Phys Med Rehabil 1989; 70: 763-766. Summary statement of the American College of Cardiology and the American Heart Association on the use of sildenafil (Viagra) in patients at clinical risk from cardiovascular effects. 10 August 1998. <http://www.americanheart.org/ Whats_News/AHA_Science_Advisories/viagra.html>. Accessed 18 February 2000. Fact sheet for prevention of myocardial infarction. National Heart Foundation of Australia. Curr Ther 1999; 39: 52. Fact sheet for angina. National Heart Foundation of Australia. Curr Ther 1999; 39: 63. Grundy SM, Balady GJ, Criqui MH, et al. Guide to primary prevention of cardiovascular diseases. A statement for healthcare professionals from the Task Force on Risk Reduction. Circulation 1997; 95: 2329-2331. Heart Foundation of Australia. Guide for the use of lipid lowering drugs in adults. Canberra: Heart Foundation of Australia, 1999. Available at <http://www. heartfoundation.com.au/include/defaultStory.asp?OwnerUID=200&Content Type=tblcategory>. Mitka M. Some men who take Viagra die -- why? [news]. JAMA 2000; 283(5). <http://jama.ama-assn.org/issues/v283n5/full/jmn0202-2.html>. Accessed 18 February 2000. (Received 8 Oct 1999, accepted 14 Feb 2000) Authors' details Keogh Institute for Medical Research, Perth, WA. K Kim Chew, FRCP(Edin), FRCP(Glas), Senior Clinical Fellow; Bronwyn G A Stuckey, MB BS, FRACP, Medical Director, and Consultant Endocrinologist, Department of Endocrinology and Diabetes, Sir Charles Gairdner Hospital, Perth, WA. University of Western Australia, Perth, WA. Peter L Thompson, FRACP, FACP, Clinical Professor of Medicine, University of Western Australia, and Consultant Cardiologist, Sir Charles Gairdner Hospital, Perth, WA. Reprints will not be available from the authors. Correspondence: Dr B G A Stuckey, Keogh Institute for Medical Research, 3rd Floor A Block, Queen Elizabeth II Medical Centre, 2 Verdun Street, Nedlands, WA 6009. rmriATwt.com.au 2: Pharmacodynamic and pharmacokinetic characteristics of sildenafil11-13 Sildenafil is about 4000-fold more selective for type 5 phosphodiesterase (PDE5) than for PDE3, which is involved in cardiac contractility. In healthy volunteers, sildenafil produced a modest decrease in blood pressure (up to 8.4 mmHg systolic and 5.5 mmHg diastolic), but no consistent orthostatic effects and no clinically relevant electrocardiographic changes. In patients receiving medications containing nitrates, the hypotensive effects of sildenafil can be severe. In a US study with isosorbide mononitrate 20 mg twice daily, 50 mg sildenafil produced maximal blood pressure reductions of 40.9 mmHg systolic sitting and 51.6 mmHg systolic standing, and 25.8 mmHg diastolic sitting and 29.3 mmHg diastolic standing, about one hour after dosing and lasting up to 6 h. Similar haemodynamic interaction occurred for about two hours after a dose of 500 µg sublingual glyceryl trinitrate (maximum reductions: systolic, 36.0 mmHg sitting; diastolic, 20.5 mmHg sitting). Sildenafil has a half-life of about 4 h. After an oral dose of 100 mg, the plasma level peaks at about 440 ng/mL within 30-120 min, and drops to about 2 ng/mL at 24 h. Sildenafil is predominantly metabolised in the liver by the cytochrome P450 3A4 system, and 18% of the dose is excreted in the urine. Increased plasma levels may result from concomitant use of a cytochrome P450 3A4 inhibitor (eg, erythromycin, fluconazole, fluoxetine, cimetidine) or reduced clearance in elderly persons (>65 years) and in those with significant hepatic and renal impairment (creatine clearance <30 mL/min). 3: US Food and Drug Administration summary of reports of death in sildenafil users Number of deaths reported130Number with cause not mentioned or unknown48Number from homicide or drowning2Number from stroke3Number from cardiovascular events Definite or suspected myocardial infarction Cardiac arrest Cardiac symptoms Coronary artery disease77 41 27 6 3Use of nitrates Took or were administered a nitrate medication Found with nitrate in their possession16 3Time of death after use of sildenafil Not stated or unknown Within 4-5 hours of using sildenafil (includes 27 during or immediately after sexual intercourse) Later the same day Next day Two days later Three to seven days later61 44 6 8 5 4Cardiovascular risk factors One or more risk factors No identified risk factors, but severe coronary artery disease found at autopsy No history of cardiac disease or risk factors No risk factors and no sexual activity Not specified 90 3 12 2 23 4: Sexual activity and adverse cardiovascular events Sexual activity, like any other physical effort, increases cardiac work and myocardial oxygen demand.17 Heart rate and blood pressure rise to an energy expenditure of 2.0-5.4 metabolic equivalents.18 Sexual activity may trigger an adverse cardiovascular event.19,20 Coital death has been reported to account for 0.6% of sudden deaths,21 although it is said to be rare in a stable sexual relationship.22 In a study of 43 cases, sudden death was found to occur primarily in patients with severe heart disease, especially coronary heart disease, and in only three cases was sexual activity involved.23 In another study, 9% of patients reported having had sexual activity in the 24 hours, and 3% in the two hours, preceding myocardial infarction. The relative risk of myocardial infarction occurring in the two hours after sexual activity was estimated to be 2.5, and was not increased in patients with a history of previous angina pectoris or myocardial infarction. The absolute risk increase was low, at one chance in a million for a healthy individual.24 Electrocardiographic abnormalities during sexual activity have been reported in 12 of 24 patients with recent myocardial infarction or revascularisation.25 In another study, 31% of men with ischaemic heart disease had ischaemia on Holter monitoring during sexual intercourse, although only 7% were symptomatic.26 However, the frequency of angina pectoris and ventricular premature beats is less during sexual intercourse than during standard laboratory exercise, and sexual relations are thought to carry no special risk for the average postinfarction patient.27 Sexual activity in the early posthospital phase of myocardial infarction is not a stronger stimulus than other activities for cardiac electrical instability.28 Anxious sexual preoccupation, frustration and avoidance may actually be greater risk factors than coitus or coital alternatives.22

Peter L Thompson

Introduction Concern about cannabis use is not new: ordinances were passed in 8th-century Egypt prohibiting use of hemp drugs, transgressors being subject to tooth extraction!1 In the 19th century, controversy over widespread cannabis use led to the Indian Hemp Drugs Commission, a large-scale investigation into the health effects (both physical and mental) of cannabis use. The Commission's wide-ranging report found that the relationship between cannabis and "mental injury" was complex, and criticised many witnesses for their obvious bias.2 Social, moral and political agendas continue to influence both cannabis research and the "cannabis debate".3 Are there grounds for concern? Cannabis is now widely used by young people in Australia, mostly intermittently. However, 7% of 17-year-old girls and 11% of boys the same age use it at least weekly.4 Furthermore, the concentration of the primary psychoactive component of cannabis, δ-9-tetrahydrocannabinol, may have increased in recent years due to hydroponic cultivation and cross-breeding,5 although the evidence for this is patchy.3 The implications for the mental health of users are unclear; the more pressing issue may be the as yet unquantified impact of a declining age of initiation into cannabis use.6There are several potential psychological harms of cannabis use (Box). The link between cannabis and psychosis remains the most intriguing. Specifically, cannabis has been noted to cause psychotic-like symptoms during intoxication, to lead to a "cannabis psychosis", to increase the relative risk of schizophrenia, and to affect the clinical course of established schizophrenia.11 Population studies, such as the Epidemiologic Catchment Area study in the United States, confirm an association between cannabis use and psychotic symptoms. In this study, daily marijuana use over a year was associated with a 2.4-times greater risk of psychotic experiences, while any use was associated with a 1.3-times greater risk for self-reported psychotic experiences when compared with non-users. The relative risk for daily users remained significant after adjustment for other substance abuse and baseline psychiatric diagnoses.12 Is there a specific "cannabis psychosis?" The case for a specific acute "cannabis psychosis" is based largely on case reports and case series that link a history of cannabis use with a psychotic presentation. Two types of presentation are described: a "toxic psychosis", in which psychotic symptoms are associated with confusion. This generally occurs after ingestion of large amounts of cannabis in someone with no significant psychiatric history, and recovery is usually rapid with abstinence;13 and a functional psychosis occurring without confusion (exacerbation of schizophrenia, and schizophreniform psychoses). The case for a true functional psychosis caused by cannabis is more vexed than the case for a "toxic psychosis", with various authors describing a heterogeneous clinical picture, often with manic features.14 The better-designed studies suggest that it is, in fact, hard to delineate a specific functional cannabis psychosis. McGuire and colleagues examined all patients presenting with psychotic symptoms at two London hospitals, comparing those with positive and negative results on urinary cannabinoid screening.15 Using structured interviews, they found a similar spread of diagnoses and illness onset among case and control patients, with similar numbers of first admissions and symptom profiles. Cannabis and schizophrenia Does cannabis cause schizophrenia? Perhaps the more worrying question is whether cannabis causes chronic psychosis, particularly schizophrenia. The work of Andreasson and others examined this question in a cohort of male Swedish conscripts, followed up through a national psychiatric case register.16 They found that having used cannabis between one and 10 times at conscription increased the relative risk of schizophrenia to 1.3, the risk rising to 6.0 for those who had used cannabis on 50 or more occasions. However, this relative risk was reduced after adjustment for factors which independently contributed to the risk of schizophrenia. While this study provides some of the strongest evidence for a link between cannabis and psychosis, methodological concerns have been raised. These include the temporal gap between self-reported cannabis use at conscription and later schizophrenia, the potential confounding role of other substance use (particularly as amphetamines were a major drug of abuse during the study period), the adequacy of psychological assessment at conscription, and the reliability of self-reported drug use at conscription.3 Nevertheless, the association between cannabis use and schizophrenia is strengthened by studies which demonstrate that cannabis is widely used among people with schizophrenia. A recent study in Newcastle examined substance use in all outpatients with schizophrenia, finding 29.9% of subjects had some use of cannabis in their lifetime, with 7.7% and 28.3% of subjects having lifetime diagnoses of cannabis abuse and dependence, respectively.17Notably, alcohol was more commonly used than cannabis, while amphetamines were the third most commonly used substance. A number of hypotheses have been proposed to explain the prevalence of cannabis use in schizophrenia. Apart from the causation/precipitation role, it has been suggested that cannabis is used as self-medication for psychotic or dysphoric symptoms, or to ameliorate the side effects of antipsychotic drugs. Alternatively, the relationship may reflect the common peaks of onset of schizophrenia and cannabis use (particularly as prevalence samples are seldom compared with age- or sex-matched general population controls), or the role of underlying factors such as demographic differences. How might cannabis use affect established schizophrenia? Clinical intuition suggests that cannabis has an adverse effect on the clinical course, a view supported by some well-designed studies. Linszen and colleagues studied 93 subjects with schizophrenia prospectively over a year, finding a higher rate of relapse in the cannabis-users than in the non-users, with a differential risk of relapse according to level of cannabis use.18 This effect persisted after adjustment for age, sex, age at first hospital admission, and alcohol use. However, this study was limited by the failure to consider the role of polysubstance use, reliance on self-report alone, and the gross measure of compliance used. Data from the Epidemiologic Catchment Area study also support a relationship between an alcohol- or cannabis-use disorder and a higher risk for hospitalisation in those with schizophrenia over a year.19 However, in this study, substance use covaried with depressive rather than psychotic symptoms, suggesting a complex relationship between substance use and symptoms in schizophrenia. Confounding issues: The relationship between cannabis and psychosis is thus far from straightforward. Cannabis is rarely the only substance used. Other factors, such as personality, may confound the relationship. One study found higher rates of schizotypy in volunteer cannabis users, raising the possibility that the relationship between cannabis and psychosis is mediated by a premorbid "psychosis prone" personality.20 Mueser and colleagues have identified antisocial personality as a common factor underlying both schizophrenia and substance-use disorders.21 Methodological concerns, such as the failure to assess and control for use of other substances and reliance on the case-study method, make it difficult to draw firm conclusions in this area. A possible neurophysiological link Recent findings in neuroscience have lent credence to a link between cannabis and psychosis. An endogenous cannabinoid system has been identified, with one type of cannabinoid receptor (CB1) found in the hippocampus, associated cortical areas, cerebellum and basal ganglia.22 In rat brain, cannabinoid receptors have been shown to collocate with dopamine D1 receptors,23 and cannabis administration increases the activity of tyrosine hydroxylase (an enzyme linked to dopamine metabolism).24 These findings hint at the mechanism through which cannabis could lead to psychosis, by modulation of dopaminergic transmission. The discovery of cannabinoid receptors has also led to the discovery of endogenous cannabinoids, among them anandamide (from the Sanskrit word for "bliss"). Such discoveries have encouraged some researchers to speak of a "cannabinoid hypothesis" of schizophrenia, likening the cognitive deficits of schizophrenia to those induced temporarily by δ-9-tetrahydrocannabinol. Indeed, a recent study found higher levels of two endogenous cannabinoids in cerebrospinal fluid in 10 patients with schizophrenia compared with 11 non-psychotic control patients.25 The authors suggested this may represent the response of the cannabinoid system to dopamine imbalance, or may reflect an underlying pathogenic "hypercannabinergic" state. This intriguing result can be viewed as preliminary only, because of the small number of subjects and the paucity of information about possible confounding factors, particularly substance use. The role of cannabinoids is being investigated in other neurological conditions, including Huntington's chorea and Tourette's syndrome. Such theories, while appealing, remain purely speculative. A hundred years have passed since the Indian Hemp Drugs Commission. The research that has followed confirms an association between cannabis and psychosis, but the nature of the connection remains elusive. We clearly see a need for further carefully controlled, prospective clinical studies, as well as insights from neuroscience, to clarify whether the relationship is causal or, in fact, due to factors common to both cannabis and psychosis. However, in the meantime, we believe it is important to inform young people with psychosis of the possible impact of ongoing cannabis use on their symptoms, particularly on the risk of relapse. References Dhunjibhoy JE. A brief resume of the types of insanity commonly met with in India, with a full description of Indian "hemp insanity" peculiar to the country. J Ment Sci 1930; 76: 254-264. Mikuriya TH. Physical, mental and moral effects: The Indian Hemp Drugs Commission Report. Int J Addict 1968; 3: 253-270. Hall W, Solowij N, Lemon J. The health and psychological consequences of cannabis use. National Drug Strategy Monograph Series No. 25. Canberra: AGPS, 1994. Donnelly N, Hall W. Patterns of cannabis use in Australia. National Drug Strategy Monograph Series No. 27. Canberra: AGPS, 1994. Adams IB, Martin BR. Cannabis: pharmacology and toxicology in animals and humans. Addiction 1996; 91: 1585-1614. Greenfield SF, O'Leary G. Sex differences in marijuana use in the United States. Harv Rev Psychiatry 1999; 6: 297-303. Fergusson DM, Horwood LJ. Early onset cannabis use and psychosocial adjustment in young adults. Addiction 1997; 92: 279-296. Yamaguchi K, Kandel DB. Patterns of drug use from adolescence to adulthood. II. Sequences of progression. Am J Public Health 1984; 74: 668-672. Solowij N. Cannabis and cognitive functioning. Cambridge: Cambridge University Press, 1998. Thornicroft G. Cannabis and psychosis. Is there epidemiological evidence for an association? Br J Psychiatry 1990; 157: 25-33. Hall W. Cannabis use and psychosis. Drug Alcohol Rev 1998; 17: 433-444. Tien AY, Anthony JC. Epidemiological analysis of alcohol and drug use as risk factors for psychotic experiences. J Nerv Ment Dis 1990; 178: 473-480. Bernhardson G, Gunne L-M. Forty-six cases of psychosis in cannabis abusers. Int J Addict 1972; 7: 9-16. Thacore VR, Shukla SRP. Cannabis psychosis and paranoid schizophrenia. Arch Gen Psychiatry 1976; 33: 383-386. McGuire P, Jones R, Harvey I, et al. Cannabis and acute psychosis. Schizophr Res 1994; 13: 161-168. Andreasson S, Allebeck P, Engstrom A, Rydberg U. Cannabis and schizophrenia. Lancet 1987; 2: 1483-1486. Fowler IL, Carr VJ, Carter NT, Lewin TJ. Patterns of current and lifetime substance use in schizophrenia. Schizophr Bull 1998; 24: 443-455. Linszen DH, Dingemans PM, Lenior ME. Cannabis abuse and the course of recent onset schizophrenic disorders. Arch Gen Psychiatry 1994; 51: 273-279. Cuffel BJ, Chase P. Remission and relapse of substance use disorders in schizophrenia. Results from a one-year prospective study. J Nerv Ment Dis 1994; 182: 342-348. Williams JH, Wellman NA, Rawlins JNP. Cannabis use correlates with schizotypy in healthy people. Addiction 1996; 91: 869-877. Mueser KT, Rosenberg SD, Drake RE, et al. Conduct disorder, antisocial personality disorder and substance use disorders in schizophrenia and major affective disorders. J Stud Alcohol 1999; 60: 278-284. Herkenham M, Lynn AB, Little MD, et al. Cannabinoid receptor localization in brain. Proc Natl Acad Sci U S A 1990; 87: 1932-1936. Herkenham M. Cannabinoid receptor localization in brain: relationship to motor and reward systems. Ann N Y Acad Sci 1992; 654: 19-32. Hernandez ML, Garcia-Gil L, Berrendro F, et al. Æ-9-tetrahydrocannabinol increases activity of tyrosine hydroxylase in cultured fetal mesencephalic neurons. J Mol Neurosci 1997; 8: 83-91. Leweke FM, Giuffrida A, Wurster U, et al. Elevated endogenous cannabinoids in schizophrenia. Neuroreport 1999; 10: 1665-1669. Authors' details Manly Hospital, Sydney, NSW. Diana R McKay, FRANZCP, Psychiatrist. University of Sydney Academic Psychiatry, Royal North Shore Hospital, Sydney, NSW. Christopher C Tennant, MD, FRANZCP, Professor of Psychiatry. Potential psychological harms of cannabis use Cannabis has been associated with: early school leaving (the strength of the association lessening after adjusting for cannabis users' poor school performance before cannabis use);7possible progression to other illicit drug use;8an "amotivation" syndrome (now felt to represent chronic intoxication in heavy long-term users); dependence;3subtle impairments in information processing;9 and psychosis.10

Diana R McKay · Christopher C Tennant

Abstract In 1995, the National Health and Medical Research Council (NHMRC) announced its commitment to developing evidence-based clinical practice guidelines "to promote best practice linked to outcomes and effective cost management". To date, resources for dissemination and implementation have been identified for only two guidelines developed by the NHMRC as part of that commitment. Clinical practice guidelines for the management of men with lower urinary tract symptoms (LUTS) were launched in 1997. We were members of the working party that developed these guidelines, and here we give our personal account, offering insights into the tensions and contradictions impeding the translation of political commitment to evidence-based medicine into policy and practice. Introduction For at least a decade, there has been vocal support in Australia for the concept of evidence-based medicine and its most obvious progeny, clinical practice guidelines (CPGs). Yet, we perceive a "healthy cynicism" is emerging. Perhaps too much has been inferred from weak evidence. CPGs have been perceived as "cook book" medicine. Health and medical research might be criticised for focusing more on "academic" outputs than on generating evidence relevant to the needs of practising doctors. As participants in an initiative to develop evidence-based CPGs -- Clinical practice guidelines for the management of uncomplicated lower urinary tract infections in men1 -- under the auspices of the National Health and Medical Research Council (NHMRC), we describe our experiences in the light of these concerns. What is LUTS? Lower urinary tract symptoms (LUTS) in men are very common. Nearly 40% of men aged over 45 years admit to having some lower urinary tract symptoms, and this prevalence increases with age.2 For many years, the public has been encouraged to equate symptoms with disease. Given media interest in prostate cancer and the consequent community awareness, it is also not surprising that Australian men worry that LUTS might signal malignancy.3,4LUTS was previously known as BPH (benign prostatic hyperplasia or benign prostatic hypertrophy) or prostatism, implying that the underlying cause of the symptoms was a disorder of the prostate.5 However, as the prevalence of LUTS is the same for men and women in age-matched cohorts, it is plausible that LUTS may partly be explained by ageing. Invasive and expensive investigations, medical treatments and surgical procedures may have inadvertently reinforced the perception that LUTS is potentially life-threatening.3,4 Why develop evidence-based CPGs in Australia about LUTS, and how? In 1994, the United States Agency for Health Care Policy and Research (AHCPR) published CPGs entitled Benign prostatic hyperplasia: diagnosis and treatment.5 This evidence-based document heralded changes in fundamental concepts about LUTS in men and questioned accepted practice. As Australian guidelines seemed timely, the NHMRC saw merit in addressing this issue. In 1995, a 44-page document entitled Guidelines for the development and implementation of clinical practice guidelines was published by the NHMRC to help those developing CPGs.6 It advised that CPG working parties should be comprised of representatives of all relevant stakeholders, but gave no direction on how to select them. Instead, the NHMRC relied on the perspicacity of the stakeholder organisations to understand the complexity of guideline development and the likely controversies. The multidisciplinary members of the LUTS working party worked cohesively, even though their selection was more accidental than planned. The principal medical organisations with a legitimate interest in the management of men with LUTS are the Royal Australian College of General Practitioners (RACGP) and the Urological Society of Australasia (USA). The Consumers' Health Forum was also asked to nominate a representative. Evidence-based medicine emphasises the importance of consumer participation in health decision-making at all levels.7 Articulate, well-informed and effective consumer representation is crucial in the development of CPGs. Yet such representation raises many confronting questions. In our case, ought the consumer representative have LUTS? Could one individual reflect the vast and diverse needs and views of the affected consumer group? Might consumers self-nominate or be selected because they have an "axe to grind"? How might consumer representatives react to discrepancies between the evidence they access during the deliberations of the working party compared with previous treatment they may themselves have received? We also were concerned about conflict of interest. While the CPG development process should focus exclusively on the evidence, every member of the LUTS working party (including ourselves) potentially had a conflict of interest. Representatives of professional groups were expected to convey the concerns and preferences of their constituency. Yet the influence of pecuniary interests might be more subtle. For example, specialists may benefit financially from an aggressive approach to referral, investigations and treatments. Bureaucrats might put cost savings or short term political gain before outcomes. Searching for evidence to answer clinical questions An immediately apparent and perennial problem for the LUTS working party was the impoverished nature of the evidence available. There was very little evidence from well-designed randomised-controlled trials (level II evidence) or from meta-analyses of such trials (level I evidence). Having to consider evidence predominantly at lower levels of the NHMRC evidence taxonomy, the LUTS working party soon found this taxonomy too coarse to classify the nature of the evidence about LUTS treatments. At that time (1995), the NHMRC included within one broad band of evidence (classified as level IV) those "opinions of respected authorities, based on clinical experience, descriptive studies, or reports of expert committees".6 Of necessity, we devised a more detailed taxonomy whereby three subcategories of level IV evidence could be distinguished as follows: IV-1 Evidence from descriptive studies including case series, case reports and cross-sectional studies. IV-2 Published policies, recommendations or opinions of recognised experts, organisations or learned colleges, including endorsement of IV-3 evidence by recognised Australian bodies such as the RACGP and the Urological Society of Australasia. IV-3 Consensus opinion of the working party not endorsed formally by recognised bodies.1 The NHMRC has since excluded anecdote and expert opinion entirely from its more recent classification of evidence,8 under which level IV evidence now only includes case series (either post-test only or pretest-post-test). Furthermore, the NHMRC taxonomy had been designed to classify the evidence for treatments. Neither its original nor its revised taxonomy is appropriate for the appraisal of the evidence for diagnostic investigations and prognostic factors. Even the desirability of level I or II evidence can be challenged if it fails to pass the test of "clinical relevance". For example, randomised-controlled trials of the use of the drug finasteride for LUTS showed a statistically significant improvement in symptom score points from 2.5 to 2.8 in men treated with this drug compared with controls. As changes of at least three such points are required for men to experience a subjective sense of benefit in their quality of life, this evidence failed the "test of clinical relevance".1 To quote Gertrude Stein, "for a difference to be a difference it has to make a difference"! Another difficult task for the working party was making recommendations when published evidence was non-existent or inconclusive. Guided by Davidoff,9 the working party recognised two typical reactions in response. An interventionist approach assumes that patients are best served by providing a service for which benefit is as yet speculative, on the basis that treatment is better than no treatment. By contrast, a conservative approach demands a greater certainty of benefit, assuming there is always a risk of iatrogenic harm from intervention in the absence of definitive evidence of benefit. The LUTS working party soon realised that the clinical management of LUTS often represented a choice between conservative and interventionist interpretations of inadequate and circumstantial evidence. The annual healthcare cost of treating men with LUTS was estimated to be $177 million.10 Adopting a conservative rather than an interventionist approach in response to an impoverished evidence base would readily save $27 million per year, without compromising patient outcomes.5 Prostate cancer screening ... again At the time of the LUTS working party's deliberations, the Australian Health Technology Advisory Committee had already established a lack of compelling evidence for prostate cancer screening for the symptomless.11 However, it vacillated about the issue of prostate-specific antigen (PSA) testing of men with LUTS. The AHCPR guidelines included seven pages about this issue, but made no recommendation.5 The LUTS working party "bit the bullet". It found no evidence that men with LUTS were at any greater risk of prostate cancer than men with no symptoms,1 and advised against PSA testing in men presenting with LUTS. It added the caveat that, if a doctor or his patient chose to disregard this recommendation, such testing should only be done with properly informed consent.1 Because of the somewhat entrenched positions about PSA testing by various groups, the working party correctly anticipated this might be a controversial recommendation, but we did not anticipate the longevity of this controversy. As recently as September 1999, the working party was polled about proposed changes to its recommendation as a result of ongoing "significant concerns" conveyed to the NHMRC. However, the evidence for an association between LUTS and early prostate cancer remains dubious.12 The consultation process The NHMRC requires all draft guidelines be submitted for public consultation.6 To fulfil this requirement, it is sufficient to place an advertisement inviting interested parties to request a copy of the draft document. The LUTS working party decided to mail copies of the draft guidelines to all urologists in Australia, all identifiable general practitioner organisations, a selection of consumers and to six international urological referees. Responses from the Australian urological community varied widely. Some were very supportive: many were negative about CPGs in general and the recommendations of the LUTS draft in particular. Urologists also expressed anxiety about the purpose and outcome of the economic analysis of implementing the CPGs. By contrast, the international urological referees were uniformly supportive and complimentary (see Box 1). Our recollection, however, was that this consultation process largely failed to bring to the attention of the working party any objective evidence of which it was unaware. Perhaps it was a predominantly political process. On reflection, we recommend the development of a standard proforma for submissions in the mandatory CPG consultation process that reinforces rather than undermines its evidence-based principles. Derivative guidelines for GPs and consumers The LUTS working party attempted to establish the most appropriate form of information for general practitioners by conducting a focus group with GPs from diverse backgrounds. Opinions were divided. Some preferred an "in-depth" reference, while others preferred a single-sided laminated sheet listing key messages as bullet points. As there is no body of empirically derived insights regarding optimal formats for derivative GP guidelines, we had nothing but our creativity and experience to guide our development of the GP version of the guidelines.13 Similar frustrations beset the development of derivative information for consumers. An author experienced in writing for a lay readership was needed, but such a person was not present from the beginning of the CPG development process to understand fully the nuances of the many complex issues. The NHMRC's guideline for guidelines6 did not provide detailed and empirically based advice for "transforming" information from a technical CPG into appropriate consumer information. We did our best,14 attempting a humorous approach to appeal to the target group in response to findings of qualitative research.15 The NHMRC subsequently realised the importance of more practical advice, commissioning a consumer information "toolkit".16 Implementation: who really cares? The cost of developing the LUTS CPGs was approximately $160 000 (C Lutton, NHMRC, personal communication). Original correspondence between the NHMRC and the chair of the LUTS working party suggested a shared and genuine commitment to implementation.17 The NHMRC commissioned research to establish baselines against which dissemination and implementation could be evaluated.1Accordingly, the working party developed a list of performance indicators with which to evaluate the effectiveness of such efforts (Box 2). Not unreasonably, the working party expected that wide dissemination and implementation of the derivative documents -- for general practitioners and consumers -- would be similar to that for the NHMRC early breast cancer guidelines.18 Instead, we were advised that distribution should be revenue-neutral, or possibly even revenue-generating. All versions of the LUTS CPGs had to be purchased from the Australian Government Publishing Service. As only 14% of GPs currently have access to the Internet in their practices,19 Internet availability would not have compensated for this restricted distribution. In any case, only the consumer version of the CPGs can be downloaded from the NHMRC publications website (<http://www.health.gov.au/nhmrc/ publicat/cp-home.htm>, accessed January 2000). Between the initiation of the LUTS working party and the launch of the LUTS CPGs in April 1997, both the federal government and the membership of the NHMRC changed. The NHMRC newsletter included a cover story about the LUTS CPGs two years after the launch in 1997,20 and, to our knowledge, this was the only prominent promotion of the guidelines by the NHMRC. Awareness among GPs of this newsletter and the proportion prompted to obtain the LUTS CPGs in response are unknown. Recently, it was claimed that the LUTS CPGs "have not been widely adopted" by urologists, generating consensus among the Australian Prostate Cancer Collaboration against active promotion of a "LUTS education message in the community".21 Yet an international publishing house readily agreed to publish an evidence-based book to address men's questions.22 An application by one of us (G H L H) to the Strategic Research Development Committee of the NHMRC in response to its call for proposals in relation to its Evidence-Based Clinical Practice Research Program (EBCPRP) was unsuccessful in obtaining funds for an implementation trial. Of the seven indicators for dissemination and implementation proposed in the guidelines (Box 2), none has been measured since their release. Updating guidelines: the need for commissioned research CPG working parties inevitably identify important issues that require research. Arguably, organisations sponsoring guideline development ought to support mechanisms to "fast track" priority research. The LUTS working party identified 10 research issues that it considered would add significantly to currently available evidence.1 To our knowledge, this research has not been pursued. The guideline development process also called for "regular" review of guidelines to ensure that their recommendations remain current.6 The LUTS guidelines are due to be updated. Yet our reflections on their development suggest this could also be a disappointing exercise. With the benefit of hindsight, we recommend that all working parties developing guidelines regularly compare their experiences against a comprehensive checklist (Box 3). If, at any point, a specific mismatch is identified, steps could be taken quickly to rectify the situation. This did not happen with the LUTS guidelines, and it appears their time has passed. We remain disappointed that the genuine effort of every member of the LUTS working party to produce sound and influential guidelines has largely "come to naught".17 We have no evidence to suggest practice has improved. Conclusions An effective organisation is needed in Australia for the synthesis, propagation and generation of evidence in response to priority health issues. As suggested elsewhere, it could be a reformed NHMRC,23 or the proposed National Institute for Clinical Studies.24 Until there is such an organisation, better outcomes promised through the development and implementation of evidence-based CPGs will not be realised. References Clinical practice guidelines for the management of uncomplicated lower urinary tract symptoms in men. Canberra: National Health and Medical Research Council, 1997. Ward JE, Sladden M. Urinary symptoms in older men, their investigation and management: is there an epidemic of undetected morbidity in the waiting room? Family Practice 1994; 11: 251-259. Pinnock CB, Marshall VR. Troublesome lower urinary tract symptoms in the community: a prevalence study. Med J Aust 1997; 167: 72-75. Ward JE, Hughes A-M, Hirst GHL, Winchester L. Men's estimate of prostate cancer risk and self-reported rates of screening. Med J Aust 1997; 167: 250-253. Benign prostatic hyperplasia: diagnosis and treatment. Clinical practice guideline no. 8. Agency for Health Care Policy and Research, US Department of Health and Human Services, 1994 (AHCPR Publication No. 94-0582). Guidelines for the development and implementation of clinical practice guidelines. Canberra: National Health and Medical Research Council, 1995. McDonald J. Evidence-based health care from the consumer perspective. Aust Health Consumer 2000; 1 (Summer): 8-10. A guide to the development, implementation and evaluation of clinical practice guidelines. Canberra: National Health and Medical Research Council, 1999. Davidoff F. Evangelists and snails redux: the case of cholesterol screening. Ann Int Med 1996; 124: 513-514. Butler J. Economic aspects of lower urinary tract symptoms in men and their management. Working paper 37. Canberra: National Centre for Epidemiology and Population Health, Australian National University, 1996. AHTAC. Prostate cancer screening. Canberra: AGPS, 1996. Young J, Muscatello D, Ward J. Are men with lower urinary tract symptoms at increased risk of prostate cancer? A systematic review and critique of the available evidence. Br J Urol Internat 2000. In press. "Is it my prostate Doc?": a guide for general practitioners. Canberra: National Health and Medical Research Council, 1997. To pee . . . or not to pee . . . A guide for men about their urinary symptoms. Canberra: National Health and Medical Research Council, 1997. Pinnock C, O'Brien B, Marshall V. Older men's concerns about their urological health: a qualitative study. Aust N Z J Public Health 1998; 22: 368-373. How to present the evidence for consumers: preparation of consumer guidelines. Canberra: National Health and Medical Research Council, 2000. Hirst G. Clinical practice guidelines: to what end? Med J Aust 1997; 167: 288. Clinical practice guidelines: the management of early breast cancer. Canberra: National Health and Medical Research Council, 1995. Young J, Ward J. General practitioners' use of evidence databases. Med J Aust 1999; 170: 56-58. Hirst G. Men and their urinary symptoms: guidelines support a conservative approach to the management of many men. NHMRC News. 1999; 1(2): 3-5. Pinnock C. Report from the first national meeting of the Australian Prostate Cancer Collaboration Education Group. Cancer Forum 1999; 23: 165-167. Hirst G, Wilde S. Your prostate, your choices: the news may be more reassuring than you think. Sydney: Bantam Books, 1999. Health and Medical Research Review Strategic Review. The virtuous cycle: working together for health and medical research. Canberra: AGPS, 1998. Australian Department of Health and Aged Care. The 1999-2000 Health Budget in detail. Major boost to medical research takes Australia into the century of healing. Budget media release, 11 May, 1999. <http://www.health.gov.au/pubs/ budget99/media/mrmw991.htm> (accessed 15 February, 2000). Authors' details Department of Urology, Mater Hospitals, Brisbane, QLD. Geoffrey H L Hirst, MB BS, FRACS, Urologist. Needs Assessment & Health Outcomes Unit, Central Sydney Area Health Service, Sydney, NSW. Jeanette E Ward, PhD, FAFPHM, Director, and Clinical Associate Professor, Department of Public Health & Community Medicine, University of Sydney. Reprints will not be available from the authors. Correspondence: Dr G H L Hirst, Taylor Medical Centre, 40 Annerly Road, Woolloongabba, QLD 4102. drhirst@gil.com.au 1: Examples of written responses to the draft LUTS guidelines Australian urologists "...the guidelines it recommends are not supported by appropriate evidence and are unsustainable. We do not feel this represents best practice. We propose the document be rejected as its present form is unacceptable". "I think they do inevitably suggest that part of the reason for the guidelines themselves is financial, and this heightens the concern that at some stage they will be used in a prescriptive way". Australian general practitioner organisations "Very useful". "Guidelines which update individual specialities are extremely valuable, as this one is". "I agreed with all the guidelines as stated...the literature review was impressive". "...perhaps the use of more flow charts and tables would aid in this process [as] few GPs will be willing to read through a document of this size". Consumers "It is a GREAT help to me - million thanks and much gratitude to you!!" International expert "I congratulate you and your colleagues for producing a superb document. Your clinical practice guidelines constitute an outstanding effort. Where conclusive literature does not exist, the discussion so states. Implicit judgements have been kept to a minimum and where used are appropriately identified as such".

Jeanette E Ward

Mooted changes to the Declaration on the agenda of the World Medical Association have sparked a vigorous debate on international research issues. The medical, research and ethics communities in Australia need to participate more broadly in this debate. Introduction The Nuremberg Code, which was formulated to prevent a recurrence of the horrific medical experiments carried out on humans during World War II, is unwavering in its commitment to the primacy of the human subject. It states that any person who is a research participant "should be so situated as to be able to exercise free power of choice" and that "(t)he experiment should be such as to yield fruitful results for the good of society, unprocurable by other methods or means of study, and not random and unnecessary in nature."1 The Declaration of Helsinki2was the World Medical Association's (WMA's) response to the Nuremberg Code and its goal was to safeguard research subjects. However, in declaring the need to weigh the importance of the research objective against the risk to the subject (Article I.4), the Declaration was seen as a subtle retreat from the Code.3 Some fear that changes to the Declaration currently under consideration by the WMA would substantially "water down" the basic principles of ethical human research. The Declaration does not specifically deal with international collaborations. In 1993, the Council for International Organizations of Medical Sciences (CIOMS) developed the International ethical guidelines for biomedical research involving human subjects,4 which address issues pertinent to the conduct of research in developing countries. Despite the existence of the Declaration and other documents, it is apparent that the application of safeguards to protect research subjects is far from uniform, especially among impoverished or marginalised people.5,6 Proposed revisions to the Declaration of Helsinki A number of revisions are currently proposed to the Declaration to make it more relevant to researchers (some of whom, it has been suggested, commonly breach its provisions).7 Those who oppose the amendments fear that research participants will be made more vulnerable to harm in order to make research more efficient and perhaps expedient (see Box).8Two proposals have generated a great deal of discussion and controversy. One concerns the abolition of the distinction between "therapeutic" and "non-therapeutic" research. The other (the main focus of this article) relates to provision of the best proven treatment and to use of placebo-controlled trials. "Therapeutic" v "non-therapeutic" research The introduction to the Declaration requires that "a fundamental distinction . . . be recognised between medical research in which the aim is essentially diagnostic or therapeutic for a patient, and medical research, the essential object of which is purely scientific and without implying direct diagnostic or therapeutic value to the person subjected to the research". Article II.6 states that in "therapeutic" research "The physician can combine medical research with professional care, the objective being the acquisition of new medical knowledge, only to the extent that medical research is justified by its potential diagnostic or therapeutic value for the patient." Article III.2 states that in "non-therapeutic" research "The subjects should be volunteers -- either healthy persons, or patients for whom the experimental design is not related to the patient's illness." Robert Levine, Professor of Internal Medicine at Yale University School of Medicine, argues that these Articles together rule out "all rational research on the causes of diseases or on their pathogenesis or pathophysiology".7 He notes that clinical trials may include both therapeutic and non-therapeutic agents.7 Others believe the division between therapeutic and non-therapeutic research to be firmly entrenched in research guidelines developed since and influenced by the Declaration and to be well understood in practice.9 "Best proven treatment" v "highest attainable treatment" The changes proposed to Article II.3 (see Box) have generated extensive debate. Opinions are polarised between those in favour of the "best proven treatment or method" and those for the "highest attainable and sustainable treatment or method". The problem with the "best proven treatment" approach is that it may prevent valuable research being done, as treatments and services will not be readily available in resource-poor settings and may not be provided by researchers and their sponsors. The problem with the "highest attainable treatment" goal is that it may lead to a marked decline in care -- there are no clear criteria for establishing what level of treatment is acceptable or firm safeguards for applying it. The problem in resource-poor countries Perhaps the real issues to be debated are how best to enable people to make meaningful choices and how to ensure that they are not treated poorly and without respect because of their circumstances. What should be examined are the problems created in some societies by a lack of fundamental civil and economic rights, and whether this advantages proposed research projects. To assume that all would like to be treated as people in affluent countries would, and to rely on this judgement as a basis for formulating an encompassing ethical ideal, is, to some extent, misguided. Classic egalitarian premises, upon which comparable rights documents, such as the Universal Declaration of Human Rights,10 are based fail to take account of subcultures and their priorities. Indeed, issues like gender and race are still inadequately addressed in these documents, in which the tacitly assumed "universal person" is the European white heterosexual male.11 This difficulty is accentuated when privileged cultures interact with others. Is it true to say that standards appropriate for industrialised countries are equally relevant to others? If not, what are we left with? Ruth Macklin, Professor of Bioethics at the Albert Einstein College of Medicine in New York, has analysed ethical concerns in international research according to the concept of justice.12 She states that a prominent feature of justice is that no one group should "receive disproportionate benefits or bear disproportionate burdens",12 a corollary being that like cases should be treated alike. One side argues that if the study is unethical in one place it is unethical in both. The other argues that risk-benefit ratios are different in resource-poor countries and therefore require a different response and, further, that if the benefit is actually to accrue and only to accrue to the developing country, this is ethically significant.12 Her conclusion in this debate is that both sides can claim that their arguments observe the ethical requirement of justice. Advocates of placebo-controlled trials in resource-poor countries cite local support and participation in defence of their views. Thus, in the case of trials of less expensive regimens to prevent vertical transmission of HIV, Edward Mbidde, a Ugandan physician, said in a now oft-cited statement, "(t)hese are Ugandan studies, conducted by Ugandan investigators, on Ugandans ... for the good of their people".13 It would be too simple a response to discount this comment entirely as being no answer to a breach of ethical standards.14 Yet, even when a host country agrees to allow drug trials, substantial ethical difficulties remain. One of us (J B), after working for 10 years as a clinician in Mozambique (where the local provincial health service budget was about $US3 per capita per year), believes that the basic rights of potential trial participants in some parts of Africa may be so compromised that refusal to participate is not an option: "This is the sort of health service where every clinician finds him or herself from time to time looking at the pharmacy cupboard and wondering how to divide the remaining three vials of penicillin between the five patients in the ward who need it. (Whether to give starting doses to everyone in the hope that the promised new supplies will arrive, or just give it to one seriously ill child, for whom at least it represents a curative course.) From that perspective, enrolling patients in a clinical trial will always look attractive, no matter how unethical that research may turn out to be."15 Almost any reward, even bars of soap or transistor radio batteries, is likely to ensure trial participation. Perhaps international collaborations, particularly those involving complex drug trials, should not be conducted where there is this degree of poverty. Speaking at the same symposium,15 Pascale Allotey, Lecturer in International Programs at the Key Centre for Women's Health, Melbourne University, made the point that the possibility of enhanced services or cashflow to a community will mean that community leaders will very likely agree to trials taking place, as, ostensibly, will community members. However, they may resent doing so. Fears of a diminished standard of care as a result of withdrawal from a trial are quite real in these circumstances. Where to from here? Are genuinely consensual relations possible between the research community and participants who otherwise have little or no access to healthcare or other basic rights and liberties? Can structures and criteria be implemented that promote dialogue and recognise diversity of approach, but discourage abuse of trial participants? The following suggestions were offered to the participants in the aforementioned symposium15 for consideration, and most agreed that more discussion was required to flesh out what these ideas might mean in practice (the suggestions are not entirely new and are broadly consistent with the CIOMS guidelines and draft UNAIDS guidelines16): (a) Where a population does not possess basic economic or social rights it should be regarded, prima facie, as one whose members' capacity to freely consent is gravely impaired. Research studies in such populations, especially those involving randomised trials, require special justification. An exception might be where the research goal is to work out how to apply a proven technology: for example, an assessment of whether open or covered buckets are more suitable water containers in a refugee camp (Associate Professor Michael Toole, Macfarlane Burnet Centre for Medical Research, personal communication). UNAIDS guidelines16 state that, in international collaborative programs, strategies should aim to balance inequalities by involving members of affected communities from very early on in the design and development stage, and by imposing a number of safeguards around the process of informed consent. (b) A research protocol should describe the conditions that might make a research population vulnerable to exploitation and the steps that will be taken to overcome them.16 The UNAIDS guidelines impose this requirement on research protocols. We further propose that these steps should be described in publications derived from the research, in order to give the issues greater prominence and to further this discussion. (c) In planning research in populations severely deprived of civil and political rights, agreements with governments and ethics committees are insufficient. This is especially the case when governments have demonstrated grossly repressive or corrupt behaviour, or where ethical review systems can not be regarded as independent. Research should not take place in these circumstances. This recommendation should be distinguished from ethical guidelines applying to research in emergency or refugee settings. (d) Thorough community-based consultation is required to determine local views, needs and priorities. Researchers need to establish what local research priorities exist (although, in many deprived populations, any problem area could be seen as a priority). This includes, in particular, consultation with people who have little power or are ostracised for whatever reason. Ethnographic studies could be conducted in advance of a proposed project to determine actual rather than supposed local attitudes, and debriefing could be required after completion of a trial (Deborah Zion, Centre for Human Bioethics, Monash University, personal communication). Including nationals on committees, or agreement by host governments and ethics committees, are not substitutes for community consultation. (e) An analysis should be made, in advance of a project, of the long-term consequences of the intervention. Long-term considerations should certainly include, but not be limited to, sustained access to a trial drug. One of the possible adverse consequences to consider would be the diversion of local researchers and healthcare providers into projects that are not local initiatives. Consistent with current ethical standards, if there is no prospect of benefit to the community in its terms the research should not be undertaken. At the very least, a memorandum of understanding should be prepared before the commencement of any international collaboration, indicating what each party -- community, government, research institution and sponsor -- expects prior to, during and as a consequence of the trial. Conclusion Angela Harris, Professor of Law at the University of California, Berkeley, has stated that modern human rights standards are at once indispensable and inadequate.17 The same may be said for ethical guidelines on medical research. The solutions are not clearcut. The WMA meeting in Tel Aviv, Israel, in October 1999, at which the Declaration of Helsinki was reconsidered, issued the following brief statement: "The meeting heard of widespread support for retaining the existing structure of the Declaration of Helsinki. It was agreed that the working group set up to consider amendments to the Declaration should report back with a proposed revision at next year's annual General Assembly meeting in Edinburgh, Scotland (3 October 2000)." It is to be hoped that the Australian medical research community and other interested groups will debate the issues and arrive at a consensus. Acknowledgements The authors warmly thank the following for their valuable contributions to "The Declaration of Helsinki: a symposium to review current proposals for change", held at the Victorian Institute of Forensic Medicine, Melbourne, on August 30 1999, and for their contributions to the discussion: Stephen Cordner and Helen McKelvie (Victorian Institute of Forensic Medicine), Norman Swan (Radio National), Sandra Hacker (Australian Medical Association), Louis Waller (Faculty of Law, Monash University), Pascale Allotey (Key Centre for Women's Health, Melbourne University), Deborah Zion and Justin Oakely (Centre for Human Bioethics, Monash University), Lyn Gillam (Centre for the Study of Health and Society, Melbourne University) and Ian Kerridge (Faculty of Medicine, Newcastle University). The views expressed in this article are those of the authors. References The Nuremberg Code. From: Trials of war criminals before the Nuremberg Military Tribunals under Control Council Law No. 10. Nuremberg, Oct 1946-Apr 1949. Washington, DC: US Government Printing Office, 1949. World Medical Association. Recommendations guiding physicians in biomedical research involving human subjects. As adopted by the 18th World Medical Assembly, Helsinki, June 1964 (the "Declaration of Helsinki"). Katz J. The consent principle in the Nuremberg Code: its significance then and now. In: Annas G, Grodin M. The Nazi doctors and the Nuremberg Code: human rights in human experimentation. New York: Oxford University Press, 1992: 227-239. Council of International Organizations of Medical Sciences, in collaboration with the World Health Organization. International ethical guidelines for biomedical research involving human subjects. Geneva: CIOMS, WHO, 1993. French H. AIDS research in Africa: juggling risks and hopes. New York Times October 9, 1997: A1, A14. Allotey P. Clinical trials in developing countries: bringing people into the debate. Monash Bioethics Review 1999; 18: 18-23. Levine R. The need to revise the Declaration of Helsinki. N Engl J Med 1999; 341: 531-534. Brennan T. Proposed revisions to the Declaration of Helsinki -- will they weaken the ethical principles underlying human research? N Engl J Med 1999; 341: 527-531. British Medical Association. BMA response to the proposed revision of the WMA Declaration of Helsinki. London: BMA, August 1999. Universal Declaration of Human Rights. Adopted and proclaimed by the United Nations General Assembly on December 10, 1948. Geneva: Office of the United Nations High Commissioner for Human Rights. Otto D. Rethinking the "universality" of human rights law. Columbia Human Rights Law Rev 1997; 29: 1-46. Macklin R. Justice in international research. In: Kahn J, Mastroianni A, Sugarman J, editors. Beyond consent: seeking justice in international research. New York: Oxford University Press, 1998: 131-146. Varmus H, Satcher D. Ethical complexities of conducting research in developing countries. N Engl J Med 1997; 337: 1003-1005. Bayer R. The debate over maternal-fetal HIV transmission prevention trials in Africa, Asia, and the Caribbean: racist exploitation or exploitation of racism? Am J Public Health 1998; 88: 567-570. Proceedings of a symposium. The Declaration of Helsinki: a symposium to review current proposals for change. Victorian Institute of Forensic Medicine; 1999 Aug 30; Monash University, Melbourne, VIC. Joint United Nations Programme on HIV/AIDS. Ethical considerations in HIV preventive vaccine research: UNAIDS guidance document. Geneva: UNAIDS, February 2000. Harris A. Foreword: the jurisprudence of reconstruction. California Law Rev 1994; 82: 741-785. Authors' details Department of Epidemiology and Preventive Medicine, Monash University Medical School, Alfred Hospital, Melbourne, VIC. Bebe Loff, MA, LLB, NHMRC Public Health Scholar. Jim Black, MB BS, MCommH, DTM&H, NHMRC Medical Scholar. Reprints will not be available from the authors. Correspondence: Ms B Loff, Department of Epidemiology and Preventive Medicine, Monash University, Melbourne, VIC. Bebe. LoffATmed.monash.edu.au Some proposed revisions to the Declaration of HelsinkiCurrent statement I.8: Reports of experimentation not in accordance with principles laid down in this Declaration should not be accepted for publication. Proposed revision: Variances from these principles should be explained and justified in the report. Editors are obligated to consider carefully the justification for any variances from these principles in deciding whether to accept or reject the report for publication. Current statement I.10: ...the informed consent should be obtained by a physician who is not engaged in this investigation and who is completely independent of this official relationship. Proposed revision: In some cases of this type, it may be preferable if the informed consent were to be obtained by a qualified person who is not engaged in the investigation, independent of the dependent relationship, or both. Current statement II.3: In every medical study, every patient, including those of a control group, if any, should be assured of the best proven diagnostic and therapeutic method. Proposed revision: In any biomedical research protocol every patient-subject, including those of a control group, if any, should be assured that he or she will not be denied access to the best proven diagnostic, prophylactic or therapeutic method that would otherwise be available to him or her. Current statement II.3: This does not exclude the use of inert placebo in studies where no proven diagnostic or therapeutic method exists. Proposed revision: This principle does not exclude the use of placebo or no-treatment control groups if such are justified by a scientifically and ethically sound research protocol. When outcome measures are neither death nor disability, placebo or other no-treatment controls may be justified on the basis of their efficiency. Current statement II.5: If the physician considers it essential not to obtain informed consent, the specific reasons for this proposal should be stated in the experimental protocol for transmission to the independent committee. Proposed revision: When permitted by applicable law, the requirement for informed consent may be waived by the independent research ethics committee. Such a waiver may be appropriate in research that presents little or no threat to the rights and welfare of research subjects as exemplified by use of anonymous tissue samples for research purposes and in certain other types of research in such fields as epidemiology and policy evaluation. It may be justified in research in emergency situations in which patient-subjects have temporary or enduring loss of decisional capacity and interventions or procedures must be initiated before informed consent can be obtained from patient-subjects or their legally authorised representatives. In the latter case the research ethics committee may require special procedures to protect the rights and welfare of subjects.

Bebe Loff · Jim Black