Volume 166 Issue 9

Urinary diagnosis of gonorrhoea and chlamydia in men in remote Aboriginal communities Steven J Skov, Penny Miller, Wayne Hateley, Ivan B Bastian, Jenny Davis and Peter W Tait MJA 1997; 166: 468 For editorial comment see Waddell Readers may print a single copy for personal use. No further reproduction or distribution of the articles should proceed without the permission of the publisher. For permission, contact the Australasian Medical Publishing Company Journalists are welcome to write news stories based on what they read here, but should acknowledge their source as "an article published on the Internet by The Medical Journal of Australia <http://www.mja.com.au/>". Abstract - Introduction - Methods - Results - Discussion - Acknowledgements - References - Authors' details - ©MJA1997 Abstract Aims: (1) To evaluate the acceptability and validity of an intervention based on urine tests for diagnosis and treatment of gonorrhoea and chlamydia in men in remote Aboriginal communities. (2) To provide a prevalence estimate of these infections in the male population in the surveyed communities. Methods: First-void urine samples from 460 men in remote communities and 33 men in the Alice Springs Gaol were tested for gonorrhoea and chlamydia with at least one of polymerase chain reaction (PCR), enzyme immunoassay (EIA) and culture (gonorrhoea only). Results: One hundred and three men (20.9%) were infected with gonorrhoea or chlamydia. The prevalence of infection for gonorrhoea only was 11.7%, for chlamydia only 4.1% and for dual infection 5.1%. Eighty-eight infected men and 45 of their sexual partners were recorded as having been treated within two months of testing. PCR tests detected the largest number of infections and were the easiest to use. Conclusions: The prevalence of these infections was higher than anticipated. Urine PCR tests were acceptable to men and are well suited to the remote-community setting. As an effective alternative to urethral swabs, they permit a range of community-based strategies to address high rates of infection with gonorrhoea and chlamydia. MJA 1997; 166: 468-471 Introduction The rates of sexually transmitted diseases (STDs) in central Australia, particularly gonorrhoea and chlamydia, are among the highest in Australia and indeed the world,1,2 despite a modest decline in gonorrhoea over the past 15 years (according to notifiable diseases data from the health departments of South Australia, Western Australia and the Northern Territory). Although these infections are more common in this region than in the rest of Australia among both Aboriginal and non-Aboriginal people, most of the excess morbidity occurs among Aboriginal people (health departments' data). Many factors contribute to this situation. Dispossession, unemployment and their sequelae are the fundamental determinants of the poor health of Aboriginal people and also underlie the observed high levels of STDs. The serious under-resourcing of central Australian health services, the difficulty in maintaining confidentiality in small communities, the stigma attached to STDs, the unpleasant nature of urethral and endocervical swabs and, in particular, the frequent lack of male health care workers -- all militate against effective STD programs.3-5 Urine tests for the detection of gonorrhoea and chlamydia are now commercially available. Enzyme immunoassay (EIA) tests have sensitivities above 80% and specificities above 93%,6-9 while polymerase chain reaction (PCR) tests have reported sensitivities of 90%-100% and specificities of 98%-100%.10-14 These technologies offered exciting possibilities to address some of the difficulties in delivering STD services, so, after receiving approval from the Alice Springs Institutional Ethics Committee, the Tri-State STD/HIV Project (TSP), in collaboration with Nganampa Health Council (NHC) and the Central Australian Aboriginal Congress (CAAC), both organisations controlled by Aboriginal communities, undertook the programs described in this article. Methods The study was undertaken in four parts. Phase I: NHC provides health services to several remote Aboriginal communities and conducts annual syphilis serology screening programs in these communities. NHC maintains a population register, and all persons between the ages of 12 and 40 years who are in the communities during the screening programs are sought out and asked to participate. In April 1995, 189 men participating in this program in three communities serviced by NHC were asked to give a first-void urine sample in addition to the blood sample for the syphilis test. Phase II: In June 1995, another 218 men in the remaining three NHC communities had urine tests only; all males over the age of 12 who were present in the communities at the time were asked to participate. Phase III: In October 1995, a further 53 men were tested who had not been previously tested. Phase IV: CAAC, as part of its many functions, provides the health service to the Alice Springs Gaol. In June 1995, 33 new inmates were asked to give a first-void urine sample as part of their comprehensive health check on admission. Diagnostic tests used for gonorrhoea were culture, EIA (Abbott Gonozyme, Chicago, IL) and PCR (Roche Amplicor CT/NG, Branchburg, NJ) and for chlamydia were EIA (SYVA MicroTrak II, Palo Alto, CA) and PCR (Roche Amplicor, Branchburg, NJ) (Box 1, below). If there was insufficient urine for all tests, the order of priority was to do PCR first, then culture, chlamydia EIA and finally gonococcal EIA. To check for cross-contamination between PCR specimens, all Phase III specimens that provided sufficient urine were divided before any PCR procedures were performed. When a positive result was obtained the other half of the original sample was retested. Chocolate and/or Thayer-Martin agar plates incubated in candle jars or CO2 gaspacks were used for gonococcal culture. BioCult GC tubes (Orion Diagnostica, Espoo, Finland), which contain a modified Thayer-Martin dipslide and a CO2-generating tablet, were also trialled. Culture media were inoculated with sediment from 10 mL of centrifuged urine and incubated for up to 72 hours at 35oC on-site and/or at Western Pathology in Alice Springs. Neisseria gonorrhoeae was identified by standard methods: typical morphology of colonies, oxidase paper test, microscopy with Gram stain, and latex agglutination (Phadebact GC, Boule Diagnostics AB, Huddinge, Sweden). Urine samples for EIA and PCR were prepared on-site according to the test manufacturer's specifications and then sent to Western Pathology in Alice Springs for processing. Male Aboriginal health workers and registered nurses in each clinic were responsible for liaison with the community and collection of specimens. Each man was asked to give 25-40 mL of first-void urine in a sterile collection jar. No urethral swabs were sought. Initial preparation of urine specimens for transport to the laboratory was completed within three hours of collection. A diagnosis of infection was made and treatment was initiated if any test was positive. Resident clinic staff were responsible for offering treatment, further investigation and safe-sex education to infected men and their sexual partners. Results Study population We tested urine samples from 493 men, of whom 460 lived in six remote communities and 33 were gaol inmates. Acceptability to clients was high, with fewer than 10 men (exact count not possible) declining testing. The 460 men from remote communities represented about 60% of the male population over the age of 12 in the area serviced by the participating clinics (Box 2, below). The proportion of men tested in different communities varied from 28% to 85%. The rate of infection Based on all methods of diagnosis combined, 103 of the men tested (20.9%) were found to be infected with either gonorrhoea or chlamydia. The rate of infection in different communities varied between 14.5% and 26%. Among the gaol inmates, who originated from all parts of the southern NT, seven (21.2%) were infected. Most infections (90.3%) occurred in men aged 15-39 years and the five-year age-specific prevalence of infection in this group varied from 22% to 27% (Box 2, above). Treatment and follow-up Eighty-eight men were treated within two months of testing. The average delay between testing and treatment was 18 days. Forty-five sexual partners of infected men were also recorded as having been treated. Performance of the tests used Different combinations of tests were used in different phases. The volume of urine received from each man varied, so it was not always possible to do all the intended tests. A summary of the numbers of specimens subjected to the various tests and their relative performance is shown in Box 3 (below). During Phase II, culture for gonorrhoea using both BioCult GC tubes and standard Thayer-Martin plates was performed on 193 urine specimens. Thirteen diagnoses were made with the BioCult GC tubes and only seven were made with standard Thayer- Martin plates: all cases positive on Thayer-Martin plates were also positive on the BioCult GC tubes. During Phase III, 19 of the 20 initially positive PCR results were retested (in one case there was insufficient urine). In 18 of these cases there was complete agreement between the first and second test. One chlamydia-positive specimen was negative on the second testing. However, because of delays in transport, the tests on this specimen were done at six and eight days after collection, well in excess of the four days maximum recommended by the test manufacturers. Discussion These results highlight the need for improvements in current STD control, including surveillance programs. Based on routine clinic activity and notifications, NHC reported only 38 men with either gonorrhoea or chlamydia during the whole of 1994 (health departments' notifiable diseases data), compared with the 96 cases detected in this study. Other regions of Australia that have comparable notification rates of these infections1 may have similar problems in underdetection. The variation in participation rates in different communities and age-groups was largely attributable to the numbers of people who happened to be present at the time of the study (people in these remote communities being highly mobile), the working relationship between community members and the health staff, and the assiduousness of the health staff in conducting the program. In central Australia it is often considered, without specific evidence, that people who do not participate in such programs may be at higher risk of infection. Recent work by NHC showed no difference in rates of syphilis infection between those tested during the main body of a screening program and those tested later as part of an effort to test non-participants (Dr Penny Miller, unpublished data). The screening programs in our study were intended to identify the most effective and practical tests for everyday use in remote clinics. Collection of urethral swabs was avoided because it would not have been acceptable to the community. Hence, the study lacked a diagnostic "gold standard" and could not formally evaluate sensitivities and specificities. The published specificities of all the tests were high -- above 98% for urine PCR tests.10-14 The procedures used in Phase III demonstrated no problems with cross-contamination in PCR testing. Several studies have suggested that urine PCR tests for chlamydia are more sensitive than urethral-swab culture and highly specific.10-13 Less work has been done on urine PCR for diagnosing gonococcal infection, but two studies indicate sensitivities above 90% and specificities of 100%.13,14 In ideal conditions, the sensitivities of urethral-swab microscopy and culture are 90%-98%,15,16 but are likely to be much less in remote communities because of high staff turnover, the use of non-nutritive transport media, and frequently prolonged transport times. In such circumstances, urine PCR may be as good as or better than urethral-swab microscopy and culture for diagnosis of gonorrhoea. In any event, according to local management protocols,17 all men presenting to clinics with urethritis are offered immediate treatment for both gonorrhoea and chlamydia. In terms of practical application, PCR was superior to culture for gonorrhoea and EIA for both infections. PCR tests detected more infections with either organism than did the other tests (Box 3). Assuming that the high published specificities for the PCR tests held under field conditions, this increased rate of detection would be attributable to a superior sensitivity. The PCR test was also the easiest to use: urine is simply stored and transported at 4¡-8¡C in the same jar used to collect it. The urine must then be processed in the laboratory within four days (test manufacturer's instructions), which is usually feasible in most remote situations. In contrast, both EIA tests used require centrifuging before transport, and the SYVA MicroTrak II EIA required addition of a transport buffer. However, PCR tests do not provide information about antibiotic susceptibility. If PCR technology is to be used as the principal diagnostic tool for gonorrhoea, there would need to be accompanying sentinel systems for gonococcal culture and antibiotic susceptibility. We found that culture of first-void urine was positive in 72% of diagnoses of gonorrhoea. Sensitivities of 70%-100% for urine culture of gonorrhoea have been reported.18,19 On this basis, urine culture in addition to urine PCR for diagnosis of gonorrhoea could be performed routinely at sentinel sites and during similar programs to those reported here in order to maintain antibiotic-sensitivity surveillance. In the field situation, the BioCult GC tubes detected more gonorrhoea than standard Thayer-Martin plates and were easier and more convenient to use. The accuracy, ease of use and acceptability to men of urine PCR tests suggest several strategies to make clinical services more accessible to people, reduce the amount of disease in the community and identify individuals who are in need of safe-sex education. The use of urine tests for routine diagnosis may encourage more men to present with an STD, even if there is no male practitioner present. Health services could adopt active case-finding strategies such as opportunistic testing when people present for other reasons, including annual comprehensive health checks or community surveys. Such strategies are under consideration by health services in central Australia. For example, CAAC is seeking funding to establish an outreach program via a mobile clinic to make comprehensive well-men's check-ups, including STD checks, accessible to Aboriginal men in Alice Springs. We also examined screening as an STD control strategy. With fewer than 10 men refusing to participate, these programs resulted in 88 men who had either gonorrhoea or chlamydia and at least 45 of their sexual partners being treated. We did not have the resources to determine whether the men were symptomatic at the time of the test. However, none of them had presented to the clinic for treatment. When these findings are considered in the light of routine notifications and our own local experience, it is likely that most of these people would not have been diagnosed and treated outside these programs. Health services catering to remote Aboriginal communities in other parts of Australia may wish to consider this new technology and its usefulness in comprehensive STD education and control programs. Acknowledgements The Tri-State STD/HIV Project (TSP) is jointly funded and managed by the Commonwealth Department of Health and Family Services, Territory Health Services, the South Australian Health Commission and the Western Australian Health Department. The work would not have been possible without the support of the clinical staff of Nganampa Health Council (NHC) and the Central Australian Aboriginal Congress (CAAC). Technical advice and material resources were contributed by the TSP, NHC, CAAC, Western Pathology, Alice Springs Hospital laboratory, the Australian Army, the Northern Territory AIDS/STD Unit, Roche Diagnostics, and the Institute of Medical and Veterinary Science. John Kaldor, Russell Waddell, Frank Bowden and John Boffa all commented on earlier drafts of this paper. References National Notifiable Diseases Surveillance System. Annual report of the National Notifiable Diseases Surveillance System. Commun Dis Intell 1994; 18: 518-548 . De Schryver A, Meheus A. Epidemiology of sexually transmitted diseases: the global picture. Bull World Health Organ 1990; 68 (5): 639-654. Scrimgeour D, Rowse T. Evaluation of STD control activities in central Australia. Menzies School of Health research report. Alice Springs: Menzies School of Health, 1992. Warchivker I. Variations in health care expenditure in the Alice Springs Rural District in 1993-94. Aust N Z J Public Health 1996; 20: 11-13. McDermott R, Beaver C. Models of horizontal equity in resource allocation in Aboriginal health. Aust N Z J Public Health 1996; 20: 13-15. Roongpisuthipong A, Lewis JS, Kraus SJ, Morse SA. Gonococcal urethritis diagnosed from enzyme immunoassay of urine sediment. Sex Trans Dis 1988; 15: 192-195. Schachter J, Pang F, Parks RM, et al. Use of gonozyme on urine sediment for diagnosis of gonorrhoea in males. J Clin Microbiol 1986; 23: 124-125. Moncada J, Schachter J, Shafer MA, et al. Detection of Chlamydia trachomatis in first catch urine samples from symptomatic and asymptomatic males. Sex Trans Dis 1994; 21: 8-12. Sanders JW, Hook EW, Welsh LE, et al. Evaluation of an enzyme immunoassay for detection of Chlamydia trachomatis in urine of asymptomatic men. J Clin Microbiol 1994; 32: 24-27. Jaschek G, Gaydos CA, Welsh LE, Quinn TC. Direct detection of Chlamydia trachomatis in urine specimens from symptomatic and asymptomatic men by using a rapid polymerase chain reaction assay. J Clin Microbiol 1993; 31: 1209-1212. Bianchi A, Scieux C, Brunat N, et al. An evaluation of the polymerase chain reaction Amplicor Chlamydia trachomatis in male urine and female urogenital specimens. Sex Trans Dis 1994; 21: 196-200. Bauwens JE, Clark AM, Loeffelholz MJ, et al. Diagnosis of Chlamydia trachomatis urethritis in men by polymerase chain reaction assay of first-catch urine. J Clin Microbiol 1993; 31: 3013-3016. Mahony JB, Luinstra KE, Tyndall M, et al. Multiplex PCR for detection of Chlamydia trachomatis and Neisseria gonorrhoeae in genitourinary specimens. J Clin Microbiol 1995; 33: 3049-3053. Komeda H, Deguchi T, Yamamoto H, et al. Detection of Neisseria gonorrhoeae in first-voided urine sediments from male urethritis patients by polymerase chain reaction. Kansenshogaku Zasshi 1992; 66: 1209-1212. Judson FN. Gonorrhoea. Med Clin North Am 1990; 74: 1353-1366. Lind I. The laboratory diagnosis of gonorrhoea. In: Facklam R, Laurell G, Lind I, editors. Recent developments in laboratory identification techniques. Amsterdam: Excerpta Medica, 1979. Central Australian Rural Practitioners' Association. The CARPA Standard Treatment Manual. 2nd ed. Alice Springs: Institute for Aboriginal Development, 1994. Woods ER, Galvez LM, Talis AL, Jean Emans S. First catch urine sediment for Chlamydia trachomatis and Neisseria gonorrhoeae in adolescent males with pyuria. J Adolesc Health 1991; 12: 329-334. Feng WC, Medeiros AA, Murray ES. Diagnosis of gonorrhoea in male patients by culture of uncentrifuged first-voided urine. JAMA 1977; 289: 896-898. (Received 24 June, accepted 17 Oct, 1996) Authors' details Tri-State STD/HIV Project, Alice Springs, NT. Steven J Skov, MPH, FAFPHM, Medical Officer, Tri-State STD/HIV Project. Nganampa Health Council, Alice Springs, NT. Penny Miller, MB BS, STD/HIV Program Coordinator. Wayne Hateley, STD/HIV Aboriginal Health Worker. Royal Darwin Hospital, NT. Ivan B Bastian, MB BS, MSc, Microbiology Registrar. Australian Army, Darwin, NT. Jenny Davis, Medical Technician. Central Australian Aboriginal Congress, Alice Springs, NT. Peter W Tait, MB BS, DipRACOG, FRACGP, Acting Senior Doctor. Reprints will not be available. Correspondence: Dr Penny Miller, Nganampa Health Council, PO Box 2232, Alice Springs, NT 0871. - To top of article - ©MJA 1997 <URL: http://www.mja.com.au/> © 1997 Medical Journal of Australia.

Steven J Skov · Penny Miller · Wayne Hateley · Ivan B Bastian · Jenny Davis · Peter W Tait

Is screening of Australian blood donors for HTLV-I necessary? Gordon S Whyte MJA 1997; 166: 478 For editorial comment see Kaldor Subsequently cited in Wong et al. Should we be screening blood donors for hepatitis G virus? The case against screening. MJA 1998; 169: 375-377 Readers may print a single copy for personal use. No further reproduction or distribution of the articles should proceed without the permission of the publisher. For permission, contact the Australasian Medical Publishing Company Journalists are welcome to write news stories based on what they read here, but should acknowledge their source as "an article published on the Internet by The Medical Journal of Australia <http://www.mja.com.au/>". Abstract - Introduction - The decision to screen blood for HTLV-I - Blood donor data collection - Extrapolation of State data to national data - New and repeat Australian donors - Screening of donors - Prevalence and incidence of HTLV-I - Risk of HTLV-I transmission - Discussion - Acknowledgement - References - Authors' details - ©MJA1997 Abstract Objective: To re-examine the 1992 decision by Australian Red Cross for its blood banks to screen blood donors for antibody to human T-cell lymphotropic virus type I (HTLV-I) by determining the risk of its transmission by blood transfusion. Methods: Data on patterns of return behaviour by repeat blood donors in Victoria were modelled to deduce the number of donors giving repeat donations in Australia from March 1993 to December 1995. Data on annual donor and issued cellular blood products from 1992 to 1995 were obtained from national Red Cross statistics. From the numbers of donations given by repeat donors, together with the number of new donors, the number tested for HTLV-I was deduced. The number and characteristics of donors screened positive for HTLV-I antibody were collated. The crude prevalence of HTLV-I was calculated by dividing the number of donors with HTLV-I by the total number of donors (repeat donors and new donors). The incidence of HTLV-I was calculated by dividing the number of seroconversions in repeat donors by the cumulative period of donor exposure. Results: Sixteen homologous and five autologous donors were found to be positive for HTLV-I; none seroconverted and no clear risk factors for HTLV-I were identified. The prevalence of HTLV-I in Australian donors is 1 in 100 000 and the incidence less than 1 in 1 million person-years. In the absence of HTLV-I screening, the calculated risk of a transfused patient developing HTLV-I infection is 1 in 370 000, with a risk of developing HTLV-I disease of 1 in 9 to 15 million. Conclusion: Three possible future courses of action for screening for HTLV-I are to screen every donation, to screen only new donors or to discontinue screening altogether. Using the information in this study, public discussion should be encouraged to assist stakeholders to agree on an acceptable level of risk and an appropriate level of screening for HTLV-I in Australia. MJA 1997; 166: 478-481 Introduction Human T-cell lymphotropic virus type I (HTLV-I) is endemic in all continents including Australia,1,2 where in some Aboriginal communities it is present in up to 14% of individuals.3-6 HTLV-I was first described in cases of aggressive T-cell leukaemia in southern Japan7 and in a T-cell lymphoma in a Jamaican patient.8 In areas endemic for the virus, 2% to 4% of those infected rapidly develop fatal leukaemia, with a peak incidence in the sixth decade of life.9 Progressive spastic myelopathy (or tropical spastic paraparesis), first reported in association with HTLV-I infection in Caribbean patients,9 occurs infrequently (0.25%).9 Transmission of HTLV-I is primarily by sexual contact and by infected cellular blood products (packed red cells, platelet-rich plasma and whole blood); breast milk is a major route of HTLV-I transmission in communities where the virus is endemic.10 In predominantly white communities, occasional cases of HTLV-I infection, in the absence of any risk factors, may represent transmission across several generations in a family.11 In a survey of 11 121 Northern Territory blood donors in 1991-1992, only one donor (with no evident risk factors) was confirmed as HTLV-I seropositive.6 Since March 1993, all Australian blood donors have been screened for antibody to HTLV-I. The decision to implement screening was made by the National Executive of the Australian Red Cross in February 1992, despite a lack of agreement among some government advisory bodies. The decision to screen for a rare transfusion-transmitted disease can be re-examined in the light of the results of HTLV-I screening of Australian blood donors from 1993 to 1995. The decision to screen blood for HTLV-I in Australia In October 1986, the American Red Cross proposed that, when a test became available, blood donors should be screened for HTLV-I as this retrovirus could be spread by blood transfusion and had been detected in blood donors.12 In November 1988, the United States Food and Drug Administration recommended testing of whole blood and cellular blood products for antibodies to HTLV-I,13 which was followed by a public review of the issues involved.14 Universal screening was also introduced in Canada, France, the Netherlands and Sweden. In Australia, after recommendations by the Red Cross National Blood Transfusion Committee (in September 1989 and later in March 1991) for HTLV-I screening to be funded, the then Federal Minister for Health sought advice from the National Health and Medical Research Council (NHMRC) in January 1992. In July 1991, the Communicable Diseases Standing Committee of the NHMRC had decided that the costs of screening Australian blood supplies (for HTLV-I) outweighed any public health benefits. This position was endorsed in 1992 by the Public Health Committee of the NHMRC15 and published by the Executive of the NHMRC in December 1992.16 However, the NHMRC also stated ". . . decisions on screening may have to be made on other than public health grounds".16 In February 1992, in the absence of a decision from NHMRC and acting on legal advice, the National Blood Transfusion Committee of the Australian Red Cross recommended that cellular blood products be screened from May 1992.17 By March 1993, although some State governments refused to fund HTLV-I screening, all Red Cross Blood Banks had commenced routine testing for HTLV-I under instruction from the Red Cross National Executive. A report by the Australian Health Ethics Committee, which was endorsed by the NHMRC in November 1993, stated that the cost of screening for HTLV-I was considerable and the risk to the community was low, and that "the spectre of large damages (litigation) . . . probably had a significant influence on the reasoning leading to implementation".17 It considered that the risk of transfusing HTLV-I-infected blood, although real, was rare. The report concluded that a decision not to screen all blood in Australia for HTLV-I would not be unethical.17 Blood donor data collection Data on the number of individuals who donated whole blood in Victoria between March 1994 and December 1995 and the interval from the date of the previous donation were extracted from Victorian Red Cross Blood Bank records. Patients attending for autologous, directed or therapeutic donations were excluded, as were donors returning for repeat testing or counselling only. Plasma donors were excluded because HTLV-I is not transmitted by plasma. The 1994-1995 attendance pattern was comparable with attendance patterns of repeat donors attending in June and July between 1993 and 1996 and was therefore applicable. Data on the number of whole blood collections in Australia and the number of new donors each year were obtained from the annual statistics of the Australian Red Cross Society. Data on the number of blood donations and the number of issues of cellular blood products in Australia for the financial years 1992-95 were obtained from national Red Cross statistics. Extrapolation of State (Victorian) data to national data Donor attendance patterns Repeat donors give many donations, so it is necessary to deduce the number of donors tested for HTLV-I from the total number of donations. The total number of repeat donations is used to calculate the number of repeat donors giving the donations by using a hypothetical model, together with the number of donations given by new donors. Using this information, the prevalence of HTLV-I in donors can then be estimated. Iterative model: The pattern of return after previous donations by Victorian repeat donors was applied to a hypothetical model in which 1000 donations were given each month over 34 months. Using an iterative spreadsheet model, the number of donors giving 1000 repeat donations each month from March 1993 to December 1995 was calculated (Figure, below). It was assumed that all repeat donors in Australia had a similar pattern of repeat donation. By analogy, the proportion of repeat donors contributing the 34 000 donations from repeat donors in the model was applied to the total number of donations from repeat donors in Australia over the 34-month period. To apply the model, it was assumed that the same number of repeat donations was given each month and that the discounting effect of donor rejection was constant over time. The number of Victorian donors and the intervals between donations were extrapolated to all repeat donations in Australia between March 1993 and December 1995 to derive a figure for donor exposure in person-years. "Donor exposure" is the sum of the time between one donation and the next for all donation intervals during the period. The sex and age distribution of Victorian donors was extrapolated to all Australian donors. Prevalence, incidence and risks of HTLV-I The number and characteristics of donors confirmed positive for HTLV-I were provided by Red Cross blood banks in each State and Territory. Donors were confirmed positive if their plasma reacted in triplicate with one of seven HTLV-1 enzyme-linked immunosorbent assay screening tests (Genetic Systems; Abbott; Cambridge recombinant/Ortho; Serodia particle-agglutination; Murex; Sanofi Platelia new; Organon Teknica), as approved by the National Reference Laboratory, and showed a diagnostic pattern on a western blot. The crude prevalence of HTLV-I was calculated by dividing the number of donors with HTLV-I by the total number of donors tested (repeat donors and new donors). The incidence of HTLV-I was calculated by dividing the number of seroconversions in repeat donors by the cumulative period of donor exposure. The risk of disease transmission was calculated from the infectivity rate for HTLV-I and the long-term risk of HTLV-I disease in patients. New and repeat Australian donors The return patterns for repeat donors over the period remained much the same. No donors returned within 12 weeks of donating blood. Of the donors who returned after making a previous donation, 0.29, 0.74, 0.84, 0.89, 0.92 and 0.94 had returned after successive quarters and 0.06 had returned after 18 months. Across Australia, first-time donors gave 335 183 whole blood donations and repeat donors gave 2 038 927 whole blood donations between March 1993 and December 1995. The proportion of first-time donations over the four fiscal years from July 1992 was 0.14, 0.14, 0.14, and 0.13, respectively. Screening of donors for HTLV-I By applying the pattern of return for repeat donors in Victoria to a hypothetical figure of 1000 donations a month over 34 months from repeat donors, it was concluded that the 34 000 donations would have been given by 19 197 donors in Victoria who had returned within 18 months. Furthermore, 6% (2040) would have been given by repeat donors whose previous donation was more than 18 months previously. By applying the Victorian model to the national statistics, it was concluded that over the 34 months 2 038 927 donations were given by 1 273 550 repeat donors. In addition, there were 335 183 new donors and donations. Therefore, a total of 1 608 733 individuals had been screened for HTLV-I. Some States had begun testing for HTLV-I before March 1993, but the high proportion of repeat donors in Australia, the small number of infected donors and the absence of serconversion permit the assumption that they would have been identified if screening had been delayed to March 1993. Prevalence and incidence of HTLV-I in blood donors To December 1995, 21 donors had been confirmed positive for HTLV-I in Australia, each on the first occasion the donor was tested. Five of the blood collections were for autologous transfusion. There were no seroconversions (a change in serological status from negative to positive) during the study period. Two donors were identified before June 1992, with 5, 8 and 6 in each subsequent 12 months. More men and more first-time donors were positive than expected, but the age distribution matched that of the general donor population (Box, below). There were no clear patterns of disease acquisition, although four of the five donors born in endemic areas were aged less than 40. Therefore, the crude prevalence of HTLV-I in Australian blood donors was 16 in 1 608 733, or 1 in 100 546. In Victoria, repeat whole blood donors gave 358 332 donations between March 1994 and December 1995, with intervals from the previous donation of up to 23.5 years. The repeat donors represent 11 851 014 person-weeks of exposure, or 227 904 person-years. By extrapolation, 2 038 927 repeat donations in Australia represent 1 296 785 person-years of exposure. There were no seroconversions. Therefore, the crude incidence of HTLV-I in Australian blood donors was less than 1 in 1 000 000 person-years. Risk of HTLV-I transmission via blood transfusion The risk of transmitting HTLV-I in Australia by blood transfusion over the study period was calculated from the number of donations given by 16 donors in 2 374 110 donations. The spreadsheet calculations showed that 1.6 million donors gave 2.4 million donations, so 16 donors would have given 24.6 donations. Therefore, the risk of receiving blood infected with HTLV-I before testing was about 1 in 100 000. The infectivity rate has been recently reported as 0.27,18 so only 1 in 370 000 transfusion recipients would become infected. Relatively few (2.5%-4%) people with HTLV-I infection not acquired by blood transfusion risk developing disease after 10 to 30 years.19 Therefore, the risk of developing HTLV-I disease from blood transfusion in Australia without testing would have been 1 in 9 to 15 million. Transfusion recipients, particularly those who are immunocompromised, may have a shorter incubation period,20 and infants of infected mothers have a 25% chance of becoming infected.21 If universal screening were discontinued, then the risk of transfusing infected blood would progressively return to the pretesting situation because of the recruitment of new donors from a population with the same characteristics as at present, as well as the retirement of repeat donors who have already been screened. In the absence of seroconversion, if only previously untested donors are screened for HTLV-I then there will be a zero risk of transmitting HTLV-I by blood transfusion. However, seroconversion has been reported in Dutch, French and American studies.22-24 If only the 14% of donations by new donors are screened, then the costs to Red Cross and the community would be significantly reduced. Discussion Our results have shown that the prevalence of HTLV-I in Australian donors is 1 in 100 000. By comparison, the prevalence of hepatitis C virus in new Victorian donors is 1 in 560; of hepatitis B virus, 1 in 650; and of HIV, 1 in 27 000. However, the introduction of universal blood screening for these diseases has reduced the risk to the transfused population for hepatitis C virus to 1 in 150 000, for hepatitis B virus to 1 in 150 000 and for HIV to 1 in 1.3 million.25 The risk of transfusing HTLV-I-infected blood would have been 1 in 100 000 without screening. In the United Kingdom, HTLV-I has been found in 1 in 20 000 donors; in the United States, in 1 in 6000; and in Sweden and the Netherlands, in 1 in 50 000.19 The decision by the Australian Red Cross to commence testing of all blood donations for HTLV-1 was contentious. Red Cross had shown that HTLV-I was present in the Australian blood supply,9 and believed that testing should be undertaken to ensure the safety of the blood supply as well as its own credibility.26 Our findings have shown that in the three years since screening began 16 Australian blood donors were found with HTLV-I. Without screening, the risk of viral transmission by blood transfusion would have been 1 in 100 000. On the other hand, the NHMRC and some State governments believed that universal screening was not justified on public health grounds. The findings in this study show that the risk of a blood transfusion recipient developing HTLV-I-related disease as a result of transfusion is about 1 in 10 million; these data were not available when Red Cross made their decision for universal screening. From the perspective of Red Cross and transfusion recipients, screening affords the certain benefit27 of the removal of the threat of HTLV-I infection from transfusion whatever the future risk of developing leukaemia. From a public health perspective, the certain benefit is the prevention of the very low risk of leukaemia or spastic paraparesis. Stakeholders (Australian Red Cross Blood Service, State and Federal governments and the community) would be assisted by public discussion of an acceptable level of risk and appropriate level of screening for rare transfusion-transmitted diseases; HTLV-I provides a suitable test case. It may be appropriate to screen only new donors for HTLV-I (at a lower cost) now that the donor base has been repeatedly screened. However, it is likely that Red Cross would need a form of statutory defence, such as that provided in Victoria for HIV and hepatitis C virus,28 if it were to apply less-than-universal screening for HTLV-I and other conditions of low risk to public health. Acknowledgement I thank the Directors of each State Blood Transfusion Service for providing the figures for HTLV-I and for constructive and critical comment. References Doherty RR. HTLV-I in Australia and Oceania: long term resident or recent immigrant? Med J Aust 1996; 164: 84-86. Gallo RC. A surprising advance in the treatment of viral leukemia. N Engl J Med 1995; 332: 1783-1784. Ascher D, Goudsmit J, Poeroy K, et al. Antibodies to HTLV-I in populations of the south western Pacific. J Med Virol 1988; 26: 339-351. Bastian I, Hinuma Y, Doherty RR. HTLV-I among Northern Territory Aborigines. Med J Aust 1993; 159: 12-16. Bastian I, Gardner J, Webb D, Gardner I. Isolation of a human T-lymphotropic virus type I strain from Australian Aboriginals. J Virol 1993; 67: 843-851. Bastian I, Dent J, McFarlane R, et al. HTLV-I among Northern Territory blood donors. Med J Aust 1993; 159: 7-12. Hinuma Y. Natural history of the retrovirus associated with a human leukemia. Bio essays 1985; 3: 205-209. Poeisz B, Ruscetti F, Gazdar A, et al. Detection and isolation of type C retrovirus particle from fresh and cultured lymphocytes of a patient with cutaneous T-cell lymphoma. Proc Natl Acad Sci U S A 1980; 77: 7415-7419. Hollsberg P, Hafler DA. Pathogenesis of diseases induced by human lymphtropic virus type 1 infection. N Engl J Med 1993; 328: 1173-1182. Nakano S, Ando Y, Saito K, et al. Primary infection of Japanese infants with adult T-cell leukemia associated retrovirus (ATLV): evidence for viral transmission from mothers to children. J Infect 1986; 12: 205-212. May JT, Stent G, Schnagl RD. Antibody to human T-cell lymphotropic virus type I in Australian Aborigines [letter]. Med J Aust 1988; 149: 104. Barnes D. HTLV-I: to test or not to test. Science 1988; 242: 372-373. Parkman PD. HTLV-I antibody testing; instruction of all registered blood establishments. Bethesda: Food and Drug Administration, 1988. Public Health Service Working Group. Licensure of screening tests for antibody to human lympotropic virus type I. MMWR Morb Mortal Wkly Rep 1988; 37: 736-747. Anon. Blood tests would drain funds: medics. Courier Mail 1992 June 5: 15. National Health and Medical Research Council, Communicable Diseases Standing Committee. HTLV-I screening: outcome of consideration by the Executive. Canberra: NHMRC, 4 December 1992. National Health and Medical Research Council. Case study of screening blood donations for human T-cell lymphotropic virus type I: Report of the Australian Health Ethics Committee. Canberra: NHMRC, November 1993. Donegan E, Lee H, Operskalski GM, et al. Transfusion transmission of retroviruses: human T-lymphotropic virus type I and II compared with human immunodeficiency virus type I. Transfusion 1994; 34: 478-483. Pagliuca A, Pawson R, Mufti GJ. HTLV-I screening in Britain. BMJ 1995; 311: 1313-1314. Kaplan JE, Litchfield B, Roualt C, et al. HTLV-I associated myelopathy associated with blood transfusion in the United States. Neurology 1991; 41: 192-197. Sugiyama H, Doi H, Yamaguchi K, et al. Significance of post-natal mother-to-child transmission of human T-lymphotropic virus type-1 on the development of adult T-cell leukemia/lymphoma. J Med Virol 1986; 20: 253-260. Vrielink H, van der Poel CL, Reesink HW. Efficacy of selected versus random blood donor screening for anti HTLV-I antibodies [letter]. Vox Sang 1995; 68: 251-252. Courouce AM, Pillonel J. Transfusion-transmitted viral infections. N Engl J Med 1996; 335: 1609-1610. Schreiber GB, Busch MP, Kleiman SH, Korelitz JJ. The risk of transfusion-transmitted viral infections. N Engl J Med 1996; 334: 1685-1690. Whyte G S, Savoia H F. The effectiveness of donor selection for reducing the risk of HCV, HBV and HIV in new blood donors in Victoria. Med J Aust . In press. Wylie B. HTLV-I: is donor screening really necessary? [editorial] Med J Aust 1993; 159: 4-5. McGuire A, Henderson J, Mooney G. The economics of health care. London: Routledge, 1994: 112. Health Act 1958 (Vic.), s. 139. Melbourne: The Law Printer, 1995.(Received 5 Aug 1996, accepted 18 Feb 1997) Authors' details Australian Red Cross, Blood Bank of Victoria, Southbank, VIC. Gordon S Whyte, FRACP, FRCPA, Director. Reprints will not be available. Correspondence: Dr G S Whyte, PO Box 354, Southbank, VIC 3205. E-mail: rcbb AT peg.apc.org ©MJA 1997 <URL: http://www.mja.com.au/> © 1997 Medical Journal of Australia.

Gordon S Whyte