among people consulting for HIV antibody testing: a pilot surveillance study in ..... (the characteristic that had the strongest association with seropositivity) is a ...
ORIGINAL RESEARCH * NOUVEAUTES EN RECHERCHE
Risk factors for HIV seropositivity among people consulting for HIV antibody testing: a pilot surveillance study in Quebec Michel Alary, MD, MSc; Josette Castel, MD The surveillance of AIDS (acquired immune deficiency syndrome) through case reporting only reflects the epidemiologic features of HIV (human immunodeficiency virus) transmission a few years earlier and not the prevalence of HIV seropositivity. HIV infection is not a notifiable condition in Quebec. We were asked by the ministere de la Sante et des Services sociaux du Quebec to perform a pilot project for the surveillance of HIV seropositivity using a network of sentinel physicians. From May 15, 1988, to Sept. 30, 1989, physicians from four collaborating centres collected data on the serologic status, demographic characteristics and risk factors for 4209 patients who underwent HIV antibody testing. Of the 3899 subjects included in the study 7.9% were HIV positive. Through logistic regression analysis the following variables were found to be significantly associated with HIV seropositivity: presence of HIV-related symptoms (prevalence odds ratio [POR] 36.5), origin from an endemic area (POR 9.1), homosexuality or bisexuality (POR 8.4), intravenous drug use (POR 4.2), male sex (POR 2.8), previous HIV antibody testing (POR 2.5) and previous sexually transmitted disease (POR 1.8). Over the study period we found a large increase in HIV seroprevalence among intravenous drug users (4.2% in 1988 to 19.0% in 1989) (p = 0.02). This increase might reflect a recent change in the epidemiologic pattern of HIV transmission in Quebec. Surveillance of HIV seropositivity through a network of sentinel physicians may be a reasonable alternative to mandatory reporting.
Puisque la surveillance du SIDA (syndrome d'immunodeficience acquise) par la declaration des cas reflete seulement l'epidemiologie de la transmission du VIH (virus immunodeficitaire humain) avec quelques annees de retard, il est important de mettre en place des methodes de surveillance de la seropositivite au VIH. L'infection a VIH n'est pas a declaration obligatoire au Quebec. Cependant, le ministere de la Sante et des Services sociaux du Quebec nous a demande de realiser un projet pilote de surveillance de la seropositivite au VIH a l'aide d'un reseau de medecins sentinelles. De 15 mai 1988 a 30 sept. 1989 les medecins de quatre centres collaborateurs ont recueilli des donnees sur le statut serologique, les caracteristiques demographiques et les principaux facteurs de risque chez 4209 sujets soumis a une serologie anti-VIH. Parmi les 3899 d'entre eux que nous etudions, la seropositivite au VIH se situe a 7,9%. Par la regression logistique on met les variables suivantes en rapport significatif avec la seropositivite au VIH: les sympt6mes d'une infection a VIH (risque relatif [RR] 36,5), la provenance d'une region
From the Departement de sante communautaire, h6pital du Saint-Sacrement, and the Equipe de recherche en epidemiologie, Departement
de medecine sociale et preventive, Faculte' de m'decine, universite' Laval, Quebec Reprint requests to: Dr Michel Alary, Equipe de recherche en epidemiologie, Departement de medecine sociale et preventive, Faculte de medecine, universite Laval, Quebec, PQ GIK 7P4 CAN MEDASSOCJ 1990; 143(1)
25
endemique (RR 9,1), l'homosexualite ou bisexualite (RR 8,4), l'utilisation de drogues injectables (RR 4,2), le sexe masculin (RR 2,8), les serologies anterieures a l'egard du VIH (RR 2,5) et la notion de maladies a transmission sexuelle anterieures (RR 1,8). Au cours de ce travail, la seroprevalence du VIH a augmente de facon importante chez les utilisateurs de drogues injectables (4,2% en 1988 et 19,0% en 1989) (p = 0,02). Ce dernier resultat pourrait refleter un changement recent dans l'epidemiologie de la transmission du VIH au Quebec. La surveillance de la seropositivite au VIH a I'aide d'un reseau de medecins sentinelles semble donc ere une alternative raisonnable a la declaration obligatoire de cette condition.
T he worldwide surveillance of AIDS (acquired immune deficiency syndrome) is currently done through case reporting. As of Oct. 31, 1989, the World Health Organization had received reports of 186 803 AIDS cases,' of which 2867 were from Canada. In Quebec 920 cases were reported as of Sept. 15, 1989.2 However, the surveillance system has some limitations. All cases are not necessarily reported.3 In addition, because of the long incubation period of AIDS (median 5 to 12 years4-6) the reports actually reflect the epidemiologic features of human immunodeficiency virus (HIV) transmission a few years earlier, whereas the true epidemic is the current propagation of HIV infection. To overcome these limitations new strategies have been developed to implement methods for the surveillance of HIV seropositivity. For instance, anonymous unlinked seroprevalence studies in different populations are under way, and the preliminary results have been published.7'8 Some or all jurisdictions in at least 25 countries require that all cases of HIV infection be reported to public health officials.9 In the United States in October 1988 reporting of HIV infection was required in 24 states, with nominal reporting in 16.10 In Canada reporting is currently mandatory in eight provinces. In Quebec an AIDS advisory committee to the minister of health did not recommend the inclusion of HIV infection as a notifiable disease;1' such a regulation could cause people with risk factors to fear consultation for HIV antibody testing and counselling. Instead, the committee recommended that alternative methods of HIV surveillance be studied and implemented over the next few years. Because of our previous experience with surveillance of sexually transmitted diseases (STDs) through a network of sentinel physicians,'2 the ministere de la Sante et des Services sociaux du Quebec (MSSS) asked us to apply a similar method for the surveillance of HIV seropositivity. In a pilot project we sought collaboration from physicians providing care to people at high risk for HIV infection. We report data on the risk factors associated with HIV seropositivity among patients for whom HIV antibody testing was requested by the physicians. In addition, we examine the trends in risk factors over 16 months. 26
CAN MED ASSOC J 1990; 143 (1)
Methods Data collection In May 1988 we asked physicians at four centres in Quebec to participate in the project. Three of the centres were public AIDS prevention and screening clinics designated by the MSSS (two in Montreal and one in Quebec City), and the fourth was a large, private STD clinic in Montreal. All of the physicians agreed to participate. Data collection started on May 15, 1988, and is still under way. We examined only the data received by Sept. 30, 1989. The physicians were asked to fill out an anonymous information form for all patients for whom HIV antibody testing was requested. Informed consent was obtained before testing was done. Most of the patients were willing to give consent because they consulted their physicians specifically for testing. In other instances the testing, prescribed on clinical or epidemiologic grounds, was agreed to after discussion with the physician. The following variables, which are routinely collected for all people attending the participating centres, were recorded at the time of consultation: sex, age, country of birth, place of residence, reason for testing, sexual orientation, number of sexual partners during the preceding year, history of STDs (including gonorrhea, chlamydial infection, syphilis, genital herpes, condyloma and hepatitis B), history of HIV antibody testing and presence of other risk factors for HIV infection (including intravenous drug use, hemophilia, receipt of blood or blood products before 1985 and occupational hazards). Each form was identified with an anonymous code number. The coding system was determined at each centre; identification was possible for people who were tested repeatedly within each centre but not for those who were tested at more than one centre. Once the test results were received the physicians were asked to add the information to the form and send it to the project coordinator.
Laboratory procedures The serum samples were sent to one of eight microbiology laboratories designated by the MSSS
to perform enzyme-linked immunosorbent assay (ELISA) for HIV antibody detection. All of the laboratories used an ELISA kit designed by Genetic Systems Co., Seattle; this kit has been shown to be slightly superior to four others.'3 If the first result was positive the test was repeated twice. Samples found to be positive on two separate testings were sent to the Laboratoire de sante publique du Quebec, Ste-Anne-de-Bellevue, for confirmation. The first confirmatory test was indirect immunofluorescent assay (IFA) on Molt4-T4 cells infected with the lymphadenopathy-associated virus strain of HIV type 1. The serum samples were first diluted 1:4 and then incubated for 30 minutes with the infected cells. The cells were washed and then incubated with the F(ab' )2 fraction of goat antihuman IgG and a counter-colorant. The slides were examined for viral inactivation; a panel of serum samples known to be positive, weakly positive and negative were included for quality control. Any samples showing a positive reaction were considered confirmed as positive. Those with a weakly positive or negative reaction were tested by means of radioimmunoprecipitation assay;'4 samples reacting to
(92.6%) of which were included in the study. Of the 310 forms excluded, 294 were duplicates and 16 did not contain serologic results. According to the confirmatory test results the rate of HIV seroprevalence was 7.9% (308 of 3899). Table 1 summarizes the distribution of the demographic characteristics and risk factors. The mean age was slightly higher among the men than among the women (31.8 v. 30.5 years). Of the 3882 subjects for whom age was given 3185 (82.0%) were 20 to 39 years old. However, the seropositivity rate was highest among those 40 to 49 years of age (47 [9.7%] of 485); the next highest rates were among those aged 30 to 39 (130 [9.1] of 1427) and those aged 20 to 29 (121 [6.9%] of 1758). Only two of the patients younger than 20 and seven of those over 50 were seropositive (age was unavailable for one seropositive patient). Of the 3687 subjects for whom place of residence was given 3377 (91.6%) were living in the Montreal area; 274 (94.2%) of the 291 seropositive people for whom place of residence was recorded were also in that area. Sex and sexual orientation were both available for 3803 of the subjects. Homosexual and gp 160/120 HIV glycoproteins were considered posi- bisexual people accounted for 41.9%; the proportion was much higher among the men tive. All of the procedures were routinely.used by the (1526 [59.8%] of 2551) than among the women laboratories. (67 [5.4%] of 1252).
Statistical analysis SAS (SAS Institute Inc., Cary, NC) was used for statistical analysis. The distribution of risk factors for HIV infection and other variables was compared according to serologic status. Association was measured with the prevalence odds ratio (POR). We tested the statistical significance of the different crude PORs using the chi-squared or Fisher's exact test (bilateral p values are reported) and examined the independent effect of the variables on the prevalence of HIV seropositivity using logistic regression analysis. '5 The time trends of seroprevalence and of the distribution of risk factors were examined. For this analysis we included only the subjects who were tested before Sept. 1, 1989. Indeed, the delay in reporting positive results may be as long as 1 month, whereas the delay in reporting negative results is generally 2 to 4 days.
Results
Study population From May 15, 1988, to Sept. 30, 1989, 4209 anonymous information forms were received, 3899
Table 1: Characteristics of 3899 patients who underwent human immunodeficiency virus (HIV) antibody testing in a pilot surveillance project in Quebec
Characteristc Male sex, % (n = 3875) Mean age, yr (n = 3882) Non-Canadian origin, % (n - 3318) Origin from an endemic area, % (n = 3734) Hemophilia or blood transfusion, % (n = 3734) Intravenous drug use, % (n = 3734) Homosexuality or bisexuality, % (n = 3811) Sexual partner with risk factors, %
Value
66.8 31.3 7.7 1.4 0.7 3.7 41.9
(n = 3734)
20.2
>2
67.3 12.2
No. of sexual partners,* % (n = 3364) 10 Previous sexually transmitted disease (STD), % (n = 3623) Occupational risk factor, % (n = 3734) No known risk factor, % (n = 3734) Presence of HIV-related symptoms, % (n = 3831) Previous HIV antibody testing, % (n = 3833) *
40.3
1.0 16.3 3.5 25.3
*Number of sexual partners within 1 year before HIV antibody testing. CAN MED ASSOC J 1990; 143 (1)
27
Risk factors for HI V seropositivity Of the subjects 608 (16.3%) did not report any risk factors for HIV infection; 6 (1.0%) of them were HIV positive. In a bivariate analysis seropositivity was statistically associated with presence of symptoms related to HIV infection, homosexuality or bisexuality, male sex, previous HIV antibody testing, sexual contact with a seropositive person, previous STD, 10 or more sexual partners during the preceding year and intravenous drug use (Table 2). To separate these effects we designed a logistic regression model that included the risk factors associated with seropositivity in the bivariate analysis, age, origin from an endemic area and place of residence (Montreal area or other). In this model the association of seropositivity with number of sexual partners and sexual contact with an HIV-infected person was no longer significant, whereas the other associations remained significant. In addition, origin from an endemic area (Haiti or central Africa) emerged as a strong risk factor. However, the adjusted PORs were generally lower than those in the bivariate analysis (Table 3, model 1). Since the presence of HIV-related symptoms (the characteristic that had the strongest association with seropositivity) is a consequence rather than a cause of HIV infection we also designed a logistic regression model that excluded this characteristic. The PORs were essentially similar to those in model 1 for the other risk factors (Table 3, model 2). Logistic regression models were also built after
stratification for sexual orientation. Among the homosexual subjects HIV seropositivity was significantly associated with previous STD (POR 1.9, p < 0.0001) and previous HIV antibody testing (POR 1.8, p = 0.000 1). Among the heterosexual subjects the significant risk factors were intravenous drug use (POR 13.3. p < 0.0001), origin from an endemic area (POR = 14.7, p < 0.000 1) and previous HIV antibody testing (POR = 17.8, p < 0. 000 1).
Time trend analysis Between May 1988 and August 1989 the monthly seroprevalence rates fluctuated between 4.2% and 10.2%, but no specific trend was observed. However, a major change occurred with respect to the relative importance of risk factors for seropositivity; intravenous drug users accounted for only 2 (1.8%) of the 110 HIV-positive subjects in 1988, whereas they accounted for 16 (9.3%) of the 172 HIV-positive subjects identified between January and August 1989 (p = 0.01, Fisher's exact test). Table 4 shows the prevalence of HIV seropositivity in 1988 and 1989 for all of the heterosexual subjects, the intravenous drug users, the homosexual subjects and the heterosexual nonusers of intravenous drugs. Significant differences in the prevalence
between 1988 and 1989 were found for all heterosexual subjects and intravenous drug users. However, since 15 of the 18 seropositive intravenous drug users were heterosexual (9 men, 6 women) no time trend was observed among the heterosexual nonusers
Table 2: Distribution of risk factors for HIV seropositivity among seropositive and tive subjects
seronega..
HIV status; no. (and %) of subjects Risk factor
Presence of HIV-related
symptoms Homosexuality orbisexuality Male sex Previous HIV antibody
testing Seropositive sexual partner Previous STO
Seropositive 96 (31 .9) (n = 301) 268 (87.9) (n 305) 290 (94.8) (n :- 306) 175 (57.8) (n = 303) =
39 (13.3)
(n
Onigin from an endemic area i.ntravenous drug use
293)
170 (57.8)
294) 50 (21 .4) (n = 233) 7 (2.4) (n = 293) (n
0 sexual partners
=
=
18
(6.1)
(n -293)
Seronegative
r-atio (PORY
40 (1.1) (n 3530) 1327 (37.8) (n 3.1-506) 2300 (64.4) (n = 3569) 795 (22.5) (.n 3530) 202 (5.9) (n = 3441)
40.9
=
z-
1290 (38.8)
(n =3329) 352 (11.2) (n 3131)
CAN MED ASSOC J 1990; 143 (1)
2-5
2.22 '.9
=
44
(1.3)
3441) 119 (3.4) (n 3441)
(n
=
:.8
=
*All PORs had p values less thani 0(1001 by chi-squared analysis except for intravenous drug uise (p from an endemic area (P= 0.1). 28
1(10-
0 02) and origiu
of intravenous drugs. A significant positive linear trend was observed when we analysed the monthly seroprevalence rates among intravenous drug users from May 1988 to August 1989 (X2 = 6.42, p = 0.01 1).
Discussion This pilot project provided data on HIV seropositivity among high-risk people, all of whom either asked to be tested or were offered testing because of clinical or epidemiologic considerations. However, even in this high-risk group, which is not representative of the general population, the main risk factors for AIDS (male homosexuality, intravenous drug use, origin from an endemic area and previous STD) emerged as risk factors for HIV seropositivity. The lack of association with high number of sexual partners may be because this information was collected for only 1 year before HIV antibody testing. Recent changes in sexual behaviour might have masked high-risk practices that were present a few years earlier. The association of seropositivity with history of STD probably reflects, at least to
some extent, the relation between HIV infection and high number of sexual contacts over a longer period. The seropositivity rate among the subjects who did not report any risk factor for HIV infection was higher than that expected in the general population. Some of these subjects may have been reluctant to report high-risk behaviour to their physicians. Nevertheless, the data suggest that a significant number of people may require HIV antibody testing despite a self-reported low level of risk. We received forms for only 4209 (4.3%) of the 96 788 tests that were requested by Quebec physicians during the study period.'6 Whereas the seropositivity rate in our study was 7.9% the rate for all tests done in Quebec was 3.2% (3092 of 96 788). All of the physicians who collaborated in our pilot project practise in the two largest cities of Quebec (Montreal and Quebec City), from which over 90% of the AIDS cases in Quebec are reported.2 Therefore, this geographic concentration of collaborating physicians may explain, at least to some extent, the difference in the seroprevalence rates. The increase in HIV seroprevalence among intravenous drug users as well as the increased propor-
Table 3: Independent risk factors for HIV seropositivity after adjustment by logistic regression
Model 2t
Model 1 * Risk factor
POR
p value
Presence of HIV-related symptoms Origin from an endemic area Homosexuality or
36.5 9.1
< 0.0001
-
-
0.0001
8.1
< 0.0001
8.4 4.2
< 0.0001 < 0.0001
8.5 3.4
< 0.0001 0.0002
2.8
0.003
2.5
0.028
2.5 1.8
< 0.0001
2.4 1.6
< 0.0001
bisexuality Intravenous drug use Male sex Previous HIV antibody testing Previous STD *lncludes all the variables.
POR
0.0002
p value
0.0005
tincludes all the variables except for HIV-related symptoms.
Table 4: HIV seroprevalence in 1988 and 1989 by group Year; no. (and %) of HIV-positive subjects
Group Heterosexual subjects
Intravenous drug users Homosexual subjects Heterosexual nonusers of intravenous drugs
1988 7 (0.8) (n = 818) 2 (4.2) (n = 48) 111 (18.4) (n = 604) 5 (0.7) (n = 757)
1989 30 (2.3) (n = 1327) 16 (19.0) (n = 84) 147 (16.0)
p value* 0.02
0.02 0.24
(n=921)
15 (1.2) (n 1210)
0.25
=
*Two-tailed Fisher's exact test used to compare seroprevalence between years. CAN MED ASSOC J 1990; 143 (1)
29
tion of intravenous drug users among seropositive people might reflect a recent change in the local pattern of HIV transmission that was not detected through AIDS surveillance. In the United States intravenous drug use was related to 27% of AIDS cases reported in 1988,8 whereas in Canada it was related to only 3.6% of the cases reported by Sept. 25, 1989.17 The rates in Quebec were similar to the national rates and have been stable over time.2 Since we cannot exclude a bias related to self-selection the trend in our study should be seen as the capability of this surveillance system to give alarm signals. More specific studies should be done to confirm the presence of a new problem. In the case of intravenous drug use the situation may be urgent, as this mode of transmission may quickly lead to high seroprevalence rates among people with this risk factor. 18 The main limitation of any HIV infection surveillance system based on cases identified in clinical settings is the incomplete ascertainment of HIVinfected people. Since most infected people are still asymptomatic, ascertainment is mainly linked to self-selection by those who ask for testing and to clinical suspicion by physicians who request the test. Despite this limitation mandatory reporting of HIV infection is required in many developed countries.9 '0 We believe that the surveillance method used in this pilot project could be a reasonable alternative to mandatory reporting of HIV infection. Our method may be socially more acceptable than mandatory reporting and has the advantage of collecting data about seronegative as well as seropositive people. This type of data collection allows the comparison of prevalence rates. Thus, although the rates are biased because of self-selection, comparisons (through PORs or trend analysis) between prevalence rates should not be severely biased. Indeed, to bias these types of association the reasons for self-selection would have to be very different among seropositive and seronegative subjects. If the surveillance system were to collect only data about HIV-infected people any trend detection in the proportion of infected subjects in a given risk group could be severely biased. For instance, the large increase in the proportion of seropositive intravenous drug users from 1988 to 1989 could only be due to a rise in the number of intravenous drug users who underwent HIV antibody testing. However, to bias an increase of seroprevalence among intravenous drug users the reasons for self-selection would have to change in a different way over time among seropositive and seronegative people. This is unlikely because the distribution of other risk factors (including presence of HIV-related symptoms) did not change over time. Therefore, as in the case of intravenous drug use one could eventually suspect changes in heterosexual 30
CAN MED ASSOC J 1990; 143 (1)
transmission patterns or other changes in the epidemiologic features of HIV infection. The main disadvantages of our system are the possible underreporting with respect to the total number of HIV antibody tests requested by physicians and the problem of identifying people who have been tested more than once. Some of these disadvantages can be overcome. The number of physicians collaborating in the surveillance system should be enlarged to represent a greater proportion of all serologic tests performed in Quebec. In addition, a unique anonymous identifier could be used to detect duplications. Furthermore, by using the same coding procedure for all tests one could couple the database of the sentinel physicians with the database of the HIV diagnostic program. This coupling would allow comparison of basic demographic factors (age, sex, place of residence) between the sentinel physicians' patients and all people undergoing HIV antibody testing in Quebec. The use of a standard coding system would be much more difficult in anonymous testing sites, because patients do not identify themselves when attending these premises; however, in Quebec less than 2% of the tests are currently performed at anonymous testing sites. Because of these limitations our proposed surveillance method cannot be used to monitor the prevalence of HIV infection in the general population. To be complete, HIV infection surveillance should include anonymous, unlinked seroprevalence studies. Two such studies are currently being performed in Quebec; one involves sentinel hospitals and the other childbearing women.19 Over the next few months the current network of sentinel physicians will be broadened. In addition, improved services for counselling and psychologic support of HIV-infected people will be implemented.
References 1. World Health Organization: Acquired immunodeficiency syndrome (AIDS) - data as at 31 October, 1989. Wkly Epidemiol Rec 1989; 64: 337-338 2. Surveillance des cas du syndrome d'immunodeficience acquise, Quebec. Cas cumulatifs 1979-1989 (no 89-5), Bureau regional des maladies infectieuses, Montreal, 1989 3. Piot P, Plummer F, Mhalu F et al: AIDS: an international perspective. Science 1988; 239: 573-579 4. Melbye M, Biggar RJ, Ebbensen P: Long-term seropositivity for human T-lymphotropic virus type III in homosexual men without the acquired immunodeficiency syndrome - development of immunologic and clinical abnormalities: a longitudinal study. Ann Intern Med 1986; 104: 496-500 5. Medley GF, Billar L, Cox DR et al: The distribution of the incubation period of the acquired immunodeficiency syndrome (AIDS). Proc R Soc Lond [Biol] 1988; 233: 367-377 6. Anderson RM, Medley GF: Epidemiology of HIV infection and AIDS: incubation and infectious periods, survival and vertical transmission. AIDS 1988; 2 (suppl 1): S57-S63 7. Hoff R, Berardi V, Weiblen BJ et al: Seroprevalence of human immunodeficiency virus among childbearing women:
8. 9.
10.
11. 12.
13.
estimation by testing samples of blood from newborns. N Engl J Med 1988; 318: 525-530 AIDS and human immunodeficiency virus infection in the United States: 1988 update. MMWR 1989; 38 (suppl 4): S1S38 Berkelman RL, DeKoven J: The role of HIV infection reporting in public health. In Abstracts from the Vth International Conference on AIDS, Montreal, June 4-9, 1989, International Development Research Centre, Ottawa, 1989: 59 Stehr-Green J, Berkelman RL, Van Patten H et al: HIV infection reporting in the United States. Ibid: 27 Groupe de travail sur le SIDA: Rapport synthese, ministere de la Sante et des Services sociaux du Quebec, Quebec, 1988: 94-103 Alary M, Joly JR, Poulin C: Incidence of four sexually transmitted diseases in a rural community: a prospective study. Am JEpidemiol 1989; 130: 547-556 Ozanne G, Fauvel M: Performance and reliability of five commercial enzyme-linked immunosorbent assay kits for
Conferences continued from page 16 Sept. 13, 1990: Canadian Association of Physical Medicine and Rehabilitation Annual Symposium - "Research Methods in Rehabilitation Medicine" Prince of Wales Hotel, Niagara-on-the-Lake, Ont. Dr. Pat Nance, Nova Scotia Rehabilitation Medicine Centre, 1341 Summer St., Halifax, NS E3H 4K4
Sept. 13-15, 1990: New Brunswick Medical Society Annual General Meeting Hotel Beausejour, Moncton Ms. Judy Orem, annual general meeting coordinator, New Brunswick Medical Society, 176 York St., Fredericton, NB E3B 3N8; (506) 458-8860
Sept. 13-15, 1990: Ontario Medical Association and Canadian Anaesthetists' Society Annual Fall Meeting Niagara Falls, Ont. Dr. F. Halliday, OMA/CAS Annual Fall Meeting, c/o Greater Niagara General Hospital, PO Box 1018, Niagara Falls, Ont. L2E 6X2; (416) 358-0171, ext. 474 Sept. 14-16, 1990: Canadian Hospital Association 7th Annual Invitational Seminar on Health Care Directives Millcroft Inn, Alton, Ont. Conferences, Canadian Hospital Association, 100- 17 York St., Ottawa, Ont. KIN 9J6; (613) 238-8005, FAX (613) 238-6924 Sept. 14-17, 1990: Royal College of Physicians and Surgeons of Canada Annual Meeting (held in conjunction with the Annual Meeting of the Canadian Pediatric Society and Canadian Society for Clinical Investigation) Metro Toronto Convention Centre Anna Lee Chabot, coordinator, Royal College of Physicians and Surgeons of Canada, 74 Stanley St., Ottawa, Ont. KIM 1P4; (613) 746-8177, FAX (613)746-8833
14. 15. 16.
17. 18.
19.
anti-human immunodeficiency virus antibody in high-risk subjects. J Clin Microbiol 1988; 26: 1496-1500 Barin F, McLane MF, Allan JS et al: Virus envelope protein of HTLV-III represents major target antigen for antibodies in AIDS patients. Science 1985; 228: 1094-1096 Kleinbaum DG, Kupper LL, Morgenstern H: Epidemiologic Research. Principles and Quantitative Methods, Van Nos Reinhold, New York, 1982: 447-491 Programme quebecois de diagnostic des infections HIV. Rapport periodique, Laboratoire de sante publique du Quebec, Ste-Anne-de-Bellevue, 27 oct 1989 Surveillance Update: AIDS in Canada, Federal Centre for AIDS, Ottawa, Sept 25, 1989 Phanuphak P, Poshyachinda V, Un-eklabh T et al: HIV transmission among intravenous drug abusers. In Abstracts from the Vth International Conference on AIDS, Montreal, June 4-9, 1989, International Development Research Centre, Ottawa, 1989: 981 Hankins C, Lapointe N: Epidemiology of HIV infection. Can J Public Health 1989; 80 (suppl 1): S21 -S23
Sept. 14-17, 1990: 67th Annual Meeting of the Canadian Paediatric Society (held in conjunction with the Annual Meeting of the Royal College of Physicians and Surgeons of Canada) Metro Toronto Convention Centre Dr. Victor Marchessault, executive vice-president, Canadian Paediatric Society, 401 Smyth Rd., Ottawa, Ont. K1H 8L1; (613) 737-2728
Sept. 15-23, 1990: British Medical Association Annual Scientific Meeting Edinburgh Meetings Department, PO Box 8650, Ottawa, Ont. KIG OG8; 1-800-267-9703, FAX (613) 731-9013 Le 20-22 sept. 1990: Le septieme congres scientifique de l'Association des mddecins d'urgence du Qudbec (AMUQ) L'h6tel Bonaventure, Montrdal Dr F. Bertrand, directeur, Septieme congres scientifique, AMUQ, 208 Niagara, Kirkland, PQ H9J 3W7
Sept. 21-23, 1990: Dermatology '90: Therapeutic Update New World Harbourside, Vancouver Dermatology '90, 204-402 W Pender St., Vancouver, BC V6B 1T6; (604) 669-7175, FAX (604) 669-7083 Sept. 22-28, 1990: 23rd International Congress on Occupational Health: Sharing Solutions Montreal Convention Centre Secretariat, 23rd International Congress on Occupational Health, 2-58 de Brdsoles St., Montreal, PQ H2Y 1V5; (514) 499-9835, FAX (514) 288-4627 Le 30 sept.-le 3 oct. 1990: 4e Congres international francophone de gdrontologie Palais des congrbs, Montrdal Les services de congrbs GEMS, 100-4260 Girouard, Montrdal, PQ H4A 3C9; (514) 485-0855, FAX (514) 487-6725
continued on page 37 CAN MED ASSOC J 1990; 143 (1)
31
