Final Report: Estimating the Number of Persons co-infected with Hepatitis C Virus and Human Immunodeficiency Virus in Canada

3. Results

3.1 HIV prevalence in Canada as of December 1999

Table 1 shows the estimates of HIV by exposure category for December 1999. Overall, there were an estimated 49,800 HIV-infected persons in Canada, an increase of 24% compared to 1996. The most important groups with respect to absolute numbers of HIV infections were (in descending order): MSM, IDUs and persons infected by heterosexual contact; these three groups accounted for 95% of all HIV infections. IDU and MSM alone accounted for 80% of HIV infections in Canada. The plausible ranges derived from the Monte Carlo simulation for each exposure category were: MSM 26,00-33,600, MSM-IDU 1,600-2,600, IDU 7,900-11,700, heterosexual (HIV-endemic and other heterosexual contact) 6,100-10,100, and Other (recipients of clotting factors and blood transfusion) 330-520. The plausible range for HIV prevalence in Canada as a whole was 45,200-54,600.

3.2 H3CV prevalence

3.2.1 Literature search

The literature yielded many studies about HCV prevalence among HIV-infected persons, especially among IDUs and MSM. A more limited amount of information was available for hemophilia patients and multiply transfused patients. Due to lack of good data, the other HIV-related exposure categories are subject to greater uncertainty. However, due to the smaller numbers involved and the lesser importance of HCV infection, this will have limited impact on the final result.

With respect to prisoners in particular, we obtained several useful articles. We identified only one study on HCV infection among Haitans.

One consistent finding across almost all studies was the substantially higher prevalence of HCV among HIV-infected persons compared to HIV-uninfected subjects.

3.2.1.1 HCV infection among HIV-infected men who have sex with men

Approximately 15 studies of HCV infection within this population were identified and analyzed. In addition, as previously noted, an analysis of data from a study of homosexual men in Vancouver was kindly provided by Kevin Craib of the British Columbia Centre for Excellence on HIV. The results are presented below and summarized in Table 2.

The results of the review as a whole provided relatively consistent estimates of HCV among HIVinfected MSM. A large study of 1,038 homosexual men carried out in Sydney, Australia in 1984-85 <24> observed an HCV prevalence of 12% among HIV-infected MSM and 4.0% among those uninfected by HIV.

A study by Wormser <25> in New York City examined 31 HIV-infected MSM, and found none infected with HCV. However, the sample size was limited; the upper confidence interval of the HCV prevalence was 11%.

A large study of 1,058 MSM in Pittsburgh, USA by Ndimbe <26> examined HIV and HCV prevalence in 1984-85. The prevalence of hepatitis C markers among the 207 HIV-infected MSM was 5.8% versus 2.2% among the 851 HIV-negative subjects.

A study by Francisci in Perugi, Italy of HIV-positive MSM patients from 1985 to 1992 <27> observed a prevalence of HCV of 6.7%.

Particularly relevant to the present project is a study carried out by Corona <28> in an STD clinic in Rome, Italy in 1989. He studied several different groups of patients at risk for HIV. Among the 195 MSM examined, 4 (2.1%) were infected with HCV. HCV prevalence was a function of age: none of the 99 men under age 35 were HCV-positive, compared to the 4 (4.2%) of 96 men 35 years of age or older. Corona eliminated from these analyses any MSM who had injected drugs. None of the 24 subjects who were HIV-positive were also HCV-positive.

Sonnerborg carried out a study of 107 MSM in Stockholm, Sweden in 1988-89 <29>. He found an HCV prevalence of 14% among 59 HIV-infected men and 0.0% HCV prevalence among 48 HIVnegative men.

A laboratory-based study by Anand and colleagues examined sera submitted for HIV testing at a provincial laboratory in Canada <30>. Overall, 8 (7.9%) of the 101 specimens received from MSM were HCV-positive. Five (9.3%) of the 54 HIV-positive men were also HCV-positive, compared to 3 (6.4%) of the 47 HIV-uninfected men. The difference observed was not statistically significant.

Finally, we reviewed the preliminary results of an important study by Craib among the original Vancouver Lymphadenopathy Study (VLAS) participants enrolled from 1982 to 1984 <31>. HIV and HCV test results were available for 662 of 729 recruited subjects. Overall, 5.9% or 39 of the 662 men were HCV-positive. However, HCV prevalence was substantially higher among HIV-infected men compared to HIV-negative men: 31 (8.8%) of 352 HIV-infected men were HCV-infected, compared to 8 (2.6%) of 310 HIV-negative men. The 3.6 fold difference was statistically significant (p <0.001). Interestingly, on further examination, Craib found that a small but significant number of subjects (41) in his study admitted to injecting drugs at some time during their lifetime. Stratifying HCV prevalence by lifetime history of drug injection yielded an estimate of HCV prevalence of 3.1% among those who had never injected drugs. HIV-stratfied results among MSM who had never injected drugs were 4.3% among the 322 HIV-positive men and 1.7% of the 299 seronegative men. In contrast, among MSM who had injected drugs during their lifetime, HCV prevalence was 49%, 57% among the 30 HIV-positive men and 27% among the 11 HIV-negative men.

In summary, the HCV prevalence among HIV-infected MSM appears relatively consistent and, for the studies examined, is in the range 3 to 14%. The HCV prevalence among non-IDU MSM observed by Craib among the VLAS participants was somewhat lower than the results from most other studies. This may have two possible explanations: HCV prevalence may be somewhat lower in Vancouver than in the other study populations examined such as in Spain and the United States where HCV prevalence is higher than in Canada. Second, it is possible that, in many of the other studies, some MSM included in the analysis had injected drugs. Given the high HCV prevalence associated with injection of drugs, this could well have upwardly biased their results. It should also be noted that HCV prevalence in British Columbia is somewhat higher than elsewhere in Canada. Thus, a reasonable mid-point for HCV prevalence among non-injecting HIV-infected MSM in Canada would be 5%, similar to that observed by Craib in his Vancouver study.

The results of the studies of HCV infection among HIV-infected MSM we reviewed are summarized in Table 5.

3.2.1.2 HCV infection among HIV-infected injection drug users

A large number of studies of HCV prevalence among IDUs stratified on HIV prevalence were identified. We selected 14 for presentation in this review. The results were relatively consistent across different countries and over different periods and provide relatively convergent estimates for the final estimate of number of HCV-HIV co-infections. As with MSM, HCV prevalence among HIV-infected IDUs was higher than for HIV-negative IDUs in all studies reviewed.

A review by Crofts of studies carried out from 1970 to the 1990's in Australia <32> revealed a range in HCV prevalence of 18% to 94%. The majority of the 14 studies cited yielded estimates in the range of 50% to 70%. Nevertheless, these studies did not stratify on HIV infection; therefore HCV prevalence among IDUs who were HIV-infected would be substantially higher than the crude HCV prevalence observed. In fact, a study by the same author among 311 IDUs in Victoria, Australia recruited through social networks in 1990-91 <33> observed an HCV prevalence of 79% among 14 HIV-infected IDUs, compared to 68% of 297 HIV-uninfected IDUs. It should be noted, however, that the number of HIV-infected subjects was small.

A study of a similar order of magnitude was carried out by Rodriguez <34> among IDUs in Spain at a drug treatment unit in 1993-94. Rodriguez found that 92% of the 26 HIV-infected IDUs studied were HCV-positive compared to 86% of the 95 HIV-negative subjects.

A small early study by Wormser in New York City <25> among 58 IDUs recruited beginning in 1987 found a rate of 66% among 50 HIV-infected IDUs.

Coppola studied 137 IDUs in Sardinia, Italy in 1992-93 <35>. He observed an HCV prevalence of 91% among the 32 HIV-infected IDUs, but only 78% among the 105 HIV-uninfected IDUs. A somewhat larger study by Francisi <27> of IDUs, also in Italy, examined HCV serostatus among 351 HIV-infected IDUs from 1985 to 1992. In this study, 252 (72%) of HIV-infected subjects were HCV-infected.

Sonnerborg, in the same study cited above for MSM, examined 99 IDUs in Stockholm in 1988-89 <29>. The prevalence of HCV was 94% among the 52 HIV-positive subjects and 79% among the 47 HIV-negative subjects.

A large study among IDUs in Baltimore, Maryland, USA was carried out by Thio in 1988-89 <36>. She found that 98% of 559 HIV-positive IDUs studied were HCV-infected compared to 87% of 944 HIV-negative IDUs. Overall, the prevalence of HCV in this population was 91%.

Weinstock recently published a study among a large sample of HIV-infected patients in suburban New York City <37>. Sera were apparently collected from 1989 to 1995. This study which included 582 IDUs found that 474 (81%) were HCV-positive.

A seroepidemiologic study among IDUs was carried out in Cape Breton Island, Nova Scotia by Health Canada in collaboration with the local and provincial health departments <38>. Subjects were recruited from both the community and a local correctional facility. Among 92 IDUs tested, 43 (47%) were HCV-positive, and 5 (5.4%) were HIV-positive. The joint prevalence of HCV and HIV was not specifically presented.

A similar study, also conducted by Health Canada, was carried out in Prince Albert, Saskatchewan <39 > with similar results. Of the 188 IDUs studied, 93 (49%) were HCV-positive and two (1.1%) were infected with HIV. The joint prevalence of HCV and HIV was not presented in this study either. However, in both of these studies, the number of HIV-infected IDUs was small and would have yielded only limited information about HCV prevalence among HIV-infected IDUs.

In summary, essentially all studies identified and reviewed for the purposes of the present study observed an extremely high rate of HCV infection among injection drug users, and an even higher rate among those who were HIV-infected. Rates ranged from approximately 50% to 98%, with most results in the 70% to 95% range.

The results of the studies reviewed on HCV infection among HIV-infected IDUs are summarized in Table 3.

3.2.1.3 HCV infection among HIV-infected MSM-IDU

Few of the studies reviewed stratified MSM into those who were and were not injection drug users. Nor, conversely, did many identified studies examine sexual orientation among IDUs. However, the study by Craib <31> did examine HCV prevalence among the HIV-positive men who had a history of injection drug use. It must be noted that the analysis was based on a question about having ever injected, although such behaviours may have been sporadic and remote. Nevertheless, Craib found that 20 (49%) of the 41 persons who had ever injected were HCV positive, compared to 3.2% of those who had not. In that study, HCV prevalence among the HIV-positive MSM who had injected was 57%.

3.2.1.4 HCV infection among HIV-infected persons from HIV-endemic countries

Only one study of HCV infection was identified among persons from HIV-endemic countries. Allain <40> examined the prevalence of antibodies to HIV, HTLV-1 and HCV in three populations of Haitians: symptomatic outpatients, surgical patients and pregnant women. Among a subset of patients tested for HCV, HIV rates were relatively high. HIV infection was found among 39% of symptomatic outpatients, 6.1% of surgical patients and 4.0% of pregnant women. HCV prevalence was, however, relatively low: 1.5%, 0.9% and 0.4%, respectively. Somewhat surprisingly, no association of HCV antibody was observed with HIV or HTLV seropositivity in any of the three populations. Thus, the rate of HCV in this population was in the range of 0.5% to 1.0%, similar to that among Canadians as a whole.

3.2.1.5 HCV infection among HIV-infected persons infected by heterosexual contact

This group is difficult to define and estimates from studies carried out in populations outside Canada may have limited relevance for the present study. The group is also heterogeneous in the sense that the context in which one may become HIV-infected by this route differs greatly from population to population. Consequently, HCV prevalence in this group may also vary greatly. In particular, HCV prevalence may be much higher among sexual partners of IDUs; some of these sexual partners may have injected drugs and not reported this behaviour. In spite of these reservations, results from several studies identified in this review are worth noting.

Two seroepidemiologic studies undertaken by Health Canada of IDUs in Canada also examined the seroprevalence of markers among their sexual partners. Among the 80 sexual partners studied in Cape Breton Island <38>, only one subject was seropositive for HIV (prevalence 1.3%) and one was HCVpositive (prevalence 1.3%). The report did not indicate whether the HCV and HIV infections were in the same person. Nevertheless, an estimate of HCV prevalence among HIV-infected sexual partners could not be derived from this study given the limited number of HIV-infected partners (i.e. one). Similarly, in an outbreak investigation carried out in Prince Albert, Saskatchewan <39>, none (0.0%) of 48 sexual partners had serologic evidence of HIV infection, and 3 (6.3%) were HCV-positive. The same limitations apply to this study.

Several studies carried out outside Canada are of interest. Quaranta recruited 272 patients in a French multicentre study (called SEROCO) in 1988-91 <41>. Among the patients studied were 82 persons infected by HIV through heterosexual contact. Fifteen (18%) were HCV-positive, with 95% confidence limits of 11 and 28%.

Dorucci analysed data from a prospective cohort of HIV seroconverters from 16 centres in Italy <42>. Among the 416 HIV-positive subjects were 81 persons infected through heterosexual intercourse, mainly from HIV-infected IDU sexual partners. HCV prevalence in this group was 20% (16/81), with 95% confidence limits of 12% to 30%.

Ockenga recruited HIV-infected outpatients in Hannover, Germany in 1993-94 <43>. Of the 33 patients infected by heterosexual transmission, three (9.1%) were also infected with HCV; the 95% confidence limits were 1.9 to 24%.

The results of the studies of HCV infection among HIV-infected persons infected by heterosexual contact are summarized in Table 4.

3.2.1.6 HCV infection among HIV-infected hemophilia patients

It has long been widely recognized that the majority of hemophilia patients treated with factor concentrates (clotting factors prepared from plasma) became infected with A non-A non-B hepatitis @ (as hepatitis C was referred to until the viral antigens were characterized and serologic tests became available in 1989). This conclusion was based on the clinical observation of acute symptoms and laboratory signs of hepatitis following the onset of treatment with factor concentrates before HIV screening of plasma donors and viral inactivation of concentrate were implemented.

Studies carried out since the HCV antibody test became available have supported the belief that most hemophilia patients treated before 1985 became HCV-infected. Weinstock examined 35 HIV-infected hemophilia patients in suburban New York City and observed an HCV prevalence of 90% <37>. Similarly, Ockenga studied HIV-infected hemophilia patients in Hannover, Germany and observed an HCV prevalence of 77% <43>.

A study carried out by Brenner examined HCV-HIV joint marker prevalence in patients treated at two outpatient clinics in Israel <44>. Among the hemophilia patients, about two-thirds had received non heat-treated factor concentrates and the rest only cryoprecipitate; 33 (67%) of the patients were HCVpositive. However, 9 (88%) of the 11 HIV-infected hemophilia patients were HCV-positive compared to 24 (63%) of the 38 HIV-negative patients.

The results of the reviewed studies of HCV infection among HIV-infected hemophilia patients are summarized in Table 5.

3.2.1.7 HCV infection among HIV-infected among blood transfusion recipients

Several studies have examined HCV prevalence among HIV-infected transfusion recipients. In the period 1988-91, Quaranta tested 16 HIV-infected clinic patients in France who had been transfused and found that five (31%) were also infected with HCV <41>. The confidence limits for this prevalence were 11% to 59%.

The study noted in Section 3.1.6 above by Brenner <44> in Israel also examined 63 multiply transfused patients. He found an HCV prevalence of 21%. None of these transfusion recipients was, however, HIV-positive.

3.2.1.8 HCV infection among HIV-infected prisoners

Several studies were identified which helped to shed light on the issue of HCV infection and HCV-HIV co-infection among prisoners.

Pallas <45> studied prisoners in Cantabria in northern Spain in the period 1991-94 and observed an overall HCV prevalence of 41% among the 675 subjects tested. 255 (38%) prisoners reported having had injected heroin. HCV prevalence among the non-IDUs was 13% and 85% among those who had injected heroin. The prevalence of HCV infection was a function of duration of use: 70% among those who had injected heroin for five or fewer years, and 91% among prisoners who had injected for more than five years.

In a second related study conducted apparently in the same two institutions limited to prisoners who had injected drugs within the previous 12 months <46>, HCV prevalence was 94% among HIV-infected IDUs and 90% among HIV-uninfected IDUs.

A study of female inmates by Ford et al <47> in Kingston Ontario shed some light on the problem of HCV infection among prisoners in Canada. The study achieved a high participation rate. Of the 113 women examined, 45 (40%) were HCV-positive, and one (0.9%) was HIV-positive. No information on risk behaviours was available. However, it is very likely that the vast majority of HCV infections observed were in women who had injected drugs. With only one HIV-infected subject, this study cannot provide useful information specifically on HCV prevalence among HIV-infected prisoners.

A seroepidemiologic study among 415 male residents in federal correctional institutions in British Columbia was carried out in 1990 <48>. Of those tested, 106 (26%) were confirmed positive for HCV. No testing for HIV was done in the context of this study.

3.2.2 Consultation on HCV prevalence in Canada

Table 6 shows the best estimates and plausible ranges for the prevalence of HCV in 1999 among HIVinfected persons established by consensus during the teleconference described above.

With respect to the interprovincial weighting, we decided that the transfusion-related HCV cases should be subject to the same variation as the estimates derived from the HCV Working Group report of June 1998 <49>. We decided, however, that those infected by clotting factors should not be subject to interprovincial variation since the same products were distributed from a central source to all provinces. Likewise for persons from HIV-endemic countries, we did not vary the HCV prevalence among HIVinfected persons by region. For MSM, the key informants were of the opinion that a reasonable span of HCV prevalence would be 0.80 to 1.20, from the lowest to the highest HCV prevalence. For MSM, MSM-IDU, and persons infected through heterosexual contact, we decided the span should be 0.88 to 1.12 of that of the national estimated HCV prevalence.

Table 7 presents the HCV prevalence estimates among HIV-infected persons by exposure category and geographic region as used in the analysis. As seen in this table, the highest HCV prevalences are among injection drug users and persons who received clotting factors. Men who have sex with men who are also injection drug users are similar to heterosexual injection drug users. We believed that HCV prevalence among MSM should be only slightly higher than background HCV prevalence.

Back