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

4. Discussion

The purpose of the first phase of the study was to determine the number and distribution of persons infected with HIV infection in Canada, and to establish a functional foundation for the estimate of HCVHIV co-infection. The number of HIV-infected persons in 1999 so derived represents an increase of 24% over that estimated for three years earlier. The increase is significant from a public health perspective and is related in part to the fact that HIV incidence did not appear to decrease during this period. Among IDUs, HIV incidence remains substantial due in part to the explosive spread of HIV in Vancouver and Ottawa and the continuing high rate of transmission in Montreal. In addition, HIV incidence among MSM may be increasing, especially in Toronto. In combination with the stable or increasing HIV incidence rate, the decrease in mortality rates associated with the advent of highly active antiretroviral therapy (HAART) in 1995-96 has also been a factor in the rise of HIV prevalence over this three-year period.

To establish a plausible range of HCV infection rates among HIV-infected persons in the seven major exposure categories (namely MSM, MSM-IDU, IDU, persons from HIV-endemic countries, other persons infected by heterosexual contact, patients with hemophilia, and transfusion recipients), we reviewed studies from many industrialized countries. The results of these studies cannot be applied to the present study for Canada without taking into account several important factors:

(1) The subjects in most of these studies were not recruited using random selection from the population as a whole. Biases may have been introduced in this regard, and must be considered in the interpretation of the results.

(2) Several of the studies reviewed (especially those carried out in the early 1990s) did not use a second, confirmatory test when measuring HCV prevalence. Though the confirmation rate is generally high in serologic testing, it is not 100%. Therefore, I have selected, where possible, only studies where confirmed results were presented.

(3) As noted above, of the studies reviewed on MSM, only one specifically indicated that IDUs had been excluded from the analysis. The results of that study yielded an HCV prevalence lower than almost all of the studies reviewed. No specific mention of injection drug use is made in most of these reports. It is possible that, at least for some, a proportion of subjects also injected drugs. Given the extremely high rate of HCV infection related to injection drug use, the inclusion of even a few persons who had injected drugs could have inflated the observed prevalence.

(4) Finally, HCV prevalence varies to a great extent from country to country and from region to region. This reality must be taken into consideration in the application of the results of the review in the final model. The results must be modulated in the context of observations made in Canadian populations, even where joint HCV-HIV prevalence have not been available.

There are several limitations to the modeling techniques used here to estimate HIV prevalence and HCV-HIV co-infection. As for any study of this kind, we were limited by the sparseness of data for some populations. The use of key informants is a reasonable approach in the face of limited data but the results so derived must be considered only an approximation. The results of the analysis of HIV prevalence and HCV-HIV co-infection among Aboriginal populations and prisoners must also be interpreted with caution. Nevertheless, the Monte-Carlo simulations allow for the determination of a plausible range of estimates which incorporate the uncertainties in the values of the input parameters and, in some ways, is more valid than the point estimates.

With respect to the Aboriginal population specifically, using the proportion of AIDS cases to determine the overall number of HIV-infected persons would be misleading, since the HIV epidemic began later in this population. An initial estimate of 1,245 using this approach almost certainly underestimates the true number. Obtaining the estimate based on the proportion of the AIDS cases which occurred in more recent years presents the difficulty of not being able to adjust for biases due to the differential uptake of antiretroviral therapy. In 1999, the most recent year for which data was available, AIDS cases in the Aboriginal population represented 15% of AIDS cases overall. The method actually used also took into account the number of HIV infections diagnosed among the Aboriginal population and produced a more plausible estimate of 2,740. Nevertheless, it must be noted that the estimate of HCV-HIV co-infected persons among the Aboriginal is subject to considerable uncertainty.

With respect to prisoners incarcerated in federal and provincial prisons, we estimate that, as of December 1999, 867 persons, or 2.6% of the prison population, were infected with HIV. In December 1999, 196 HIV-infected prisoners were known to federal prison authorities, which represents a prevalence of 1.4%. Extrapolating the same proportion of persons incarcerated in the provincial prison system would yield a total number of 470 or about 54% of our estimated number.

However, it is not surprising that only 50% of HIV infections among prisoners would be known to prison authorities. Overall, in Canada, about 30% of infections are not diagnosed <55 >, and there are likely additional reasons for resistance to HIV testing and to sharing information about HIV-positive serostatus in the prison setting.

Back