Characterizing viral subtypes to assess patterns of HIV transmission

INTRODUCTION

As international travel and migration facilitate the global spread of HIV, an increasing proportion of new HIV infections in Europe and North America are due to non-B HIV subtypes. ^1–4 In Canada, 88% of new HIV diagnoses included in the Canadian HIV Strain and Drug Resistance Surveillance Program (SDRP) from 1997 to 2005 were subtype B. ¹ HIV testing of immigration applicants was implemented in January 2002 and over 50% of immigrants to Canada settle in Ontario; ⁵ approximately 45% of HIV cases reported annually in Canada are diagnosed in Ontario. ⁶ However, prior to Ontario's participation in the SDRP in October 2003, little was known about the distribution of HIV subtypes in this province. Monitoring HIV subtypes may also help provide valuable insights to assess HIV transmission patterns. We characterized the distribution of HIV-1 subtypes among SDRP participants in Ontario and assessed predictors of infection with a non-B HIV subtype.

METHODS

Study subjects were persons newly diagnosed with HIV-1 at Ontario's Public Health Laboratories – Toronto (PHL-T) from 1 October, 2003 to 31 March, 2005. PHL-T performs essentially all diagnostic HIV testing in Ontario using a standard algorithm of screening by enzyme immunoassay and confirmation by Western blot. Eligible subjects were first-time HIV-positive, antiretroviral treatment naïve residents of Ontario aged 18 years or older. Subjects were recruited through the physician who prescribed their HIV test and provided consent to have their residual diagnostic specimen submitted for genotyping at the National HIV and Retrovirology Laboratories, Ottawa. HIV RNA was reverse transcribed and amplified by nested polymerase chain reaction using in-house pol specific primers encompassing the entire protease gene and the first 321 amino acids of reverse transcriptase. Purified DNA was sequenced on a 3130xl Genetic Analyser (Applied Biosystems, Foster City, CA, USA). Subtype determination was based upon results returned from the Stanford University HIV Drug Resistance Database (http://hivdb.stanford.edu/index.html) and the Rega HIV-1 Subtyping Tool v1.0 (http://jose.med.kuleuven.be/subtypetool/html/index.html). The study protocol was approved by the Research Ethics Board of the Faculty of Medicine, University of Toronto and the Health Research Ethics Board of Health Canada.

Epidemiological data were obtained from the HIV laboratory requisition. Supplementary data on risk factors, region of birth, race ethnicity and antiretroviral treatment were collected using questionnaires sent with the test result as part of an ongoing laboratory data enhancement program. Cases were classified into mutually exclusive exposure categories according to a hierarchy of risk factors reflecting the most likely mode of HIV transmission, as follows: men who have sex with men (MSM), MSM and injection drug use (MSM-IDU), injection drug use (IDU), birth in an HIV-endemic country, and heterosexual transmission. (HIV-endemic countries are those where estimated HIV prevalence in the adult population is high [≥1.0%] and heterosexual transmission predominates [≥50%].) Unadjusted odds of infection with non-B HIV subtype were examined with respect to age, sex, region of birth, ethnicity, exposure category and Ontario health region. Factors that were found to be associated (P < 0.10) were assessed in a multivariate logistic regression model to identify significant independent (P < 0.05) predictors of non-B HIV subtype. Region of birth was not included in the multivariate analysis due to a strong correlation with ethnicity and exposure category. All analyses, including model fitting by backward selection, were carried out using SAS version 9.1 (SAS Institute, Cary, NY, USA). Adjusted odds ratios (aOR) and 95% confidence intervals (95% CI) were also determined.

RESULTS

Of 1560 persons aged 18 years or older diagnosed with HIV from October 2003 to March 2005, 204 were included in this study. Reasons for exclusion were: unreturned questionnaire/consent, 787 (58.0%); insufficient sample, 150 (11.1%); previous HIV diagnosis, 147 (10.8%); refusal to participate, 106 (7.8%); no genotyping result, 101 (7.4%); residency outside Ontario, 41 (3.0%) and antiretroviral treatment, 24 (1.8%). There were no statistically significant differences in age or gender in included cases compared with excluded cases. Study subjects were predominantly men (78.9%), born in Canada (52.9%), Caucasian (54.9%), tested in Toronto (68.1%) and aged 35 years or older (63.7%); median age was 38.0 years.

DISCUSSION

This first study of HIV subtypes in Ontario revealed a substantial proportion (23.0%) of non-B strains: this compares with earlier estimates of 12% non-B HIV-infection reported in Canada from 1997 to 2005. ¹ Not unexpectedly, we observed a high proportion of non-B subtypes among persons born in an HIV-endemic region. We also found an unexpectedly high prevalence of non-B subtypes in women overall, particularly among women infected heterosexually who were not from HIV-endemic countries, providing interesting insights into patterns of sexual exposure to HIV.

Immigration from regions where non-B subtypes are predominant likely plays an important role in the introduction of non-B HIV subtypes into Ontario. Our findings are consistent with studies from other countries with considerable immigration from HIV-endemic areas; these have documented increased subtype diversity ^3,7,8 and an increased proportion of non-B subtypes in heterosexually acquired HIV infections, ^8,9 especially among women. ^10–13 Because a substantial proportion of African immigrants settle in Ontario ⁶ and a lower proportion of infected persons from HIV-endemic countries appear diagnosed (59%; 49% in men and 75% in women) in comparison to other groups (67% of MSM and 70% of IDU), ¹⁴ our study may have underestimated the true prevalence of non-B subtypes in Ontario. In 2004, the prevalence of non-B subtypes in other provinces participating in SDR surveillance varied from 9% to 36%. ^15,16

Although we distinguished persons born in HIV-endemic regions from others infected heterosexually, our findings should be interpreted in light of our study limitations. While data on ethnicity and region of birth were not available for excluded subjects, study subjects differed significantly from excluded cases with respect to exposure category (P < 0.0001): a higher proportion of study subjects were MSM or MSM-IDU (55.4% versus 48.9%) and a lower proportion were in the heterosexual exposure category (10.3% versus 21.5% of excluded subjects). This, and the predominance of male, Canadian-born and MSM subjects further suggests that non-B subtypes in Ontario may be more prevalent than found in our study. Low participation, especially among persons in the heterosexual exposure category, may have introduced bias in our study sample and, therefore, the association between non-B HIV subtype and heterosexual exposure warrants further investigation. We found that heterosexual women were significantly more likely than MSM/MSM-IDU to have a non-B subtype while heterosexual men were not, suggesting that among women, heterosexually acquired HIV infection was more likely to be related to sexual contact with persons from non-B regions, either in Canada or through international travel.

In Ontario, persons in the heterosexual exposure category and those from HIV-endemic countries accounted for 14% and 16% of modelled prevalent HIV infections in recent years, respectively. ¹⁴ New HIV diagnoses are predominantly among MSM and IDU; however, an increasingly important component of the HIV epidemic in Ontario involves ongoing transmission through heterosexual contact. A recent study among persons from Africa and the Caribbean diagnosed with HIV in Ontario estimated that the majority acquired their HIV infection in their country of origin. ¹⁷ While most heterosexual contact among persons from HIV-endemic regions presumably occurs within these groups, evidence from the UK, Quebec and Toronto suggest that persons not from HIV-endemic countries may also acquire HIV to a significant extent through heterosexual contact with persons from HIV-endemic countries. ^10,18,19 Among 183 persons diagnosed with AIDS in Quebec from 1979 to 2001 not from HIV-endemic countries who acquired HIV through heterosexual contact and for whom the risk factors of the source sexual partner were known, 100 (55%) reported having a partner from an HIV-endemic country. ¹⁹ As HIV-1 subtypes A, C and D are particular to Africa and Asia from which a high proportion of Ontario immigrants originate, ²⁰ these strains may become increasingly useful as markers for HIV transmission between groups.

Continued monitoring of HIV-1 subtypes is critical for epidemiological surveillance of HIV transmission, which, in turn, serves to guide and evaluate prevention programs. HIV prevention strategies aimed at the heterosexual population, especially women, may be needed to address low-risk perception in these groups. With increasing prevalence of heterosexually acquired HIV infection in Ontario, ¹⁴ ensuring effective prevention programmes among persons from HIV-endemic countries is equally important. In addition to informing vaccine development, HIV subtyping is a valuable tool to better understand trends in HIV transmission patterns and can complement other HIV surveillance activities and special studies.