HIV in Denmark and Greenland, 1995–2004, the effect of highly active antiretroviral therapy and characteristics of the HIV-infected population: an observational study

Objectives. The PhD thesis is composed of five articles (1-5) and a review. The work leading to the thesis began in 2003. At that time, the dramatic decrease in mortality in HIV-infected individuals after the introduction of highly active antiretroviral therapy (HAART) in 1995-1996 was evident. However, concerns were raised that the effec tiveness of HAART might wane due to long-term drug toxicity and emergence of drug-resistant virus, leading to multi-drug class treat ment failure. Furthermore, the number of people receiving HAART in Denmark and Greenland was unknown, as was the disease progres sion stage at time of diagnosis. Aims of this thesis were therefore 1) to examine temporal trends in deaths, new infections, and individual patients characteristics among the Danish HIV-infected population, 2) to describe demographic characteristics and examine the impact of HAART on immune status and mortality among the HIV-infected population in Greenland, 3) to examine the level of complete plasma HIV RNA (viral load) suppression during the initial period after HAART initiation as a predictor for long-term viral suppression, CD4 count increase, and mortality, 4) to examine the incidence, preva lence, and predictors for triple-class antiretroviral drug failure, and 5) to estimate acquisition of antiretroviral drug-resistant virus over time and examine temporal trends in the prevalence of patients at risk of transmitting drug-resistant HIV.

Study design. Observational epidemiological studies.

Methods. The Danish HIV Cohort Study is a prospective cohort study covering all clinics treating HIV in Denmark and Greenland, including all prevalent HIV cases as of 1 January 1995 and all incident cases since then. Types of data collected are individual characteristics, biochemical test results, treatment history, and clinical events. We compared individual characteristics using chi-square test, Student’s t-test, and one-way analysis of variance. Spearman’s rank correllation, linear regression, Poisson regression, logistic regression and Cox proportional hazards regression were used to compare outcomes. To maximize control of confounding we used matching in design, restric tion in design, restriction in analysis, stratification, and multivariate modelling.

Results. We found a decrease in the annual number of deaths in Denmark after introduction of HAART, and a stable number of new HIV infections with no sign of an increase in the proportion of patients with advanced HIV disease or young age. In Greenland, the HIV-infected population was middle-aged, mainly living in Nuuk and Sisimiut, and predominantly infected by heterosexual contact. Even in patients receiving HAART, mortality was more than ten times higher than in the general population, and only 40% had fully suppressed viral load. In another study, we found that the proportion of time with viral load suppression 6-18 months after HAART initiation was asso ciated with long-term survival and clinical improvement. Episodes of non-suppression increased the risk of death and poor immunologic outcome and might be markers of poor adherence; treatment inter ruptions further enhanced the risk. We also found that the incidence rate of triple-class drug failure declined over calendar time, and that the prevalence among patients on HAART was stable. Finally, we found that the number of HIV-infected individuals at risk of transmit ting drug-resistant virus was declining, and that improved treatment appeared to be responsible for this positive development. Although the prevalence of potential drug resistance decreased overall, potential resistance to non-nucleoside analogue reverse transcriptase inhibitors (NNRTI) increased in patients who were mono- or dual-drug antiret roviral therapy-naïve before HAART initiation.

Conclusions. The primary goal in the day-to-day treatment of HIV patients is to obtain an undetectable viral load, and the positive effects observed in these studies reflected clinical success. The declining prevalence of drug failure and risk of transmission could be referred directly back to the continuous increase in the proportion of patients with successful viral suppression. The importance of achieving an undetectable viral load was underscored by the positive long-term outcome in patients with complete viral suppression in the first period of HAART treatment, and by the negative outcome among those who did not reach this goal. Treatment improvements over the years seem to have suspended and possibly stopped the threatening accumulation of patients with multi-class drug failure. They also seem to have warded off the accumulation of infectious patients with a history of virological failure, despite high drug pressure on HIV at the population level. A continuation of these trends would preserve initial therapeutic options for the majority of newly infected individuals in the future. Drug resis tance from less successful therapeutic strategies used in earlier years would not be carried on, and the HIV-infected population would be able to fully benefit from new drugs on the market. However, drug failure – and NNRTI failure in particular – must be closely monitored in the future.

Keywords: HIV, cohort study, Greenland, Denmark, temporal trends, therapeutic effectiveness