Correspondence: Dr Marie Helleberg, Department of Infectious Diseases, Rigshospitalet, Blegdamsvej 9, 2100 Copenhagen Ø, Denmark. Tel: 004535457726; fax: 004535456648; e-mail: firstname.lastname@example.org
Recent studies have reported faster progression of HIV infection than anticipated based on results from earlier studies. The aim of the present study was to examine if the virulence of HIV-1 infection changed in the period 1995–2010 among chronically HIV-infected individuals in Denmark.
We included all patients registered in the Danish HIV Cohort Study, who were diagnosed in 1995–2009, had a CD4 count > 100 cells/μL at diagnosis and had at least two CD4 measurements prior to initiation of antiretroviral therapy (ART). Changes in viral set point and rate of CD4 cell decline from enrolment until the initiation of ART by calendar year of HIV diagnosis were analysed. Time to first CD4 count < 350 cells/μL was compared among patients diagnosed in 1995–2000, 2001–2005 and 2006–2010.
We followed 1469 HIV-infected patients for a total of 5783 person-years. The median viral set point was 4.27 log10 HIV-1 RNA copies/mL [interquartile range (IQR) 3.58–4.73 log10 copies/mL]. The median CD4 cell decline per year was 57 cells/μL (IQR 10–139 cells/μL). In analyses adjusted for age, gender, origin, route of transmission and CD4 count at diagnosis, there were no associations between year of diagnosis and viral set point or CD4 cell decline. Time to first CD4 count < 350 cells/μL did not change in the study period [incidence rate ratio (IRR) 0.90 (95% confidence interval (CI) 0.76–1.06) for 2001–2005 and 1.09 (95% CI 0.79–1.34) for 2006–2010 compared with 1995–2000].
We found no evidence of changing trends in viral set point, CD4 cell decline or time to CD4 count < 350 cells/μL during the period 1995–2010 in a cohort of chronically HIV-infected individuals.
HIV is a zoonosis originating from chimpanzees . The virulence of zoonotic viruses may change over time during the process of adaptation to the human host. Theoretically, decreasing virulence of HIV might facilitate transmission through an increase in survival of the host and thereby a longer period with potential for onward transmission. Increasing virulence could lead to an increase in transmission rates as a result of higher viral loads in the early stages of the disease, where the majority of onward transmissions may occur . With the widespread use of combination antiretroviral therapy (ART) the balance might change so that HIV strains that cause very high viral loads in the early stages of disease may be more likely to be transmitted than strains that cause low-grade viraemia and have a more protracted course.
Two recent studies, examining the potential benefit of ART in primary and early HIV infection, showed surprisingly high rates of progression of HIV disease among individuals randomized to the deferred treatment arm [3, 4]. One of these trials was halted prematurely because the progression of HIV disease was faster than anticipated based on results of previous studies . It was speculated that this faster rate of progression could be attributable to an increase in HIV virulence. The aim of the present study was to examine if the virulence of HIV-1 changed in the period 1995–2010 in Denmark.
We analysed the viral set point and rate of CD4 cell decline from enrolment until initiation of ART among treatment-naïve patients prospectively enrolled in the Danish HIV Cohort Study (DHCS) in the period 1995–2010. Changes in these markers of HIV virulence by calendar year of HIV diagnosis were analysed. Time to ART initiation and to first CD4 count < 350 cells/μL was compared among patients diagnosed in three periods: 1995–1999, 2000–2004 and 2005–2009.
Denmark has an estimated HIV prevalence of approximately 0.09% in the adult population. HIV-infected individuals are treated in one of the country's eight specialized medical centres, and were, during the study period, seen on an out-patient basis with measurement of CD4 count and viral load at intended intervals of 12 weeks. ART is provided free of charge to all HIV-infected residents only at these centres and cannot be purchased from pharmacies or private health care providers. Treatment is prescribed according to national guidelines. The guidelines and adherence to them are described in Petersen et al. .
The DHCS, which is described in detail elsewhere , is a population-based nationwide cohort study of all HIV-infected individuals who have been treated at Danish HIV centres after 1 January 1995. Individuals are consecutively enrolled. Data are updated yearly and include demographics, date and route of transmission of HIV infection, AIDS-defining events and ART. CD4 cell counts and HIV RNA measurements are extracted electronically from laboratory data files.
Data on migration and vital status were obtained from the Danish Civil Registration System , which is a national register, established in 1967, that contains the demographic data and vital status of all Danish citizens.
All antiretroviral-naïve, HIV-1-infected individuals aged ≥ 16 years, diagnosed in the period 1995–2010 and enrolled in the DHCS were eligible for this study. Exclusion criteria were: 1) CD4 count at HIV diagnosis < 100 cells/μL and 2) fewer than two CD4 cell measurements with an interval of at least 3 months available prior to initiation of ART.
Outcome and definitions
We analysed three study outcomes: 1) viral set point, defined as the (log10) viral load measured at the date closest to 6 months, and within 3–9 months, after HIV diagnosis, 2) CD4 cell decline per year from enrolment in the DHCS until initiation of ART and 3) time to first CD4 count < 350 cells/μL.
Viral set point and CD4 decline
We transformed data to log10 for viral loads and 3√ for CD4 counts to achieve normal distributions. CD4 cell decline per year was calculated for each patient by linear regression including all CD4 counts from diagnosis until ART initiation. Changes in viral set point and CD4 cell decline by calendar year of HIV diagnosis were estimated using linear regression. Confounders included in the adjusted regression analyses were: gender, origin (Danish, African, Asian or other), age and 3√CD4 count at HIV diagnosis, of which the last two were included as continuous variables. In the model analysing CD4 cell decline, viral load at HIV diagnosis was also included. As previous studies have shown effect modification by route of HIV transmission, analyses were stratified by this parameter. Similarly, linear regression was used to analyse correlations between CD4 cell decline and viral set point with confounders listed above included in the adjusted analyses.
Time to ART initiation and to CD4 count < 350 cells/μL
We used Kaplan–Meier methods to estimate time to initiation of ART and time to CD4 count < 350 cells/μL and to construct survival curves. In analyses of ART initiation, time was calculated from HIV diagnosis until the initiation of ART, emigration, death or 31 August 2010, whichever came first. In the analysis of time to first CD4 count < 350 cells/μL, time was calculated from the date of measurement of the first CD4 count until the date of the first CD4 measurement < 350 cells/μL or the last CD4 measurement before ART initiation, emigration, death or 31 August 2010, whichever came first. In this analysis only individuals with CD4 count > 350 cells/μL at the time of diagnosis were included (n = 1295). Incidence rate ratios (IRRs) were estimated by Poisson regression analyses adjusted for age and 3√CD4 count at HIV diagnosis.
The study was approved by the Danish Data Protection Agency. Ethics approval and individual consent are not required by Danish legislation governing this type of study.
spss statistical software, Version 15.0 (Norusis; SPSS Inc., Chicago, IL) and stata, Version 8.0 (Stata Corporation, College Station, TX) were used for data analysis.
A total of 3708 HIV-1-infected individuals enrolled in the DHCS, aged ≥ 16 years and ART naïve, were diagnosed with HIV infection in the period 1995–2010. We excluded 694 individuals with CD4 count < 100 cells/μL at HIV diagnosis and 1545 for whom two CD4 measurements with an interval of at least 3 months were not available before ART initiation, leaving 1469 individuals in the study, who were followed for a total of 5783 person-years.
Among individuals diagnosed in 1995–2000, 326 and 512 of 1399 (60%) were excluded because they had a CD4 count < 100 cells/μL at HIV diagnosis and insufficient CD4 measurements available, respectively. The corresponding numbers were 217 and 459 of 1189 (57%) in 2001–2005 and 151 and 574 of 1120 (65%) in 2006–2010.
Characteristics of the study population
The majority of individuals were male and of Danish origin (Table 1). Among individuals diagnosed in the most recent part of the study period, a larger proportion were homosexuals, whereas the proportion of injecting drug users (IDUs) had decreased. The median age and CD4 count at HIV diagnosis increased during the study period (Table 1). Despite increasing CD4 counts, the median time from HIV diagnosis to initiation of ART decreased from 5.0 years [95% confidence interval (CI) 4.5–5.4 years] in 1995–2000 and 4.6 years (95% CI 4.1–4.9 years) in 2001–2005 to 2.7 years (95% CI 2.3–3.2 years) in 2006–2010.
Table 1. Characteristics of the study population and individuals excluded from the study, stratified by year of HIV diagnosis
Characteristics of individuals excluded from the study
Individuals excluded from the study had a higher median age and a lower median CD4 count at diagnosis than the study population and the proportions of heterosexuals and immigrants were larger (Table 1). The CD4 count at diagnosis increased during the study period, reflecting the fact that a larger proportion of individuals were excluded because two CD4 counts with an interval ≥ 3 months were not available prior to ART initiation or the end of the study and a smaller proportion were excluded because they had a CD4 count < 100 cells/μL at diagnosis.
Markers of HIV virulence
The median viral set point was 4.18 [interquartile range (IQR) 3.65–4.62] in 1995–2000, 4.33 (IQR 3.68–4.81) in 2001–2005 and 4.28 (IQR 3.34–4.77) in 2006–2010. There were a median of 5 (IQR 2–11) CD4 measurements per patient contributing to the estimation of the CD4 cell decline before ART initiation. In the periods 1995–2000, 2001–2005 and 2006–2010, the numbers were 6 (IQR 3–13), 6 (IQR 3–11) and 3 (IQR 2–6), respectively. The median CD4 cell decline per year was 56 cells/μL (IQR 14–125 cells/μL) in 1995–2000, 59 cells/μL (IQR 13–137 cells/μL) in 2001–2005 and 58 cells/μL (IQR 7–169 cells/μL) in 2006–2010. The date of diagnosis was associated with neither viral set point nor CD4 cell decline [slope – 0.01 (95% CI – 0.02–0.01) and 0.01 (95% CI – 0.01–0.02), respectively] (Table 2). In the analysis of viral load at diagnosis instead of viral set point, results were almost identical (data not shown). The median change in viral load prior to ART initiation was 0.08 log10 copies/mL/year (IQR – 0.21 to 0.44 log10 copies/mL/year).
Table 2. Change in viral set point and CD4 cell decline by date of HIV diagnosis for all study subjects and in analyses stratified by route of HIV transmission
Linear regression analyses
Change per year later diagnosis (95% CI)
Viral set point was defined as the (log10) viral load measured at the date closest to 6 months, and within 3–9 months, after HIV diagnosis, and CD4 cell decline per year was measured from enrolment until initiation of antiretroviral therapy. In regression analyses, CD4 counts were transformed to cubic root. Analyses were adjusted for the following variables: gender, age, origin and CD4 count at HIV diagnosis. Analyses of CD4 cell decline were also adjusted for viral load at HIV diagnosis.
Viral set point (log10 copies/ml)
CD4 decline (cells/μL per year)
Viral set point (log10 copies/ml)
CD4 decline (cells/μL per year)
Viral set point (log10 copies/ml)
CD4 decline (cells/μL per year)
Injecting drug use
Viral set point (log10 copies/ml)
CD4 decline (cells/μL per year)
Analyses stratified by route of HIV transmission
In analyses stratified by route of transmission, the median CD4 count at HIV diagnosis was 520 cells/μL (IQR 393–680 cells/μL) among homosexuals, 480 cells/μL (IQR 360–630 cells/μL) among heterosexuals and 560 cells/μL (IQR 397–710 cells/μL) among IDUs. We observed no association between year of diagnosis and markers of HIV virulence (Table 2).
Correlation between CD4 cell decline and viral set point
There was a negative correlation between 3√CD4 cell decline and (log10) viral set point [coefficient -0.22 (95% CI – 0.26 to – 0.18)]. This correlation did not differ among the three study periods [coefficient – 0.21 (95% CI – 0.28 to – 0.13) in 1995–2000, – 0.24 (95% CI – 0.30 to – 0.18) in 2001–2005 and – 0.20 (95% CI – 0.28 to – 0.11) in 2006–2010].
Time to CD4 count < 350 cells/μL
The median time from HIV diagnosis until first CD4 count < 350 cells/μL among individuals diagnosed in 1995–2000, 2001–2005 and 2006–2010 was 2.5 years (95% CI 2.3–3.0 years), 3.3 years (95% CI 2.8–3.9 years) and 2.4 years (95% CI 2.4–3.1 years), respectively (Fig. 1). In the adjusted analysis there were no differences among the three study periods [IRR 0.90 (95% CI 0.76–1.06) for 2001–2005 and 1.09 (95% CI 0.79–1.34) for 2006–2010 compared with 1995–2000].
In this study examining changes in HIV virulence in the period 1995–2010, we found no evidence of changing trends in viral set point, CD4 cell decline or risk of progression to CD4 count < 350 cells/μL.
Previous studies of changes in HIV virulence over time, summarized in a recent meta-analysis , have reached various conclusions. Our findings are in agreement with studies from the Swiss HIV Cohort of mainly chronically infected individuals , the United States Multicenter Cohort Study of men who have sex with men (MSM)  and a study from the French Agence Nationale de Recherche sur le SIDA (ANRS) PRIMO Cohort of individuals with primary HIV infection , whereas findings from the CASCADE (Concerted Action on SeroConversion to AIDS and Death in Europe) collaboration and the Italian HIV Cohort were consistent with an increase in HIV virulence [12, 13]. In the Italian cohort, 46.3% of individuals were IDUs and the increasing trend of HIV virulence was primarily seen in this transmission group, whereas there was no significant increase in HIV virulence among MSM. In the most recent period of the present study, < 5% of individuals were infected through IDU and 53% were MSM, which may explain the divergent conclusions of the two studies. Several factors such as study populations, periods and methods are likely to contribute to divergent findings between studies. In the pre-highly active antiretroviral therapy (HAART) era, virulence was estimated by analysing the time to AIDS or death [14, 15], which makes comparisons with recent studies difficult. Analyses of changes in virulence among individuals with primary HIV infection vs. chronically infected individuals may also yield differential results. Further, different subtypes of HIV-1 may not have equivalent virulence. An increase in the proportion of transmitted drug resistance might cause a trend towards diminished virulence as some resistance mutations reduce the fitness of the virus . Transmitted drug resistance is rare in Denmark . Factors related to both the pathogen and the host may influence disease progression. In our study population, the median rate of CD4 cell decline was 57 cells/μL per year, which is similar to findings from the Swiss and Italian cohort studies [9, 12]. Individuals of more advanced age have lower CD4 counts at HIV diagnosis and faster rates of CD4 cell decline [18, 19]. Differences in age rather than in HIV virulence may explain why the rate of HIV progression was much faster in the set point trial than anticipated based on results from a study by the CASCADE collaboration. The median ages at seroconversion were 36 and 30 years, respectively, in these two studies [4, 20].
There is cumulating evidence of the benefits associated with earlier treatment initiation in terms of the rate of HIV transmission  and AIDS related-  and non-AIDS-related morbidity and mortality among HIV-infected individuals [23-25]. US guidelines now recommend ART for all HIV-infected individuals regardless of CD4 count . The impact of possible changes in HIV virulence will diminish as guidelines for ART are modified in favour of earlier initiation. However, if the virulence of HIV is increasing, timely diagnosis and treatment will become even more important.
A major strength of the study is the rather large study population from a nationwide cohort where loss to follow-up is very limited , which improves the generalizability of the results and reduces bias related to differences in disease severity and health-seeking behaviour. The study has some limitations. In the most recent years of the study period, treatment guidelines changed, favouring earlier initiation of ART. Therefore, the median observation time was shorter than in earlier years, and a higher proportion of HIV-infected individuals diagnosed in that period were excluded because of a lack of two CD4 measurements with an interval of at least 3 months prior to ART initiation. We may therefore have failed to detect an increase in HIV virulence. However, when viral load among excluded individuals was analysed, we found no increase during the study period. The date of seroconversion was unknown for the majority of study subjects and thus we are unable to assess whether the initial drop in CD4 count or peak viraemia after seroconversion changed during the study period, and our estimates of the virulence of HIV are limited to the chronic stage of infection. As national guidelines recommend ART for individuals presenting with primary HIV infection and because of our criterion of two CD4 measurements with an interval of at least 3 months prior to ART initiation, most individuals diagnosed in the acute stage of the infection and those with advanced disease at the time of diagnosis were excluded from analyses, and therefore results cannot be generalized to these groups. However, we do not think that exclusion of late presenters, who may be more likely to be fast progressors, resulted in failure to detect an increase in virulence over calendar time, as the median CD4 count at the time of diagnosis has increased since 2001 in the DHCS . As HIV subtypes were only analysed in a minor fraction of the study population, we were unable to examine potential differences between HIV subtypes. Individuals diagnosed with HIV prior to 1995 were only retrospectively enrolled in the DHCS if they survived until 1995 and, in order to avoid a ‘healthy survivor’ bias, these patients were not included in the study. We are therefore not able to determine changes in HIV virulence before 1995. As in previous observational studies of changes in HIV virulence, the methods of measurement of viral load and CD4 count changed during the study period, which is not easily adjusted for, although it is unlikely to cause a major bias.
In conclusion, the time to ART initiation decreased significantly during the study period, 1995–2010, but we found no evidence of changing trends in viral set point, CD4 cell decline or time to CD4 count < 350 cells/μL in a cohort of chronically HIV-infected individuals.
We thank the staff of our clinical departments for their continuous support and enthusiasm. The centres involved in the Danish HIV Cohort Study are the Departments of Infectious Diseases at Copenhagen University Hospitals, Rigshospitalet and Hvidovre, Odense University Hospital, Aarhus University Hospitals, Skejby and Aalborg, Herning Hospital, Hillerød Hospital, and Kolding Hospital.
All of the authors contributed to the conception and design of the study as well as to analyses and interpretation of data. The manuscript was drafted by MH, JG and NO and was critically reviewed and subsequently approved by all authors.
Conflicts of interest
NO has received research funding from Roche, Bristol-Myers Squibb, Merck Sharp & Dohme, GlaxoSmithKline, Abbott, Boehringer Ingelheim, Janssen-Cilag, and Swedish Orphan. CP has received funding from Abbott, Roche and Merck Sharp & Dohme. JG has received research funding from Abbott, Roche, Bristol-Myers Squibb, Merck Sharp & Dohme, ViiV, Swedish Orphan and Gilead. MH, GK, CSL and GP report no conflicts of interest.
No funding sources were involved in the study design, data collection, analysis, report writing, or the decision to submit the paper.