• antiretroviral therapy;
  • CD4;
  • seroconverters;
  • survival


  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. References
  8. Appendix

Objectives  To estimate the times from HIV seroconversion to death, and to the initiation of therapy and the mean CD4 cell count at initiation.

Design and methods  Using Kaplan–Meier methods, allowing for late entry, we analysed CASCADE (Concerted Action on SeroConversion to AIDS and Death in Europe) data from HIV-infected individuals with known dates of seroconversion. We tested the association of time to initiation of therapy and of survival with: exposure category, age, sex, presentation during acute infection and calendar year at risk (as time-dependent) in Cox proportional hazards models, stratifying by study. We estimated the mean CD4 cell count at the initiation of therapy using interval regression.

Results  Of 5893 seroconverters, 1613 (27.4%) died. The risk of death was 65% lower (95% CI = 57–72%) in 1997–99 compared to previous years. Being at risk in earlier calendar years, older age and a short interval between negative and positive test dates were associated with shorter survival. At the same time from seroconversion, people at risk in 1997–99, older individuals and people with a short test interval were more likely to initiate therapy. Injecting drug users (IDUs) were less likely to initiate therapy compared to those exposed through sex between men (RR = 0.79, 95% CI = 0.69–0.89). The mean CD4 cell count at therapy initiation was 205 cells/mL, which increased significantly over time. Although the earlier initiation of therapy was consistent with longer survival in the 1997–99 period, we found no evidence of this in other calendar periods.

Conclusions  We found a significant and substantial reduction in the risk of death and a significant trend of earlier initiation of antiretroviral therapy (ART) in the 1997–99 period. Although IDUs were less likely to initiate therapy their overall survival did not appear to differ from others. The increasing tendency to initiate ART closer to seroconversion has unknown long-term consequences which require monitoring.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. References
  8. Appendix

Within 2 years of the disappointing results from clinical trials of zidovudine (ZDV) monotherapy in early HIV infection[1,2], in vitro studies and clinical trials of ZDV in combination with other nucleoside reverse transcriptase inhibitors (NRTI) had demonstrated more effective inhibition of viral replication and significant delays in disease progression[3–5]. These encouraging results were followed rapidly by even more promising drug combinations, which included protease inhibitors and non-nucleoside reverse transcriptase inhibitors (NNRTI)[6,7].

There can be little doubt that the clinical course of patients infected with HIV in industrialized countries has improved dramatically since the introduction of potent combination therapies in the mid-1990s, particularly those including protease inhibitors. Although reductions in deaths and in hospitalizations for HIV-related events in more recent time periods have been reported from a number of cohorts of patients with prevalent HIV infection[8–11], most would agree that we still know relatively little about the long-term effects of the present regimens, particularly as they are known to be toxic.

A few seroconverter cohorts have reported a delay in the development of AIDS and a lengthening in the survival of those infected since the advent of highly active anti-retroviral therapy (HAART)[12–14]. Although such studies have the advantage of being able to adjust for duration of infection when assessing the likely effects of therapy at a population level, there are relatively small numbers in the individual cohorts. Their ability to detect changes in survival expectations in the most recent time periods is therefore limited.

In this paper we describe the uptake of ART in a large cohort of seroconverters in Europe during the last decade. Using data from almost 6000 individuals who were monitored from the time of their HIV infection, we examine the time period from HIV seroconversion to the initiation of ART, the CD4 count at the time of initiation, and how these have changed over time and the factors that are associated with the uptake of ART. We also estimate survival time from seroconversion, factors associated with this time and temporal changes. We discuss how changes in the management of people with HIV infection may have influenced changes in survival observed over the same period.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. References
  8. Appendix

Funding was obtained in April 1997 through the European Union Programme for Biomedicine and Health to set up a collaboration between the investigators of cohorts of people whose time of HIV seroconversion was well estimated (seroconverters). The main aim of CASCADE is to bring together European data and expertise to address questions on HIV natural history which cannot be addressed adequately through individual studies. CASCADE is currently a collaboration between a total of 19 cohorts from the following countries: Denmark (1)[15], Greece (1)[16], Germany (1)[17], France (2)[18,19], Italy (1)[20], the Netherlands (2)[21,22], Norway (2)[23,24], Spain (3)[25–27], Switzerland (1)[9], the United Kingdom (3)[13,28,29] and Australia, through a European Union bilateral agreement (2)[30,31].

After discussions with the investigators of each cohort regarding follow-up of recruited seroconverters and the ascertainment of deaths for patients who had become lost to follow-up, an appropriate analysis cut-off date was established for each cohort.

Kaplan–Meier methods[32], allowing for late entry[33], were used to estimate time from HIV seroconversion to death from any cause. Seroconversion was estimated as: the midpoint between the first positive and last negative antibody test dates; the date of laboratory evidence of seroconversion; the estimation of the cumulative distribution of infections using mathematical techniques for interval censored data; the availability of an antibody positive test within 2 years of the start of the local epidemic; or through the presence of a seroconversion illness together with the availability of a previous antibody negative test. The association of the following variables with survival following HIV seroconversion was examined: exposure category (sex between men, sex between men and women, IDU, haemophilia, others); age at estimated seroconversion (continuous); sex; HIV test interval (within 1 calendar month, 1 month or more); and calendar period at risk as a time-dependent covariate (pre-1989, 1989–94, 1995–96, 1997–99).

The HIV test interval is the time, in months, between the negative and positive antibody test dates. Patients testing positive within 1 calendar month of the negative test are likely to have been diagnosed during the acute infection phase and may therefore have different prognosis and clinical management from seroconverters with a longer test interval[34–36].

The calendar period at risk is the calendar date during which infected individuals were observed. A person can therefore contribute to more than one period at risk. The four calendar periods were chosen to correspond approximately with available ART in Europe: none (pre-1989), monotherapy (1989–94), dual combination therapy (1995–96) and triple therapy (1997–99; usually including protease inhibitors), respectively, although this may vary slightly between countries, clinics and individuals. This is reflected in information available on uptake of therapy in the CASCADE cohort ( Fig. 1).


Figure 1. Use of antiretroviral therapy over time in 4945 HIV-infected people with known duration of infection. Note: a number of cohorts did not have complete follow-up in 1998, giving a smaller total number with ART information.

Download figure to PowerPoint

For the analyses of time from seroconversion to the initiation of ART, patients entered the risk set on the latest of three dates: seroconversion; entry into the cohort; and 1 December 1986, the earliest date that ART therapy was likely to have been prescribed in Europe. The initiation of ART was the outcome of interest and patients not known to have started ART were censored on the date they were last known to be ART-naive. Patients who had initiated ART on or before their estimated date of seroconversion were nominally given a survival time of 1 day. Cox proportional hazard models[37] were used to assess the association between time from seroconversion to the initiation of therapy and the variables described above. For this analysis, the calendar period at risk is the time during which infected ART-naive individuals were observed and could have started ART therapy. We calculated the proportion of total person-time at risk on different levels of therapy (mono, dual, three or more) in each of the four calendar periods to describe the changes in ART uptake over time. We also examined the effect of including serial CD4 cell count as a time-dependent covariate, including only time at risk with a CD4 cell count in the previous 6 months.

We used normal regression of interval censored data to estimate the mean CD4 cell count at which ART is initiated. For individuals who started ART, the outcome variable is the last CD4 cell count before start of therapy (within 6 months). For individuals known to be ART-naive, the outcome variable is an interval 0–y, where y is the CD4 cell count at the time they were last known to be ART-naive (within 6 months). For individuals who started ART without a CD4 count in the previous year, the outcome variable is the interval 0–y, where y is the last CD4 cell count. We used a square root transformation of the CD4 cell counts (variance stabilizing transformation of these data), and back-transformed to calculate predicted values. We examined the association of mean CD4 cell count at the initiation of therapy with study, exposure category, sex, age, HIV test interval and calendar year at risk. To investigate changes over calendar year at risk, CD4 and ART experience in the calendar periods described above were analysed in a multivariate marginal model, allowing for the fact the individuals can contribute correlated CD4 observations in more than one time period[38].

People aged under 15 years at seroconversion were excluded from all analyses relating to ART as the management of HIV disease in children is likely to differ from that of adults.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. References
  8. Appendix

Data on 5893 seroconverters with documented times of seroconversion from 19 studies were included in the survival analyses, of whom 1613 died ( Table 1). Of the 4280 individuals with censored observation, 477 (11%) were not seen 6 months or more before the end of the period of follow-up for each study. In all, 2090, 4708, 4410 and 3861 subjects contributed to the four calendar periods at risk: pre-1989, 1989–94, 1995–96 and 1997–99, respectively.

Table 1.  Distribution of exposure category, sex and age at estimated seroconversion of subjects included in analyses
Exposure categoryMaleFemaleTotalMedian age in years (range)
Sex between men2819281930 (15–70)
Sex between men28550579028 (13–69)
 and women
Injecting drug use1183586176924 (14–55)
Haemophilia37237221 (< 1–85)
Other/not known1093414332 (15–76)
Total47681125589327 (< 1–85)

The risk of death was significantly lower in 1997–99 than in previous years, with no evidence of a difference between the other three periods ( Table 2). As expected, older age was significantly associated with a higher risk, with a 44% increase in the risk of death (95% CI = 37–52%) for each 10-year increase in age. The increase in risk with increasing age was lower for those infected through sex between men than for other exposure categories. We estimate a median survival (95% CI) of 13.8 (12.8–14.6), 12.4 (11.6–13.3), 9.8 (8.6–11.0), 8.4 (7.7–9.5), 7.6 (6.3–9.3) and 3.5 (1.5–4.6) years for age groups 15–24, 25–34, 35–44, 45–54, 55–64 and 65 years or more, respectively. Patients with an HIV test interval of 1 month or more had a reduced risk of death compared to patients with an interval of less than 1 month ( Table 2). There was no evidence that exposure category and sex were associated with survival.

Table 2.  Effect of covariates examined on survival following HIV seroconversion *
VariableRR of death (95% confidence intervals)P-value
  • *

    Multivariate models stratified by study. P-values from log-likelihood tests.

Exposure category
 Sex between men1.000.20
 Injecting drug use1.01 (0.85–1.20) 
 Sex between men and women0.76 (0.61–0.95) 
 Haemophilia0.51 (0.18–1.42) 
 Other/unknown1.24 (0.92–1.68) 
Age (per 10 years increase)1.44 (1.37–1.52)< 0.0001
 Female0.99 (0.84–1.17) 
HIV test interval
 1 month or more1.000.02
 Within 1 month1.27 (1.03–1.56) 
Calendar year at risk
 Pre-19891.22 (0.93–1.59)< 0.0001
 1995–961.00 (0.88–1.14) 
 1997–990.35 (0.28–0.43) 

Of 4945 people for whom information on ART was known, 2016 never started ART (41%), 1983 started with one drug (40%), 632 with two drugs (13%) and 314 with three drugs or more (6%). Pre-1989, only 92 (3.2%) person-years (PY) of 2894 PY at risk were spent on monotherapy. Of the 2793 PY spent on therapy in the 1989–94 period, 91.5%, 8.4% and 0.1% were spent on mono, dual and triple (or more) combination therapy, respectively. For the period 1995–96, the corresponding figures were 57.9%, 31.4% and 10.7% (1466 PY). In the 1997–99 period the proportion of PY on monotherapy decreased to 11.5% while the proportions on dual and triple (or more) combinations increased to 37.2% and 51.3%, respectively (1441 PY).

Older age was associated with an increased risk of initiating ART ( Table 3). IDUs were significantly less likely to initiate ART compared to men who have sex with men, with relative risk 0.79 (95% CI = 0.69–0.89) after adjusting for the effect of other factors. People with an HIV test interval of less than 1 month were more likely to initiate ART than people with an interval of 1 month or more. There was no evidence that time to initiating ART differed by sex. Time to initiating ART was strongly associated with calendar year at risk, showing a strong trend over time towards the initiation of treatment closer to seroconversion (trend P < 0.0001, Fig. 2).

Table 3.  Effect of covariates on the time from HIV seroconversion to the initiation of ART *
VariableRR of initiating ART (95% CI)P-value
  • *

    Multivariate models stratified by study. P-values from log-likelihood tests.

Exposure category
 Sex between men1.000.09
 Injecting drug use0.79 (0.69–0.89) 
 Sex between men and women0.97 (0.84–1.12) 
 Haemophilia0.99 (0.39–2.53) 
 Other/unknown1.06 (0.83–1.34) 
Age at seroconversion (per1.20 (1.15–1.26)< 0.0001
 10 years increase)
 Female0.94 (0.83–1.06) 
HIV test interval
 1 month or more1.000.04
 Within 1 month1.27 (1.09–1.48) 
Calendar year at risk
 Pre-19890.52 (0.44–0.62)< 0.0001
 1995–961.61 (1.45–1.78) 
 1997–994.01 (3.57–4.50) 

Figure 2. The association between the time to therapy initiation from HIV seroconversion and calendar year at risk. Adjusted for exposure category, age, sex and HIV test interval.

Download figure to PowerPoint

The inclusion of CD4 count as a time-dependent covariate, although highly significant (P < 0.0001), had relatively little effect on the relative risk estimates for the other variables, although the evidence for a significant effect of exposure category and HIV test interval was much weaker (results not shown). This is likely to be a consequence of relatively sparse CD4 data for IDUs and people with an HIV test interval of less than 1 month. There was strong evidence (P < 0.0001) for a non-linear effect of serial CD4 cell count, and on investigating this further using natural cubic splines[39] the best-fitting model was approximated by a piecewise linear fit with one knot at CD4 count of 500 cells/μL. In this model, the relative risk of starting ART associated with a 100-cell drop in CD4 count is 1.90 (95% CI = 1.85–1.96) for those at a count below 500, and 1.33 (95% CI = 1.25–1.41) for those at counts above 500 cells/μL.

The estimated mean CD4 count at the initiation of therapy was 205 cells/μL (95% CI = 200–211). There was evidence that the mean count differed by cohort, ranging from 132 cells/μL for seroconverters in the Netherlands IDU cohort to 292 cells/μL for those in the Aquitaine cohort (P < 0.0001). After adjusting for the effect of individual cohort, we found strong evidence for a trend in higher estimated CD4 cell counts at initiation of ART with later calendar years (trend P < 0.0001, Table 4). Although there was no evidence that the estimated mean cell count at the initiation of therapy differed by age, there was marginal evidence suggesting that IDUs initiated ART at lower CD4 counts (P = 0.05), and that women and patients with an HIV test interval within 1 month initiated ART at higher CD4 counts (P = 0.05 and P = 0.04, respectively).

Table 4.  Estimated CD4 count (cells/μL) at initiation of ART: effect of explanatory variables
VariableDifference in mean CD4 cell count *95% confidence interval P
  • *

    From the reference category, defined as the Italian Seroconversion Study, exposure category sex between men, aged 30, male, HIV test interval 1 month or more, at risk for initiating ART 1989–94. Note: Based on 4758 people at risk for starting ART after 1/1/1987 and with at least one CD4 count, of whom 2849 started ART. P-values from log-likelihood tests. CD4 modelled on square root scale and then back-transformed to calculate absolute differences, with approximate 95% confidence intervals calculated using the Delta method. Multivariate model is also adjusted for study.

Exposure category
 Sex between men *− 0 0.05
 Injecting drug use− 30(− 46, − 15) 
 Sex between men and women− 16(− 34, + 2) 
 Haemophilia− 49(− 130, + 32) 
 Other/unknown− 25(− 54, + 4) 
Age at seroconversion (years)
 + 10 years+ 2(− 4, + 8)0.65
 Male *0 0.05
 Female+ 21(+ 5, + 37) 
HIV test interval
 1 month or more *0 0.04
 Within 1 month+ 30(+ 7, + 53) 
Calendar year at risk
 Pre-1989− 67(− 78, − 56)< 0.0001
 1989–94 *0  
 1995–96− 24(− 35, − 13) 
 1997–99+ 63(+ 48, + 78) 
Mean CD4 in reference category201(188, 215) 


  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. References
  8. Appendix

We found a 65% (95% CI = 57%-72%) reduction in the risk of death in the 1997–99 period, when HAART was available, compared to the 1989–94 period, when monotherapy was available. We also found a significant trend over time towards the earlier initiation of ART in relation to the time of seroconversion, with a relative rate of starting therapy of 4.01 for the calendar years 1997–99 compared to 1989–94. These two findings may appear to relate the initiation of ART closer to the time of seroconversion, with the observed survival improvements. However, a number of issues do not necessarily support this conclusion.

First, those initiating ART in more recent calendar periods are more likely to have started with more potent combinations of drugs.

Secondly, although the rate of starting therapy in 1987 and 1988 was significantly lower than that in 1992 (RR = 0.34 and 0.48, respectively), we found no evidence that survival was shorter for people at risk at that time. The relative risk of death in 1987–88 was 1.22 compared to the 1989–94 period. The delay of initiating therapy in the earlier years did not therefore appear to be detrimental to survival.

Thirdly, we found that IDUs were 21% less likely to initiate therapy at the same time from seroconversion compared to people exposed through sex between men. IDUs initiated therapy at an average of 30 cells/μL lower than men who have sex with men. There was no evidence, however, that the survival of IDUs was significantly different from that of people infected through other routes. This is similar to recent findings reported by the investigators of the Swiss cohort[40].

The finding that therapy is initiated closer to seroconversion for older individuals, but at similar CD4 counts among all age groups, is not surprising and is likely to be because the CD4 decline is faster in older individuals[41]. As their survival expectations are shorter, older individuals are more likely to initiate treatment earlier than younger individuals who are at the same time from seroconversion.

Although we observed an overall trend towards the earlier initiation of ART over time, it appears that ART was delayed in 1993 and 1994 with relative rates of 0.79 and 0.69, respectively, compared to 1992. The reason for this may be because the results of two clinical trials, AIDS Clinical Trials Group (ACTG) protocol 019[1] and Concorde[2], showing no overall survival benefit of immediate vs. deferred ZDV had become available and a disenchantment with ART may have ensued.

We found that patients with an HIV test interval of less than 1 month were more likely to initiate therapy early in their infection (P = 0.04) and at marginally higher CD4 cell counts than people with a test interval of 1 month or more. This suggests a faster decline in CD4 cell count in people with a test interval less than 1 month, whose overall survival is also significantly shorter (P = 0.02). Sero- converter studies are particularly valuable to the long-term monitoring of HIV-infected people presenting during the time of acute infection in whom treatment may be initiated early or else delayed.

We found no evidence of a difference in survival between males and females (P = 0.94) or of the time to the initiation of therapy (P = 0.47). There is evidence, however, that in women therapy is initiated an average of 21 cells/μL higher than in men. This is in agreement with findings from a European study of IDUs that, at the same time from seroconversion, women have higher CD4 cell counts than men[42].

The estimated mean CD4 cell count at which ART is initiated has risen dramatically over time. While the observed higher count at therapy initiation in the most recent calendar period appeared to be matched with an improvement in the survival of those observed in that period, evidence of a consistent association between earlier initiation of therapy and an improvement in survival is lacking in the other periods. The long-term risks and benefits of therapeutic regimens in HIV infection require constant and long-term monitoring.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. References
  8. Appendix
  • 1
    Volberding, PA, Lagakos, SW, Grimes, JM et al. A comparison of immediate with deferred zidovudine therapy for asymptomatic HIV-infected adults with CD4 cell counts of 500 or more per cubic millimeter. N Engl J Med 1995; 333, 401 407.
  • 2
    Concorde Co-ordinating Committee. Concorde. MRC/ANRS randomised double-blind controlled trial of immediate and deferred zidovudine in symptom-free HIV infection. Lancet 1994; 343, 871 881.
  • 3
    Hammer, SM, Katzenstein, DA, Hughes, MD et al. A trial comparing nucleoside monotherapy with combination therapy in HIV-infected adults with CD4 cell counts from 200 to 500 per cubic millimetre. AIDS Clinical Trials Group Study 175 Study Team. N Engl J Med 1996; 335, 1081 1090.
  • 4
    Saravolatz, LD, Winslow, DL, Collins, G et al. Zidovudine alone or in combination with didanosine or zalcitabine in HIV-infected patients with the acquired immunodeficiency syndrome or fewer than 200 CD4 cells per cubic millimetre. Investigators for the Terry Beirn Community Programs for Clinical Research on AIDS. N Engl J Med 1996; 335, 1099 1106.
  • 5
    Delta Co-ordinating Committee. Delta: a randomised double-blind controlled trial comparing combinations of zidovudine plus didanosine with zidovudine alone in HIV-infected individuals. Lancet 1996; 348, 283 291.
  • 6
    Hammer, SM, Squires, KE, Hughes, MD et al. A controlled trial of two nucleoside analogues plus indinavir in people with human immunodeficiency virus infection and CD4 cell counts of 200 per cubic millimetre or less. AIDS Clinical Trials Group 320 Study Team. N Engl J Med 1997; 337, 725 733.
  • 7
    Gulick, RM, Mellors, JW, Havlir, D et al. Treatment with indinavir, zidovudine, and lamivudine in adults with human immunodeficiency virus infection and prior antiretroviral therapy. N Engl J Med 1997; 337, 734 739.
  • 8
    Spira, R, Marimoutou, C, Binquet, C, Lacoste, D, Dabis, F, For The Groupe d'epidémiologie Clinique du, SIDA. en Aquitaine (GECSA). Rapid change in the use of antiretroviral agents and improvement in a population of HIV-infected patients. France, 199597 J Acquir Immune Defic Syndr 1998; 18, 358 364.
  • 9
    Egger, M, Hirschel, B, Francioli, P et al. Impact of new antiretroviral combination therapies in HIV infected patients in Switzerland: prospective multicentre study. Swiss HIV Cohort Study. BMJ 1997; 315, 1194 1199.
  • 10
    Mocroft, A, Vella, S, Benfield, TL et al. Changing patterns of mortality across Europe in patients infected with HIV-1. EuroSIDA Study Group. Lancet 1998; 352, 1725 1730.
  • 11
    Palella, Fj Jr, Delaney, KM, Moorman, AC et al. Declining morbidity and mortality among patients with advanced human immunodeficiency virus infection. HIV Outpatient Study Investigators. N Engl J Med 1998; 338, 853 860.
  • 12
    Detels, R, Muñoz, A, McFarlane, G et al. Effectiveness of potent antiretroviral therapy on time to AIDS and death in men with known HIV infection duration. Multicenter AIDS Cohort Study Investigators. JAMA 1998; 17, 1497 1503.
  • 13
    Uk Register of, HIV & Seroconverters Steering Committee. The, AIDS incubation period in the UK estimated from a national register of HIV seroconverters. AIDS 1998; 12, 659 667.
  • 14
    Dorrucci, M, Balducci, M, Pezzotti, P et al. Temporal changes in the rate of progression to death among Italians with known date of HIV seroconversion: estimates of the population effect of treatment. J AIDS 1999; 22, 65 70.
  • 15
    Pedersen, C, Lindhardt, BØ, Jensen, BØ et al. Clinical course of primary HIV infection: consequences for subsequent course of infection. BMJ 1989; 299, 154 157.
  • 16
    Touloumi, G, Karafoulidou, A, Gialeraki, A et al. Determinants of progression of HIV infection in a Greek haemophilia cohort followed for up to 16 years after infection. J AIDS 1998; 19, 89 97.
  • 17
    Schwartländer, B, Bek, B, Skarabis, H et al. Improvement of the predictive value of CD4+ lymphocyte count by beta 2-microglobulin, immunoglobulin A and erythrocyte sedimentation rate. AIDS 1993; 7, 813 821.
  • 18
    Marimoutou, C, Chene, G, Dabis, F, Lacoste, D, Salamon, R, For The Groupe d'epidemiologie Clinique du, SIDA. en Aquitaine (GECSA). Human immunodeficiency virus infection and AIDS in Aquitaine. 10 years' experience of a hospital information system. 198595 Presse Med 1997; 26, 703 710. [in French].
  • 19
    Carré, N, Meyer, L, Boufassa, F et al. and the SEROCO Study Group. High risk of HIV disease progression after infection through a sexual partner with AIDS. AIDS 1996; 10, 77 80.
  • 20
    Pezzotti, P, Phillips, AN, Dorrucci, M et al. Category of exposure to HIV and age in the progression to AIDS. longitudinal study of 1199 people with known dates of seroconversion. BMJ 1996; 313, 583 586.
  • 21
    Van Den Hoek, JAR, Coutinho, RA, Van Haastrecht, HJA et al. Prevalence and risk factors of HIV infections among drug users and drug-using prostitutes in Amsterdam. AIDS 1988; 2, 55 60.
  • 22
    Van Griensven, GJP, Tielman, RAP, Goudsmit, J et al. Risk factors and prevalence of HIV antibodies in homosexual men in the Netherlands. Am J Epidemiol 1987; 125, 1048 1057.
  • 23
    Eskild, A, Magnus, P, Brekke, T et al. The impact of exposure group on the progression rate to acquired immunodeficiency syndrome. Scand J Infect Dis 1997; 29, 103 109.
  • 24
    Dobloug, JH, Bruun, JN, Skaug, K. The early introduction of HIV infection among Norwegians at highest risk. Scand J Infect Dis 1990; 22, 753 754.
  • 25
    Montoro, JB, Oliveras, J, Lorenzo, JI et al. An association between clotting factor concentrates use and mortality in human immunodeficiency virus-infected haemophilic patients. Blood 1995; 86, 2213 2219.
  • 26
    Grupo de estudio de seroconvertores, CAM. Identificación de factores que afectan a la historia natural de la infección por el VIH en una cohorte de personas con fecha de seroconversión conocida. Santiago de Compostela, SEISIDA, 1999 CP5–7.
  • 27
    Perez-Hoyos, S, Avino, M, Hernandez, I et al. AIDS-free time and survival of an injecting drug users HIV seroconverters cohort. Gac Sanit 1999; 13, 337 345.
  • 28
    Brettle, RP, McNeil, AJ, Burns, S et al. Progression of HIV. Follow-Up of Edinburgh Injecting Drug Users with Narrow Seroconversion Intervals In 198385 AIDS 1996; 10, 419 430.
  • 29
    Sabin, CA, Devereux, H, Phillips, AN et al. Immune markers and viral load after HIV-1 seroconversion as predictors of disease progression in a cohort of haemophilic men. AIDS 1998; 12, 1347 1352.
  • 30
    Tindall, B, Swanson, CE, Cooper, DA. Development of AIDS in a cohort of HIV-seropositive homosexual men in Australia. Med J Aust 1990; 153, 260 265.
  • 31
    Vanhems, P, Allard, R, Cooper, A et al. Acute human immunodeficiency virus type 1 disease as a mononucleosis-like illness: is the diagnosis too restrictive. Clin Infect Dis 1997; 24, 965 970.
  • 32
    Kaplan, EL & Meier, P. Non-parametric estimation from incomplete observations. J Am Stat Assoc 1958; 53, 457 481.
  • 33
    Clayton, C & Hills, M. Statistical Models in Epidemiology. Oxford. Oxford University Press, 1993.
  • 34
    Vanhems, P, Lambert, J, Cooper, DA et al. Acute HIV-1 illness severity and prognosis: a dose response effect. Clin Infect Dis 1998; 26, 323 329.
  • 35
    Lindbäck, S, Broström Karlsson, A, Gaines, H. Does symptomatic primary HIV-1 infection accelerate progression to CDC stage IV disease, CD4 count below 200 × 106/l, AIDS, and death from AIDS? BMJ 1994; 309, 1535 1537.
  • 36
    Keet, IPM, Krijnen, P, Koot, M et al. Predictors of rapid progression to AIDS in HIV-1 seroconverters. AIDS 1993; 7, 51 57.
  • 37
    Cox, DR. Regression models and life tables. J R Stat Soc 1972; 34, 187 220.
  • 38
    Huber, PJ. The behavior of maximum likelihood estimates under non-standard conditions. In: Proceedings of the Fifth Berkeley Symposium on Mathematical Statistics and Probability. Berkeley: University of California Press, 1967: 221 233.
  • 39
    Durrelman, S & Simon, R. Flexible regression models with cubic splines. Stats Med 1989; 8, 551 561.
  • 40
    Junghans, C, Low, N, Chan, P et al. Uniform risk of clinical progression despite differences in utilization of highly active antiretroviral therapy: Swiss HIV Cohort Study. AIDS 1999; 13, 2547 2554.
  • 41
    Phillips, AN, Lee, CA, Elford, J et al. More rapid progression to AIDS in older HIV-infected people: the role of CD4+ T-cell counts. J Acquir Immune Defic Syndr 1991; 4, 970 975.
  • 42
    Prins, M, Robertson, JR, Brettle, RP et al. Do gender differences in CD4 cell counts matter? AIDS 1999; 13, 2361 2443.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. References
  8. Appendix

Analysis and Writing Committee. Kholoud Porter, Abdel Babiker, Sarah Walker, Janet Darbyshire, Noël Gill.

Steering Committee. Valerie Beral, Roel Coutinho, Janet Darbyshire (Project Leader), Noël Gill (Chairman), Christine Lee, Laurence Meyer, Giovanni Rezza.

Co-ordinating Centre. Kholoud Porter (Scientific Co-ordinator), Abdel Babiker, Sarah Walker, Janet Darbyshire, Freya Tyrer.

Collaborators. Aquitaine cohort, France: Francois Dabis, Catherine Marimoutou; SEROCO cohort, France: Laurence Meyer, Faroudy Boufassa; German cohort, Germany: Osamah Hamouda, Monika Brunn; Italian Seroconversion Study, Italy: Patrizio Pezzotti, Giovanni Rezza; Valencia haemophilia cohort, Spain: Jose I. Lorenzo; Greek Haemophilia cohort, Greece: Giota Touloumi, Angelos Hatzakis, Anastasia Karafoulidou, Olga Katsarou; Edinburgh Hospital cohort, United Kingdom: Ray Brettle; Madrid cohort, Spain: Julia Del Amo, Jorge del Romero; Amsterdam Cohort Study among drug users, the Netherlands: Maria Prins, Roel A Coutinho; Amsterdam Cohort Study on homosexual men, the Netherlands: Birgit van Benthem, Roel A Coutinho; Copenhagen cohort, Denmark: Ole Kirk, Court Pedersen; Valencia IDU cohort, Spain: Ildefonso Hernández Aguado, Santiago Pérez-Hoyos; Oslo and Ulleval Hospital cohorts, Norway: Anne Eskild, Johan N Bruun, Mette Sannes; Royal Free haemophilia cohort, United Kingdom: Caroline Sabin, Christine Lee; UK Register of HIV Seroconverters, UK: Anne M Johnson, Andrew N Phillips, Abdel Babiker, Janet H Darbyshire, Noël Gill, Kholoud Porter; Swiss HIV cohort, Switzerland: Matthias Egger, Patrick Francioli, Martin Rickenbach; Sydney AIDS Prospective Study, Australia: David Cooper, John Kaldor; Sydney Primary HIV Infection cohort, Australia: David Cooper, John Kaldor, Jeanette Vizzard; Barcelona Haemophilia cohort, Spain: Joan M Tusell, Isabel Ruiz; Barcelona IDU cohort, Spain: Joan A Cayla, Patricia Garcia de Olalla; MRC Biostatistics Unit, Cambridge, UK: Nicholas E Day, Daniela De Angelis.