The introduction of protease inhibitors (PI) was a major step forward in the treatment of HIV-infected patients, and since their introduction the international guidelines for antiretroviral therapy have recommended combination treatment with at least three drugs, consisting of one (two) PI(s) and two nucleoside analogue reverse transcriptase inhibitors (NRTI), although the class of non-nucleoside analogue reverse transcriptase inhibitors (NNRTI) has recently become an equal alternative to the PI component [1,2].
Such guidelines are generally based on randomized clinical trials, and several of these have documented the superiority of triple therapy including a PI compared with dual therapy with two NRTIs in terms of effect on surrogate markers as well as on clinical progression [3–6]. In contrast, no randomized clinical trial comparing the efficacy of different PI-containing regimens was available when this class of drugs first became available in clinical practice.
Observational studies have documented an inferior virological response among patients starting saquinavir hard gel capsules (hgc), but the implications of this in terms of potential differences in long-term immunological and, in particular, clinical response remain to be determined [7–14]. It is unlikely that large scale, randomized trials will be performed to compare the long-term clinical efficacy of PI's. In these circumstances, observational studies of large groups of patients may provide valuable information, although in such studies there might be a potential bias in treatment comparisons [15,16]. When carefully interpreted, results from observational studies might provide unbiased information, especially if major prognostic variables are fairly similar in the groups compared .
We therefore aimed to analyse the long-term clinical response relative to the virological and immunological responses among patients starting one of the three first available PIs, ritonavir, indinavir and saquinavir hgc within the EuroSIDA study.
EuroSIDA is a prospective, observational cohort study consisting of more than 8500 HIV-infected patients from 63 clinical centres in 20 European countries (see Appendix). A predetermined number of consecutive HIV-infected patients older than 16 years and seen in out-patient clinics were enrolled from May 1994 (EuroSIDA cohort I, 3120 patients), from December 1995 (EuroSIDA cohort II, 1367 patients), from February 1997 (EuroSIDA cohort III, 2844 patients), and from March 1999 (EuroSIDA cohort IV, 1200 patients). A CD4 cell count < 500 × 106 cells/l within the last 4 months was obligatory for cohorts I–III, and only data from these three cohorts was included in the present analysis, as very limited follow-up data is yet available for cohort IV.
Information was collected from patients charts and by patients interviews and recorded on a standardized data collection form at enrolment and at half-yearly follow-up visits. Details of data collection in the study has previously been published . At each follow-up, information on all CD4 cell counts and viral loads performed over the past 6 months is requested, as are all clinical diagnoses and changes to treatment that have occurred since last patient follow-up.
Treatment was registered as the date of first time ever on a given drug and in case of permanent discontinuation, the date of stopping the treatment. Time variables were collected as month and year. All forms were checked by scientific staff at the co-ordinating office for logical errors, and the centres were monitored to ensure uniform and correct patient enrolment and data transmission from patient charts to the data collection form.
Eligible patients for the analysis were those who under prospective follow-up in the EuroSIDA study started ritonavir, indinavir or saquinavir hgc as part of their initial PI-containing regimen [highly active antiretroviral therapy (HAART), ≥ 3 drugs. Patients who started a PI as monotherapy/dual therapy, or started a PI-containing regimen prior to recruitment to the EuroSIDA study, were excluded from the analyses. In addition, a CD4 cell count in the 6 months prior to starting the PI was mandatory, as was the demand of some follow-up after starting the PI. Patients who started dual PI therapy, nelfinavir or amprenavir were not included in the analysis due to a low number of patients and insufficient follow-up time.
For the present analysis, information from 11 follow-up forms for cohort I, eight for cohort II and five for cohort III was available. The latest follow-up to be included was performed in the period December 1999–February 2000.
The baseline characteristics of the three treatment groups were compared using χ2 test and non-parametric tests such as the Kruskall–Wallis test. These parameters included demographic characteristics, CD4 cell count, plasma viral load (pVL), previous exposure to NRTIs as well as number of new antiretroviral drugs added at baseline and the NRTIs used at baseline.
The regional classification was based on an arbitrary division used in earlier EuroSIDA analyses (South: Greece, Israel, Italy, Portugal, and Spain; Central: Belgium, France, south Germany, Luxembourg, Switzerland; North: Denmark, Ireland, north Germany, Netherlands, Norway, Sweden, and United Kingdom) [18,19].
In a Kaplan–Meier analysis of time to modification of the initial PI component of the HAART regimen (i.e. stopping, adding a new PI or changing the PI component), the rates of modification were compared using the log rank test. Changes in CD4 cell count and pVL (log10-transformed) from baseline were assessed at three-monthly intervals as estimated by linear interpolation and compared between the three groups using the Kruskall–Wallis test. A 3-month time interval was chosen as this was the median time between laboratory measurements. In the analysis of pVL, only those who had a pVL in the 6-month period before baseline were included. The proportions of patients with a pVL < 500 copies/ml at each time point were compared using χ2 test. This cut-off value was chosen due to the variety of different cut-off levels in the assays used.
To compare the rate of clinical progression in the three groups, Kaplan–Meier analysis of time to clinical progression (new AIDS-defining event or death) was performed. The potential influence of the type of PI initiated on the clinical progression rate was further evaluated in Cox proportional hazards models. These models included parameters such as gender, HIV transmission category, race, region of Europe, CD4 cell count, haemoglobin, weight, development of AIDS, calendar time, and treatment status (naive, previous exposure to each of the NRTIs and treatment duration and number of new drugs added at baseline). Any factors that were significant in univariate models with P < 0.1 were then included in multivariate models. Furthermore, as there were no major differences according to region of Europe, this parameter was excluded from the model, and we used stratification by centre of EuroSIDA instead to adjust for site heterogeneity. The assumption of proportional hazards in the Cox model was formally tested, and revealed no evidence of non-proportionality. All analyses were run in accordance with the principle of intent-to-treat, and no adjustments were made for modification of treatment.
The analysis was performed in SAS, version 6.12 (SAS Institute, Cary, North Carolina, USA). P-values of less than 0.05 were considered significant, and all tests of significance were two sided.
A total of 2708 patients fulfilled the inclusion criteria of this study; of which 556 (21%) initiated ritonavir, 1342 (50%) indinavir, and 810 (30%) saquinavir hgc. The characteristics of these patients were in general very similar across the treatment groups (Table 1). Patients starting ritonavir had a slightly lower CD4 count and higher pVL than patients starting other PI-regimens (Table 1).
Approximately 90% of all patients had received antiretroviral treatment before starting their first PI. Differences in previous exposure to the specific NRTIs were observed; zalcitabine was more common in the saquinavir hgc group compared with the other groups, and there was also minor differences in the duration of NRTI treatment. Furthermore, at the time of starting the initial PI, the use of NRTIs differed between the groups, and a higher proportion of patients in the ritonavir group initiated PI therapy without starting other new drugs (Table 2).
Modification of initial PI-containing regimen
The rate of modification differed substantially between PI groups. In total, 64% of the patients who initiated ritonavir modified the PI component of the HAART regimen, whereas 56 and 71% did so in the indinavir and saquinavir hgc groups, respectively;P < 0.001. The median time to modifying the PI therapy was 20 months [95% confidence interval (CI), 17–23] for ritonavir, 24 (95% CI, 22–26) for indinavir, and 17 (95% CI, 15–18) for saquinavir hgc, P < 0.001. The rate of modification of ritonavir was initially high and subsequently slowed down, whereas the rate of modification of saquinavir hgc treatment continued to increase over time (Fig. 1).
Virological and immunological responses
The virological response is illustrated in Figure 2a. The median decrease in pVL was 1.5 (ritonavir), 1.5 (indinavir) and 1.2 log10 copies/ml (saquinavir hgc) after 12 months;P < 0.001. An analysis of the proportion of patients who had a pVL < 500 copies/ml at a given time point, resulted in a similar picture. After 12 months follow-up, 47, 54 and 41% of the patients had a pVL < 500 copies/ml;P < 0.001. These differences remained significant in a multivariate logistic regression model including the same variables as listed in Table 3 : odds ratio of a pVL < 500 copies/ml after 12 months was 1.83 (95% CI, 1.33–2.51) for ritonavir compared with saquinavir hgc and 2.39 (95% CI, 1.86–3.06) for indinavir compared with saquinavir hgc.
The changes in CD4 cells are shown in Figure 2b. At 12 months after starting HAART, the median increase in CD4 cell count was 90 (ritonavir), 96 (indinavir) and 74 × 106 cells/l (saquinavir hgc);P < 0.001. In a multivariate linear regression model, the change in CD4 cell count after 12 months differed significantly between indinavir and saquinavir hgc (P = 0.007), but not between ritonavir and saquinavir hgc (P = 0.19).
Among the 2708 patients, 447 (17%) progressed clinically; either in terms of developing a first AIDS-defining event (n = 137, 31%), a new AIDS-defining event after the initial AIDS diagnosis (n = 179, 40%) or death (n = 131, 29%). The most common AIDS-defining events were oesophageal candidosis (19%) and non-Hodgkin lymphoma (12%); all other events occurred in less than 10% of the patients. In the ritonavir group, 106 patients progressed (19%), whereas 201 (15%) and 140 (17%) did so in the indinavir and saquinavir hgc groups, respectively. There was no difference in type of event according the PI used (P = 0.55). The incidence of clinical progression was 8.4 (95% CI, 6.8–10.0) for ritonavir, 6.8 (95% CI, 5.9–7.7) for indinavir and 7.9 (95% CI, 6.6–9.2) events/100 person–years of follow-up for saquinavir hgc; (ritonavir versus indinavir:P = 0.09; ritonavir versus saquinavir hgc P = 0.63; and indinavir versus saquinavir hgc:P = 0.19). The incidences were based on 1261, 2937 and 1774 person–years of follow-up, respectively, and the overall loss to follow-up throughout the study period (not seen since January 1999) was 9.4, 7.5 and 12.6%, respectively). Further, the incidence of clinical progression did not differ between the treatment groups when assessed in CD4 and pVL strata according to the latest measurements (data not shown).
The proportion of patients who had experienced clinical progression at 12 months was 11.9% (95% CI, 9.2–14.6) for ritonavir; 9.2% (95% CI, 7.6–10.8) for indinavir; and 11.9% (95% CI, 9.7–14.1) for saquinavir hgc; log-rank test, P = 0.20 (Fig. 3). To analyse potential differences in clinical outcome between the three treatment groups, Cox proportional hazards models were established. When adjusting for differences in factors at baseline, patients starting indinavir were at a statistically significant 23% lower risk of progressing clinically compared with those starting saquinavir hgc (P = 0.043). Patients starting ritonavir were at 17% lower risk compared with the saquinavir group, although this result was not statistically significant (Table 3). Other prognostic baseline factors were age, CD4 cell count, body weight, haemoglobin, prior AIDS diagnosis, and number of new drugs initiated when starting HAART (Table 3). Factors not significantly associated with clinical progression in univariate models (P < 0.10) were not included in the final multivariate model. In particular, HIV transmission category was not a significant prognostic factor and did not change the results of the multivariate model if included.
In the highly selected subset of patients with a pVL measurement in the 6 months prior to starting the PI, this covariate was also included in the model. Adjusting for the same factors as in the original model and additionally pVL only changed the relative risks (RH) slightly (although it changed the P-values quite substantially due to smaller numbers). When pVL was included, RH for ritonavir compared with saquinavir hgc was 0.92 (95% CI, 0.59–1.45;P = 0.72); and for indinavir compared with saquinavir hgc, 0.89 (95% CI, 0.60–1.30;P = 0.53); and the corresponding relative hazards without inclusion of pVL were 0.92 (95% CI, 0.59–1.44;P = 0.71) and 0.88 (95% CI, 0.60–1.29;P = 0.53), respectively. Baseline pVL was not a significant prognostic factor for clinical progression (RH = 1.12; 95% CI, 0.95–1.31 per log10 higher;P = 0.17).
Again including all patients, the RH of clinical progression in the first 12 months of HAART and thereafter was also investigated. These models were adjusted for identical factors as in Table 3. In the first 12 months of HAART, patients starting ritonavir or indinavir were at a 22 and 32% lower risk of clinical progression, RH = 0.78 (95% CI, 0.54–1.12;P = 0.17) and 0.68 (95% CI, 0.50–0.93;P = 0.02), respectively compared with those starting saquinavir hgc. After 12 months of HAART, the risk of clinical progression did not differ significantly, RH = 1.00 (95% CI, 0.61–1.63;P = 0.99) and 1.01 (95% CI, 0.65–1.58;P = 0.95), respectively.
When using the latest values of CD4 cell count, weight, and haemoglobin in order to include changes in these markers which occurred after starting HAART, there was no longer any difference (see Table 3 right column for comparison) in the risk of clinical progression according to the type of PI initiated: RH (ritonavir versus saquinavir hgc) = 0.93 (95% CI, 0.67–1.29;P = 0.66), and RH (indinavir versus saquinavir hgc) = 0.94 (95% CI, 0.70–1.26;P = 0.70). When the latest pVL was included (in the subset in which this was available), the RHs were 0.97 (95% CI, 0.67-1.42;P = 0.89), and 1.02 (95% CI, 0.74-1.43;P = 0.89) respectively. Therefore, after adjustment for the poorer initial virological and immunological response seen among patients starting saquinavir hgc, there were no differences in the risk of clinical progression between ritonavir, indinavir and saquinavir hgc.
We have documented that use of saquinavir hgc as part of a HAART regimen is associated with a significantly inferior clinical outcome compared with indinavir which was consistent with and in the same direction as the differences in CD4 cell count and pVL responses. Differences in the marker responses explained fully the higher rate of clinical progression seen among patients starting saquinavir hgc.
Other studies have reported on the inferior virological outcome of saquinavir hgc capsules compared with other PIs [7–14]. The consistency of this finding across several independent observational studies, where patterns of saquinavir hgc use have varied widely, together with the known low bioavailability of saquinavir hgc capsules , indicates that the association is largely due to the inferior potency of saquinavir hgc and not to bias due to confounding by indication. Thus this seems to be an example of where observational studies have reliably identified a difference in efficacy between drug regimens. This finding does not include the better-absorbed saquinavir soft gel capsules which have recently replaced saquinavir hgc nor the ritonavir-boosted saquinavir hgc therapy where the pharmacological properties are dramatically improved [21–23].
No randomized trials and observational studies have reported on long-term immunological and clinical outcome according to the PI used. In our study, the CD4 response corresponded in general to the pVL response. In a multivariate Cox model of clinical progression, patients starting indinavir were at a 23% lower risk of clinical progression relative to those starting saquinavir hgc. The difference between ritonavir and saquinavir hgc was nearly at the same level as that of indinavir versus saquinavir hgc, but the difference was not significant, perhaps due to the smaller number of patients starting ritonavir.
Our results are not fully in agreement with a previous study which did not find a tendency of an inferior clinical response among patients starting Saquinavir hgc relative to other PIs in spite of significant differences in virological response . There is no obvious reason for this discrepancy between the studies, but chance may play a role. Providing a substantially longer follow-up with a median follow-up of 30 months after starting HAART, the present study allows for estimating longer-term clinical outcome and for a better precision of the size of effect, RH = 0.77 (95% CI, 0.60–0.99) in the present study compared with 1.10 (95% CI, 0.81–1.49) for saquinavir hgc versus indinavir . In our study, the difference in clinical progression seemed to get weaker with increasing time, as a post-hoc analysis fond that the effect was only present within the first 12 months of HAART. However, there was no significant interaction between the effects of calendar time and type of PI on the clinical outcome, and the CD4 cell response remained consistently lower in the saquinavir group compared with the indinavir group. Therefore, to further analyse the longer-term differences, continuous monitoring is warranted.
It is possible that antiretroviral therapy is modified according to changes in surrogate markers, so that patients who experience inferior immunological and/or virological responses are switched to a more potent drug combination before experiencing clinical progression. This would limit our ability to detect real clinical differences between the PI groups. In accordance with this, the differences in virological response decreased with increasing time on HAART. Further, the CD4 cell count, which showed less pronounced differences between the PI groups than the pVL, might be more important for the prognosis than pVL [24–26].
Our results are based on the principle of intent-to-treat, which is a trial concept, though in many clinical trials even an intent-to-treat analysis will exclude patients after they discontinue treatment, as such patients are not routinely assessed for viral load or CD4 cell counts [1,27,28]. In an observational study as EuroSIDA, this information is readily available, and in addition, the loss to follow-up rate is very low in EuroSIDA (< 5% per year). Thus, patients who modify treatment are still included in the analyses, rather than excluding their data or registering them as a failure . Inclusion of data after discontinuation of the initial PI provides important information, as a given treatment might influence the outcome of a subsequent therapy due to the development of resistance. Factors such as adherence, poor absorption and potency of a given antiretroviral drug might therefore affect the outcome of a subsequent regimen, and the inferior immunological and virological responses of Saquinavir hgc might lead to compromised responses to the following regimens .
Several reservations in the interpretation of the findings in the present study are necessary. First of all, this is not a randomized clinical trial, and there are many different reasons governing the decision to start an individual PI which would not be a bias in a randomized trial, but which may affect the results of an observational study [15,16,31]. For example, experience with PI treatment has rapidly advanced and patients starting a PI regimen in 1998 may be more likely to start three new drugs compared to patients starting in 1996. We have attempted to adjust for such factors in our analyses, but other potential confounders may remain that we were not able to adjust for. Additional adjustment for pVL in the subgroup of patients with available pVL at the time of starting the initial PI only caused minor changes in the risk of clinical progression in the three treatment arms. As an important point, the key baseline factors for the three treatment groups were fairly well balanced, and this was particularly so between the groups of indinavir and saquinavir hgc. Therefore the statistical adjustments are likely to remove the influence of confounding from these factors . Furthermore, as discussed above, our results concerning virological outcome are consistent with those from several other cohorts .
No randomized trial has been conducted to analyse these potential differences in long-term clinical prognosis of different PI-containing regimen [1,2]. Therefore, observational cohorts might be the only source for providing data on this issue . Being a large observational cohort, representing a variety of centres over most of Europe with heterogeneity in the approach to therapy and a substantial total follow-up time, the EuroSIDA study has the capacity of addressing the question of potential differences in the clinical outcome as well as the immunological and virological outcome according to which treatment the patient has received .
The patients in EuroSIDA make up an unselected part of the patient population at the participating centres, representing most parts of Western Europe, which contrasts the often highly selected patients included in randomized trials. However, it should be kept in mind that although the patients in EuroSIDA probably better reflect the general HIV-infected population than the patients of the randomized clinical trials, many of the EuroSIDA centres are large clinics and often university-associated. Such centres might have earlier access to new therapies, access to HAART through clinical trials and expanded access, more regular follow-up visits and assessments, and are therefore likely to represent a high quality of clinical practice.
In conclusion, among patients starting their initial PI regimen, saquinavir hgc is associated with inferior virological and immunological responses relative to indinavir. We documented consistent differences in the long-term clinical response with those in virological and immunological responses. The poorer clinical outcome among patients starting saquinavir hgc could be fully explained by changes in markers on HAART.
Sponsorship: The European Commission BIOMED 1 (CT94-1637) and BIOMED 2 (CT97-2713) programmes were the primary sponsors of the EuroSIDA study. Unrestricted grants were also provided by Pharmacia Upjohn, Glaxo Wellcome, Roche, and Merck. The participation of centres from Switzerland was supported by a grant from the Swiss Federal Office for Education and Science.
1. Anon. British HIV Association guidelines for antiretroviral treatment of HIV seropositive individuals.BHIVA Guidelines Co-ordinating Committee.
Lancet 1997, 349: 1086 –1092.
2. Carpenter CC, Cooper DA, Fischl MA. et al
. Antiretroviral therapy in adults: updated recommendations of the International AIDS Society-USA Panel. JAMA 2000, 283: 381 –390.
3. Hammer SM, Squires KE, Hughes MD. et al
. A controlled trial of two nucleoside analogues plus indinavir in persons with human immunodeficiency virus infection and CD4 cell counts of 200 per cubic millimeter or less.AIDS Clinical Trials Group 320 Study Team.
N Engl J Med 1997, 337: 725 –733.
4. 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.
5. Cameron DW, Heath-Chiozzi M, Danner S. et al
. Randomised placebo-controlled trial of ritonavir in advanced HIV-1 disease.Advanced HIV Disease Ritonavir Study Group.
Lancet 1998, 351: 543 –549.
6. Collier AC, Coombs RW, Schoenfeld DA. et al
. Treatment of human immunodeficiency virus infection with saquinavir, zidovudine, and zalcitabine.AIDS Clinical Trials Group.
N Engl J Med 1996, 334: 1011 –1017.
7. Paredes R, Mocroft A, Kirk O. et al
. Predictors of virological success and ensuing failure in HIV-positive patients starting highly active antiretroviral therapy
in Europe: results from the EuroSIDA study. Arch Intern Med 2000, 160: 1123 –1132.
8. Fatkenheuer G, Theisen A, Rockstroh J. et al
. Virological treatment failure of protease inhibitor therapy
in an unselected cohort of HIV-infected patients. AIDS 1997, 11: F113 –F116.
9. Wit FW, van Leeuwen R, Weverling GJ. et al
. Outcome and predictors of failure of highly active antiretroviral therapy
: one-year follow-up of a cohort of human immunodeficiency virus type 1-infected persons. J Infect Dis 1999, 179: 790 –798.
10. Casado JL, Perez-Elias MJ, Antela A. et al
. Predictors of long-term response to protease inhibitor therapy
in a cohort of HIV-infected patients. AIDS 1998, 12: F131 –F135.
11. d'Arminio MA, Testa L, Adorni F. et al
. Clinical outcome and predictive factors of failure of highly active antiretroviral therapy
in antiretroviral-experienced patients in advanced stages of HIV-1 infection. AIDS 1998, 12: 1631 –1637.
12. Ledergerber B, Egger M, Opravil M. et al
. Clinical progression and virological failure on highly active antiretroviral therapy in HIV-1 patients: a prospective cohort study.Swiss HIV Cohort Study.
Lancet 1999, 353: 863 –868.
13. Jensen-Fangel S, Kirk O, Larsen L. et al
. Saquinavir hard gel suppresses viral load insufficiently in HIV- infected patients naive to anti-retroviral therapy: a retrospective cohort study. Scand J Infect Dis 1999, 31: 489 –493.
14. Grabar S, Pradier C, Le Corfec E. et al
. Factors associated with clinical and virological failure in patients receiving a triple therapy including a protease inhibitor. AIDS 2000, 14: 141 –149.
15. Chalmers I. Unbiased, relevant, and reliable assessments in health care: important progress during the past century, but plenty of scope for doing better [editorial]. BMJ 1998, 317: 1167 –1168.
16. Pocock SJ, Elbourne DR. Randomized trials or observational tribulations? N Engl J Med 2000, 342: 1907 –1909.
17. Phillips AN, Grabar S, Tassie JM, Costagliola D, Lundgren JD, Egger M. Use of observational databases to evaluate the effectiveness of antiretroviral therapy for HIV infection: comparison of cohort studies with randomized trials.EuroSIDA, the French Hospital Database on HIV and the Swiss HIV Cohort Study Groups.
AIDS 1999, 13: 2075 –2082.
18. Lundgren JD, Phillips AN, Vella S. et al
. Regional differences in use of antiretroviral agents and primary prophylaxis in 3122 European HIV-infected patients.EuroSIDA Study Group.
J Acquir Immune Defic Syndr Hum Retrovirol 1997, 16: 153 –160.
19. Phillips AN, Katlama C, Barton S. et al
. Survival in 2367 zidovudine-treated patients according to use of other nucleoside analogue drugs.The EuroSIDA Study Group.
J Acquir Immune Defic Syndr Hum Retrovirol 1998, 17: 239 –244.
20. Noble S, Faulds D. Saquinavir.A review of its pharmacology and clinical potential in the management of HIV infection.
Drugs 1996, 52: 93 –112.
21. Cohen SJ, Schuurman R, Burger DM. et al
. Randomized trial comparing saquinavir soft gelatin capsules versus indinavir as part of triple therapy (CHEESE study). AIDS 1999, 13: F53 –F58.
22. Mitsuyasu RT, Skolnik PR, Cohen SR. et al
. Activity of the soft gelatin formulation of saquinavir in combination therapy in antiretroviral-naive patients.NV15355 Study Team.
AIDS 1998, 12: F103 –F109.
23. Katzenstein TL, Kirk O, Pedersen C. et al
. The Danish Protease Inhibitor Study: a randomized study comparing the virological efficacy of 3 protease inhibitor-containing regimens for the treatment of human immunodeficiency virus type 1 infection. J Infect Dis 2000, 182: 744 –750.
24. Deeks SG, Hecht FM, Swanson M. et al
. HIV RNA and CD4 cell count response to protease inhibitor therapy
in an urban AIDS clinic: response to both initial and salvage therapy. AIDS 1999, 13: F35 –F43.
25. Kaufmann D, Pantaleo G, Sudre P, Telenti A. CD4-cell count in HIV-1-infected individuals remaining viraemic with highly active antiretroviral therapy (HAART).Swiss HIV Cohort Study.
Lancet 1998, 351: 723 –724.
26. Miller V, Sabin C, Rottmann C, et al
. Prognostic value of virus load levels in patients receiving HAART: the Frankfurt cohort.Fourth International Congress on Drug Therapy in HIV Infection
. Glasgow,October 1998 [Abstract OP 6.1.2].
27. Hollis S, Campbell F. What is meant by intention to treat analysis? Survey of published randomised controlled trials. BMJ 1999, 319: 670 –674.
28. Altman DG. Practical Statistics for Medical Research
, 1st edn, 1991. London, UK: Chapman & Hall.
29. Anon. Analysis of HIV-1 clinical trials: statistical magic? The AVANTI Steering Committee. Lancet 1999, 353: 2061 –2064.
30. Hirsch MS, Brun-Vezinet F, D'Aquila RT. et al
. Antiretroviral drug resistance testing in adult HIV-1 infection: recommendations of an International AIDS Society-USA Panel. JAMA 2000, 283: 2417 –2426.
31. Barton S. Which clinical studies provide the best evidence?.The best rct still trumps the best observational study.
BMJ 2000, 321: 255 –256.
32. Kirk O, Mocroft A, Katzenstein TL. et al
. Changes in use of antiretroviral therapy in regions of Europe over time.EuroSIDA Study Group.
AIDS 1998, 12: 2031 –2039.
The EuroSIDA Study Group
The multicentre study group on EuroSIDA (national coordinators in parenthesis).
Austria (N. Vetter) Pulmologisches Zentrum der Stadt Wien, Vienna. Belgium (N Clumeck) P. Hermans, B. Sommereijns, Saint-Pierre Hospital, Brussels; R. Colebunders, Institute of Tropical Medicine Antwerp. Czech Republic (L. Machala) H. Rozsypal, Faculty Hospital Bulovka, Prague. Denmark (J. Nielsen) J. Lundgren,T. Benfield, O. Kirk, Hvidovre Hospital, Copenhagen; J. Gerstoft, T. Katzenstein, B. Røge, P. Skinhøj, Rigshospitalet, Copenhagen; C. Pedersen, Odense University Hospital, Odense. France (C. Katlama) C. Rivière, Hôpital de la Pitié-Salpétière, Paris; J.-P. Viard, Hôpital Necker-Enfants Malades, Paris; T. Saint-Marc, Hôpital Edouard Herriot, P. Vanhems, University Claude Bernard, Lyon; C. Pradier, Hôpital de l'Archet, Nice. Germany (M. Dietrich) C. Manegold, Bernhard-Nocht-Institut for Tropical Medicine, Hamburg; J. van Lunzen, Eppendorf Medizinische Kernklinik, Hamburg; V. Miller, S. Staszewski, J.W. Goethe University Hospital, Frankfurt; F.-D. Goebel, Medizinische Poliklinik, Munich; B. Salzberger, Universität Köln, Cologne; J. Rockstroh, Universitäts Klinik Bonn. Greece (J. Kosmidis) P. Gargalianos, H. Sambatakou, J. Perdios, Athens General Hospital, Athens; G. Panos, M. Astriti, 1st IKA Hospital, Athens. Hungary (D. Banhegyi) Szent Lásló Hospital, Budapest. Ireland (F. Mulcahy) St. James's Hospital, Dublin. Israel (I. Yust) D. Turner, Ichilov Hospital, Tel Aviv; S. Pollack, Z. Ben-Ishai, Rambam Medical Center, Haifa: Z. Bentwich, Kaplan Hospital, Rehovot; S. Maayan, Hadassah University Hospital, Jerusalem. Italy (S. Vella, A. Chiesi) Istituto Superiore di Sanita, Rome; F. Suter, A. Cremaschi, Ospedale Riuniti, Bergamo; R. Pristerá, Ospedale Generale Regionale, Bolzano; F. Mazzotta, F.Vichi, Ospedale S. Maria Annunziata, Florence; B. DeRienzo, A. Bedini, Università di Modena, Modena; A. Chirianni, E. Montesarchio, Presidio Ospedaliero AD. Cotugno, Naples; V. Vullo, P. Santopadre, Università di Roma La Sapienza, Rome; C. Arici, P. Franci, P. Narciso, A. Antinori, M. Zaccarelli, Ospedale Spallanzani, Rome; A. Lazzarin, R. Finazzi, Ospedale San Raffaele, Milan; A. D'Arminio Monforte, Osp. L. Sacco, Milan. Luxembourg (R. Hemmer), T. Staub, Centre Hospitalier, Luxembourg. Netherlands (P. Reiss) Academisch Medisch Centrum bij de Universiteit van Amsterdam, Amsterdam. Norway (J. Bruun) A. Maeland, Ullevål Hospital, Oslo. Poland (B. Knysz) J. Gasiorowski, Medical University, Wroslaw; A. Horban, Centrum Diagnostyki i Terapii AIDS, Warsaw; R. Rogowska-Szadkowska, Medical University, Bialystok; A. Boron-Kaczmarska, Medical Univesity, Szczecin; M. Beniowski, Osrodek Diagnostyki i Terapii AIDS, Chorzow; H. Trocha, Medical University, Gdansk. Portugal (F. Antunes) Hospital Santa Maria, Lisbon; K. Mansinho, Hospital de Egas Moniz, Lisbon; R. Proenca, Hospital Curry Cabral, Lisbon. Spain (J. González-Lahoz) R. Polo, V. Soriano, Hospital Carlos III, Madrid; B. Clotet, A. Jou, J. Conejero, C. Tural, Hospital Germans Trias i Pujol, Badalona; J.M. Gatell, J.M. Miró, Hospital Clinic i Provincial, Barcelona. Sweden (A. Blaxhult) Karolinska Hospital; B. Heidemann, Södersjukhuset; P. Pehrson, Huddinge Sjukhus, Stockholm. Switzerland (B. Ledergerber) R. Weber, University Hospital, Zürich; P. Francioli, A. Telenti, Centre Hospitalier Universitaire Vaudois, Lausanne; B. Hirschel, V. Soravia-Dunand, Hospital Cantonal Universitaire de Geneve, Geneve. United Kingdom (S. Barton) St. Stephen's Clinic, Chelsea and Westminster Hospital, London; A.M. Johnson, D. Mercey, Royal Free and University College London Medical School, London (University College Campus); A. Phillips, C. Loveday, M.A. Johnson, A. Mocroft, Royal Free and University College Medical School, London (Royal Free Campus); A. Pinching, J. Parkin, Medical College of Saint Bartholomew's Hospital, London; J. Weber, G. Scullard, Imperial College School of Medicine at St. Mary's, London; M. Fisher, Royal Sussex County Hospital, Brighton; R. Brettle, City Hospital, Edinburgh.
Steering committee: J. Nielsen (chair), N. Clumeck, M. Dietrich, J.M. Gatell, A. Horban, A.M. Johnson, C. Katlama, B. Ledergerber, C. Loveday, A. Phillips, P. Reiss, S. Vella.
Coordinating centre staff: J. Lundgren (project leader), I. Gjørup, T. Benfield, O. Kirk, A. Mocroft, D. Mollerup, M. Nielsen, A. Sørensen, H. Buch, L. Madsen, L. Teglbjærg.