The availability of combination antiretroviral therapy (cART) has resulted in immune restoration in most treated HIV-infected patients and in a dramatic decrease in AIDS-related mortality.1-3 Provided that treatment is taken daily and there is regular follow-up, most cART-treated individuals have a social life that may include working, having children, or buying a house, because life expectancy has dramatically improved. Studies considering short-term or midterm follow-up after cART initiation have shown that mortality remained higher in HIV-infected individuals in the cART period than in the general population in France,4 Switzerland,5,6 and Denmark.7 Early favorable viroimmunologic response to treatment has proven to be associated with a longer survival, however, regardless of the initial CD4 cell count and plasma HIV RNA levels.8 We hypothesized that high values of CD4 cell counts attained might allow these patients to reach the same mortality rates as those of the general population in the long term. We thus compared mortality rates in HIV-infected adults after the first cART prescription with the mortality of the general population of the same age and gender according to the level of CD4 cell count reached and according to the duration of exposure to cART.
HIV-infected adults who started cART containing a protease inhibitor (PI) for the first time in 1997 to 1999 were identified from 2 well-established cohorts. Indeed, PI-containing regimens were the first available cART, and exposure to this category of cART is therefore the longest that can be observed among treated patients. Patients selected for this study may have been previously exposed to mono- or dual therapy with nucleoside reverse transcriptase inhibitors. The Agence Nationale de Recherches sur le Sida et les hépatites virales (ANRS) CO8 APROCO-COPILOTE cohort is a prospective observational study that consecutively enrolled 1281 HIV-1-infected adults in 47 hospital departments in France who were starting a PI-containing treatment for the first time in 1997 to 1999.9 Standardized clinical and biologic data were collected at baseline, after 1 and 4 months of cART, and every 4 months thereafter. Investigators were requested to notify the data management center of patient deaths as soon as they were known, and regular monitoring rounds were organized to monitor consistency between hospital files and case report forms. The ANRS CO3 AQUITAINE Cohort was implemented in 1987 by the Groupe d'Epidemiologie Clinique du Sida en Aquitaine (GECSA) based on a public hospital surveillance system of HIV-infected adults in the Aquitaine region of southwest France.10 Standardized clinical and biologic data are collected at each hospital contact and at least every 6 months. An active search of patients lost to follow-up is performed yearly. In both cohorts, patients signed informed consent forms.
Duplicate patient records between the 2 databases were excluded. We considered data on patients who remained alive up to December 31, 2005, or until loss to follow-up or death. Death rates were calculated per 100 person-years (PYs). Standardized mortality ratios (SMRs) were estimated with reference to the 2002 French general population death rates stratified for gender and for every 10 years of age,11 and the 95% confidence intervals (CIs) of the SMRs were estimated by the Byar approximation of the Poisson method.12
CD4 cell counts during follow-up were estimated using a mixed linear model to take into account unbalanced data attributable to missing at-random measurements13 and measurement error.14 Square root of CD4-positive values were fitted using a piecewise linear model allowing for a change of slope at 4 months15 and adjusted for baseline covariates: age, clinical AIDS stage, plasma HIV RNA level, history of antiretroviral treatment, and HIV transmission (injecting drug use vs. others). For each year of age during follow-up, 3 values of CD4 cell count were estimated (every 4 months), and the lowest value of the year was classified in the following categories: ≥500 cells/mm3, 350 to 499 cells/mm3, 200 to 349 cells/mm3, or <200 cells/mm3. Death rates and SMRs were estimated for the cumulated time period spent within each category of CD4 cell count.
To identify if and when during follow-up mortality rates reached values of the general population, we performed successive selections of patients with long-term follow-up. SMRs were computed successively with truncation at each year of follow-up for the 2 highest categories of CD4 cell count (≥500 cells/mm3 and 350 to 499 cells/mm3). For instance, for the analysis of time spent with a CD4 count ≥500 cells/mm3 and truncation at 6 years, a patient still followed 8 years after cART initiation may contribute to the analysis for the time spent with CD4 a cell count ≥500 cells/mm3 only after the sixth year of follow-up.
SMRs were also estimated according to HIV transmission group (injecting drug use vs. others) and hepatitis C virus (HCV) coinfection as defined by the presence of HCV antibody or plasma HCV RNA at baseline, because similar types of analyses reported higher mortality ratios in these groups.5,6
The underlying cause of death was ascertained with data available in the hospital file according to the International Classification of Diseases 10th revision (ICD-10) rules16 and adapted to the specificities in HIV infection.17
Statistical analyses were performed using Statistical Analysis System software (SAS, version 9.1; SAS Corporation, Cary, NC).
A total of 2435 patients (1281 from the APROCO-COPILOTE cohort and 1154 from the AQUITAINE cohort) were included in the analysis. The median patient age was 36 years; 77% were men; and HIV transmission categories were homosexual or bisexual in 38%, heterosexual in 35%, and injecting drug use in 21% of cases. Overall, 29% of patients were HCV infected (88% among patients infected through injecting drug use). The median CD4 count was 270 cells/mm3 at the time of cART initiation, 16% of patients had a CD4 cell count ≥500 cells/mm3, and 19% of patients had a CD4 cell count between 350 and 499 cells/mm3. At baseline, 22% had a previous AIDS-defining clinical event; 39% had previously received antiretroviral treatment with 1 or 2 drugs; and the first PI prescribed was indinavir in 43%, nelfinavir in 31%, saquinavir in 16%, and ritonavir in 15%.
Estimated CD4 counts were ≥500 cells/mm3 in 39% of the 1949 patients still followed 3 years after cART initiation and in 49% of the 1430 patients still followed at 6 years (Fig. 1). During a median follow-up of 6.8 years (interquartile range [IQR]: 4.1 to 7.9, 13,970 PYs), 288 individuals died, 2.1 deaths per 100 PYs (95% CI: 1.8 to 2.3). Overall mortality was 7.0 times higher than in the general population, 4.8 in men and 13.0 in women, 16.3 in injecting drug users, and 13.9 in HCV-coinfected patients (Table 1). Considering the total time spent within each category of CD4 cell count, mortality remained higher than in the general population in all categories and SMRs were gradually higher when CD4 cell counts were lower (Table 2). In patients with a CD4 count ≥500 cells/mm3, however, mortality reached the level of the general population after the sixth year after initiation of cART (SMR = 0.5, 95% CI: 0.1 to 1.6; Table 3; Fig. 2). Considering the time spent in the category of a CD4 count from 350 to 499 cells/mm3, the SMR was lower after 6 years but remained around twice the mortality of the general population (Table 4, Fig. 2). Overall, the underlying cause of death was AIDS related in 35% of cases, and in 52%, 21%, 15%, and 8% when the CD4 count at the age of death was <200 cells/mm3, 200 to 349 cells/mm3, 350 to 499 cells/mm3, and ≥500 cells/mm3, respectively.
In this study with a median follow-up of 7 years after cART initiation, age- and gender-adjusted overall mortality remained 7-fold higher in HIV-infected adults than in the general population. The mortality rate became similar to that of the general population after the sixth year of follow-up among patients whose CD4 counts had reached 500 cells/mm3, however.
Because we aimed at identifying if and when during follow-up mortality rates reached values of the general population, we selected patients who had the highest CD4 cell counts and were followed long term. Therefore, we cannot exclude a survivor bias, and we acknowledge that these results only apply to a specific subgroup. Nevertheless, identifying patients with the best prognosis, regardless of their history, may help to identify therapeutic objectives and formulate guidelines. That is the reason why we did not adjust for potential confounding factors such as previous antiretroviral exposure, baseline CD4 cell count and HIV RNA level, and type of treatment received, as we would have done in a classic prognosis study.
We acknowledge that our analysis may have some limitations. For this analysis, we considered that the 2 cohorts studied were similar enough to be pooled because they were located in the same country, where guidelines for the case management of HIV-infected patients are generally well known and followed by physicians.18 Moreover, epidemiologic and biostatistical support and data management are coordinated by the same epidemiology unit.
The level of CD4 cell count is known as a strong prognostic factor for the occurrence of AIDS-defining events and death. In our study, the proportion of non-AIDS-related causes of death increased with higher CD4 cell counts. Achieving the level of 500 cells/mm3 seemed to be associated with the same rates of mortality as among the general population after 6 years after cART, whereas it was still higher during a shorter time of follow-up. Among patients who spent time with a CD4 count between 350 and 499 cells/mm3, mortality remained higher than in the general population. Several interpretations of these observations are possible. First, the level of CD4 cell count is associated with mortality, even when the CD4 count is greater than 200 cells/mm3, because of the persistence of AIDS-defining or HIV-associated morbidities. For instance non-Hodgkin lymphoma remains a frequent cause of death among treated HIV-infected patients.17 Second, non-AIDS-defining morbidities such as bacterial infections or cancers may occur more frequently at an intermediate level of CD4 cell count.19 Nevertheless, even among patients with the highest category of CD4 cell count, the reason for the long period needed to reach the same mortality rates as in the general population remains to be clarified. We hypothesize that immune restoration after HIV infection may be a long-lasting process and that time is necessary to recover immune functions able to reduce mortality to the same level as in the general population. Another hypothesis may be that patients without severe comorbidities succeed more frequently in reaching a high CD4 cell count, and perhaps a longer time of follow-up.
Published studies have reported higher mortality in HIV-infected persons compared with the general population.4,6,7 Van Sighem et al20 found higher mortality in HIV-infected patients in The Netherlands even after having selected patients followed for at least 6 months and taking into account the CD4 cell count 6 months after cART initiation. In the Swiss cohort, Jaggy et al5 reported a moderate excess of death rates, compared with the general population, when the CD4 count reached 250 cells/mm3 at least once after cART.
Our analysis does not take into account some confounding factors that might at least partly explain differences in mortality rates. First, the risk of death from cardiovascular diseases or cancer might be related to the high proportion of smokers among HIV-infected individuals.21,22 Second, injecting drug users have a higher risk of death from overdose and violence.23 They are frequently coinfected with HCV, which exposes them to cirrhosis and hepatocarcinoma. In fact, mortality was higher among injecting drug users and HCV-infected patients in our analysis, in agreement with other studies.5,6,24 None of these characteristics was available in databases of the general population, nor were socioeconomic conditions or levels of education, which are associated with higher mortality in the general population25 as well as in treated HIV-infected patients.26 Confounding factors may thus explain the higher overall mortality compared with the general population and the higher SMR in women than in men, agreeing with a previous analysis in the APROCO-COPILOTE cohort4 and with other reports.6 The less favorable prognosis observed in HIV-infected women as compared with women in the general population could reflect a less favorable sociodemographic status of HIV-infected women and a higher frequency of comorbidities (29% of women were infected through injecting drug use and 30% were HCV infected). Other time-dependent markers of HIV progression (eg, HIV RNA level) are associated with mortality in the long term. Nevertheless, to our knowledge, the additional effect of HIV RNA level on mortality in patients with a high CD4 cell count has not been reported so far and would probably be weak.
Although patients who started PI-containing cART may not be representative of patients having started cART with more recent combinations, they can be considered as representative of the large number of patients, approximately 35,000 in France, who started cART in the years 1997 to 1999, who currently have the longest follow-up under cART. We excluded the year 1996 because it was the first year of cART being available in France and an intermediate period of implementation with heterogeneous practices. We can hypothesize that patients who started cART later than 1999 may have a better prognosis, because therapeutic strategies have improved,2 and that they may reach the same mortality rate more rapidly than the general population.
These results remain to be confirmed in other populations, and cohort collaborations may address this question with a larger sample size and a longer follow-up. Nevertheless, we believe that communicating these results to patients and physicians is already crucial to assist them in maintaining their efforts to achieve and sustain high CD4 cell counts through sustained adherence to cART. We acknowledge that our results are derived from a method using a selection of subgroups of patients, and thus may only be generalized to this specific population.
In countries in which a certificate of health status is required to obtain insurance contracts and loans, HIV infection with a favorable response to treatment in the long term might no longer be considered an obstacle, based on our observations. To improve prognosis in most HIV-infected patients, medical teams should evaluate all known factors associated with suboptimal response to treatment (ie, tolerance, adherence, social support, care of depression) to achieve the goal of sustained immune reconstitution. In addition to identifying factors that may hinder this objective,27 operational tools to improve complete therapeutic success should be developed and evaluated.
The authors thank the ANRS CO8 APROCO-COPILOTE Study Group (see Appendix). The authors also thank the ANRS CO3 AQUITAINE Cohort (see Appendix).
1. Mocroft A, Ledergerber B, Katlama C, et al. Decline in the AIDS and death rates in the EuroSIDA study: an observational study. Lancet
2. Collaboration CASCADE. Determinants of survival following HIV-1 seroconversion after the introduction of HAART. Lancet
3. Palella FJ Jr, Baker RK, Moorman AC, et al. Mortality
in the highly active antiretroviral therapy
era: changing causes of death and disease in the HIV Outpatient Study. J Acquir Immune Defic Syndr
4. Lewden C, Raffi F, Chêne G, et al. Mortality
in a cohort of HIV-infected adults started on a protease inhibitor-containing therapy-standardization to the general population. J Acquir Immune Defic Syndr
5. Jaggy C, von Overbeck J, Ledergerber B, et al. Mortality
in the Swiss HIV Cohort Study (SHCS) and the Swiss general population. Lancet
6. Keiser O, Taffe P, Zwahlen M, et al. All cause mortality
in the Swiss HIV Cohort Study from 1990 to 2001 in comparison with the Swiss population. AIDS
7. Jensen-Fangel S, Pedersen L, Pedersen C, et al. Low mortality
in HIV-infected patients starting highly active antiretroviral therapy
: a comparison with the general population. AIDS
8. Chêne G, Sterne JA, May M, et al. Prognostic importance of initial response in HIV-1 infected patients starting potent antiretroviral therapy
: analysis of prospective studies. Lancet
9. Le Moing V, Thiébaut R, Chêne G, et al. Predictors of long-term increase in CD4(+) cell counts in human immunodeficiency virus-infected patients receiving a protease inhibitor-containing antiretroviral regimen. J Infect Dis
10. Chêne G, Binquet C, Moreau JF, et al. Changes in CD4+ cell count and the risk of opportunistic infection or death after highly active
antiretroviral treatment. Groupe d'Epidemiologie Clinique du SIDA en Aquitaine. AIDS
12. Breslow N, Day N. Statistical Methods in Cancer Research II. The Design and Analysis of Cohort Studies
. Lyon, France: World Health Organization-International Agency for Research on Cancer; 1987.
13. Laird N, Ware J. Random-effects models for longitudinal data. Biometrics
14. Dafni U, Tsiatis A. Evaluating surrogate markers of clinical outcome when measured with error. Biometrics
15. Thiébaut R, Chêne G, Jacqmin-Gadda H, et al. Time-updated CD4+
T-lymphocyte count and HIV RNA as major markers of disease progression in naive HIV-1-infected patients treated with a highly active antiretroviral therapy
: the Aquitaine Cohort, 1996-2001. J Acquir Immune Defic Syndr
16. World Health Organization. International Classification of Diseases
, 10th revision. Geneva, Switzerland: World Health Organization; 1993.
17. Lewden C, Salmon D, Morlat P, et al. Causes of death among HIV-infected adults in the era of potent antiretroviral therapy
: emerging role of hepatitis and cancers, persistent role of AIDS. Int J Epidemiol
18. Dormont J. Stratégies d'utilisation des antirétroviraux dans l'infection par le VIH. Rapport 1998
. Médecine sciences ed. Paris, France: Flammarion; 1998.
19. Clifford GM, Polesel J, Rickenbach M, et al. Cancer risk in the Swiss HIV Cohort Study: associations with immunodeficiency, smoking, and highly active antiretroviral therapy
. J Natl Cancer Inst
20. van Sighem A, Danner S, Ghani AC, et al. Mortality
in patients with successful initial response to highly active antiretroviral therapy
is still higher than in non-HIV-infected individuals. J Acquir Immune Defic Syndr
21. Bénard A, Tessier J-F, Rambeloarisoa J, et al. HIV infection
and tobacco smoking behaviour: prospects for prevention? ANRS CO3 Aquitaine Cohort, 2002. Int J Tuberc Lung Dis
22. Friis-Møller N, Sabin CA, Weber R, et al. Combination antiretroviral therapy
and the risk of myocardial infarction. N Engl J Med
23. Frischer M, Bloor M, Goldberg D, et al. Mortality
among injecting drug users: a critical reappraisal. J Epidemiol Community Health
24. Wang C, Vlahov D, Galai N, et al. Mortality
in HIV-seropositive versus - seronegative persons in the era of highly active antiretroviral therapy
: implications for when to initiate therapy. J Infect Dis
25. Jougla E, Rican S, Péquignot F, et al. La mortalité. In: Leclerc A, Fassin D, Grandjean H, et al, eds. Les inégalités sociales de santé
. Paris, France: Editions La découverte et Syros; 2000:147-162.
26. Lewden C, Raffi F, Cuzin L, et al. Factors associated with mortality
in human immunodeficiency virus type 1-infected adults initiating protease inhibitor-containing therapy: role of education level and of early transaminase level elevation (APROCO-ANRS EP11 study). J Infect Dis
27. Kaufmann GR, Furrer H, Ledergerber B, et al. Characteristics, determinants, and clinical relevance of CD4 T cell recovery to <500 cells/μL in HIV type 1-infected individuals receiving potent antiretroviral therapy
. Clin Infect Dis
ANRS CO8 APROCO-COPILOTE Study Group
Principal Investigators: C. Leport, F. Raffi
Epidemiology: G. Chêne, R. Salamon
Social Sciences: J-P. Moatti, J. Pierret, B. Spire
Virology: F. Brun-Vézinet, H. Fleury, B. Masquelier
Pharmacology: G. Peytavin, R. Garraffo
Other members: D. Costagliola, P. Dellamonica, C. Katlama, L. Meyer, M. Morin, D. Salmon, A. Sobel
Events Validation Committee: L. Cuzin, M. Dupon, X. Duval, V. Le Moing, B. Marchou, T. May, P. Morlat, C. Rabaud, A. Waldner-Combernoux
Project coordination: F. Collin
Observers: P. Bursacchi, JF. Delfraissy, J. Dormont, M. Garré
Clinical Research Group: V. Le Moing, C. Lewden
Clinical centers (coordinators): Amiens (Pr. J. L. Schmit), Angers (Dr. J. M. Chennebault), Belfort (Dr. J. P. Faller), Besançon (Pr. J. L. Dupond, Dr. J. M. Estavoyer, Pr. P. Humbert), Bobigny (Pr. A. Krivitzky), Bordeaux (Pr. M. Dupon, Pr. Longy-Boursier, Pr. P. Morlat, Pr. J. M. Ragnaud), Bourg-en-Bresse (Dr. P. Granier), Brest (Pr. M. Garré), Caen (Pr. R. Verdon), Compiègne (Dr. Y. Domart), Corbeil Essonnes (Dr. A. Devidas), Créteil (Pr. A. Sobel), Dijon (Pr. H. Portier), Garches (Pr. C. Perronne), Lagny (Dr. P. Lagarde), Libourne (Dr. J. Ceccaldi), Lyon (Pr. D. Peyramond), Meaux (Dr. C. Allard), Montpellier (Pr. J. Reynes), Nancy (Pr. T. May), Nantes (Pr. F. Raffi), Nice (Pr. J.P. Cassuto, Pr. P. Dellamonica), Orléans (Dr. P. Arsac), Paris (Pr. E. Bouvet, Pr. F. Bricaire, Pr. P. Bergmann, Pr. J. Cabane, Dr. G. Cessot, Pr. P. M. Girard, Pr. L. Guillevin, Pr. S. Herson, Pr. C. Leport, Pr. M. C. Meyohas, Pr. J. M. Molina, Pr. G. Pialoux, Pr. D. Salmon), Poitiers (Pr. B. Becq-Giraudon), Reims (Pr. R. Jaussaud), Rennes (Pr. C. Michelet), Saint-Etienne (Pr. F. Lucht), Saint-Mandé (Pr. T. Debord), Strasbourg (Pr. J. M. Lang), Toulon (Dr. J. P. De Jaureguiberry), Toulouse (Pr. B. Marchou), Tours (Pr. J. M. Besnier)
Data monitoring and statistical analysis: C. Alfaro, F. Alkaied, S. Boucherit, A. D. Bouhnik, C. Brunet-François, M.P. Carrieri, M. Courcoul, F. Couturier, J. L. Ecobichon, M. François, L. Iordache, V. Journot, P. Kurkdji, J. P. Legrand, E. Lootvoet, E. Pereira, M. Préau, C. Protopopescu, J. Surzyn, A. Taieb, F. Tourteau, V. Villes, H. Zouari
ANRS CO3 AQUITAINE Cohort
Epidemiology: G. Chêne, F. Dabis, C. Lewden, S. Lawson-Ayayi, R. Thiébaut, M. Winnock
Infectious Diseases-Internal Medicine: M. Dupon, P. Mercié, J. F. Moreau, P. Morlat, J. L. Pellegrin, J. M. Ragnaud, D. Neau, N. Bernard, D. Lacoste, D. Malvy
Immunology: J.-F. Moreau, P. Blanco
Virology: H. Fleury, M. E. Lafon, B. Masquelier, I. Pellegrin
Pharmacovigilance: G. Miremont
Clinical Pharmacology: D. Breilh
Monitoring, data management, and statistical analysis: E. Balestre, M.J. Blaizeau, M. Decoin, S. Delveaux, L. Dequae-Merchadou, D. Dutoit, S. Geffard, C. Hannapier, L. Houinou, S. Labarrère, V. Lavignolle-Aurillac, G. Palmer, D. Touchard, B. Uwamaliya-Nziyumvira
Clinical centers (participating physicians)
Bordeaux University Hospital: P. Morlat (N. Bernard, M. Bonarek, F. Bonnet, K. Lacombe, P. Gellie, D. Lacoste, F. Paccalin, M. C. Pertusa), M. Dupon (H. Dutronc, F. Dauchy, S. Lafarie), M. Longy-Boursier (P. Mercié, A. Aslan, D. Malvy, T. Pistonne, M.-C. Receveur, P. Thibaut), J. M. Ragnaud (D. Neau, C. Cazanave, D. Chambon, C. De La Taille, A. Ochoa), J. L. Pellegrin (J.F. Viallard, O. Caubet, C. Nouts), P. Couzigou
Dax Hospital: P. Loste (L. Caunègre)
Bayonne Hospital: F. Bonnal (S. Farbos, M.C. Gemain).
Libourne Hospital: J. Ceccaldi (S. Tchamgoué).
Mont de Marsan Hospital: S. De Witte