Background: The efficacy of antiretroviral therapy (ART) has been established through clinical trials (CTs). However, selection bias and differences can limit their applicability to the general population.
Methods: All treatment-naive HIV-infected patients who began ART in routine care (RC) between 2000 and 2008 were compared with all patients who initiated ART through a CT in terms of incidence of virological failure (VF), increase in CD4+ count, mortality rate, and loss to follow-up (LTFU).
Results: At baseline, the RC group had less years of education, higher unemployment rate, higher proportion of females (14.2 vs. 5.7%; P < 0.01), lower median CD4+ (97 vs. 158 cells/μL; P < 0.01), and lower proportion of patients with hemoglobin >12 g/dL (74 vs. 83%, P = 0.04). VF at week 48 was less frequent in the CT compared with the RC group (1.8% vs. 6.21%, P = 0.02). In multivariate analysis, participation in CT [odds ratio (OR): 0.20, 95% confidence interval (CI): 0.04 to 0.91, P = 0.03], hemoglobin >12 g/dL (OR: 0.29, 95% CI 0.09-0.89, P = 0.03), and receiving an optimal highly active antiretroviral therapy regimen (OR: 0.09, 95% CI: 0.01 to 0.52, P < 0.01) remained associated with lower risk of VF. All cause mortality was 0.017 (95% CI: 0.002 to 0.122) versus 0.094 (95% CI: 0.053 to 0.17) deaths per 1000 person-days in the CT group and in the RC group, respectively (P = 0.05). No differences were found in the proportion of patients LTFU.
Conclusions: Receiving ART through CT was associated with lower probability of VF, lower mortality (probably related to less severe clinical characteristics at baseline), and similar rates of LTFU than RC.
*Department of Infectious Diseases, Instituto Nacional de Ciencias Médicas y Nutrición Salvador Zubirán, Mexico City, Mexico
†Vanderbilt University School of Medicine, Nashville, TN.
Correspondence to: Juan Sierra-Madero, MD, Department of Infectious Diseases, Instituto Nacional de Ciencias Médicas y Nutrición Salvador Zubirán, Vasco de Quiroga 15, Col. Sección XVI, Tlalpan 14000, México, DF, Mexico (e-mail: firstname.lastname@example.org).
Presented at XVIII International AIDS Conference, July 18–23, 2009. Vienna, Austria.
Dr Crabtree-Ramírez has received grants from BMS and is speaker for Merck; Dr Sierra-Madero has received grants from Pfizer, is speaker for ViiV Healthcare, was consultant for ViiV Healthcare, and is currently consultant for MSD.
All others have no funding or other conflicts of interest to disclose.
Received July 14, 2011
Accepted November 1, 2011
Highly active antiretroviral therapy (HAART) has dramatically increased the life expectancy of individuals infected with HIV.1 The efficacy of this treatment to improve survival and control viral replication has been proven in multiple clinical trials (CTs), which has allowed HAART to become the current gold standard in the treatment of HIV infection.2–5 Additionally multiple cohort studies have corroborated the effectiveness of antiretroviral (ARV) treatment.6
However, because of potential selection biases that exist in CTs, there is concern whether the results of these trials can be paralleled in the general nontrial population. Some examples of these biases are the low participation of groups considered minorities such as Latinos in studies performed in the United States7,8 and female gender in the studies conducted in Latin America.9–12 Additionally, most trials evaluating the efficacy of different regimens in ARV-naive patients have excluded patients with advanced disease (CD4+ T-cell count <50 cells/μL or an active opportunistic infection).13,14 This issue becomes even more important in the context of low-income and middle-income countries, where a greater proportion of patients initiating HAART do so in more advanced stages of HIV infection.15,16 The purpose of this study was to determine if there are differences in the proportion of patients experiencing virological failure (VF) in the first year after initiating HAART between those participating in CTs and those receiving routine care (RC). Additionally, the changes in CD4+ cell count, mortality rate, and loss to follow-up (LTFU) were evaluated in the 2 groups.
Study Design and Population
Retrospective study in the cohort of patients with HIV infection receiving medical care at the HIV/AIDS Clinic of the Instituto Nacional de Ciencias Médicas y Nutrición Salvador Zubirán (INCMNSZ) in Mexico City, Mexico. The study population consists of all patients older than 18 years with HIV infection and no previous antiretroviral therapy who initiated treatment in the HIV/AIDS Clinic of INCMNSZ between 2000 and 2008. The population was divided into 2 groups: those who initiated HAART through CTs conducted at the HIV/AIDS Clinic and those patients who initiated HAART through RC in the same center.
The database of the HIV/AIDS Clinic records prospective data obtained through clinical questionnaires (including sociodemographic data) performed at each medical visit. Laboratory information obtained includes CD4+ cell count, HIV RNA viral load (VL), and hematology. Hospitalization, mortality, and lost to follow-up are actively collected in the database. Regular internal and external audits are performed to ensure the quality of the data.
Definition of Variables and Outcomes
The baseline levels of CD4+ cell count and VL were defined as the values determined closest to the date of HAART initiation but no more than 180 days before or 7 days after this date. The baseline weight, hemoglobin (Hb), and clinical stage were defined as the values determined closest to the date of HAART initiation but no more than 30 days after this date. HAART treatment was defined as optimal if it included 2 nucleoside reverse transcriptase inhibitors and a protease inhibitor (PI) boosted with ritonavir, a nonnucleoside reverse transcriptase inhibitor, an integrase inhibitor, or a CCR5 inhibitor. Suboptimal treatment was defined as one that contained 3 or 4 nucleoside reverse transcriptase inhibitor or regimens based on a nonboosted PI. Finally, the number of visits was determined for each patient in each group. The definition of AIDS was made based on the classification of the Centers for Disease Control and Prevention.17 The main outcome of interest was VF, which was defined as the measurement of plasma HIV-1 RNA greater than 400 copies/mL at week 48 after the initiation of HAART (confirmed with a second HIV-VL determination or a single HIV-VL > 400 copies/mL plus an entry in the patient's chart of change in ARV therapy secondary to VF). Other outcomes measured were mortality and LTFU. The change in CD4+ cell count between baseline and week 48 was calculated in both groups. For the RC group, VL and CD4+ cell count measurements made between weeks 40 and 60 were used to represent week 48 values. A patient was considered lost to follow-up if no clinic visits occurred for more than 1 year after initiating HAART. Patients who discontinued follow-up during the first year of treatment were retrospectively assigned as lost to follow-up if the last visit occurred during the first year of HAART initiation and the period between that visit and the next one was more than 365 days or did never return to the clinic. Deaths from any cause occurring during the first year after HAART initiation were recorded.
The χ2 test or Fisher exact test was used to compare categorical variables, and the student t test or the median was used to compare continuous variables. Kaplan–Meier survival analysis was used to estimate the time between HAART initiation and VF during the first year of treatment. Mortality rates between groups were compared using Cox proportional hazards model. Univariate and multivariate logistic regression models were used to determine risk factors associated with VF and death due to any cause. The statistical analysis was conducted in Stata v10 using a 0.05 level of significance.
A total of 513 naive patients were admitted to the HIV/AIDS Clinic of INCMNSZ between 2000 and 2008 of which 159 were in the CT group and 354 in the RC group. During the period of study, the following CTs for ARV-naive subjects were conducted: Alternative Antiretroviral Strategies: a Comparison of Three Initial Regimens (ALTAIR, n = 14), D4T ER versus D4T IR (n = 26); Atazanavir/ritonavir vs Lopinavir/ritonavir, 48 and 96 weeks analysis (n = 9); the Atazanavir/ritonavir vs Nevirapine and Tenofovir/Emtricitabine study (ARTEN, n = 2); DMP 450 in combination with d4t and 3TC (DMP-102 study, n = 9); Emtricitabine vs Stavudine in combination therapy (FTC-301, n = 18); Efavirenz versus Lopinavir/ritonavir in naive subjects (n = 61); Raltegravir vs Efavirenz (n = 8); Maraviroc vs Efavirenz for initial treatment HIV-naive subjects (MERIT, n = 8); Darunavir/ritonavir vs lopinavir/ritonavir (n = 4). Only 1 of these studies had as inclusion criteria the presence of advanced disease (with CD4+ count <200 cells/μL at baseline).18–28
The baseline characteristics of the patients are described in Table 1. The CT group had a smaller proportion of women (5.7 vs. 14.2%, P < 0.01) and a trend toward a younger population (32 vs. 35 years, P = 0.06) than the RC group. The proportion of employed patients at entrance was higher in the CT group (74.3 vs. 58.7, P < 0.01). With respect to clinical and laboratory data, the CT patients had a higher proportion of patients with Hb concentration >12 g/dL (83.2 vs. 74.3, P = 0.04) than the RC patients. The overall proportion of patients with AIDS (AIDS-defining events and/or CD4+ < 200 cells/μL) was high (72%). Although not statistically significant, the CT group had a smaller proportion of patients with AIDS at baseline than the RC group (68% vs. 74%, P = 0.14). The CT group also had a lower proportion of patients with baseline VL >75,000 copies/mL (59% vs. 81%, P < 0.01) and a higher median baseline CD4+ cell count (158 vs. 97 cells/μL, P < 0.01) compared with the RC group. On average, the patients of the CT group had a higher mean number of clinic visits 5.56 (SD ± 2.38) versus 4.51 (SD ± 2.24) during the first year of treatment.
The number of patients with VF at week 48 was 3 (1.8%) in the CT group and 21 (5.9%) in the RC group, P = 0.04 (Table 2), with a slight difference in the average time to VF [CT 360 days (±30) vs. RC 351 days (±50), P = 0.05; Fig. 1]. In the univariate analysis, baseline Hb concentration >12 g/dL (OR: 0.58, 95% CI: 0.35 to 0.97, P = 0.04) and receiving optimal HAART (OR: 0.23, 95% CI: 0.08 to 0.68, P < 0.01) were associated with lower risk of VF. CT participation was marginally associated with lower risk of VF (OR: 0.30, 95% CI: 0.09 to 1.03, P = 0.06). However, by multivariate analysis (controlling for gender, age, route of transmission, AIDS-defining events on admission, participation in a CT, optimal HAART, year of HAART initiation, baseline Hb, baseline CD4+ count, baseline VL greater 75,000 copies/mL, employment, and years of education at entrance to care), only participation in CT (OR: 0.20, 95% CI: 0.04 to 0.91, P = 0.04), baseline Hb >12 g/dL (OR: 0.29, 95% CI: 0.09 to 0.89, P = 0.03), and receiving optimal HAART (OR: 0.09, 95% CI: 0.01 to 0.52, P < 0.01) remained associated with lower risk for VF (Table 3).
CD4+ Changes From Baseline
The increase in CD4+ count at week 48 for the CT group was 186 cells per microliter [interquartile range (IQR): 119–279] versus 163 cells per microliter (IQR: 90–246) for the RC group, which was statistically significant (P = 0.04).
Loss to Follow-Up
In the first year after treatment initiation, there were 20 (12%) patients lost to follow-up in the CT group and 61 (17%) patients in RC group, with no statistical difference (P = 0.18).
During the first year after HAART initiation, all cause mortality was the following: 1 death in the CT group and 13 deaths in the RC group (Table 2) with a rate of 0.017 (95% CI: 0.002 to 0.122) versus 0.094 (95% CI: 0.053 to 0.17) deaths per 1000 person-days, respectively (P = 0.05). In a Cox model, the hazard ratio for mortality comparing CT versus RC was of 0.169 (95% CI: 0.02 to 1.293, P = 0.08). In the multivariate analysis (Table 4), including age at HAART initiation, gender, years of education, employment at entrance to care, men who have sex with men, presence of AIDS-defining event, participation in a CT, year of HAART initiation, change of antiretroviral therapy treatment, and VF, only Hb >12 g/dL (OR: 0.16, 95% CI: 0.03 to 0.84, P = 0.03) was significantly associated with a lower death rate during the first year of treatment.
Strict study inclusion criteria used in CTs to evaluate different HAART regimens and less acceptability to participate in trials by more vulnerable groups may influence the applicability of the results to the general population in a given region. We undertook this analysis to compare the characteristics of patients receiving first HAART regimens through CTs with those receiving it through RC.
Our results show important differences at baseline in those participating in trials compared with those not participating. At baseline, patients in trials had higher CD4+ count values, higher proportion of patients with Hb >12 g/dL, lower proportion of employment, and higher proportion of VL >75,000 copies/mL. In addition, a significantly lower proportion of female subjects were included in trials compared with RC. Including less sick patients in trials is most likely a reflection of strict study inclusion and exclusion criteria, which gives preference to patients who the investigator considers less likely to interrupt the study because of complications related to AIDS.29–33 In fact most of the CTs conducted in our center during the study period (except for the efavirenz vs. lopinavir/ritonavir study in advanced patients) excluded those with very low CD4+ counts or AIDS-defining events.18–28 However, it turns out that late presentation to care is highly prevalent in our region as was recently described,15 which is clearly reflected in the characteristics of the RC group. Another significant finding was the extremely low frequency of female subjects inclusion into trials. An explanation for this is strictly speculative and deserves further detailed exploration. A perception of higher vulnerability, less social support, and a lower acceptance rate by females to volunteer due to family commitments may be some explanations for this finding.7–9
Regarding the outcomes measured in this analysis, the proportion of patients developing VF was 3 times higher in the RC group than the CT group. This finding could be explained at least in part, by the larger proportion of patients in the high viral load stratum and more severe markers of advanced clinical stage of HIV infection in the RC group, which have been associated with lower virological responses in different studies34–41 even though this was not supported by our multivariate analysis. Additionally a bias toward enrolling patients who are more reliable and stable psychosocially for consistent follow-up and better adherence to treatment may have also contributed to the better outcomes in the CT group.
Smith et al13 evaluated patients from a cohort in the United Kingdom with findings similar to ours in terms of improved virological outcomes in patients participating in CTs (94 vs. 86%, OR: 2.71, P = 0.007), other studies have also found the same results.42 They also found that the CT group had more follow-up visits during the first year than the RC group [7 vs. 6 (IQR: 5–9), P < 0.0001]. This issue may reflect that patients in CT have better disposition to attend the clinic visits. Moreover, the study staff made efforts to keep an active follow-up of the subjects, rescheduling missed visits, and taking immediate action to identify and resolve unfavorable events which may not always happen in RC. In contrast to our results, in similarly designed studies, other authors have found a higher rate of LTFU in RC43 and no differences in virological outcomes.14
The increased mortality seen in our analysis for the RC group compared with the CT group occurred mostly during the first year. This is similar to the early mortality described in cohort studies of patients starting HAART in limited resource settings.4,44,45
The results of this study clearly point out that patients starting HAART who participate in CTs in Mexico are different in many aspects from those who do not. Extrapolating CT results to general population may thus be appropriate for certain outcomes and populations, but not for others. Underrepresentation of females and late presenters is characteristic of trial population in our country and should be considered when translating results of trials into clinical practice in the region.
Limitations of our study are its retrospective nature and the inclusion of data only from a single center. A possible bias generated by conceivably a better capture of mortality and VF in the CT group would rather underestimate the differences between both groups observed in the study and may account actually for some lessening of the differences. Another factor that may have lessen the differences observed in the outcomes of the study is the fact that subjects in the RC group received more frequently simpler nonnucleoside reverse transcriptase inhibitor–based regimens compared with higher pill burden, more complex boosted-PI–based regimens in the CT group (Table 1). Finally, we recognize that an analysis of the CT group that included patients who were rejected from trial participation because of screening failures and investigator decision may have reflected a more realistic view of this group; however, published results of CTs only analyze the outcomes and baseline characteristics of those randomized subjects.
In our region, patients initiating ARV therapy in CTs have a lower probability of viral failure during the first year and a lower mortality rate than patients in RC, which probably reflects a more advanced clinical stage at baseline, even though this finding was not confirmed in the multivariate model. More studies including other centers in the region may help understand these findings in a wider perspective.
1. Palella FJ Jr, Delaney KM, Moorman A, 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.
2. Lohse N, Hansen AB, Pedersen G, et al.. Survival of persons with and without HIV infection in Denmark, 1995–2005. Ann Intern Med. 2007;146:87–95.
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. 2006;43:27–34.
4. Mocroft A, Ledergerber B, Katlama C, et al.. Decline in the AIDS and death rates in the EuroSIDA study: an observational study. Lancet. 2003;362:22–29.
5. Murphy EL, Collier AC, Kalish LA, et al.. Highly active antiretroviral therapy decreases mortality and morbidity in patients with advanced HIV disease. Ann Intern Med. 2001;135:17–26.
6. Brainstein P, Brinkhof MW, Dabis F, et al.. Mortality of HIV-1 infected patients in the first year of antiretroviral therapy: comparison between low- income and high-income countries. Lancet. 2006;367:817–824.
7. Sullivan PS, McNaghten AD, Begley E, et al.. Enrollment of racial/ethnic minorities and women with HIV in clinical research studies of HIV medicines. J Natl Med Assoc. 2007;99:242–250.
8. Gifford AL, Cunningham WE, Heslin KC, et al.. Participation in research and access to experimental treatments by HIV-infected patients. N Engl J Med. 2002;346:1373–1382.
9. Braitstein P, Boulle A, Nash D, et al.. Gender and the use of antiretroviral treatment in resource-constrained settings: findings from a multicenter collaboration. J Womens Health (Larchmt). 2008;17(1):47–55.
10. Tuboi SH, Brinkhof MW, Egger M, et al.. Discordant responses to potent antiretroviral treatment in previously naive HIV-1-infected adults initiating treatment in resource-constrained countries: the antiretroviral therapy in low-income countries (ART-LINC) collaboration. J Acquir Immune Defic Syndr. 2007;45:52–59.
11. Diaz T, Chu SY, Sorvillo F, et al.. Differences in participation in experimental drug trials among persons with AIDS. J Acquir Immune Defic Syndr Hum Retrovirol. 1995;10:562–568.
12. Madge S, Mocroft A, Wilson D, et al.. Participation in clinical studies among patients infected with HIV-1 in a single treatment centre over 12 years. HIV Med. 2000;1:212–218.
13. Smith CJ, Sabin CA, Hohnson MA, et al.. Participation in clinical trials at the Royal Free Hospital: characteristics of those included and impact on treatment outcomes [abstract # 618]. Presented at: 12th Conference on Retroviruses and Opportunistic infections; February 22–25, 2005; Boston, MA.
14. Routman JS, Willig JH, Westfall AO, et al.. Comparative efficacy versus effectiveness of initial antiretroviral therapy in clinical trials versus routine care. Clin Infect Dis. 2010;50:574–584.
15. Crabtree-Ramírez B, Caro-Vega Y, Shepherd BE, et al.. Cross-sectional analysis of late HAART initiation in Latin America and the Caribbean: late testers and late presenters. PLoS One. 2011;6:e20272.
17. 1993 Revised classification system for HIV infection and expanded surveillance case definition for AIDS among adolescents and adults. MMWR Recomm Rep. 1992;41. (No. RR-13). Available at: http://www.cdc.gov/mmwr/preview/ind1992_rr.html.
Accessed December 8, 2011.
18. Puls RL, Srasuebkul P, Petoumenos K, et al.. Efavirenz versus boosted atazanavir or zidovudine and abacavir in antiretroviral treatment-naive, HIV-infected subjects: week 48 data from the Altair study. Clin Infect Dis. 2010;51:855–864.
19. Pollard R, Ive P, Farthing C, et al.. Stavudine XR vs stavudine IR as part of potent antiretroviral combination therapy: 24-week safety and antiviral efficacy. Presented at: 9th Conference on Retroviruses and Opportunistic Infections; February 24–28, 2002; Seattle, WA [abstract # 411-W].
20. Molina JM, Andrade-Villanueva J, Echevarria J, et al.. Once-daily atazanavir/ritonavir compared with twice-daily lopinavir /ritonavir, each in combination with tenofovir and emtricitabine, for management of antiretroviral-naive HIV-1-infected patients: 96-week efficacy and safety results of the CASTLE study. J Acquir Immune Defic Syndr. 2010;53:323–332.
21. Molina JM, Andrade-Villanueva J, Echevarria J, et al.. Once-daily atazanavir/ritonavir versus twice-daily lopinavir/ritonavir, each in combination with tenofovir and emtricitabine, for management of antiretroviral-naive HIV-1-infected patients: 48 week efficacy and safety results of the CASTLE study. Lancet. 2008;372:646–655.
22. Soriano V, Köppe S, Mingrone H, et al.. Prospective comparison of nevirapine and atazanavir/ritonavir both combined with tenofovir DF/emtricitabine in treatment-naïve HIV-1 infected patients: ARTEN study week 48 results. Presented at: Program and abstracts of the 5th International AIDS Society Conference on HIV Pathogenesis, Treatment and Prevention; 2009; Cape Town, South Africa [Abstract LBPEB07].
23. Sierra-Madero J. Antiviral activity, safety and pharmacokinetics of Mozenavir (DMP 450), a novel cyclic urea protease inhibitor, in combinations with d4t and 3TC in treatment-naïve HIV-1 infected patients (Study DMP-102). Presented at: The 1st IAS Conference on HIV Pathogenesis and Treatment; 2001; Atlanta, GA [abstract no 2].
24. Saag MS, Cahn P, Raffi F, et al.. Efficacy and safety of emtricitabine vs stavudine in combination therapy in antiretroviral-naive patients: a randomized trial. JAMA. 2004;292:180–189.
25. Sierra-Madero J, Villasis-Keever A, Méndez P, et al.. Prospective, randomized, open label trial of efavirenz vs lopinavir/ritonavir in HIV+ treatment-naive subjects with CD4+<200 cell/mm3 in Mexico. J Acquir Immune Defic Syndr. 2010;53:582–588.
26. Lennox JL, DeJesus E, Lazzarin A, et al.. Safety and efficacy of raltegravir-based versus efavirenz-based combination therapy in treatment-naive patients with HIV-1 infection: a multicentre, double-blind randomized controlled trial. Lancet. 2009;374:796–806.
27. Sierra-Madero J, Di Perri G, Wood R, et al.. Efficacy and safety of maraviroc versus efavirenz, both with zidovudine/lamivudine: 96-week results from the MERIT study. HIV Clin Trials. 2010;11:125–132.
28. Ortiz R, Dejesus E, Khanlou H, et al.. Efficacy and safety of once-daily darunavir/ritonavir versus lopinavir/ritonavir in treatment-naive HIV-1-infected patients at week 48. AIDS. 2008;22:1389–1397.
29. Bartlett JA, Chen SS, Quinn JB. Comparative efficacy of nucleoside/nucleotide reverse transcriptase inhibitors in combination with efavirenz: results of a systematic overview. HIV Clin Trials. 2007;8:221–226.
30. Staszewski S, Morales-Ramirez J, Tashima KT, et al.. Efavirenz plus zidovudine and lamivudine, efavirenz plus indinavir, and indinavir plus zidovudine and lamivudine in the treatment of HIV-1 infection in adults. N Engl J Med. 1999;341:1865–1873.
31. Moyle G, DeJesus E, Cahn P, et al.. Abacavir once or twice daily combined with once-daily lamivudine and efavirenz for the treatment of antiretroviral naïve HIV-infected adults. J Acquir Immune Defic Syndr. 2005;38:417–425.
32. DeJesus E, Herrera G, Tenfilo E, et al.. Abacavir versus zidovudine plus lamivudine and efavirenz for the treatment of antiretroviral-naïve HIV-infected adults. Clin Infect Dis. 2004;39:1038–1046.
33. DeJesus E, McCarty D, Farthing CF, et al.. Once-daily versus twice-daily lamivudine, in combination with zidovudine and efavirenz, for the treatment of antiretroviral-naïve adults with HIV infection: a randomized equivalence trial. Clin Infect Dis. 2004;39:411–418.
34. Lucas GM, Chaisson RE, Moore RD. Highly active antiretroviral therapy in a large urban clinic: risk factors for virologic failure and adverse drug reactions. Ann Intern Med. 1999;131:81–87.
35. Paris D, Ledergerber B, Weber R, et al.. Incidence and predictors of virologic failure of antiretroviral triple-drug therapy in a community based cohort. AIDS Res Hum Retroviruses. 1999;15:1631–1638.
36. Tuboi SH, Harrison LH, Sprinz E, et al.. Predictors of virologic failure in HIV-1-infected patients starting highly active antiretroviral herapy in Porto Alegre, Brazil. J Acquir ImmuneDefic Syndr. 2005;40:324–328.
37. Fisher M. Late diagnosis of HIV infection: major consequences and missed opportunities. Curr Opin Infect Dis. 2008;21:1–3.
38. Kigozi IM, Dobkin LM, Martin JN, et al.. Late disease stage at presentation to an HIV clinic in the era of free antiretroviral therapy in sub-saharan Africa. J Acquir Immune Defic Syndr. 2009;52:280–289.
39. Sterne JA, Hernán MA, Lederberger B, et al.. Long-term effectiveness of potent antiretroviral therapy in preventing AIDS and death: a prospective cohort study. Lancet. 2005;366:378–384.
40. Hogg RS, Yip B, Chan KJ, et al.. Rates of disease progression by baseline CD4 cell count and viral load after initiating triple-drug therapy. JAMA. 2001;286:2568–2577.
41. Philips AN, Staszewski S, Weber R, et al.. HIV viral load response to antiretroviral therapy according to the baseline CD4 cell count and viral load. JAMA. 2001;286:2560–2567.
42. Eg Hansen A, Gerstoft J, Kirk O, et al.. Unmeasured confounding casued slightly better response to HAART within than outside a randomized controlled trial. J Clin Epidemiol. 2008;61:87–94.
43. Philips A, Grabar S, Tassie J, et al.. Use of observational databases to evauate the effectiveness of antiretroviral therapy for HIV infection: comparison of cohort studies with randomized trials. AIDS. 1999;13:2075–2082.
44. Egger M, May M, Chene G, et al.. Prognosis of HIV-1 infected patients starting highly active antiretroviral therapy: a collaborative analysis of prospective studies. Lancet. 2002;360:119–129.
45. Tuboi SH, Schechter M, McGowan C, et al.. Mortality during the first year of potent antiretroviral therapy in HIV-1–infected patients in 7 sites throughout Latin America and the Caribbean. J Acquir Immune Defic Syndr. 2009;51:615–623.