Risk of cancers during interrupted antiretroviral therapy in the SMART study

Transcription

1 Risk of cancers during interrupted antiretroviral therapy in the SMART study Michael J. Silverberg a, Jacqueline Neuhaus b, Mark Bower c, Daniela Gey d, Angelos Hatzakis e, Keith Henry f, Jose Hidalgo g, Leonardo Lourtau h, James D. Neaton b, Giuseppe Tambussi i and Donald I. Abrams j Objective: To compare rates of AIDS-defining and non-aids-defining malignancies between patients on a CD4 T-cell-guided antiretroviral therapy (ART) strategy and continuous ART. Design: A randomized clinical trial. Methods: Malignancy rates were compared between the drug conservation arm in which ART was stopped if the CD4 T-cell count exceeded 350 cells/ml and (re)started if it fell to less than 250 cells/ml and the viral suppression arm utilizing continuous ART. Cox models were used to examine baseline characteristics including age, sex, race, cigarette use, previous malignancies, CD4 T-cell and HIV-RNA levels, hepatitis B or C, and ART duration. Results: A total of 5472 participants were randomly assigned to treatment groups, of whom 70 developed cancer: 13 AIDS-defining malignancies and 58 non-aidsdefining malignancies (one patient had both). The AIDS-defining malignancy rate per 1000 person-years was higher in the drug conservation arm (3.0 versus 0.5). Proximal CD4 T-cell and HIV RNA levels mediated much of this increased risk. The drug conservation arm also had higher rates of Kaposi s sarcoma (1.9 versus 0.3) and lymphoma (Hodgkin s and non-hodgkin s; 1.1 versus 0.3). The non-aids-defining malignancy rate was similar between the drug conservation and viral suppression arms (8.8 versus 7.1). The most common non-aids-defining malignancies were skin (n ¼ 16), lung (n ¼ 8) and prostate (n ¼ 6) cancers. Conclusion: Non-AIDS-defining malignancies were more common in this cohort than AIDS-defining malignancies. This analysis provides further evidence against the use of CD4 T-cell-guided ART because of a higher risk of AIDS-defining malignancies in addition to opportunistic infections and deaths. ß 2007 Lippincott Williams & Wilkins AIDS 2007, 21: Keywords: antiretroviral therapy, cancer, CD4 T-cell counts, HIV, randomized clinical trials From the a Division of Research, Kaiser Permanente Northern California, Oakland, California, USA, the b Division of Biostatistics, University of Minnesota, School of Public Health, Minneapolis, Minnesota, USA, the c Departments of Oncology and HIV Medicine, Chelsea and Westminster Hospital, London, UK, the d Copenhagen HIV Programme, Faculty of Health Sciences, University of Copenhagen, Panum Institute, Copenhagen, Denmark, the e Department of Hygiene and Epidemiology, Athens University Medical School, Athens, Greece, the f HIV Program, Minnesota AIDS Clinical Trials Unit, Minneapolis, Minnesota, USA, the g HIV Program, Guillermo Almenara Hospital, Lima, Peru, the h Servicio de Inmunocomprometidos, Hospital JM Ramos Mejía, Buenos Aires, Argentina, the i Clinic of Infectious Diseases, San Raffaele Scientific Institute, Milan, Italy, and the j Hematology Oncology Division, San Francisco General Hospital, University of California San Francisco, San Francisco, California, USA. Correspondence to Donald I. Abrams, MD, Hematology Oncology Division, San Francisco General Hospital, Ward 84, 995 Potrero Avenue, San Francisco, CA , USA. dabrams@hemeonc.ucsf.edu Received: 24 April 2007; revised: 7 June 2007; accepted: 14 June ISSN Q 2007 Lippincott Williams & Wilkins 1957

2 1958 AIDS 2007, Vol 21 No 14 Introduction Declines in AIDS-defining malignancies including Kaposi s sarcoma and non-hodgkin s lymphoma (NHL) have been observed in HIV-infected populations with the availability of effective antiretroviral therapy (ART) [1 7]. Conversely, given the longer survival of HIV-infected individuals in the ART era, there is growing concern about the higher incidence of non- AIDS-related disease. Several studies, for example, have demonstrated a shift in mortality to non-aids causes, with a growing proportion related to non-aidsdefining malignancies [8 11]. Several studies have found a higher incidence in the ART era of non- AIDS-defining malignancies in general [2] and specifically of Hodgkin s lymphoma [3,4,7,12], skin [2], lung [12,13], anal [2,14], kidney [4] and prostate [15] cancers. Other studies, however, have demonstrated no change in incidence in the ART era for the same malignancies [1 4,7,12,16 18]. As has been suggested for Kaposi s sarcoma, ART may reduce the incidence of AIDS-defining malignancies indirectly via improved immune surveillance for cancer cells, or directly via the inhibition of tumor development [19]. The relationship between immune status and the incidence of non-aids-defining malignancies, however, is unclear. A large AIDS and cancer registry study indicated that only Hodgkin s lymphoma, lung, penile, lip, testicular, and sarcomas were associated with immunodeficiency [20]. Recent studies indicated that the risk of fatal or non-fatal non-aids-defining malignancies was increased with lower CD4 T-cell counts [21,22]. Others, however, did not find such an association [17,23 25]. Further research on malignancies in HIV-infected individuals is warranted, particularly with respect to the association with ART and improvements in immune status. In this context, we report on the occurrence of malignancies in patients enrolled in the Strategies for Management of Antiretroviral Therapy (SMART) trial. The main objectives of the SMART trial were to compare opportunistic disease rates, death rates, and the risk of major ART-associated adverse events between patients randomly assigned to the episodic use of ART based on CD4 T-cell count levels or the continuous use of ART [26]. Episodic use of ART as used in the SMART study was inferior to the continuous use of ART for the primary endpoint of opportunistic disease or death from any cause as well as an important secondary endpoint of major cardiovascular, renal or hepatic disease. We extend this work here to an analysis of the incidence of AIDSdefining and non-aids-defining malignancies among SMART participants and assess the effect of SMART treatment strategies on this risk. We further elucidate the possible pathogenic mechanisms of malignant outcomes in HIV-infected individuals. Methods Study design and population The SMART study design is described in detail elsewhere [26]. Briefly, 5472 HIV-infected patients were recruited at 318 sites in 33 countries representing north America (57%), Europe (26%), south America (10%), Australia/ New Zealand (3%), Asia (3%), and Africa (1%). Recruitment began in 2002 with the following eligibility criteria: age greater than 13 years, not pregnant or breastfeeding, CD4 T-cell count greater than 350 cells/ml, and willing to initiate, modify, or stop ART according to the SMART trial protocol. Patients were randomly assigned into the drug conservation or the viral suppression study arms. The drug conservation treatment strategy involved CD4 T-cell guided episodic ART in which, after random selection, patients stopped or deferred the initiation of ARTuntil CD4 T-cell counts fell below 250 cells/ml, and then restarted ARTuntil CD4 T-cell count levels increased to more than 350 cells/ml. At this time ART was stopped again until CD4 T-cell count levels fell below 250 cells/ml. The viral suppression treatment strategy was based on the 2003 United States Department of Health and Human Services guidelines [27] and involved the use of ART throughout follow-up to maintain HIV RNA levels as low as possible. Investigators and participants were not blinded to the study arm. Follow-up visits for participants were scheduled for months 1, 2 and every 2 months thereafter in year 1, and every 4 months in subsequent years. On 11 January 2006 enrollment was stopped on the basis of a recommendation from the Data Safety and Monitoring Board, and participants in the drug conservation group were advised to restart ART as a result of the increased risk of both HIV-related and other serious events in the drug conservation group. Data collection At baseline, defined as within 45 days before random selection, a medical history and examination were performed for study participants. Data were collected regarding demographics, the use of cigarettes, history of malignancies, recent and nadir CD4 T-cell counts, recent and highest previous HIV RNA levels, and ART history. A medical history and examination was also performed at annual study follow-up visits, which included the measurement of CD4 T-cell and HIV RNA levels. The primary outcomes for analyses presented here were new AIDS-defining or non-aids-defining malignancies. As per the study protocol, deaths from any cause, including any malignancy, as well as opportunistic diseases, including AIDS-defining malignancies, were reviewed by an endpoint review committee that was blinded to the study arm. In the SMART study, cancers considered AIDS-defining malignancies were broadened from the 1993 Centers for Disease Control and Prevention AIDS surveillance definition, and included both Hodgkin s and non-hodgkin s

3 Malignant outcomes in the SMART study Silverberg et al lymphoma, invasive cervical cancer, and Kaposi s sarcoma [28,29]. All other cancers were considered non-aidsdefining malignancies. Pathology reports for non-fatal non-aids-defining malignancies were further reviewed by the collaborating oncologist (D.I.A.). Statistical methods Malignancy incidence rates per 1000 person-years were computed for all malignancies combined (i.e. AIDSdefining or non-aids-defining malignancies), AIDSdefining and non-aids-defining malignancies, as well as for individual malignancies. Cox proportional hazards models were used to compare the drug conservation and viral suppression study arms. Patients were censored at the time of first AIDS-defining or non-aids-defining malignancy, time of death, or end of study follow-up. Next, Cox models were constructed separately for AIDSdefining and non-aids-defining malignancies, with adjustment for baseline characteristics including age, sex, race/ethnicity, current smoking status, history of malignancy, baseline, recent and nadir CD4 T-cell counts, baseline and recent HIV RNA levels, duration of previous ART at study entry, and evidence of hepatitis B or C infection. The duration of HIV infection was examined, but not included in final models because of collinearity with the ART parameter. Low event rates also precluded a multivariable analysis for individual malignancies. Cox models were also used to examine the effect of study arm with and without adjustment for proximal (i.e. time updated) CD4 T-cell count and HIV RNA levels. We also performed subgroup analyses by age and cigarette use. Results Baseline characteristics and study follow-up A total of 5472 participants were enrolled and followed between 8 January 2002 and 11 January A total of 2720 participants were randomly assigned to the drug conservation arm and contributed 3666 person-years. A total of 2752 participants were randomly assigned to the viral suppression arm and contributed 3701 person-years. As described previously [26] and in Table 1, baseline demographic and clinical characteristics were similar for the drug conservation and viral suppression study arms. Overall, few participants had a previous history of non- AIDS-defining malignancies (2.4%) or AIDS-defining malignancies (2.6%), with similar percentages by study arm. Overall AIDS-defining and non-aids-defining malignancy event rates During follow-up, 70 participants developed a malignant endpoint. AIDS-defining malignancies occurred in 13 subjects and three patients experienced AIDS-defining malignancy-related deaths. Non-AIDS-defining malignancies occurred in 58 participants and 16 participants experienced non-aids-defining malignancy-related deaths. One participant was diagnosed with both an AIDS-defining malignancy and a non-aids-defining malignancy. Compared with other major non-aids endpoints of the SMART trial, the number of malignant events was similar to the number of cardiovascular events (n ¼ 79), but was more frequent than the number of cases of serious hepatic (n ¼ 17) or renal diseases (n ¼ 11). As presented in Table 2, the drug conservation strategy was associated with a higher risk of AIDS-defining malignancies, with a hazard ratio (HR; drug conservation/viral suppression) of 5.5 [95% confidence interval (CI) 1.2, 25.0; P ¼ 0.03]. The overall non-aids-defining malignancy rate was similar between the drug conservation and viral suppression arms with an HR (drug conservation/ viral suppression) of 1.3 (95% CI 0.7, 2.1; P ¼ 0.40). Death rates from AIDS-defining malignancies, non- AIDS-defining malignancies, or both, were approximately twofold higher in the drug conservation arm compared with the viral suppression arm, although the differences were not statistically significant (P > 0.05). Individual AIDS-defining and non-aids-defining malignancy event rates Event rates for individual malignancies by study arm are presented in Table 3. The 13 AIDS-defining malignancies consisted of eight cases of Kaposi s sarcoma and five cases of lymphoma (three non-hodgkin s and two Hodgkin s). Participants in the drug conservation group had a higher risk of Kaposi s sarcoma with a HR (drug conservation/ viral suppression) of 7.0 (95% CI 0.9, 57.1; P ¼ 0.07). No statistically significant difference in the rate of lymphoma was noted by treatment arm, although the number of events was low. The most common non-aids-defining malignancies were cancers of the skin (n ¼ 16), lung Table 1. SMART baseline characteristics. Parameter Drug conservation strategy Viral suppression strategy Total No. of patients Median age, years (IQR) (38 50) Current smokers, n (%) 1123 (41.3) 1090 (39.6) 2213 (40.4) History of any malignancy, n (%) 142 (5.2) 123 (4.5) 265 (4.8) History of non-aids-defining malignancies, n (%) 74 (2.7) 57 (2.1) 131 (2.4) History of AIDS-defining malignancies, n (%) 71 (2.6) 70 (2.5) 141 (2.6) IQR, Interquartile range.

4 1960 AIDS 2007, Vol 21 No 14 Table 2. Cancer endpoints for drug conservation and viral suppression arms. Rate per 1000 py Endpoints No. of patients with events DC VS HR (95% CI) a P value a Malignancy incidence ADM or NADM b ( ) 0.08 ADM ( ) 0.03 NADM ( ) 0.40 Malignancy death ADM or NADM ( ) 0.11 ADM Too few events NADM ( ) 0.31 ADM, AIDS-defining malignancy; CI, confidence interval; DC, drug conservation; HR, hazard ratio; NADM, non-aids-defining malignancy; py, person-year; VS, viral suppression. a Based on Cox proportional hazards model. b One patient had both an ADM and NADM. (n ¼ 8) and prostate (n ¼ 6). The HR (drug conservation/ viral suppression) for skin, lung and prostate cancers were 1.0 (95% CI 0.4, 2.7; P ¼ 0.98), 3.0 (95% CI 0.6, 15.0; P ¼ 0.18), and 1.0 (95% CI 0.2, 5.0; P ¼ 0.99), respectively. The rates of other non-aids-defining malignancies were similar by study arm, although the number of events was low. Predictors of AIDS-defining and non-aidsdefining malignancies The strongest predictor of AIDS-defining malignancies was a previous history of AIDS-defining malignancies with an HR of (95% CI ; P < 0.001). Of the 13 AIDS-defining malignancies occurring during Table 3. Incidence rates for individual AIDS-defining and non- AIDS-defining malignancies for drug conservation and viral suppression arms. Endpoints Rate per 1000 py No. of patients with events DC VS ADM Kaposi s sarcoma Lymphoma Non-Hodgkin s lymphoma Hodgkin s lymphoma NADM Skin Lung Prostate Colorectal Renal Anal squamous Breast Head and neck Hepatocellular Pancreas Esophageal Cervical (non-invasive) Unknown primary ADM, AIDS-defining malignancy; DC, drug conservation; NADM, non-aids-defining malignancy; py, person-year; VS, viral suppression. the study, seven had a previous AIDS-defining malignancy; of these seven, four had the same AIDS-defining malignancy (three Kaposi s sarcoma cases had previous Kaposi s sarcoma and one lymphoma case had a previous lymphoma), and three had a different AIDS-defining malignancy (three lymphoma cases had previous Kaposi s sarcoma). The only other statistically significant predictor of AIDS-defining malignancies, besides treatment group, was time since the first prescribed ARTwith an HR of 0.8 (95% CI ; P ¼ 0.03) per one year of additional ART. Age, sex, race, baseline cigarette use, baseline and nadir CD4 T-cell counts, baseline HIV RNA levels, and evidence of hepatitis B or C infection were not significant predictors for AIDS-defining malignancies. Statistically significant predictors of non-aids-defining malignancies were age, with an HR of 2.2 per 10 years older (95% CI ; P < 0.001), current cigarette use at baseline, with an HR of 2.5 (95% CI ; P ¼ 0.001), previous history of non-aids-defining malignancy, with an HR of 2.5 (95% CI ; P ¼ 0.03), and baseline HIV RNA level of 400 copies/ml or less, with an HR of 2.1 (95% CI ; P ¼ 0.02). Sex, race, baseline and nadir CD4 T-cell counts, time since first prescribed ART, and evidence of hepatitis B or C infection were not significant predictors for non- AIDS-defining malignancies. Analyses stratified by treatment arm indicated that cigarette use was a significant predictor only in the viral suppression group and a history of non-aids-defining malignancy was a significant predictor only in the drug conservation group (data not shown). HR for the risk of non-aids-defining malignancy by study arm are presented in Table 4 within strata defined by age and cigarette use. In general, the effect of study arm was strongest within lower risk strata (i.e. younger patients and non-smokers). The HR (drug conservation/ viral suppression) among those aged 50 and less and over 50 years were 1.7 and 0.9, respectively (P for interaction 0.25). Similarly, the HR (drug conservation/viral

5 Malignant outcomes in the SMART study Silverberg et al Table 4. Predictors of AIDS-defining and non-aids-defining malignancies in SMART a. ADM NADM HR (95% CI) P value HR (95% CI) P value Baseline predictors of ADM and NADM b DC study arm 7.3 ( ) ( ) 0.51 Previous malignancy c ( ) < ( ) 0.03 Age (10 years) 1.2 ( ) ( ) < Male 3.7 ( ) ( ) 0.93 Black race 1.3 ( ) ( ) 0.73 Cigarette use 0.5 ( ) ( ) CD4 T-cell count (100 cells/ml) 1.2 ( ) ( ) 0.18 Nadir CD4 T-cell count (100 cells/ml) 1.4 ( ) ( ) 0.66 HIV RNA 400 copies/ml 1.2 ( ) ( ) 0.02 Hepatitis B or C Undefined 0.9 ( ) 0.65 Years ART use (1 year) 0.8 ( ) ( ) 0.83 Hazard ratios for DC versus VS group by age and cigarette use d Age, years > 50 Too few events 0.9 ( ) Too few events 1.7 ( ) 0.17 P value for interaction 0.25 Cigarette use Yes Too few events 0.8 ( ) 0.47 No Too few events 2.3 ( ) 0.05 P value for interaction 0.04 Hazard ratios for DC versus VS group with and without adjustments for proximal CD4 T-cell counts and proximal HIV-RNA levels d Unadjusted 5.5 ( ) ( ) 0.40 Proximal CD4 T-cell count 3.6 ( ) ( ) 0.66 Proximal HIV RNA 3.4 ( ) ( ) 0.09 Proximal CD4 T-cell count and HIV RNA 2.8 ( ) ( ) 0.22 ADM, AIDS-defining malignancy; ART, antiretroviral therapy; CI, confidence interval; DC, drug conservation; HR, hazard ratio; NADM, non- AIDS-defining malignancy; VS, viral suppression. a Results based on Cox proportional hazards models. b Cox model adjusted for all variables. c Refers to previous ADM for analysis of new ADM and previous NADM for analysis of new NADM. d Reference group for hazard ratio is viral suppression study arm. suppression) among non-smokers and smokers were 2.3 and 0.8, respectively (P for interaction 0.04). CD4 T-cell counts and HIV RNA levels during follow-up in the two treatment groups may have partly accounted for differences in malignancy rates noted by study arms. Table 4 includes HR (drug conservation/viral suppression) with and without adjustment for proximal CD4 T-cell count and proximal HIV RNA levels. For AIDS-defining malignancies, adjustment for proximal CD4 T-cell counts and HIV RNA levels resulted in a decrease in unadjusted HR (drug conservation/viral suppression) of 5.5 to an HR of 2.8 (95% CI 0.6, 13.9; P ¼ 0.21), representing a 49% reduction in the magnitude of the HR and loss of statistical significance compared with unadjusted results. Similar adjustment resulted in smaller changes in HR for non-aids-defining malignancies, with no change in statistical significance. Discussion The major aims of this analysis were to compare rates of malignancies in the SMART trial comparing CD4 T-cell-guided intermittent use of ART with continuous ART, and to examine predictors of these malignancies. In a relatively short time of 16 months average follow-up, the CD4 T-cell-guided intermittent ART strategy resulted in a significant increase in the incidence of AIDS-defining malignancies. HIV disease stage accounted for much of the increased risk of AIDS-defining malignancies in the intermittent ART study arm. Non-AIDS-defining malignancies were more common than AIDS-defining malignancies, and were the most common underlying cause of death in the SMART trial. The incidence of non-aidsdefining malignancies was not, however, influenced by the treatment arm. Older age, cigarette use, baseline undetectable HIV RNA level and previous non-aids-defining malignancy were the strongest risk factors for new non- AIDS-defining malignancies. The increased risk of AIDS-defining malignancies was largely mediated by lower proximal CD4 T-cell counts and higher HIV RNA levels in the drug conservation arm. This is consistent with earlier studies that have shown a higher risk of Kaposi s sarcoma and non-hodgkin s lymphoma with decreased immune function [3,6,7,24,25,30], as well as Hodgkin s lymphoma [3,20,31,32], although the same association has not been shown by others for Hodgkin s

6 1962 AIDS 2007, Vol 21 No 14 lymphoma [18] or for cervical cancer [3,24]. In this study, no invasive cervical cancers were observed, which is consistent with the low cervical cancer risk in the ARTera [33]. It should be noted, however, that only 27% of SMART participants were women. In general, our study supports continuing ARTuse for the prevention of AIDSdefining malignancies. Non-AIDS-defining malignancies were not different according to the study arm; however, as a result of the shortened duration of the study, we cannot rule out the possibility of a modest effect of ART interruption on non-aids-defining malignancies (i.e. the upper bound of 95% CI is 2.1). There was a trend towards a threefold increased risk of lung cancer in the drug conservation versus the viral suppression study arm, but this did not reach statistical significance. As expected, age and smoking status were strong predictors of non-aidsdefining malignancies. We also observed a higher risk of non-aids-defining malignancies in participants with undetectable HIV RNA at baseline independent of age and duration of ART use. Additional analyses indicated that the higher risk of non-aids-defining malignancies with HIV RNA levels less than or equal to 400 copies/ml at baseline was not confined to either study group, nor diminished after adjustment for recent HIV RNA or CD4 levels (data not shown). The reasons for this finding are thus not known and may be due to chance. Despite the lack of a difference between the study arms overall, subgroup analyses indicated that there may be an increased risk of non-aids-defining malignancies in the drug conservation study arm among groups with a lower overall non-aids-defining malignancy risk, including younger subjects and non-smokers. It is therefore possible that the trend towards an increased risk of non-aids-defining malignancies in these subgroups and for lung cancer may be caused by lower CD4 T-cell counts during follow-up resulting from the intermittent use of ART in the drug conservation study arm. The small number of events, however, precluded further study. Earlier studies of non-aids-defining malignancies have shown an association between immunodeficiency and the following cancers: lung [20], penile [20], lip [20], testicular [20,31], sarcomas [20,31,34], myeloma [31], and brain [31]. Others did not, however, find such an association with non-aids-defining malignancies [3,7,17,23 25,35,36]. Few studies utilized CD4 T-cell count levels as a measure of immunodeficiency [3,35,36], whereas most relied on proximity to the AIDS diagnosis as a crude proxy for immunodeficiency [20,23,31,32,34]. Recent studies of fatal non-aids-defining malignancies in the D:A:D study [22] and fatal or non-fatal non-aidsdefining malignancies in the FIRST trial [21] indicated a strong relationship between current CD4 T-cell counts and non-aids-defining malignancies. However, since these studies included fatal non-aids-defining malignancies, it can not be determined whether low CD4 T-cell counts increased the risk of new malignancies, accelerated progression of existing malignancies, or a combination of both mechanisms. The main strength of this study is the randomized design. Several recent randomized clinical trials have examined alternative ART treatment strategies, primarily CD4 cellguided, versus continuous ART, but none of them specifically focused on malignancies [26,37 41]. Our study of 5472 patients was the largest of these and indicated that episodic ART resulted in an increased risk of opportunistic disease and death, and did not reduce the adverse effects associated with ARTuse [26]. A strength of SMART is the large population sample that included patients from 318 sites in 33 countries, thus ensuring the high generalizability of the study findings. Nevertheless, the study has several limitations. Despite the large size, the study is not sufficiently powered for the study of less common events, including site-specific cancers. The results for site-specific cancers were, however, fairly consistent within general categories of AIDS-defining and non-aids-defining malignancies. The low event rates also precluded multivariable analyses for each sitespecific malignancy. Findings from the SMART study do not support the use of CD4 cell-guided intermittent ART in view of the higher risk of AIDS-defining malignancies and the already described increased risk of opportunistic disease and death [26]. Malignant outcomes were frequent in this cohort, with non-aids-defining malignancies more commonly reported than AIDSdefining malignancies. Traditional risk factors such as age and cigarette use were predictive of non-aids-defining malignancies rather than immune suppression. Standard cancer preventive measures should thus be encouraged, including smoking cessation programmes for HIVinfected patients. Sponsorship: This study was supported by grants from the NIAID (U01AI and U01AI46362). Conflicts of interest: None. References 1. International Collaboration on HIV and Cancer. Highly active antiretroviral therapy and incidence of cancer in human immunodeficiency virus-infected adults. J Natl Cancer Inst 2000; 92: Bedimo R, Chen RY, Accortt NA, Raper JL, Linn C, Allison JJ, et al. Trends in AIDS-defining and non-aids-defining malignancies among HIV-infected patients: Clin Infect Dis 2004; 39: Clifford GM, Polesel J, Rickenbach M, Dal Maso L, Keiser O, Kofler A, et al. Cancer risk in the Swiss HIV Cohort Study: associations with immunodeficiency, smoking, and highly active antiretroviral therapy. JNatlCancerInst2005; 97: Engels EA, Pfeiffer RM, Goedert JJ, Virgo P, McNeel TS, Scoppa SM, et al. Trends in cancer risk among people with AIDS in the United States AIDS 2006; 20:

7 Malignant outcomes in the SMART study Silverberg et al Grulich AE, Li Y, McDonald AM, Correll PK, Law MG, Kaldor JM. Decreasing rates of Kaposi s sarcoma and non-hodgkin s lymphoma in the era of potent combination antiretroviral therapy. AIDS 2001; 15: Mocroft A, Katlama C, Johnson AM, Pradier C, Antunes F, Mulcahy F, et al. AIDS across Europe, : the EuroSIDA study. Lancet 2000; 356: Engels E, Goedert J, Hall I, Cross H, Crutchfield A, Finch J, et al. Cancer risk in 57,350 HIV-infected individuals in the US ( ): relationship with immunosuppression and changes over time. In: 14th Conference on Retroviruses and Opportunistic Infections. Los Angeles, CA, February Abstract Crum NF, Riffenburgh RH, Wegner S, Agan BK, Tasker SA, Spooner KM, et al. Comparisons of causes of death and mortality rates among HIV-infected persons: analysis of the pre, early, and late HAART (highly active antiretroviral therapy) eras. J Acquir Immune Defic Syndr 2006; 41: Lewden C, Salmon D, Morlat P, Bevilacqua S, Jougla E, Bonnet F, et al. Causes of death among human immunodeficiency virus (HIV)-infected adults in the era of potent antiretroviral therapy: emerging role of hepatitis and cancers, persistent role of AIDS. Int J Epidemiol 2005; 34: Selik RM,Byers RH Jr, Dworkin MS. Trends in diseases reported on US death certificates that mentioned HIV infection, J Acquir Immune Defic Syndr 2002; 29: Lewden C, May T, Rosenthal E, Burty C, Bonnet F, Costagliola D, et al. Causes of death among HIV-infected Adults in France in 2005 and Evolution since In: 14th Conference on Retroviruses and Opportunistic Infections. Los Angeles, CA, February Abstract Herida M, Mary-Krause M, Kaphan R, Cadranel J, Poizot-Martin I, Rabaud C, et al. Incidence of non-aids-defining cancers before and during the highly active antiretroviral therapy era in a cohort of human immunodeficiency virus-infected patients. J Clin Oncol 2003; 21: Bower M, Powles T, Nelson M, Shah P, Cox S, Mandelia S, Gazzard B. HIV-related lung cancer in the era of highly active antiretroviral therapy. AIDS 2003; 17: Diamond C, Taylor TH, Aboumrad T, Bringman D, Anton- Culver H. Increased incidence of squamous cell anal cancer among men with AIDS in the era of highly active antiretroviral therapy. Sex Transm Dis 2005; 32: Rimland D, Guest JL. Increasing incidence of prostate cancer in the Atlanta VA cohort study (HAVACS). In: 14th Conference on Retroviruses and Opportunistic Infections. Los Angeles, CA, February Abstract Hessol NA, Seaberg EC, Preston-Martin S, Massad LS, Sacks HS, Silver S, et al. Cancer risk among participants in the women s interagency HIV study. J Acquir Immune Defic Syndr 2004; 36: Burgi A, Brodine S, Wegner S, Milazzo M, Wallace MR, Spooner K, et al. Incidence and risk factors for the occurrence of non-aids-defining cancers among human immunodeficiency virus-infected individuals. Cancer 2005; 104: Bonnet F, Petoumenos K, Calmy A, Lewden C, Hoy J, Dabis F, et al. HAART is not associated with an increased risk of Hodgkin s lymphoma in HIV-infected patients. In: 14th Conference on Retroviruses and Opportunistic Infections. Los Angeles, CA, February Abstract Sgadari C, Monini P, Barillari G, Ensoli B. Use of HIV protease inhibitors to block Kaposi s sarcoma and tumour growth. Lancet Oncol 2003; 4: Frisch M, Biggar RJ, Engels EA, Goedert JJ. Association of cancer with AIDS-related immunosuppression in adults. JAMA 2001; 285: Baker J, Peng G, Rapkin J, Abrams D, Silverberg M, Cavert W, et al. HIV-related immune suppression after ART predicts risk of nonopportunistic diseases: results from the FIRST Study. In: 14th Conference on Retroviruses and Opportunistic Infections. Los Angeles, CA, February Abstract D Arminio Monforte A, Abrams D, Pradier C, Weber R, Bonnet F, De Wit S, et al. HIV-induced immunodeficiency and risk of fatal AIDS-defining and non-aids-defining malignancies: results from the D:A:D Study. In: 14th Conference on Retroviruses and Opportunistic Infections. Los Angeles, CA, February 2007, Abstract Biggar RJ, Kirby KA, Atkinson J, McNeel TS, Engels E. Cancer risk in elderly persons with HIV/AIDS. J Acquir Immune Defic Syndr 2004; 36: Mbulaiteye SM, Biggar RJ, Goedert JJ, Engels EA. Immune deficiency and risk for malignancy among persons with AIDS. J Acquir Immune Defic Syndr 2003; 32: Phelps RM, Smith DK, Heilig CM, Gardner LI, Carpenter CC, Klein RS, et al. Cancer incidence in women with or at risk for HIV. Int J Cancer 2001; 94: El-Sadr WM, Lundgren JD, Neaton JD, Gordin F, Abrams D, Arduino RC, et al. CD4R count-guided interruption of antiretroviral treatment. N Engl J Med 2006; 355: US Department of Health and Human Services. Guidelines for the Use of Antiretroviral Agents in HIV-1-Infected Adults and Adolescents, Washington, DC: Department of Health and Human Services Green LA, Rhame FS, Price RW, Perlman DC, Capps LG, Sampson JH, et al. Experience with a cross-study endpoint review committee for AIDS clinical trials. Terry Beirn Community Programs for Clinical Research on AIDS. AIDS 1998; 12: Lifson AR, Rhame FS, Belloso WH, Dragsted UB, El-Sadr WM, Gatell JM, et al. Reporting and evaluation of HIV-related clinical endpoints in two multicenter international clinical trials. HIV Clin Trials 2006; 7: Kirk O, Pedersen C, Cozzi-Lepri A, Antunes F, Miller V, Gatell JM, et al. Non-Hodgkin lymphoma in HIV-infected patients in the era of highly active antiretroviral therapy. Blood 2001; 98: Goedert JJ, Cote TR, Virgo P, Scoppa SM, Kingma DW, Gail MH, et al. Spectrum of AIDS-associated malignant disorders. Lancet 1998; 351: Grulich AE, Wan X, Law MG, Coates M, Kaldor JM. Risk of cancer in people with AIDS. AIDS 1999; 13: Massad LS, Seaberg EC, Watts DH, Hessol NA, Melnick S, Bitterman P, et al. Low incidence of invasive cervical cancer among HIV-infected US women in a prevention program. AIDS 2004; 18: Grulich AE, Li Y, McDonald A, Correll PK, Law MG, Kaldor JM. Rates of non-aids-defining cancers in people with HIV infection before and after AIDS diagnosis. AIDS 2002; 16: Chaturvedi AK, Pfeiffer RM, Chang L, Goedert JJ, Biggar RJ, Engels EA. Elevated risk of lung cancer among people with AIDS. AIDS 2007; 21: Engels EA, Brock MV, Chen J, Hooker CM, Gillison M, Moore RD. Elevated incidence of lung cancer among HIV-infected individuals. J Clin Oncol 2006; 24: Ananworanich J, Gayet-Ageron A, Le Braz M, Prasithsirikul W, Chetchotisakd P, Kiertiburanakul S, et al. CD4-guided scheduled treatment interruptions compared with continuous therapy for patients infected with HIV-1: results of the Staccato randomised trial. Lancet 2006; 368: Ananworanich J, Siangphoe U, Hill A, Cardiello P, Apateerapong W, Hirschel B, et al. Highly active antiretroviral therapy (HAART) retreatment in patients on CD4-guided therapy achieved similar virologic suppression compared with patients on continuous HAART: the HIV Netherlands Australia Thailand Research Collaboration study. J Acquir Immune Defic Syndr 2005; 39: Cardiello PG, Hassink E, Ananworanich J, Srasuebkul P, Samor T, Mahanontharit A, et al. A prospective, randomized trial of structured treatment interruption for patients with chronic HIV type 1 infection. Clin Infect Dis 2005; 40: Danel C, Moh R, Minga A, Anzian A, Ba-Gomis D, Kanga C, et al. CD4-guided structured antiretroviral treatment interruption strategy in HIV-infected adults in west Africa (Trivacan ANRS 1269 trial): a randomised trial. Lancet 2006; 367: Maggiolo F, Ripamonti D, Gregis G, Quinzan G, Callegaro A, Suter F. Effect of prolonged discontinuation of successful antiretroviral therapy on CD4 T cells: a controlled, prospective trial. AIDS 2004; 18: