Association between older age and adverse outcomes on antiretroviral therapy: a cohort analysis of programme data from nine countries
Greig, Janea; Casas, Esther C.b; O’Brien, Daniel P.b,c,d; Mills, Edward J.e; Ford, Nathana,f
aMédecins Sans Frontières, London, UK
bMédecins Sans Frontières, Amsterdam, The Netherlands
cDepartment of Infectious Diseases, Geelong Hospital, Geelong, Victoria
dDepartment of Medicine and Infectious Diseases, Royal Melbourne Hospital, University of Melbourne, Melbourne, Victoria, Australia
eFaculty of Health Sciences, University of Ottawa, Ottawa, Ontario, Canada
fCentre for Infectious Disease Epidemiology and Research, University of Cape Town, Cape Town, South Africa.
Correspondence to Jane Greig, Médecins Sans Frontières, 67–74 Saffron Hill, London EC1N 8QX, UK. Tel: +44 207 404 6600; fax: +44 207 404 4466; e-mail: firstname.lastname@example.org
Received 12 December, 2011
Accepted 6 April, 2012
Context: Recent studies have highlighted the increased risk of adverse outcomes among older patients on antiretroviral therapy (ART). We report on the associations between older age and adverse outcomes in HIV/AIDS antiretroviral programmes across 17 programmes in sub-Saharan Africa.
Methods: We included data from nine countries: Central African Republic, Côte d’Ivoire, Democratic Republic of Congo, Ethiopia, Nigeria, Republic of Congo, Uganda, Zambia and Zimbabwe. We describe survival probability for progression to death and loss to follow-up for patients initiating ART aged less than 50 years and at least 50 years. Multivariate Cox proportional hazards models were used to assess the association between age (15–39, 40–49, 50–59, 60–69 and 70–94 years) and adverse outcomes adjusting for confounders identified a priori.
Results: Our analysis included 17 561 patients followed for a median of 12 months. The majority (65%) were female and 6672 (38%) were severely immunosuppressed at baseline. Median age at ART initiation was 36.0 years (interquartile range 30.1–42.8); 11.4% of patients were aged at least 50 years. Median gain in CD4 cell count at 6 and 12 months was significantly higher in patients less than 50 years old compared with those at least 50 years (134 vs. 112 cells/μl at 6 months; 170 vs. 139 cells/μl at 12 months; both P < 0.001). In multivariate analysis, there was a significant increased risk of mortality beyond 3 months after ART initiation in all age groups of at least 40 years of age compared with less than 40 years [40–49 years adjusted hazard ratios (aHRs) 1.59, P < 0.001; 50–59 years aHR 1.58, P = 0.002; 60–69 years aHR 2.63, P < 0.001; 70–94 years aHR 3.64, P = 0.004).
Conclusion: Older age groups represent an important proportion of the overall treatment cohort in these sub-Saharan Africa programmes, and risk of mortality increased as age increased. Future research should be directed at further understanding the reasons for higher mortality, and defining simple interventions that are feasible in highly under-resourced settings to allow for adapted follow-up and care approaches for older age groups.
Increased access to antiretroviral therapy (ART) in resource-limited settings over the last decade has resulted in substantial reductions in AIDS-related illness and death [1,2] for people living with HIV/AIDS.
Concern about heightened vulnerability within specific age groups has tended to focus on the lower extreme, particularly infants and children , and more recently adolescents . Vulnerability at the other age extreme has received less attention. The need to focus on older age groups has recently been highlighted by reports from sub-Saharan Africa showing higher mortality in people aged 50 years and older [5–9]. This has led to calls for greater attention to be paid to the specific vulnerabilities of older populations in HIV/AIDS programmes .
In this study, we report on the association between older age and adverse outcomes in HIV/AIDS programmes across nine countries in sub-Saharan Africa.
This analysis is based on data from 17 programmes from nine countries in sub-Saharan Africa: Central African Republic, Côte d’Ivoire, Democratic Republic of Congo, Ethiopia, Nigeria, Republic of Congo, Uganda, Zambia and Zimbabwe. At each site, data were routinely collected by clinical staff using standardized forms and entered into an electronic database (FUCHIA, Epicentre, Paris, France). This database has been described in detail elsewhere . Data for all adult patients (15 years or older) initiating ART between September 2003 and March 2010 were included in this study.
Descriptive analyses were based on percentages and frequencies for categorical variables and medians and interquartile ranges (IQRs) for continuous variables. Baseline characteristics were described for the cohort overall and separately for patients aged less than 50 years and at least 50 years. Continuous variables were assessed for skew and as all demonstrated nonnormality they were compared using the Wilcoxon rank-sum test. Proportions were compared using the χ2-test.
We used Kaplan–Meier estimates to describe the cumulative probability of progression to death and loss to follow-up for patients aged less than 50 years and at least 50 years across all projects. Loss to follow-up was defined as missing a scheduled appointment by 2 months or more. Multivariate Cox proportional hazards models were used to model the association between age and adverse outcomes (death and loss to follow-up). Age was assessed as a categorical variable (15–39, 40–49, 50–59, 60–69 and 70–94 years). These models adjusted for the following potential confounders defined a priori: sex (male or female), BMI (kg/m2) at ART initiation, clinic setting (urban or rural), tuberculosis (TB) at ART initiation and severe immunosuppression at ART initiation (defined as WHO stage IV or CD4 cell count ≤100 cells/μl). Models were segmented by time of follow-up (mortality >0–≤3 and >3–≤42 months of follow-up; loss to follow-up >0–≤12, >12–≤24 and >24–≤42 months of follow-up) and hazard proportionality assessed by analysis of scaled Schoenfeld residuals. We performed sensitivity analyses to assess the potential influence of a specific programme on the overall mortality estimate by leaving out each programme in separate regressions, and assessing the potential effect of misclassification of mortality among patients lost to follow-up using a competing risks framework . All reported P values are exact and two-tailed, and for each analysis P < 0.05 was considered significant. All analyses were carried out in STATA 10.1 and 11.0 (StataCorp, College Station, Texas, USA).
The analysis was based on routine clinical data; therefore, ethical review and individual patient consent were not sought. All patient information was entered into a database using coded identification numbers, and no information that could reveal patient identity was entered into the databases.
Characteristics of included population
We included 17 561 patients who were followed on ART for a median of 12 months, and a median ART start date of 30 October 2007. The majority (65%) were female, and 6672 (38%) were considered severely immunosuppressed at baseline (Table 1). Median age at ART initiation was 36.0 years (IQR 30.1–42.8). Most patients (65.8%) were aged less than 40 years, whereas 3995 patients (22.8%) were aged 40–49 years, 1566 patients (8.9%) were aged 50–59 years, 367 patients (2.1%) were aged 60–69 years and 78 patients (0.44%) were aged 70–94 years. The proportion of patients aged at least 50 years was similar across countries (Table 2).
More adult patients aged less than 50 years were female (66 vs. 51%; P < 0.001), had a diagnosis of TB at ART initiation (8 vs. 6%; P < 0.001), and were treated in an urban clinic setting (45 vs. 40%; P < 0.001) than patients aged at least 50 years (Table 1). Patients in both age groups were recorded as having similar rates of severe immunosuppression (38 vs. 37%; P = 0.14).
Immune restoration, mortality and loss to follow-up
Patients were followed for an average of 1 year (IQR 0.34–1.92 years) with patients aged less than 50 followed for slightly less time than those at least 50 years (1.00 vs. 1.08 years, P = 0.007). The crude mortality rate for patients aged at least 50 years was significantly higher than for those aged less than 50 years [rate ratio (RR) 1.22, 95% confidence interval (CI) 1.03–1.43, P = 0.016], whereas the rate of loss to follow-up was not significantly different (RR 1.07, 95% CI 0.96–1.18, P = 0.19). The median gain in CD4 cell count at 6 and 12 months was significantly higher in patients less than 50 years old compared with those at least 50 years (134 vs. 112 at 6 months; 170 vs. 139 at 12 months; both P < 0.001). Only 28 and 21% of patients overall (similar by age group) had CD4 cell count measurements at both baseline and 6 or 12 months, respectively, and although most variables differed between those who did and did not have CD4 cell count measurements, there was no significant difference in median age (results not shown).
Cox proportional hazards analysis found that all patients aged at least 40 years were at increased mortality risk from more than 3 to 42 months on ART after adjusting for sex, current TB, BMI, severe immunosuppression and clinic setting [40–49 years: adjusted hazard ratio (aHR) 1.59, 95% CI 1.29–1.96, P < 0.001; 50–59 years: aHR 1.58, 95% CI 1.19–2.11, P = 0.002; 60–69 years: aHR 2.63, 95% CI 1.66–4.16, P < 0.001; 70–94 years: aHR 3.64, 95% CI 1.51–8.77, P = 0.004; Table 3; Fig. 1a]. In addition, there was a significantly increased risk of mortality in those aged 70–94 years compared with patients aged 15–39 years during the first 3 months on ART (aHR 3.18, 95% CI 2.12–4.77; P < 0.001). Severe immunosuppression at ART initiation was most strongly associated with mortality (>0–3 months follow-up: aHR 2.19, 95% CI 1.83–2.63, P = 0.008; >3–42 months follow-up: aHR 1.83, 95% CI 1.52–2.21, P < 0.001), and each unit increase in BMI reduced risk of mortality (aHR 0.88 and 0.91 for >0–3 months and >3–42 months follow-up, respectively, P < 0.001). Rural clinic setting was protective against mortality in the first 3 months of ART (aHR 0.80, 95% CI 0.68–0.95, P = 0.009) but a risk thereafter (aHR 1.27, 95% CI 1.05–1.52, P = 0.013). Conversely, increasing age was not associated with risk of loss to follow-up in any of the periods of follow-up assessed after adjusting for sex, current TB, BMI, severe immunosuppression and clinic setting (Table 4; Fig. 1b). The most significant factor associated with loss to follow-up was the clinic being in a rural setting, with risk increasing with greater follow-up time (Table 4).
The association between increasing age and mortality in the adjusted model was robust to leaving out any single country (Table 2), and a regression that assessed the potential for loss to follow-up as a competing risk (subhazard ratio 1.15, 95% CI 1.08–1.21).
All of the models displayed proportional hazards except for loss to follow-up up until 12 months on ART (Schoenfeld residual P = 0.002, Table 4), primarily because of the BMI variable, even after applying a range of statistical contingencies to attempt to obtain proportionality, including categorization, transformation, removal of data outliers and assessment of possible interactions.
HIV/AIDS has resulted in a dramatic decline in life expectancy in sub-Saharan Africa,  but with ART, people living with HIV/AIDS can enjoy almost normal life expectancy . The neglect of older age groups in the HIV/AIDS response in sub-Saharan Africa is partly due to a perception that populations in the most affected countries are unlikely to achieve old age anyway and so a focus on younger patients represents a more reasonable public health approach. However, our study shows that older age groups represent an important proportion of the overall ART population. More than one in 10 people on ART in this study were found to be aged 50 years or over, a proportion consistent across most of the nine sub-Saharan Africa countries under review.
Previous studies have reported that people aged 50 or over were at higher risk of mortality compared with younger adults [5–7]. In this study, we found that mortality risk increased as age increased after 3 months on ART, suggesting that a simple age cut-off may not be sufficient for defining older groups at heightened risk of death.
In developed countries, the increased risk of death among HIV patients in older age groups is explained by the increased risk of development of non-AIDS-defining illnesses such as cardiovascular and liver disease, renal impairment and malignancies . A recent study from Botswana found increased risk of non-AIDS-defining illnesses among higher age groups, with standardized rates higher than reported in the United States . Although this is likely to explain some of the excess mortality observed in our cohorts, it is important to note that older age groups are also at greater risk of infectious diseases, particularly TB, as a result of declines in immunity associated with older age .
The reason for higher mortality among older people with HIV/AIDS is multifactorial, with the natural aging process, immune senescence, antiretroviral drug toxicities and HIV infection all suggested to play a role [18,19]. We found that patients aged at least 50 years had a significantly lower CD4 gain at 6 and 12 months on ART, a finding consistent with those from a large observational cohort study from 33 European cohorts which found that the probability of immunological response reduced in those aged 60 years or older . This gives further cause for concern about potential vulnerability to incident opportunistic infections and suggests a need for rapid ART initiation as soon as clinically eligible.
Our results are based on data pooled from different programmes in which HIV care was provided both as a disease-specific programme and integrated into general healthcare, in urban and rural areas and across a number of countries. This allows some confidence in the generalizability of our findings to other programmes in sub-Saharan Africa. Leaving out any single country did not importantly influence the role of age on mortality. We were able to control for a number of patient and programme-level variables, but cannot exclude the possibility of residual confounding. Despite various attempts to obtain proportionality, one of the Cox hazards models did not adhere to the assumption of proportional hazards; however, BMI was retained in the model as a clinically important factor related to poor outcomes. The likely effect of a nonproportional predictor is that the power of the test for this variable is reduced, such that any association may be missed, and the concurrent model predictors that do satisfy proportionality also suffer from decreased power associated with a poorer overall model fit . Finally, we were unable to report on a number of important variables due to the limits of routine programme reporting. In common with reports from other routine programmes in sub-Saharan Africa , CD4 cell counts were not available for a substantial number of patients. Thus, we are unable to report on the association between excess mortality and age-related decline in immune status; nevertheless, higher mortality at older age groups was still seen after controlling for a composite measure of immunosuppression combining CD4 cell count and WHO stage. Specific causes of death could also not be reported because this was not routinely reported and links to death registries were not established in these settings. We could not assess association of HIV status and expected risk of mortality with age in these settings, as data was from routine monitoring of dedicated HIV programmes.
In conclusion, this study provides further evidence that older age groups are a vulnerable group in ART programmes and suggests that, although it is important to identify groups in need of more careful follow-up and earlier initiation on ART, the application of crude age categories may result in patients at higher risk being overlooked. Future research is now needed to understand the reasons for higher mortality in older age groups, and defining simple interventions that are feasible in highly under-resourced settings to allow for adapted follow-up and care approaches for older age groups.
The authors would like to thank Leslie Shanks for encouraging comments on an earlier draft.
Conflicts of interest
There are no conflicts of interest.
1. Bhaskaran K, Hamouda O, Sannes M, Boufassa F, Johnson AM, Lambert PC, et al. Changes in the risk of death after HIV seroconversion compared with mortality in the general population. JAMA 2008; 300:51–59.
2. Herbst AJ, Cooke GS, Barnighausen T, KanyKany A, Tanser F, Newell ML. Adult mortality and antiretroviral treatment roll-out in rural KwaZulu-Natal, South Africa. Bull World Health Organ 2009; 87:754–762.
3. Davies MA, Egger M, Keiser O, Boulle A. Paediatric antiretroviral treatment programmes in sub-Saharan Africa: a review of published clinical studies. Afr J AIDS Res 2009; 8:329–338.
4. Bakanda C, Birungi J, Mwesigwa R, Nachega JB, Chan K, Palmer A, et al. Survival of HIV-infected adolescents on antiretroviral therapy in Uganda: findings from a nationally representative cohort in Uganda. PLoS One 2011; 6:e19261.
5. Negin J, Wariero J, Cumming RG, Mutuo P, Pronyk PM. High rates of AIDS-related mortality among older adults in rural Kenya. J Acquir Immune Defic Syndr 2010; 55:239–244.
6. Boulle A, Van Cutsem G, Hilderbrand K, Cragg C, Abrahams M, Mathee S, et al. Seven-year experience of a primary care antiretroviral treatment programme in Khayelitsha, South Africa. AIDS 2010; 24:563–572.
7. Bakanda C, Birungi J, Mwesigwa R, Ford N, Cooper CL, Au-Yeung C, et al. Association of aging and survival in a large HIV-infected cohort on antiretroviral therapy. AIDS 2011; 25:701–705.
8. Mutevedzi PC, Lessells RJ, Rodger AJ, Newell ML. Association of age with mortality and virological and immunological response to antiretroviral therapy in rural South African adults. PLoS One 2011; 6:e21795.
9. Negin J, van Lettow M, Semba M, Martiniuk A, Chan A, Cumming R. Antiretroviral treatment outcomes among older adults in Zomba District, Malawi.PLOS One 2011; 6:e26546.
10. Mills EJ, Rammohan A, Awofeso N. Ageing faster with AIDS in Africa. Lancet 2011; 377:1131–1133.
11. FUCHIA: a free computer program for the monitoring of HIV/AIDS medical care at the population level. In: International Conference on AIDS; July 2002; Epicentre, Paris, France. [Abstract C11029]
12. Putter H, Fiocco M, Geskus RB. Tutorial in biostatistics: competing risks and multistate models. Stat Med 2007; 26:2389–2430.
13. Cotton D. Life expectancy in Africa: back to the future?. Ann Intern Med 2011; 155:265–266.
14. Mills EJ, Bakanda C, Birungi J, Chan K, Ford N, Cooper CL, et al. Life expectancy of persons receiving combination antiretroviral therapy in low-income countries: a cohort analysis from Uganda. Ann Intern Med 2011; 155:209–216.
15. Sackoff JE, Hanna DB, Pfeiffer MR, Torian LV. Causes of death among persons with AIDS in the era of highly active antiretroviral therapy: New York City. Ann Intern Med 2006; 145:397–406.
16. Wester CW, Koethe JR, Shepherd BE, Stinnette SE, Rebeiro PF, Kipp AM, et al. Non-AIDS-defining events among HIV-1-infected adults receiving combination antiretroviral therapy in resource-replete versus resource-limited urban setting. AIDS 2011; 25:1471–1479.
17. Schaaf HS, Collins A, Bekker A, Davies PD. Tuberculosis at extremes of age. Respirology 2010; 15:747–763.
18. Effros RB, Fletcher CV, Gebo K, Halter JB, Hazzard WR, Horne FM, et al.Aging and infectious diseases: workshop on HIV infection and aging: what is known and future research directions. Clin Infect Dis 2008; 47:542–553.
19. Grabar S, Kousignian I, Sobel A, Le Bras P, Gasnault J, Enel P, et al.Immunologic and clinical responses to highly active antiretroviral therapy over 50 years of age. Results from the French Hospital Database on HIV.AIDS 2004; 18:2029–2038.
20. Sabin CA, Smith CJ, d’Arminio Monforte A, Battegay M, Gabiano C, Galli L, et al. Collaboration of Observational HIV Epidemiological Research Europe (COHERE) Study GroupResponse to combination antiretroviral therapy: variation by age. AIDS 2008; 22:1463–1473.
21. Bellera CA, MacGrogan G, Debled M, de Lara CT, Brouste V, Mathoulin-Pelissier S. Variables with time-varying effects and the Cox model: some statistical concepts illustrated with a prognostic factor study in breast cancer. BMC Med Res Methodol 2010; 10:20.
22. Keiser O, Anastos K, Schechter M, Balestre E, Myer L, Boulle A, et al. Antiretroviral therapy in resource-limited settings 1996 to 2006: patient characteristics, treatment regimens and monitoring in sub-Saharan Africa, Asia and Latin America. Trop Med Int Health 2008; 13:870–879.
antiretroviral therapy; elderly; HIV/AIDS; older adults; survival
© 2012 Lippincott Williams & Wilkins, Inc.
Highlight selected keywords in the article text.