Original article Life expectancy after initiation of combination antiretroviral therapy in Thailand

Transcription

1 Antiviral Therapy 2017; 22: (doi: /IMP3121) Original article Life expectancy after initiation of combination antiretroviral therapy in Thailand Sirinya Teeraananchai 1,2 *, Suchada Chaivooth 3, Stephen J Kerr 1,2,4, Sorakij Bhakeecheep 3, Anchalee Avihingsanon 1,5, Achara Teeraratkul 6, Petchsri Sirinirund 7, Matthew G Law 2, Kiat Ruxrungtham 1,5 1 HIV-NAT, Thai Red Cross AIDS Research Centre, Bangkok, Thailand 2 Kirby Institute, University of New South Wales, Sydney, Australia 3 The HIV/AIDS, Tuberculosis and Infectious Diseases Program, National Health Security Office (NHSO), Bangkok, Thailand 4 Department of Global Health, Academic Medical Center, University of Amsterdam, Amsterdam Institute of Global Health and Development, Amsterdam, the Netherlands 5 Department of Medicine, Faculty of Medicine, Chulalongkorn University, Bangkok, Thailand 6 Thailand MOPH - US CDC Collaboration, Nonthaburi, Thailand 7 Ministry of Public Health, Nonthaburi, Thailand *Corresponding author sirinya.t@hivnat.org Background: Access to combination antiretroviral therapy (cart) has decreased mortality in HIV-positive people. We aimed to estimate the expected additional years of life in HIV-positive Thai people after starting cart through the National AIDS Program (NAP), administered by the Thai National Health Security Office (NHSO). Methods: The NHSO database collects characteristics of all Thai HIV-infected patients through the National AIDS Program, including linkage with the National Death Registry for vital status. This study included patients aged 15 years at cart initiation between 2008 and The abridged life table method was used to construct life tables stratified by sex and baseline CD4 + T-cell count. Life expectancy was defined as the additional years of life from age at starting cart. Results: 201,688 eligible patients were included in analyses, contributing 618,837 person-years of followup. Median CD4 + T-cell count was 109 cells/mm 3 and median age 37 years. The overall life expectancy after cart initiation at age 20 was 25.4 (95% CI, 25.3, 25.6) years and 20.6 (95% CI, 20.5, 20.7) at age 35 years. Life expectancy at baseline CD4 + T-cell count 350 cells/mm 3 was 51.9 (95% CI, 51.0, 52.9) years for age 20 years and 43.2 (95% CI, 42.4, 44.1) years for age 35 years, close to life expectancy in the general Thai population. Conclusions: Increasing life expectancy with higher baseline CD4 + T-cell counts supports the guideline recommendations to start cart irrespective of CD4 + T-cell count. These results are beneficial to forecast the treatment cost and develop health policies for people living with HIV in Thailand and Asia. Introduction Expanded access to combination antiretroviral therapy (cart) has led to large reductions in morbidity and mortality, enabling HIV-positive people to continue to contribute productively to their families, communities and economies [1]. Quantitating life expectancy is commonly used to measure how developments in treatment and clinical care programmes have improved the lives of people living with HIV/AIDS (PLWHIV) [2]. In the USA, the average life expectancy after an HIV diagnosis increased from 10.5 to 22.5 years from 1996 to 2005 [3]. The Antiretroviral Therapy Cohort Collaboration reported that crude mortality rates decreased from 16.3 deaths per 1,000 person-years between 1996 and 1999 to 10.0 deaths per 1,000 person-years between 2003 and 2005, and life expectancy at age 20 years increased from 36.1 years to 49.4 years [4]. In the UK, life expectancy in patients who started cart at age 20 years with CD4 + T-cell counts <350 cells/mm 3 increased from 30.0 years to 45.8 years between 1996 and 2008 [5]. Similar improvements, with projected economic benefits have been noted in resource-limited settings [6,7]. In a cohort of 22,315 subjects initiating cart in Uganda, investigators reported favourable life expectancy compared with the national average, 2017 International Medical Press (print) (online) 393

2 S Teeraananchai et al. with an additional 27.9 years expected when starting cart at age 35 years. Similar increases were noted in a Rwandan cohort with an expected additional 29.9 years when starting cart at 20 years compared to a life expectancy of 51 years for the general population [8]. Most population-based studies of life expectancy of PLWHIV have been conducted in the USA, Europe and sub-saharan Africa. There are currently no published estimates of life expectancy in Asia where absolute numbers of people living with HIV is also high. Thailand was the first Asian country to document cases of HIV [9] and currently has the highest prevalence in the region, with approximately 1% of the population affected [10]. Since 2008, Thailand has provided universal coverage for HIV Treatment through the National AIDS Program (NAP) administered by the National Health Security Office (NHSO). The aim of our study was to assess the impact of cart initiation on life expectancy in HIV-positive Thai patients, since the implementation of this universal coverage programme. Methods Patient population Thailand s NHSO is an administrative database used for recording clinical and laboratory investigations and drug treatment provided under the Thailand Universal Coverage Program. It consists of the pilot National Access to ARV for People Living with HIV/AIDS (NAPHA) which then expanded to NAP. The scheme provides free cart with CD4 + T-cell count testing and safety laboratory results twice a year. HIV RNA testing is performed 6 months after initiating cart until it is undetectable and it is done annually thereafter. The hospitals and clinical sites that participate in the programme must enter laboratory results, clinical information and cart dispensing records into the NHSO electronic database to be reimbursed from NHSO for the costs incurred. The system has been linked with the National Death Registry twice a month since 2008, using the Thai National Identity number to confirm vital status of PLWHIV treated through the NHSO system. Data was extracted for patients aged 15 years at cart initiation between January 2008 and 14 November Patients who started cart and died on the same date were excluded (n=41). The last data update from the National Death Registry for our analysis was on 5 November Lost to follow up (LTFU) was defined as not attending a scheduled clinic visit for 6 months from their previous visit. Patients who did not die were censored on the date of the last Death Registry data transfer. Patients who died were imputed as failures on their death date. Baseline was defined as date of cart initiation. Baseline CD4 + T-cell count was measured within a 6-month window of cart initiation; when more than one result fell within the window period, the measurement closest to the date of cart initiation was used. cart was defined as a multiclass regimen containing at least three drugs, including a non-nucleoside reverse transcriptase inhibitor (NNRTI) or protease inhibitor (PI) and 2 3 nucleoside reverse transcriptase inhibitors (NRTI). The study was approved by the Institutional Review Board (IRB) of the Institute for Development of Human Research Protection, Ministry of Public Health. The corresponding author was given access to de-identified study data to perform the analysis. Statistical analysis Patient characteristics at baseline were reported using descriptive statistics. Mortality rates were calculated by dividing the total number of deaths by the total number of person-years of follow-up, and expressed as deaths per 100 person-years. The Cox proportional hazards model and Kaplan Meier methods were used to assess significant predictors of mortality during follow-up. Covariates included age at cart, sex, year of starting cart, baseline CD4 + T-cell count, first cart regimen and LTFU. Variables with a P-value of less than 0.10 in the univariate analysis were include in multivariate models and a P-value <0.05 was considered statistically significant. Abridged life tables, constructed from age-specific mortality rates using Chiang s method [11], were used to predict life expectancy. These abridged life tables depicted the lifetime mortality experience of a single cohort of persons based on the mortality schedules on which the table was constructed. Life expectancy calculated by this method represents the average number of additional years of life which was expected to survive after starting cart at a particular age. We calculated life expectancy from age 15 to 55 years in 5-year intervals with 95% CIs, overall and stratified by sex, baseline CD4 + T-cell count status and year of cart initiation. The estimated life expectancy by CD4 + T-cell thresholds (<50 versus 50, <150 versus 150, <200 versus 200, <250 versus 250 and <350 versus 350) were also analysed to evaluate the additional years of life according to treatment guidelines. Since we noted high mortality rates during the first 6 months after ART initiation, and fewer of person-years of follow-up for patients who initiated ART in more recent years of the study period, we also calculated life expectancy of those who survived at least 6 months after starting cart as a sensitivity analysis. Data were managed and analysed using SAS software, version 9.4 (SAS Institute, Cary, NC, USA) and Stata 14 (Stata Corp, College Station, TX, USA) International Medical Press

3 Life expectancy after cart in Thailand Results Characteristics A total of 211,488 patients started cart from 2008, the year from which linkage to the death registry was available. Of these, 201,688 patients initiated ART at aged 15 years (Additional file 1). Characteristics at ART initiation are shown in Table 1. Median age at baseline was 37 (IQR 31 43) years, 108,569 (54%) patients were female. 92% were treated with NNRTIbased cart and 15,035 (7%) with PI-based cart. For HIV clinical stage, 27% of them (n=54,891) were Table 1. Characteristics at cart initiation Characteristics n=201,688 Female gender 108,569 (54) HIV clinical stage Asymptomatic HIV 54,891 (27) Symptomatic HIV 32,643 (16) AIDS 39,643 (20) Unknown 74,511 (37) First cart regimen NNRTI-based cart 186,323 (92) PI-based cart 15,035 (7) Others (3 NRTIs, PI+ NNRTI+NRTI) 330 (1) Median age at starting cart, years (IQR) 37 (31 43) Age groups at starting cart years 4,373 (2) years 13,032 (6) years 25,059 (13) years 40,376 (20) years 44,229 (22) years 34,533 (17) years 20,603 (11) years 10,724 (5) 55+ years 8,759 (4) Year of cart initiation ,639 (15) ,027 (13) ,521 (14) ,918 (14) ,699 (16) ,132 (17) ,752 (11) CD4 + T-cell count at baseline 127,468 (63) <50 cells/mm 3 41,518 (21) 50 <100 cells/mm 3 19,533 (10) 100 <150 cells/mm 3 13,562 (7) 150 <250 cells/mm 3 24,315 (12) 250 <350 cells/mm 3 15,051 (7) 350 cells/mm 3 13,489 (7) Unknown 74,220 (36) Median baseline CD4 + T-cell count, cells/mm 3 (IQR) 109 (33 232) Data are n (%) unless otherwise indicated. cart, combination antiretroviral therapy; NNRTI, non-nucleoside reverse transcriptase inhibitor; NRTI, nucleoside reverse transcriptase inhibitor; PI, protease inhibitor. asymptomatic, while 39,643 (20%) patients had clinical AIDS at cart initiation. Of the 127,468 (63%) patients with a baseline CD4 + T-cell count in the defined window period, the median baseline CD4 + T-cell count was 109 (IQR ) cells/mm 3. A large proportion of patients started ART with very advanced disease, with 21% (n=41,518) having a baseline CD4 + T-cell count less than 50 cells/mm 3. Over a total of 618,837 person-years of follow-up, there were 22,021 (11%) deaths and 23,964 (12%) LTFU. Of those who were LTFU, 3,377/23,964 (14%) patients died. The crude mortality rate was 3.56 (95% CI, 3.51, 3.61) deaths per 100 person-years and the LTFU rate was 3.87 (95% CI, 3.82, 3.92) per 100 person-years. The median duration on cart was 2.8 (IQR, ) years. Viral load was available before patients were censored in 152,740 (61%). Of those with a last viral load available, 126,765 (83%) patients had viral load undetectable (<40 copies/ml). Last CD4 + T-cell count was available in 191,029 (95%) patients with a median (IQR) of 367 cells/mm 3 ( ; Additional file 2). The multivariate Cox regression model showed that men were at a higher risk of death than women. Compared to the reference group of patients aged years, there was a decreased risk of death for younger patients but it was an increased risk of death for patients aged 35 years. Lower CD4 + T-cell counts were associated with higher mortality risk, as were people starting cart earlier in the study period. Furthermore, patients who were lost to follow-up had a significantly higher risk of mortality (Table 2). The Kaplan Meier estimates of death by age and baseline CD4 + T-cell count showed that patients who started ART at older ages and patients with a lower CD4 + T-cell count had the highest risk of mortality during first 6 months of cart initiation (Additional files 3 and 4). Life expectancy From , overall life expectancy after cart initiation was estimated to be 25.4 (95% CI, 25.3, 25.6) years for a patient who started cart at age 20 years and 20.6 (95% CI, 20.5, 20.7) years for a patient who started cart at age 35 years. Men were estimated to have shorter life expectancy than women. Life expectancy in a man who started cart at age 20 years was 22.1 (95% CI, 21.9, 22.3) years and it was 30.6 (95% CI, 30.4, 30.8) years for a woman; in a man who started cart at age 35 years life expectancy was 17.9 (95% CI, 17.8, 18.0) years and it was 25.4 (95% CI, 25.2, 25.5) years for a woman. We found that life expectancy increased substantially with increasing baseline CD4 + T-cell counts (Table 3). For a patient who started cart at CD4 + <50 cells/mm 3, the estimated life expectancy was 16.1 (95% CI, 15.8, 16.3) years and it was 13.4 Antiviral Therapy

4 S Teeraananchai et al. Table 2. Cox proportional hazards model of death after starting cart (baseline) Univariate Multivariate Characteristics (n=201,688) HR (95% CI) P-value ahr (95% CI) P-value Male (versus female) 1.60 (1.55, 1.64) < (1.36, 1.44) <0.001 Age at baseline (years) <0.001 < (0.63, 0.80) 0.87 (0.77, 0.99) (0.73, 0.84) 0.80 (0.75, 0.86) (0.91, 1.01) 0.94 (0.90, 0.99) Ref Ref (1.03, 1.12) 1.09 (1.04, 1.13) (1.07, 1.17) 1.15 (1.10, 1.21) (1.18, 1.30) 1.30 (1.24, 1.37) (1.42, 1.59) 1.60 (1.51, 1.70) (1.92, 2.16) 2.21 (2.09, 2.35) Baseline CD4 + T-cell count <0.001 <0.001 <50 cells/mm (4.97, 6.06) 5.18 (4.69, 5.73) 50 <100 cells/mm (3.88, 4.76) 4.01 (3.62, 4.46) 100 <150 cells/mm (2.98, 3.69) 3.08 (2.76, 3.43) 150 <250 cells/mm (2.01, 2.49) 2.14 (1.93, 2.39) 250 <350 cells/mm (1.33, 1.70) 1.48 (1.31, 1.68) 350 cells/mm 3 Ref Ref Unknown 2.54 (2.30, 2.81) 2.36 (2.14, 2.61) First regimen <0.001 <0.001 NNRTI-based HAART 2.94 (2.70, 3.21) 2.20 (2.01, 2.40) PI-based HAART Ref Ref Others (3 NRTIs, NNRTI+PI+NRTI) 2.69 (1.91, 3.79) 2.16 (1.53, 3.04) Year of cart initiation <0.001 < (0.97, 1.07) 1.06 (1.01, 1.11) (1.01, 1.10) 1.02 (0.98, 1.07) (0.96, 1.05) 0.99 (0.95, 1.04) 2011 Ref Ref (0.80, 0.89) 0.84 (0.80, 0.88) (0.77, 0.86) 0.78 (0.74, 0.82) (0.70, 0.82) 0.74 (0.69, 0.80) Lost to follow-up 3.26 (3.15, 3.39) < (4.54, 4.90) <0.001 cart, combination antiretroviral therapy; NNRTI, non-nucleoside reverse transcriptase inhibitor; NRTI, nucleoside reverse transcriptase inhibitor; PI, protease inhibitor. (95% CI, 13.3, 13.5) years at ages 20 and 35 years, respectively. This increased to a life expectancy of 51.9 (95% CI, 51.0, 52.9) years in patients who started cart at CD4 + T-cell count of 350 cells/mm 3 at age 20 years and it was 43.2 (95% CI, 42.4, 44.1) years at age 35 years. The increased life expectancy according to higher CD4 + T-cell counts at cart initiation is illustrated across various CD4 + T-cell thresholds for cart initiation at age 20 and age 35 years in Figure 1. When estimating life expectancy by year of cart, we found that there was decreasing trend with a shorter life expectancy more recently (Table 3 and Additional file 5). We believe this apparent short life expectancy more recently is an artefact generated by the high mortality in the first 6 months following cart initiation (Additional files 3 and 4) combined with shorter followup thereafter. Sensitivity analyses, where we included only patients who survived more than 6 months after cart initiation, were summarized in Table 4. Excluding early deaths after cart increased overall life expectancy in a patient who started cart at age 20 to 38.4 (95% CI, 38.2, 38.6) years. We also found that life expectancy increased with more recent cart initiation: life expectancy in a person starting cart at age 20 years increased by nearly 7 years from to There was a strong increasing trend of additional years of life across the three periods assessed. Similar increases in life expectancy in women compared to men and with increasing CD4 + T-cell counts at cart initiation were also seen in this analysis. Discussion This is the first study in Thailand and from Asia to estimate the additional years of life in a National Treatment Program after starting cart, adjusted for International Medical Press

5 Life expectancy after cart in Thailand Table 3. Life expectancy (95% CI) of all HIV-infected patients from (n=201,688) Age, years years years years years years years years years years a,b General Thai life expectancy (2013 estimates) [12] Overall Males Females Age at cart initiation, years n 4,373 13,032 25,059 40,376 44,229 34,533 20,603 10,724 8,759 Overall 27.1 (26.9, 27.4) 25.4 (25.3, 25.6) 23.7 (23.6, 23.8) 22.2 (22.1, 22.3) 20.6 (20.5, 20.7) 18.9 (18.8, 19.0) 17.2 (17.1, 17.3) 15.6 (15.5, 15.7) 14.4 (NA) Men 23.1 (22.7, 23.5) 22.1 (21.9, 22.3) 20.5 (20.4, 20.7) 19.2 (19.1, 19.3) 17.9 (17.8, 18.0) 16.4 (16.3, 16.5) 14.8 (14.7, 14.9) 13.4 (13.3, 13.5) 12.2 (NA) Women 32.4 (32.1, 32.7) 30.6 (30.4, 30.8) 28.8 (28.7, 29.0) 27.3 (27.1, 27.5) 25.4 (25.2, 25.5) 23.5 (23.3, 23.6) 21.6 (21.4, 21.7) 19.6 (19.5, 19.8) 18.3 (NA) Year of cart initiation (27, 27.9) 27.4 (27.1, 27.6) 26.4 (26.2, 26.6) 25.1 (24.9, 25.2) 23.4 (23.2, 23.5) 21.6 (21.5, 21.8) 19.8 (19.6, 19.9) 18.1 (17.9, 18.2) 16.9 (NA) (27.6, 28.2) 25.7 (25.4, 25.9) 23.6 (23.4, 23.8) 22.1 (21.9, 22.3) 20.6 (20.4, 20.8) 19.1 (18.9, 19.3) 17.5 (17.3, 17.6) 16 (15.9, 16.2) 14.8 (NA) (21.8, 22.8) 19.7 (19.4, 20.0) 17.3 (17.1, 17.6) 16 (15.8, 16.3) 14.5 (14.3, 14.7) 13.4 (13.2, 13.6) 12.1 (11.9, 12.3) 10.7 (10.5, 10.9) 9.6 (NA) Baseline CD4 + T-cell count <50 cells/mm (15.1, 16.1) 16.1 (15.8, 16.3) 15.5 (15.3, 15.6) 14.6 (14.5, 14.7) 13.4 (13.3, 13.5) 12.3 (12.2, 12.4) 10.9 (10.8, 11.1) 9.6 (9.5, 9.8) 8.5 (NA) 50 <100 cells/mm 3 21 (20.2, 21.8) 20.1 (19.7, 20.5) 18.7 (18.4, 19) 17.5 (17.3, 17.7) 15.9 (15.7, 16.1) 14.3 (14.1, 14.5) 12.6 (12.4, 12.9) 10.9 (10.7, 11.1) 9.3 (NA) 100 <150 cells/mm (24.9, 26.9) 25.2 (24.7, 25.8) 23.7 (23.4, 24.1) 21.5 (21.1, 21.8) 19.8 (19.5, 20.1) 18.0 (17.7, 18.3) 16.3 (16.0, 16.6) 14.6 (14.4, 14.9) 13.6 (NA) 150 <250 cells/mm (35.0, 36.3) 32.6 (32.1, 33) 30.5 (30.1, 30.8) 28.4 (28.1, 28.7) 26.1 (25.8, 26.4) 24.0 (23.7, 24.2) 22.2 (21.9, 22.5) 20.3 (20.1, 20.6) 18.7 (NA) 250 <350 cells/mm (39.6, 41.2) 37.5 (37, 38.1) 34.0 (33.5, 34.6) 31.1 (30.6, 31.6) 28.3 (27.9, 28.8) 25.6 (25.1, 26) 23.5 (23.0, 23.9) 21.3 (20.9, 21.7) 19.5 (NA) 350 cells/mm (55.2, 57.2) 51.9 (51, 52.9) 48.4 (47.5, 49.3) 45.9 (45.1, 46.8) 43.2 (42.4, 44.1) 40.9 (40.1, 41.8) 38.8 (37.9, 39.7) 36.7 (35.9, 37.6) 37.4 (NA) Unknown 31.0 (30.5, 31.4) 30.0 (29.7, 30.2) 28.4 (28.2, 28.6) 26.8 (26.6, 26.9) 25.0 (24.8, 25.2) 23.1 (23, 23.3) 21.1 (20.9, 21.3) 19.3 (19.2, 19.5) 18.1 (NA) a Life expectancy of general population at age years was only estimated from 55 to 59 years. b Age at combination antiretroviral therapy (cart) initiation is 55+ and 95% CI is not available because it is not a closed interval. NA, not applicable. Antiviral Therapy

6 S Teeraananchai et al. Figure 1. Average life expectancy by baseline CD4 + T-cell count thresholds from 20 and 35 years of age Additional year of life, years <50 50 < < < < Baseline CD4 + T-cell count (cells/mm 3 ) from 20 years of age Life expectancy of Thai HIV patients (n=12,957) Life expectancy after surviving beyond 6 months (n=9,435) Additional year of life, years <50 50 < < < < Baseline CD4 + T-cell count (cells/mm 3 ) from 35 years of age Life expectancy of Thai HIV patients (n=43,951) Life expectancy after surviving beyond 6 months (n=37,325) sex and CD4 + T-cell counts. Our analysis of 201,688 HIV-infected Thai patients initiating cart between 2008 and 2014, with median CD4 + T-cell count at baseline of 109 (IQR ) cells/mm 3, in the Thai National AIDS program clearly demonstrates that life expectancy improves with increasing CD4 + T-cell counts, and as the programme has evolved over time. In this analysis, life expectancy was higher among women than men and it was also higher among patients who initiated cart at CD4 + T-cell counts 350 cells/mm International Medical Press

7 Life expectancy after cart in Thailand Table 4. Life expectancy (95% CI) of HIV-infected patients who survived beyond 6 months after starting cart (n=166,256) Age at cart initiation years years years years years years years years 55+ years a n 2,959 9,496 20,242 34,071 37,549 29,059 17,165 8,784 6,931 Overall 39.5 (39.1, 39.8) 38.4 (38.2, 38.6) 36.7 (36.6, 36.9) 35.0 (34.9, 35.2) 32.9 (32.8, 33.1) 30.9 (30.7, 31) 28.6 (28.5, 28.8) 26.7 (26.6, 26.8) 25.3 (NA) Men 34.4 (33.9, 34.9) 33.6 (33.3, 33.9) 31.6 (31.4, 31.8) 29.8 (29.6, 30) 28.0 (27.8, 28.2) 26.2 (26, 26.3) 24.1 (23.9, 24.2) 22.3 (22.1, 22.5) 21.1 (NA) Women 47.7 (47.1, 48.2) 46.9 (46.5, 47.2) 45.9 (45.7, 46.2) 44.4 (44.1, 44.7) 42.1 (41.8, 42.3) 39.7 (39.4, 39.9) 37.1 (36.9, 37.4) 34.9 (34.6, 35.1) 33.2 (NA) Year of cart initiation (35.6, 36.8) 36.5 (36.2, 36.8) 35.7 (35.4, 35.9) 33.9 (33.7, 34.1) 31.6 (31.4, 31.8) 29.4 (29.2, 29.6) 27.1 (26.9, 27.3) 24.9 (24.8, 25.1) 23.4 (NA) (40, 41.1) 39.0 (38.7, 39.4) 37.0 (36.7, 37.3) 35.4 (35.2, 35.7) 33.6 (33.3, 33.8) 31.7 (31.4, 32) 29.6 (29.4, 29.9) 28.0 (27.8, 28.3) 26.9 (NA) (44.5, 47) 43.3 (42.5, 44.2) 40.4 (39.7, 41.1) 37.9 (37.3, 38.5) 35.7 (35.2, 36.3) 33.5 (32.9, 34.1) 31.4 (30.9, 32) 29.4 (29, 29.9) 27.4 (NA) Baseline CD4 + T-cell count <50 cells/mm (27.4, 29.2) 29.4 (28.9, 29.8) 29.0 (28.7, 29.3) 28.0 (27.8, 28.3) 26.4 (26.1, 26.7) 24.9 (24.6, 25.2) 22.8 (22.5, 23.1) 21.4 (21.1, 21.7) 20.3 (NA) 50 <100 cells/mm (30.8, 33.1) 31.2 (30.6, 31.7) 28.9 (28.5, 29.3) 27.2 (26.9, 27.6) 25.0 (24.7, 25.3) 22.7 (22.4, 23.0) 20.1 (19.8, 20.4) 17.8 (17.5, 18.0) 15.8 (NA) 100 <150 cells/mm (34.9, 37.6) 35.3 (34.5, 36.1) 34.0 (33.5, 34.6) 31.4 (30.9, 31.9) 29.4 (28.9, 29.9) 27.5 (27.0, 27.9) 25.4 (24.9, 25.9) 23.3 (22.9, 23.7) 21.8 (NA) 150 <250 cells/mm (44.5, 46.2) 42.6 (42.0, 43.2) 40.4 (40.0, 40.9) 38.2 (37.7, 38.6) 35.5 (35.1, 35.9) 33.1 (32.7, 33.5) 30.9 (30.6, 31.3) 28.7 (28.3, 29.0) 27.1 (NA) 250 <350 cells/mm (47.7, 50) 46.5 (45.7, 47.2) 42.6 (41.9, 43.3) 39.0 (38.4, 39.7) 35.8 (35.2, 36.4) 32.8 (32.2, 33.4) 30.2 (29.6, 30.8) 28.1 (27.6, 28.6) 26.4 (NA) 350 cells/mm (63.2, 65.8) 60.8 (59.6, 62) 57.8 (56.7, 58.9) 55.2 (54.2, 56.2) 52.0 (51.0, 53.0) 49.3 (48.3, 50.4) 46.4 (45.4, 47.4) 44.2 (43.3, 45.1) 43.8 (NA) Unknown 44.2 (43.6, 44.9) 44.2 (43.8, 44.5) 42.7 (42.4, 43.0) 41.0 (40.7, 41.3) 39.0 (38.7, 39.2) 36.9 (36.6, 37.2) 34.7 (34.5, 35.0) 32.8 (32.5, 33.0) 31.6 (NA) a 95% CI is not available because it is not a closed interval. cart, combination antiretroviral therapy; NA, not applicable. Antiviral Therapy

8 S Teeraananchai et al. Given the well-known high mortality rate during the first 6 months of cart, we also estimated life expectancy conditional on patients surviving beyond this 6 month period. We found that life expectancy at age 20 years increased by 50% with any baseline CD4 + T-cell cell count and again it substantially increased over the study period. Conditional on surviving the first 6 months of cart, we estimate that a 20-year old HIVpositive Thai women had life expectancy around 84% of the general Thai female population (67.0 versus 80.0 years) and life expectancy in HIV-positive men was around 73% of the general male population (53.6 versus 73.2 years) [12]. In recent years, Thailand has rapidly scaled up to access cart and continuously expanded access to treatment. Prior to 2006, only 50% of PLWHIV in the NHSO system received cart. After 2008, over 70% of patients enrolled in this system initiated cart during the study period [13]. We believe that it is mainly due to the roll out of effective treatments and care provided through the Universal Coverage programme. The higher baseline CD4 + T-cell cutoff is eligible to initiate cart in the country. A baseline CD4 + T-cell count of <200 cells/mm 3 was used prior to 2010 to be eligible for cart and this CD4 + T-cell count cutoff was increased to 350 in From 2015 onwards, cart was recommended for all PLWHIV regardless of CD4 + T-cell counts [14]. The overall life expectancy in Thailand was more similar to estimates from other developing countries in Africa namely Uganda [7], Rwanda [8] and South Africa [15]. When we stratified by baseline CD4 + T-cell count, Thai HIV life expectancy for those starting cart at CD4 + T-cell count of 250 cell/mm 3 aged 20 was 37 years in our cohort and a Ugandan cohort [7], but for those starting ART the life expectancy was 31 years in the Ugandan cohort but only 28 years in our cohort. In contrast, our estimated life expectancies for patients starting ART at CD4 + T-cell counts at 350 cells/mm 3 aged 20 and 35 years were 9 and 10 years higher respectively, than those from Rwanda [8]. Increasing life expectancy with CD4 + T-cell count and time trends were also seen in the same way. Furthermore, the increase in life expectancy over time that we observed has also been seen in high-income countries [4,5] and in a recent meta-analysis including low middle and high income countries [16]. This latter meta-analysis also found significantly different life expectancies between males and females in low-middle income countries. Life expectancy of men in our study who started cart at 20 years was 22.1 years, and 22.9 years in the meta-analysis, but for those aged 35 years, life expectancy was 17.9 years in our study and 22.3 years in the meta-analysis. In our study, life expectancy for women starting cart at age 20 and 35 years were lower by 3 and 4 years respectively, compared to the combined estimates from the women in the low-middle income countries in the meta-analysis [16]. Our study has a number of limitations relating to the method of estimating life expectancy. First, the method is based on the number of deaths and person years of follow-up classified by age groups. High death rates in the first 6 months after cart initiation, combined with limited follow-up for patients who started cart in recent years ( ) meant that life expectancy could not be reliably estimated. To partially mitigate this limitation, we also estimated life expectancy for patients who survived for at least 6 months in a sensitivity analyses. Second, the method also does not control for behavioural factors that influence life expectancy, such as smoking or alcohol consumption, exercise and cart adherence. Since this analysis is based on data collected through an administrative database, these variables were not recorded for analysis, and the effect of these behavioural factors on life expectancy in HIV-positive Thai people is not known. Third, this analysis included all patients starting cart, all disease stages including some with unknown baseline CD4 + T-cell count so that overall estimates of life expectancy could be imprecise. Nevertheless, our Cox regression models provided relative survival estimates taking into account these unknown covariate values. Last, there is no information for patients who transferred to private hospitals or who may have moved overseas in our database. We believe that the number of patients is low and these patients would have been imputed as LFTU after they missed a clinic visit for >6 months. There were 3,377 (2%) patients who died after LTFU in this study. A key strength of this analysis is confirmation of vital status through linkage with the National Death Registry. According to Thai law, the household is required to report the death of a family member to the local or district registration office with 24 h or 7 days in remote areas. In recent report, the completeness of death registration was 95% in 2005, and had improved annually since the Law was enacted in 1991 [17]. The vital status was ascertained even if a small proportion may have moved overseas or transferred to private health care. The accuracy of this linkage is supported by the adjusted hazard ratios for mortality for those patients who were lost to follow-up, which as expected showed increased mortality in these patients. In addition, the large number of patients participating in the NAP, and the prospective and systematic collection of data result in precise estimates of life expectancy in an Asian context when compared with the number of patients from the Ugandan [7] and Rwandan [8] cohorts which were smaller. In conclusion, we found strong evidence of increasing life expectancy with increasing baseline CD International Medical Press

9 Life expectancy after cart in Thailand T-cell counts, which support WHO guidelines and Thai National Treatment guidelines to start cart irrespective of CD4 + T-cell count recently. Continually expanding access to cart not only has individual benefits and economic benefits for Thailand by increasing life expectancy, but also has public health benefits by reducing HIV transmission. Our results will be used to forecast the cost of the treatment programme and develop health policy for the HIV population in Thailand as they live longer. Acknowledgements The dataset was provided by the Thai NHSO and the Ministry of Health in Thailand. The Kirby Institute is funded by the Australian Government Department of Health and is affiliated with the Faculty of Medicine, UNSW Australia. The content of this publication is solely the responsibility of the authors, and does not necessarily reflect the views of the Australian Government. The HIV Netherlands Australia Thailand Research Collaboration (HIV-NAT), The Thai Red Cross AIDS Research Centre. For author contributions, ST, SJK, MGL, AA and KR created the study concept and study design. SC, SB, AA, AT, PS and KR were responsible for data collection or oversaw programme implementation. ST conducted the analysis. SK, ML and KR advised on the analysis. ST, SK, ML and KR interpreted the data. ST, SJK, MGL and KR drafted the manuscript. All authors critically reviewed the manuscript and approved the manuscript for submission. Disclosure statement KR has received the Senior Research Scholar from Thailand Research Fund (TRF) and he received honoraria or consultation fees from Merck, Roche, Janssen Cilag, Tibotec, Mylan and GPO (Governmental Pharmaceutical Organization, Thailand). He also has participated in a company sponsored speaker s bureau from Abbott, Gilead, Bristol Myers Squibb, Merck, Roche, Janssen Cilag, GlaxoSmithKline and GPO (Governmental Pharmaceutical Organization). MGL has received unrestricted grants from Boehringer Ingelheim, Gilead Sciences, Merck Sharp & Dohme, Bristol Myers Squibb, Janssen Cilag, ViiV HealthCare, DSMB sitting fees from Sirtex Pty Ltd, and consultancy and presentation fees from Gilead Sciences. The rest of the authors declare no conflict of interest. Additional files Additional file 1: A figure of study profile can be found at Teeraananchai_Addfile1.pdf Additional file 2: A table of primary treatment outcomes of HIV-infected patients can be found at Teeraananchai_Addfile2.pdf Additional file 3: A figure of Kaplan Meier probability of death by baseline age groups can be found at Teeraananchai_Addfile3.pdf Additional file 4: A figure of Kaplan Meier probability of death by baseline CD4 + T-cell count groups can be found at documents/4008_teeraananchai_addfile4.pdf Additional file 5: A table of crude mortality rate of HIV-infected patients can be found at Teeraananchai_Addfile5.pdf References 1. 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: Nakagawa F, May M, Phillips A. Life expectancy living with HIV: recent estimates and future implications. Curr Opin Infect Dis 2013; 26: Harrison KM, Song R, Zhang X. Life expectancy after HIV diagnosis based on national HIV surveillance data from 25 states, United States. J Acquir Immune Defic Syndr 2010; 53: Antiretroviral Therapy Cohort Collaboration. Life expectancy of individuals on combination antiretroviral therapy in high-income countries: a collaborative analysis of 14 cohort studies. Lancet 2008; 372: May M, Gompels M, Delpech V, et al. Impact of late diagnosis and treatment on life expectancy in people with HIV-1: UK Collaborative HIV Cohort (UK CHIC) Study. BMJ 2011; 343:d Resch S, Korenromp E, Stover J, et al. Economic returns to investment in AIDS treatment in low and middle income countries. PLoS One 2011; 6:e Mills EJ, Bakanda C, Birungi J, et al. Life expectancy of persons receiving combination antiretroviral therapy in low-income countries: a cohort analysis from Uganda. Ann Intern Med 2011; 155: Nsanzimana S, Remera E, Kanters S, et al. Life expectancy among HIV-positive patients in Rwanda: a retrospective observational cohort study. Lancet Glob Health 2015; 3:e169 e Ruxrungtham K, Brown T, Phanuphak P. HIV/AIDS in Asia. Lancet 2004; 364: UNAIDS. HIV in Asia and the Pacific (Accessed 19 November 2015.) Available from org/sites/default/files/media_asset/2013_hiv-asia-pacific_ en_0.pdf 11. Chiang CL. On constructing current life tables. J Am Stat Assoc 1972; 67: Life tables by World Bank Income Group (Accessed 2 November 2015.) Available from countries/tha/en/ 13. Chasombat S, McConnell MS, Siangphoe U, et al. National expansion of antiretroviral treatment in Thailand, : program scale-up and patient outcomes. J Acquir Immune Defic Syndr 2009; 50: Antiviral Therapy

10 S Teeraananchai et al. 14. Manosuthi W, Ongwandee S, Bhakeecheep S, et al. Guidelines for antiretroviral therapy in HIV-1 infected adults and adolescents 2014, Thailand. AIDS Res Ther 2015; 12: Johnson LF, Mossong J, Dorrington RE, et al. Life expectancies of South African adults starting antiretroviral treatment: collaborative analysis of cohort studies. PLoS Med 2013; 10:e Teeraananchai S, Kerr SJ, Amin J, Ruxrungtham K, Law MG. Life expectancy of HIV-positive people after starting combination antiretroviral therapy: a meta-analysis. HIV Med 2017; 18: Vapattanawong P, Prasartkul P. Under-registration of deaths in Thailand in : results of cross-matching data from two sources. Bull World Health Organ 2011; 89: Accepted 6 December 2016; published online 5 January International Medical Press