Prognostic significance of the interval between the initiation of antiretroviral therapy and the initiation of anti-tuberculosis treatment in HIV/tuberculosis-coinfected patients: results from the TREAT Asia HIV Observational Database

Authors


Abstract

Objectives

We evaluated the effect of the time interval between the initiation of antiretroviral therapy (ART) and the initiation of tuberculosis (TB) treatment on clinical outcomes in HIV/TB-coinfected patients in an Asian regional cohort.

Methods

Adult HIV/TB-coinfected patients in an observational HIV-infected cohort database who had a known date of ART initiation and a history of TB treatment were eligible for study inclusion. The time interval between the initiation of ART and the initiation of TB treatment was categorized as follows: TB diagnosed while on ART, ART initiated ≤ 90 days after initiation of TB treatment (‘early ART’), ART initiated > 90 days after initiation of TB treatment (‘delayed ART’), and ART not started. Outcomes were assessed using survival analyses.

Results

A total of 768 HIV/TB-coinfected patients were included in this study. The median CD4 T-cell count at TB diagnosis was 100 [interquartile range (IQR) 40-208] cells/μL. Treatment outcomes were not significantly different between the groups with early ART and delayed ART initiation. Kaplan−Meier analysis indicated that mortality was highest for those diagnosed with TB while on ART (3.77 deaths per 100 person-years), and the prognoses of other groups were not different (in deaths per 100 person-years: 2.12 for early ART, 1.46 for delayed ART, and 2.94 for ART not started). In a multivariate model, the interval between ART initiation and TB therapy initiation did not significantly impact all-cause mortality.

Conclusions

A negative impact of delayed ART in patients coinfected with TB was not observed in this observational cohort of moderately to severely immunosuppressed patients. The broader impact of earlier ART initiation in actual clinical practice should be monitored more closely.

Introduction

Tuberculosis (TB) is the most common opportunistic disease and cause of death in patients with HIV infection in developing countries [1]. The two diseases are closely intertwined, and the number of coinfected patients continues to grow rapidly [2]. The optimal time to begin antiretroviral therapy (ART) in patients with both TB and HIV infection has been carefully investigated. Immune reconstitution inflammatory syndrome (IRIS), pharmacological interactions, and high pill burden have previously argued against simultaneous therapy for both HIV and TB [3-6]. In contrast, a delay in starting ART, specifically in severely immunosuppressed patients, is associated with disease progression and higher mortality [7, 8].

Previous World Health Organization (WHO) guidelines recommended that ART should be started between 2 and 8 weeks after the start of TB therapy in persons with CD4 T-cell counts < 200 cells/μL, but the commencement of ART may be delayed for patients with CD4 T-cell counts > 200 cells/μL [9]. Prospective studies have evaluated the optimal time for initiating ART in HIV/TB-coinfected persons [4], including randomized clinical trials in South Africa (SAPiT), Cambodia (CAMELIA), and several different countries (STRIDE) [4, 10, 11]. Those studies demonstrated a mortality reduction among those starting ART during the early stages of TB therapy compared with those who started in the later stages or after the completion of TB therapy. Current WHO guidelines now recommend that ART be initiated as soon as TB therapy is tolerated, ideally as early as 2 weeks and not later than 8 weeks after initiation of TB therapy, regardless of CD4 T-cell count [9]. However, it is not clear how ART timing in HIV/TB-coinfected patients impacts clinical outcomes in ‘real-life,’ nontrial settings. Another issue is the risk of starting ART in a patient with underlying but quiescent TB, which may lead to unmasking of TB disease and IRIS in the absence of anti-TB treatment.

The objective of this analysis was to evaluate the effects of the time interval between the initiation of ART and the initiation of TB treatment on clinical outcomes in a regional observational cohort of HIV/TB-coinfected Asian patients.

Methods

Study population

We analysed data from the TREAT Asia HIV Observational Database (TAHOD), a prospective, observational cohort study of adults with HIV infection from 18 sites in the Asia-Pacific region [12]. The structure of the database and standardized mechanisms for data collection and quality control have previously been described [12]. Data on additional TB-related variables, such as diagnostic site of TB, diagnostic methods, TB treatment, mycobacterial resistance and treatment outcomes, were retrospectively collected for this analysis using a standardized form. All coinfected patients in TAHOD whose dates of TB diagnosis, TB therapy initiation and ART initiation were known, and who were aged 18 years or older at TB diagnosis were eligible for inclusion in this analysis.

Variables and definitions

The following variables were assessed: age at TB diagnosis; sex; reported route of HIV infection; prior AIDS diagnoses before TB; hepatitis B or C virus coinfection; HIV and TB treatment regimens, dates of starting and stopping, adverse events, and outcomes; CD4 T-cell count (cells/μL) and HIV viral load (copies/mL); development of IRIS. The most advanced US Centers for Disease Control and Prevention (CDC) clinical category recorded was used as the clinical status for the analysis [13]. TB treatment outcomes were defined according to WHO TB reporting forms [14]. A patient who was initially culture- or smear microscopy-positive at the beginning of TB treatment but who was smear-negative in the last month of treatment and on at least one previous occasion was considered as ‘cured.’ ‘Treatment failure’ was defined as (1) a patient who was culture- or smear-positive at 5 months or later during initial TB treatment, or who was switched to a regimen including second-line TB drugs because of culture results showing multidrug-resistant (MDR) TB, or (2) a previously treated patient who was culture- or smear-positive at the end of a re-treatment regimen or who was switched to a regimen including second-line TB drugs because of culture results showing MDR TB. A patient who completed treatment but did not meet criteria to be classified as ‘cured’ or a ‘treatment failure’ was considered as ‘treatment completed.’ ‘Died’ meant a patient who died from any cause during the course of TB treatment. A patient was considered as having ‘defauIted’ if TB treatment was interrupted for two or more consecutive months. A patient was considered as ‘transferred out’ if the patient was transferred to another health facility and the TB treatment outcome was not known. Immunological response (IR) at 12 months after ART initiation was defined as a rise in CD4 T-cell count of at least 100 cells/μL. The severity of adverse events of TB therapy was determined using a modified WHO toxicity grading scale [15]. IRIS was diagnosed using the consensus case definitions of the International Network for the Study of HIV-associated IRIS [16]. The exposures of interest were the time intervals between TB therapy initiation and ART initiation, which were categorized into four groups according to the dates of TB therapy and ART initiation: TB diagnosed while on ART, ART initiated within 90 days after TB therapy initiation (‘early ART’), ART initiated later than 90 days after TB therapy initiation (‘delayed ART’), and ART not started. Death was confirmed by local medical staff and reported using standardized Cause of Death (CoDe) forms. AIDS-defining illnesses were defined according to the modified 1993 CDC definitions [13].

Data analysis

Demographic and clinical characteristics were compared among patients in the four groups. All-cause mortality following TB diagnosis was examined using survival analysis methods. The survivor functions for deaths in each group were compared using the Kaplan−Meier curve and the Cox proportional hazards model. The overall survival time was calculated from TB diagnosis to the time of death or the last follow-up visit. The final multivariate model included all covariates that remained significant at the 0.10 level (two-sided). All analyses were performed using sas (version 9.1; SAS Institute Inc., Cary, NC) and stata (version 10.1; StataCorp, College Station, TX).

Results

Demographic and clinical characteristics of subjects

A total of 768 HIV/TB-coinfected patients were included in the analysis, of whom 191 were diagnosed with TB while on ART, 238 had early ART initiation, 280 had delayed ART initiation, and 59 had not started ART during the data collection period. The time between the initiation of ART and the initiation of TB therapy varied (Table 1). Overall, 609 patients (79%) were male, the median age was 34 years [interquartile range (IQR) 29–39 years], the median CD4 T-cell count at TB diagnosis was 100 (IQR 40-208) cells/μL, and the median HIV viral load was 81 650 (IQR 499-330 000) HIV-1 RNA copies/mL (Table 2). The most common HIV exposure category was heterosexual contact (69%), followed by injecting drug use (16%). The majority of patients had no prior AIDS-related illness reported (86%). Nonnucleoside reverse transcriptase inhibitor (NNRTI)-based first-line regimens were prescribed to 615 (80%) patients and PI-based regimens to 39 (5%) patients. TB cases included pulmonary (42%), extrapulmonary (23%), and both (10%). In 25% of cases, the site of TB infection could not be identified. Age, sex and reported route of infection were not different among the four groups (Table 2). The rate of prior AIDS diagnosis at TB diagnosis was highest in the group diagnosed while already on ART (29%), followed by the early ART group (13%). The CD4 T-cell counts at TB diagnosis were highest in the group that had not yet started ART, followed by the group with TB diagnosed while on ART, the delayed ART group, and the early ART group.

Table 1. Time interval between the initiation of antiretroviral therapy (ART) and the initiation of tuberculosis (TB) treatment among HIV/TB-coinfected patients (n = 768)
Time between initiation of ART and initiation of TB treatment (days)n (%)
ART after TB treatment518 (67)
> 36579 (10)
181–36590 (12)
91–180110 (14)
61–9071 (9)
31–6076 (10)
15–3041 (5)
0–1451 (7)
ART before TB treatment191 (25)
> 36583 (11)
181–36517 (2)
91–18024 (3)
61–9014 (2)
31–6020 (3)
15–3011 (1)
0–1421 (3)
ART was not started59 (8)
Table 2. Characteristics of patients with HIV/tuberculosis (TB) coinfection, by antiretroviral therapy (ART) initiation category
CharacteristicTotal (n = 768)TB diagnosed while on ART (n = 191)Early ART (n = 238)Delayed ART (n = 280)ART not started (n = 59)
  1. IQR, interquartile range; NRTI, nucleoside/nucleotide reverse transcriptase inhibitor; NNRTI, nonnucleoside reverse transcriptase inhibitor; PI, protease inhibitor.
Age at TB diagnosis (years) [median (IQR)]34 (29, 39)35 (30, 41)33.5 (29, 39)33 (29, 38.5)32 (30, 37)
Sex, male [n (%)]609 (79)149 (78)206 (86)207 (74)47 (80)
Reported exposure route of HIV infection [n (%)]     
Heterosexual contact530 (69)129 (68)163 (68)197 (70)41 (69)
Homosexual contact67 (9)27 (14)18 (8)18 (7)4 (7)
Injecting drug use128 (16)24 (12)43 (18)51 (18)10 (17)
Other/unknown43 (6)11 (6)14 (6)14 (5)4 (7)
Prior AIDS-defining diagnosis at TB diagnosis [n (%)]     
No662 (86)136 (71)208 (87)263 (94)55 (93)
Yes106 (14)55 (29)30 (13)17 (6)4 (7)
Hepatitis B virus coinfection [n (%)]     
No465 (61)106 (56)136 (57)188 (67)35 (59)
Yes46 (6)10 (5)15 (6)20 (7)1 (2)
Not tested257 (33)75 (39)87 (37)72 (26)23 (39)
Hepatitis C virus coinfection [n (%)]     
No303 (39)72 (38)92 (39)123 (44)16 (27)
Yes116 (15)27 (14)32 (13)51 (18)6 (10)
Not tested349 (46)92 (48)114 (48)106 (38)37 (63)
CD4 count at TB diagnosis (cells/μL) [median (IQR)]100 (40, 208)151 (50, 243)66 (28, 142)119 (48, 238)357 (125, 505)
HIV viral load at TB diagnosis (copies/mL) [median (IQR)]81 650 (499, 330 000)707.5 (< 400, 80 027)206 902 (42 400, 4 880 000)170 000 (11 104, 552 216)132 518 (41 135, 466 718)
Initial ART [n (%)]     
NRTI+NNRTI615 (80)145 (76)213 (90)257 (92)
NRTI + PI39 (5)19 (10)5 (2)15 (5)
Other combination114 (15)27 (14)20 (8)8 (3)
Type of TB [n (%)]     
Pulmonary321 (42)84 (44)91 (38)125 (45)21 (36)
Extrapulmonary178 (23)57 (30)50 (21)58 (21)13 (22)
Both74 (10)15 (8)25 (11)30 (11)4 (6)
Unknown195 (25)35 (18)72 (30)67 (24)21 (36)
Total duration of TB therapy (days) [median (IQR)]275 (186, 427)256 (183, 379.5)275.5 (185, 411)303.5 (214, 534)190 (111, 281)
Time between TB therapy initiation and ART initiation (days) [median (IQR)]Before TB therapy: 223 (51, 790)After TB therapy: 42 (17, 64)After TB therapy: 212 (137.5, 407.5)

TB treatment outcomes

Within the TB treatment outcome categories, 240 (31%) individuals were considered cured, 435 (57%) completed treatment, three (0.4%) were treatment failures, 21 (3%) died, 29 (4%) defaulted, and nine (1%) transferred out (Table 3). Overall survival was 91% during the follow-up period, with the highest death rate per 100 person-years in the group diagnosed with TB while on ART (3.77/100 person-years), followed by the group not started on ART (2.94/100 person-years), the early ART group (2.12/100 person-years), and then the delayed ART group (1.46/100 person-years). The incidence of new AIDS-defining illnesses excluding TB was highest in the group not started on ART (13.66/100 person-years), followed by the group diagnosed with TB while on ART (5.06/100 person-years), the early ART group (4.24/100 person-years), and then the delayed ART group (2.91/100 person-years). The rate of favourable outcomes (e.g. cured or treatment completed) was lowest in the group not started on ART. IRIS was most common in the early ART group.

Table 3. Clinical outcomes of patients with HIV/tuberculosis (TB) coinfection, by antiretroviral therapy (ART) initiation category
OutcomeTotal (n = 768)TB diagnosed while on ART (n = 191)Early ART (n = 238)Delayed ART (n = 280)ART not started (n = 59)
  1. CI, confidence interval; IRIS, immune reconstitution inflammatory syndrome. *CD4 T-cell count increase of 100 cells/μL from ART initiation up to 12-month visit (testing range 6–18 months).
  2. **HIV viral load < 500 copies/mL (6–18 months).
TB treatment outcome [n (%)]     
Cured240 (31)61 (32)79 (33)92 (33)8 (14)
Treatment completed435 (57)99 (52)137 (58)171 (61)28 (47)
Treatment failure3 (0)0 (0)0 (0)1 (0)2 (3)
Died during TB treatment21 (3)10 (5)3 (1)1 (0)7 (12)
Defaulted29 (4)8 (4)7 (3)6 (2)8 (14)
Transferred out9 (1)4 (2)2 (1)2 (1)1 (2)
Current treatment without failure20 (3)6 (3)7 (3)5 (2)2 (3)
Missing11 (1)3 (2)3 (1)2 (1)3 (5)
Mortality     
Follow-up (person-years)29736118961025441
Deaths (n)7023191513
Rate per 100 person-years (95% CI)2.35 (1.86, 2.98)3.77 (2.50, 5.67)2.12 (1.35, 3.32)1.46 (0.88, 2.43)2.94 (1.71, 5.07)
New AIDS-defining illness, excluding TB     
Follow-up (person-years)2531533777824395
New AIDS cases (n)13827332454
Rate (95% CI) per 100 person-years5.45 (4.62, 6.44)5.06 (3.47, 7.38)4.24 (3.02, 5.97)2.91 (1.95, 4.34)13.66 (10.46, 17.83)
AIDS-defining illness (excluding TB) or death     
Follow-up (person-years)2531533777824395
New AIDS cases or deaths (n)18946483263
Rate (95% CI) per 100 person-years7.47 (6.48, 8.61)8.62 (6.46, 11.51)6.17 (4.65, 8.19)3.88 (2.75, 5.49)15.93 (12.45, 20.40)
IRIS [n (%)]34 (4)8 (4)14 (6)11 (4)1 (2)
Immunological response at 12 months after ART initiation* [n (%)]     
No167 (33)52 (40)49 (30)66 (31)
Yes338 (67)77 (60)115 (70)146 (69)
Incomplete information204627468
Virological response at 12 months after ART initiation** [n (%)]     
Yes38 (18)13 (25)7 (12)18 (17)
No178 (82)40 (75)50 (88)88 (83)
Test not done493138181174
Toxicities of TB therapy greater than grade 322 (3)4 (2)3 (1)13 (5)2 (3)

Effects of time interval between the initiation of ART and the initiation of TB treatment on overall survival

We evaluated the prognostic significance of the time interval between the initiation of ART and the initiation of TB therapy in Cox proportional hazards analysis of all-cause mortality (Table 4). There were 70 deaths reported during 2973 person-years of follow-up [an incidence of 2.35 per 100 person-years; 95% confidence interval (CI) 1.86–2.98]. Patients with early ART or delayed ART had a lower risk of mortality than those in the group with TB diagnosed while on ART (P < 0.01). However, the difference between early ART and delayed ART was not significant (P = 0.46). In the multivariate model, injecting drug users were at significantly higher risk of death compared with other HIV exposure risks (adjusted hazard ratio 1.96; 95% CI 1.05–3.66; P = 0.035). The interval between the initiation of ART and the initiation of TB therapy did not have a significant impact on all-cause mortality in the multivariate model. The Kaplan−Meier curve summarizing time to death indicated that mortality was highest for the group diagnosed with TB while on ART; the prognoses of other groups were not significantly different (Fig. 1).

Figure 1.

Time to death based on the interval between the initiation of antiretroviral therapy (ART) and the initiation of tuberculosis (TB) treatment.

Table 4. Cox proportional hazards analysis of all-cause mortality for HIV/tuberculosis (TB)-coinfected patients (n = 768)
VariableFollow-up (person-years)DeathsUnadjusted HRP-valueAdjusted HR95% CIP-value
  1. Bold text denotes adjusted hazard ratios with significant P-values.
  2. ART, antiretroviral therapy; CI, confidence interval; HR, hazard ratio.
Age at ART initiation       
< 30 years854191 1  
30–41 years1463310.960.8861.14(0.63, 2.04)0.667
> 41 years657201.380.3111.62(0.84, 3.15)0.151
Sex       
Male2332611 1  
Female64190.530.0750.52(0.25, 1.06)0.072
Reported exposure route of HIV infection       
Heterosexual contact2235441 1  
Homosexual contact29040.700.5030.55(0.19, 1.53)0.252
Injecting drug use311172.270.0051.96(1.05, 3.66)0.035
Other/unknown13651.720.2521.42(0.56, 3.61)0.464
Prior AIDS diagnosis at ART initiation       
No2547571 1  
Yes426131.390.2841.68(0.90, 3.12)0.102
Hepatitis B virus coinfection       
No1804301 1  
Yes18751.640.3061.47(0.57, 3.83)0.425
Not tested982352.170.0022.60(1.58, 4.28)< 0.001
Hepatitis C virus coinfection       
No1304231 1  
Yes320101.470.3100.70(0.28, 1.71)0.431
Not tested1349371.510.1181.05(0.47, 2.38)0.902
CD4 count at TB diagnosis       
≤ 100 cells/μL716221 1  
≥ 101 cells/μL893150.550.0760.69(0.35, 1.36)0.282
Missing1364330.890.6790.96(0.55, 1.67)0.890
HIV viral load at TB diagnosis       
< 500 copies/mL11341 1  
≥ 500 copies/mL2611.170.8861.52(0.17, 13.74)0.707
Missing2833650.740.5520.84(0.30, 2.32)0.733
Type of TB       
Pulmonary1116361 1  
Extrapulmonary730210.940.8251.02(0.57, 1.84)0.937
Both25550.590.2690.65(0.25, 1.65)0.360
Not tested or reported87180.320.0030.31(0.14, 0.70)0.005
Time interval between initiation of TB therapy and ART initiation       
TB diagnosed while on ART611231 1.00  
ART not yet started441130.630.1970.80(0.39, 1.66)0.551
Early ART after TB therapy initiation896190.610.1140.69(0.37, 1.28)0.239
Delayed ART after TB therapy initiation1025150.510.0500.57(0.29, 1.12)0.101

Discussion

TB is highly prevalent in Asia, accounting for 55% of all global cases in 2008 [16]. Our group previously reported that TB was the most common AIDS-defining diagnosis (45%) in TAHOD patients, and that survival after TB diagnosis was 50% lower than that of patients without an AIDS-defining illness [17].

Clinical trial data have shown that the optimal time to begin ART in HIV/TB-coinfected patients is earlier than has previously been recommended [10, 18]. The CAMELIA trial in Cambodia showed that early ART initiation (within 2 weeks of initiation of TB therapy) improved survival in patients with CD4 T-cell counts < 50 cells/μL [18]. These results were confirmed by two additional trials: the global STRIDE study and the South African SAPiT trial [11, 19]. In these trials, early or immediate ART initiation was also associated with an increased risk of IRIS, but did not result in poorer overall survival. Although the risk−benefit balance of earlier initiation of ART after initiation of TB treatment is not clear for patients with higher CD4 T-cell counts, WHO changed TB/HIV management guidelines on the basis of these robust results.

Our observational study showed that the treatment outcomes and overall mortality of patients with TB in whom ART was initiated early did not significantly differ from those of patients with delayed ART initiation. The lower overall mortality rate in our cohort of 2.35 per 100 person-years and different endpoints are possible reasons why we could not detect a difference between the groups. Specifically, the rates of death in the CAMELIA trial were 8.28 per 100 person-years in the early ART group and 13.77 per 100 person-years in the delayed ART group [18]. The primary composite endpoint of the global STRIDE study was new AIDS-defining illness or death, which demonstrated the benefit of early ART initiation among patients with CD4 T-cell counts < 50 cells/μL [11]. Because of the small numbers of cases in our cohort within the multiple treatment categories and CD4 strata, we did not perform subgroup analyses breaking down the CD4 T-cell count categories.

Patient data for our study were collected before regional HIV/TB treatment recommendations were changed and reflect how TB management was practically implemented for patients with moderate to severely suppressed immune systems. To our knowledge, this study is the first in our region to compare the clinical outcomes of nontrial patients who were diagnosed with TB after ART initiation and those who had TB prior to ART initiation. Interestingly, our data showed that those diagnosed with TB while on ART may have a poorer prognosis than those who had TB prior to ART initiation. High incidence rates of TB have been reported shortly after ART initiation both in developed countries and in resource-limited settings [20]. Among 191 individuals who developed TB while on ART in our cohort, 108 (56%) had TB treatment within 1 year after initiating ART. Incident TB during ART can arise because of residual immunodeficiency, and some cases of previously subclinical disease may manifest because of restoration of TB antigen-specific immune responses. A subset of these cases may have inflammatory symptoms consistent with IRIS [21]. While the use of ART decreases the risk of developing TB by 70–90% [22-24], our data reinforce the importance of latent TB screening and treatment and the recommendation to carefully exclude TB prior to ART initiation to avoid unmasking TB and its associated morbidity and mortality [25-27].

It is unclear why patients diagnosed with TB while on ART had a poorer prognosis compared with the other groups with ART after TB treatment. The patients who started ART before TB diagnosis were more likely to have prior AIDS diagnoses, suggesting that they could have had a history of greater immunodeficiency compared with those who started ART after TB diagnosis. Some of those who had previously initiated ART also had first-line ART failure at the time of TB diagnosis, which could have resulted in poorer outcomes than for ART-naïve patients starting a suppressive antiretroviral regimen. TB as unmasking IRIS may have a worse prognosis than TB without IRIS or TB with paradoxical IRIS.

The study's limitations include the observational nature of the data, and the retrospective collection of data on some of the TB-related variables. Another issue is generalizability to other clinical centres in the region. TAHOD participating sites are generally urban referral centres, and patients are enrolled who are considered likely to remain in long-term follow-up. The site and patient selection criteria consequently prevent broad generalizations, but also offer the opportunity to gather reliable longitudinal data in a nontrial setting. Furthermore, only 40% of TB cases were microbiologically diagnosed, and testing for TB drug susceptibility was not a consistent practice. We arbitrarily defined early ART with a cut-off of 90 days, although previous reports have used different cut-offs such as 14 or 60 days. However, when we compared the outcomes between 60 and 90 days, the results were similar with regard to the hazard ratios for death (data not shown).

In summary, our data demonstrate that treatment outcomes and overall mortality of HIV/TB-coinfected patients who started ART within 90 days of TB treatment did not differ from those of patients who started ART later in this observational cohort of moderately to severely immunosuppressed patients. In addition, overall mortality was highest among patients who were diagnosed with TB while on ART. TB screening efforts prior to ART initiation should be reinforced in HIV care settings in the region. As revised guidelines recommending earlier initiation of ART for coinfected patients are more widely implemented, further studies will be needed to evaluate the broader clinical impact of early ART initiation for patients with TB in Asia.

Acknowledgements

The authors would like to thank participating TAHOD sites, steering committee members (see Appendix), and the patients who participated in this study. This work was supported by the US National Institutes of Health's National Institute of Allergy and Infectious Diseases, Eunice Kennedy Shriver National Institute of Child Health and Human Development, and National Cancer Institute, as part of the International Epidemiologic Databases to Evaluate AIDS (IeDEA; U01AI069907). The TREAT Asia HIV Observational Database is an initiative of TREAT Asia, a programme of amfAR, The Foundation for AIDS Research. The Kirby Institute is funded by the Australian Government Department of Health and Ageing, and is affiliated with the Faculty of Medicine, The University of New South Wales. The content of this publication is solely the responsibility of the authors and does not necessarily represent the official views of any of the institutions mentioned above. JYC involvement was supported by Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education, Science and Technology (NRF-2013R1A1A2005412).

Conflicts of interest

The authors have no conflicts of interest to declare.

Appendix: Appendix: The TREAT Asia HIV Observational Database (TAHOD) study group

C. V. Mean, V. Saphonn* and K. Vohith, National Center for HIV/AIDS, Dermatology & STDs, Phnom Penh, Cambodia; F. J. Zhang*, H. X. Zhao and N. Han, Beijing Ditan Hospital, Capital Medical University, Beijing, China; P. C. K. Li* and M. P. Lee, Queen Elizabeth Hospital, Hong Kong, China; N. Kumarasamy*, S. Saghayam and C. Ezhilarasi, YRG Centre for AIDS Research and Education, Chennai, India; S. Pujari*, K. Joshi and A. Makane, Institute of Infectious Diseases, Pune, India; T. P. Merati*, D. N. Wirawan and F. Yuliana, Faculty of Medicine, Udayana University & Sanglah Hospital, Bali, Indonesia; E. Yunihastuti*, D. Imran and A. Widhani, Working Group on AIDS, Faculty of Medicine, University of Indonesia/Cipto Mangunkusumo Hospital, Jakarta, Indonesia; S. Oka*, J. Tanuma and T. Nishijima, National Center for Global Health and Medicine, Tokyo, Japan; J. Y. Choi*, S. H. Han and J. M. Kim, Division of Infectious Diseases, Department of Internal Medicine, Yonsei University College of Medicine, Seoul, South Korea; C. K. C. Lee*, B. L. H. Sim and R. David, Hospital Sungai Buloh, Kuala Lumpur, Malaysia; A. Kamarulzaman* and A. Kajindran, University of Malaya Medical Centre, Kuala Lumpur, Malaysia; R. Ditangco*, E. Uy and R. Bantique, Research Institute for Tropical Medicine, Manila, Philippines; Y. M. A. Chen*, W. W. Wong and L. H. Kuo, Taipei Veterans General Hospital and AIDS Prevention and Research Centre, National Yang-Ming University, Taipei, Taiwan; O. T. Ng*, A. Chua, L. S. Lee and A. Loh, Tan Tock Seng Hospital, Singapore; P. Phanuphak*, K. Ruxrungtham and M. Khongphattanayothin, HIV-NAT/Thai Red Cross AIDS Research Centre, Bangkok, Thailand; S. Kiertiburanakul*, S. Sungkanuparph and N. Sanmeema, Faculty of Medicine, Ramathibodi Hospital, Mahidol University, Bangkok, Thailand; T. Sirisanthana*, R. Chaiwarith and W. Kotarathititum, Research Institute for Health Sciences, Chiang Mai, Thailand; V. K. Nguyen*, V. H. Bui and T. T. Cao, National Hospital for Tropical Diseases, Hanoi, Vietnam; T. T. Pham*, D. D. Cuong and H. L. Ha, Bach Mai Hospital, Hanoi, Vietnam; A. H. Sohn*, N. Durier* and B. Petersen, TREAT Asia, amfAR – The Foundation for AIDS Research, Bangkok, Thailand; D. A. Cooper, M. G. Law*, J. Zhou* and A. Jiamsakul, The Kirby Institute, The University of New South Wales, Sydney, Australia.

  1. *TAHOD Steering Committee member; †Steering Committee Chair; ‡Co-Chair.

Ancillary