HHS Public Access

Author manuscript
Int J STD AIDS. Author manuscript; available in PMC 2018 November 01.
Author Manuscript

Published in final edited form as:

Int J STD AIDS. 2017 November ; 28(13): 1282–1291. doi:10.1177/0956462417699538.

Trends in CD4 count response to first-line antiretroviral

treatment in HIV-positive patients from Asia, 2003–2013: TAHOD-
LITE
Nicole L. De La Mata1,§, Penh Sun Ly2, Oon Tek Ng3, Kinh Van Nguyen4, Tuti Parwati
Merati5, Thuy Thanh Pham6, Man Po Lee7, Jun Yong Choi8, Annette H. Sohn9, Matthew G.
Law1, and Nagalingeswaran Kumarasamy10
Author Manuscript

1KirbyInstitute, UNSW Australia, Sydney, NSW, Australia 2National Center for HIV/AIDS,
Dermatology & STDs, Phnom Penh, Cambodia 3Tan Tock Seng Hospital, Singapore 4National
Hospital for Tropical Diseases, Hanoi, Vietnam 5Faculty of Medicine Udayana University &
Sanglah Hospital, Bali, Indonesia 6Bach Mai Hospital, Hanoi, Vietnam 7Queen Elizabeth Hospital,
Hong Kong, China 8Department of Internal Medicine and AIDS Research Institute, Yonsei
University College of Medicine, Severance Hospital, Seoul, South Korea 9TREAT Asia, amfAR -
The Foundation for AIDS Research, Bangkok, Thailand 10YRG CARE, Chennai, India

Abstract
Introduction—Antiretroviral treatment (ART) guidelines have changed over the past decade,
Author Manuscript

recommending earlier initiation and more tolerable regimens. The study objective was to examine
the CD4 response to ART, depending on the year of ART initiation, in HIV-positive patients in the
Asia-Pacific.

Methods—We included HIV-positive adult patients who initiated ART between 2003–2013 in
our regional cohort from eight urban referral centres in seven countries within Asia. We used
mixed-effects linear regression models to evaluate differences in CD4 response by year of ART
initiation during 36 months of follow-up, adjusted a priori for other covariates.

Results—Overall, 16962 patients were included. Patients initiating in 2006–09 and 2010–13 had
an estimated mean CD4 count increase of 8cells/μL and 15cells/μL, respectively, at any given time
during the 36 month follow-up, compared to those in 2003–05. The median CD4 count at ART
initiation also increased from 96 cells/μL in 2003–05 to 173 cells/μL in 2010–13.
Author Manuscript

Conclusions—Our results suggest that the CD4 response to ART is modestly higher for those
initiating ART in more recent years. Moreover, fewer patients are presenting with lower absolute

§
Corresponding author: Nicole L. De La Mata, The Kirby Institute, UNSW Australia, Wallace Wurth Building, Sydney, 2052,
Australia, (02) 9385 9012, ndelamata@kirby.unsw.edu.au.
Competing Interests
The authors do not have any competing interests to declare.
Authors’ contributions
NLD and ML contributed to the concept development. KN, PSL, OTN, KVN, TPM, TTP, MPL and JYC contributed data for the
analysis. NLD performed the statistical analysis and wrote the first draft of the manuscript. All authors commented on the draft
manuscript and approved of the final manuscript.
 De La Mata et al. Page 2

CD4 counts over time. This is likely to reduce their risk of opportunistic infections and future non-
Author Manuscript

AIDS defining cancers.

Keywords
Asia; HIV; epidemiology; CD4 trends; immunological response; ART

Introduction
CD4 counts are used as prognostic markers of HIV disease progression [1, 2]. Untreated
HIV-infected persons have a gradual depletion in CD4 cell levels, leading to increased risk
of AIDS-defining illnesses and death [3–6]. Combination antiretroviral therapy (ART) has
been highly effective in preventing HIV disease progression and restoring CD4 cell levels as
well as reducing viral replication and lowering rates of HIV-associated morbidity and
Author Manuscript

mortality [7–9].

Initial World Health Organization (WHO) treatment guidelines, released in 2002,

recommended ART initiation for those in advanced stages of HIV or in asymptomatic stages
with CD4 count <200 cells/μL [10]. Delayed ART initiation was earlier suggested in stable
patients to reduce the risk of developing and transmitting drug resistant HIV caused by
suboptimal adherence [11, 12]. However, recent research has shown strong evidence to
support earlier ART initiation at higher CD4 cell levels is beneficial in preventing disease
progression and transmission, and also prevents the incidence of opportunistic infections
(OIs) [8, 13–15]. After subsequent guideline revisions steadily increased the CD4 threshold
for ART, in 2015, WHO treatment guidelines recommended ART initiation among all adults,
regardless of CD4 count [16].
Author Manuscript

Others changes to the WHO treatment guidelines have also occurred over time including the
use of more tolerable and convenient ART regimens, increased support and counselling
services for patients, and routine monitoring of CD4 count and HIV viral load [16, 17].
These changes have been accompanied by improvements in patient outcomes, with reduced
mortality rates for patients receiving care in recent years [18–20]. Although part of these
improvements has been attributed to earlier ART initiation at higher CD4 counts, year of
ART initiation has also shown an independent association with improved overall survival
[21, 22].

A greater CD4 count response has previously been associated with younger age, female sex,
lower pre-ART HIV viral load and CD4 count [23–25]. Yet, there has been little exploration
as to whether CD4 count response has improved in recent years of ART initiation [23, 26].
Author Manuscript

The changes to treatment guidelines and patient management over time could result in an
improved CD4 count response for patients initiating ART in recent years. Specifically, the
move towards newer ARV drugs, associated with fewer side effects, for patients receiving
care in Asia could lead to greater patient adherence [21]. In addition, certain ARV drugs
classes, such as protease inhibitor-based regimens, may also evoke an increased CD4 count
response [26–28]. Our study objective was to examine the time trends in and factors
associated with CD4 response to first-line ART, by calendar year of ART initiation, in HIV-
positive patients receiving care in an Asian regional observational cohort study.

Int J STD AIDS. Author manuscript; available in PMC 2018 November 01.
 De La Mata et al. Page 3

Methods
Author Manuscript

Data collection and Participants

The TREAT Asia HIV Observational Database Low Intensity Transfer (TAHOD-LITE)
cohort is a sub-study of the TREAT Asia HIV Observation Database (TAHOD) and
currently consists of eight sites from the Asia-Pacific region including Cambodia, Hong
Kong, India, Indonesia, Singapore, South Korea and Vietnam. TAHOD collects detailed
patient data on a subset of patients seen at 20 treatment sites in the Asia-Pacific region [29].
Conversely, TAHOD-LITE collects routine clinical data on all patients seen at the 8
participating treatment sites. Thus, TAHOD-LITE is representative of the entire clinical
population within our participating sites. Data are collected routinely when patients attend
care at the treatment sites and include patient demographics, hepatitis serology, HIV-related
laboratory test results and ART history. A more detailed description of TAHOD-LITE has
Author Manuscript

been described elsewhere [21]. After being anonymized, data are transferred electronically
to the Kirby Institute, University of New South Wales and are subjected to quality control
procedures. Data include patient follow-up until May 2014. TAHOD-LITE was granted
ethical approvals from Institutional Review Boards (IRB) at each participating clinical site,
the University of New South Wales and the coordinating center at TREAT Asia/amfAR.
Written consent was not obtained unless required by the site-specific IRBs.

The data selected for this analysis included all patients who were aged over 18 years when
they initiated an ART regimen, consisting of three or more drugs, between 01 January 2003
and 31 December 2013, and had at least one subsequent visit after the date of ART
initiation. There were also site-based exclusions where patients were excluded if they had
initiated ART prior to: 2006 for Singapore; 2010 for Vietnam; and 2004 for Cambodia. Prior
to these years, sites were unable to provide data on all patients that had been seen at the
Author Manuscript

clinic.

Statistical analyses
The primary study objective was to evaluate CD4 count changes and factors associated with
CD4 count change over 36 months of ART. A pseudo intention-to-treat approach was taken
whereby any changes to treatment after ART initiation, including treatment interruptions
were ignored. All patients were censored at the last clinic visit or date of death or 36 months
from ART initiation, whichever occurred earlier. Pre-ART laboratory measurements,
including CD4 count and HIV viral load, were defined as those within 6 months prior, and
closest to or on the date of ART initiation.

Data were modelled using repeated-measures, random-intercept linear regression using

Author Manuscript

generalized least squares estimation to evaluate differences in the CD4 response between the
year periods of ART initiation (2003–05, 2006–09, 2010–13). Covariates, selected a priori,
included clinical site, age at ART initiation, sex, mode of HIV exposure, pre-ART HIV viral
load (copies/mL), pre-ART CD4 count (cells/μL), first ART regimen, hepatitis B and
hepatitis C co-infection, time from ART initiation and squared time from ART initiation.
These covariates were selected based on previous literature and available patient data
collected in TAHOD-LITE. Continuous variables, including age at ART initiation, pre-ART

Int J STD AIDS. Author manuscript; available in PMC 2018 November 01.
 De La Mata et al. Page 4

HIV viral load and pre-ART CD4 count, were categorized in the model. First ART regimen
Author Manuscript

was categorized based on the drug classes included. The squared time from ART initiation
was included to allow for the predicted CD4 count to be modelled as a quadratic curve from
ART initiation. We also evaluated whether there was an interaction between the year period
of ART initiation and time from ART initiation in a sensitivity analysis. This model was
selected for the analysis as it includes all CD4 count measurements during the 36 months of
follow-up and determines whether certain factors influence the CD4 response over the entire
follow-up time rather than at one time point (eg. 12 months from ART initiation). As we
modelled the CD4 count change from ART initiation, patients without a CD4 count result
within 6 months prior to ART initiation (i.e. without a pre-ART CD4 count) were excluded
from the model.

CD4 count response was also summarized by the median CD4 count, with interquartile
range (IQR), and the proportion of patients within each CD4 count category (≤50, 51–100,
Author Manuscript

101–200, 201–350, 351–500 and ≥501 cells/μL) every 6 months up to 36 months from ART
initiation, by year of ART initiation, overall and for each country. For these crude
summaries, we only included CD4 count measurements that were closest to and within ±3
months of the given time point.

Data were analysed using Stata version 12 (Stata Corporation, College Station, Texas, USA)
and SAS (version 9.4 for Windows).

Results
A total of 18 441 patients aged over 18 years had initiated ART between 1 January 2003 and
31 December 2013. Of these, 777 patients were excluded for not attending the clinic after
ART initiation and 702 patients were excluded due to site-based exclusions (see Methods;
Author Manuscript

Singapore, n=70; Vietnam, n=568; Cambodia, n=64). The remaining 16 962 were included
in the analysis.

Patient Characteristics
A summary of the patient characteristics across all countries by year of ART initiation is
given in Table 1. Briefly, the majority of patients were male (2003–05: 75%; 2006–09: 69%;
2010–13: 66%), reported heterosexual mode of HIV exposure (2003–05: 88%; 2006–09:
85%; 2010–13: 73%), initiated in recent years (2003–05: 17%; 2006–09: 37%; 2010–13:
46%) and had a first ART regimen consisting of nucleoside reverse transcriptase inhibitors
(NRTIs) and a non-nucleoside reverse transcriptase inhibitor (NNRTI) (2003–05: 92%;
2006–09: 97%; 2010–13: 98%). The median age at ART initiation was relatively consistent
Author Manuscript

between periods of ART initiation (2003–05: 35 years, IQR: 30–40; 2006–09: 36 years,
IQR: 31–42; 2010–13: 36 years, IQR: 30–43). Over 80% of the patients had a pre-ART CD4
count measurement, regardless of year of ART initiation. A minority of patients had a pre-
ART HIV viral load measurement (2003–05: 14%; 2006–09: 15%; 2010–13: 29%) and the
median pre-ART HIV viral load increased from 106 000 copies/mL (IQR: 32 000–261 000
copies/mL) in 2003–05 to 110 564 (IQR: 30 563–402 000 copies/mL) in 2010–13.

Int J STD AIDS. Author manuscript; available in PMC 2018 November 01.
 De La Mata et al. Page 5

Summary of the CD4 count response

Author Manuscript

Of the 14 448 patients with pre-ART CD4 count measurements, ART was initiated in 2003–
05 for 2 421 patients, in 2006–09 for 5 281 patients and in 2010–13 for 6 746 patients. The
median follow-up time for patients initiating in 2003–05 was 2.6 years, in 2006–09 was 2.5
years and in 2010–13 was 1.6 years.

The median CD4 count increases from ART initiation were: in 2003–05, from 96 cells/μL
(IQR: 45–171 cells/μL) at ART initiation to 374 cells/μL (IQR: 237–561 cells/μL) at 36
months; in 2006–09, from 128 cells/μL (IQR: 52–201 cells/μL) at ART initiation to 401
cells/μL (IQR: 263–575 cells/μL) at 36 months; and in 2010–13, from 173 cells/μL (IQR:
53–286 cells/μL) at ART initiation to 418 cells/μL (IQR: 274–577 cells/μL) at 36 months
(Figure 1). Overall, there was an increasing trend where those initiating in 2010–13 had a
higher median CD4 count at ART initiation follow-up compared to prior year periods.
Author Manuscript

However, this was not found be significant in the Kruskal-Wallis median test (p value =
0.368). Similar trends were observed when examined by country (Appendix 1).

The proportion of patients within each CD4 count category (≤50, 51–100, 101–200, 201–
350, 351–500 and ≥501 cells/μL) up to 36 months from ART initiation is summarized in
Figure 2. Largely as a result of ART initiation at higher CD4 counts, there was an increasing
trend where those initiating in 2010–13 had a greater proportion of patients at higher CD4
counts than in previous year periods. The proportion of patients with CD4 count ≥201 cells/
μL increased for patients initiating: in 2003–05, from 16% at ART initiation to 81% at 36
months; in 2006–09, from 25% at ART initiation to 85% at 36 months; and in 2010–13,
from 44% at ART initiation to 85% at 36 months. This trend was also apparent by country
(Appendix 2).
Author Manuscript

Modelling the CD4 count response up to 36 months

The model indicated that several factors were significantly associated with the CD4 count
response over time (Table 2). In the univariate analysis, year period of ART initiation was
significantly (p value <0.001) associated with the CD4 cell response. Those initiating in
2006–09 and 2010–13 had a mean CD4 count that at any given time during follow-up, was 8
cells/μL (95% CI: 3 to 13 cells/μL) and 13 cells/μL (95% CI: 8 to 19 cells/μL) higher than
those initiating in 2003–05. In the multivariate model, the year period of ART initiation
remained significant while adjusting for clinical site and other relevant covariates (p value
<0.001). Here, the mean CD4 cell count was higher at any given time during follow-up, for
those initiating in 2006–09 and 2010–13 by 8 cells/μL (95% CI: 3 to 13 cells/μL) and 15
cells/μL (95% CI: 9 to 20 cells/μL), respectively, compared to those initiating in 2003–05.
Author Manuscript

We conducted several sensitivity analyses to evaluate the robustness of our results. First, we
evaluated whether there was an interaction between the year period of ART initiation and
time from ART initiation. This interaction term was significant for those initiating in 2006–
09 compared to 2003–05 (p value=0.002). The estimated mean CD4 count difference was
not significantly higher for those initiating 2006–09 compared to 2003–05 from ART
initiation, except at 3 months (Appendix 3). There was little evidence to suggest an
interaction between time from ART initiation and those initiating in 2010–13, compared to

Int J STD AIDS. Author manuscript; available in PMC 2018 November 01.
 De La Mata et al. Page 6

2003–05 (p value=0.527). The estimated mean CD4 count difference was significantly
Author Manuscript

higher for those initiating in 2010–13, compared to 2003–05, up to 15 months and 27 to 36

months from ART initiation (Appendix 3). Second, we evaluated whether the inclusion of
pre-ART CD4 count in the model biased our estimated CD4 count change. Using a mixed
model approach with random intercept and random slope for time from ART initiation, we
found minimal differences in the parameter estimates when pre-ART CD4 count was
excluded (Appendix 4). We also found excluding other covariates with large proportions of
missing data, such as pre-ART HIV viral load, HBV and HCV status, did not significantly
affect the estimated mean CD4 count change for the covariates (Appendix 5). Third, we
evaluated whether a mixed linear model with random intercept and random slope for time
from ART initiation produced significantly different parameter estimates. The parameter
estimates from this model was not substantially different from the primary analysis. The
mean CD4 count difference was significantly higher, at any given time during follow-up, for
Author Manuscript

those initiating in 2006–09 (p value=0.001) and 2010–13 (p value <0.001) compared to

those initiating in 2003–05 (Appendix 5).

Other factors in the multivariate model significantly associated with a higher CD4 count
response, at any given time during follow-up, included younger age, female gender,
homosexual contact (compared to heterosexual contact), higher pre-ART HIV viral load,
lower pre-ART CD4 count, and HBV or HCV negative (compared to positive).

Discussion
In this analysis consisting of 16 962 HIV-positive patients receiving care in the Asia-Pacific
region, our findings have shown that long-term CD4 response to ART is greater in those
initiating in 2010–13 and 2006–09 compared to those initiating in 2003–05, regardless of
Author Manuscript

CD4 count at ART initiation. There was an increasing trend with those initiating in more
recent years having a greater CD4 response compared to previous years. Over the follow-up
period, the median CD4 count was also consistently higher, and the proportion of patients
with higher CD4 counts increased in more recent years.

Similar temporal trends in CD4 count at ART initiation has been shown in other studies. A
large multiregional comparison of HIV-positive adults initiating ART between 2002 and
2009 found a steady increase in the median CD4 count at ART initiation in most countries.
This trend was apparent regardless of the income status of the country, although the greatest
increases were seen in low-income and middle-income countries rather than in high-income
countries and, was also higher in females than males [30]. In contrast, a meta-analysis of 44
studies did not find a significant increasing trend in the mean CD4 count at presentation for
newly presenting HIV-positive adults. The annual estimated change in CD4 count at
Author Manuscript

presentation was 1.6 cells/μL (95% CI: −4.4 to 5.4 cells/μL) which was not significant
(p>0.05), adjusting for study inclusion criteria, data type and study location [31].

Overall, there were few studies that explored whether the year of ART initiation influenced
the CD4 count response in HIV-positive patients. One study based in a London hospital
showed an association between calendar year of ART initiation and CD4 response from
ART initiation [23]. During the first 3 months of ART initiation, this association was not

Int J STD AIDS. Author manuscript; available in PMC 2018 November 01.
 De La Mata et al. Page 7

significant. However, beyond 3 months, patients initiating from 1997 to 2003 had a yearly
Author Manuscript

CD4 count increase that was 84 cells/μL (95% CI: −48 to −120 cells/μL) higher than those
initiating in 1996 and before.

It is also difficult to ascertain which factors of patient care have contributed to the improved
CD4 count response in recent years. A move towards greater adherence in patients, either
due to physician advice, support services, or more tolerable and convenient regimens, could
have played an important role [17, 32]. Previous studies have highlighted that patients who
are more adherent have greater and more sustained gains in CD4 count than non-adherent
patients [33–35]. Other predictors significantly associated with improved CD4 count
recovery are also consistent with former studies, including older age, female sex, pre-ART
HIV viral load, HBV and HCV co-infection [36–38].

The clinical implications of our findings are fairly limited as those initiating in 2010–13
Author Manuscript

were only 15 cells/μL higher compared to those in 2003–05. But, our analysis has
highlighted that the proportion of patients at high range CD4 counts has drastically increased
over time, in particular, at ART initiation but also through to 36 months follow-up.
Therefore, over time, fewer patients are being exposed to lower CD4 counts where they are
at higher risk of OIs [39] and non-AIDS defining cancers (NADCs) [40, 41]. Patients are
also experiencing shorter durations at lower CD4 counts, which leads to a better overall
prognosis [42].

An advantage of our study was the large patient sample size yet, our patient data were
limited to a few variables. As such, we were unable to explore other important trends
relating to lower CD4 counts, including the occurrence of NADCs or OIs. In addition, HIV
viral load was not routinely collected at the clinical sites and had large proportions of
Author Manuscript

missing data. Hence, we could not expand the scope of our analysis to also examine the HIV
viral load response by year of ART initiation. Our model estimates for the pre-ART HIV
viral load may also be bias and caution is advised when interpreting these findings.

We used observational data on CD4 counts collected during routine clinic visits for HIV-
positive adults presenting between 2003 and 2013. Patients lost to follow-up (LTFU) can
introduce potential bias that can impede on the analysis because it is unclear how many
remain in care elsewhere or have died. The LTFU rate previously reported in TAHOD-LITE
was relatively low and consistent between the years of ART initiation (2003–05: 2.1 per 100
person-years; 2006–09: 2.9 per 100 person-years; 2010–13: 2.8 per 100 person-years) [21].
We also had 8 clinical sites represent 7 countries across the region, and hence, our results are
reflective of trends occurring within the clinical sites rather than their respective countries.
The presence of country-level differences in when patients present for care, the available
Author Manuscript

treatment options and the patient care provided, as well as other unmeasured confounding
factors could have also contributed to heterogeneity. However, our analysis by clinical site
has shown similar trends to the overall analysis where there is an increasing trend in CD4
response over time by the year of ART initiation. Furthermore, the model used in our
analysis is adjusted for clinic site to account for these differences between sites.

Int J STD AIDS. Author manuscript; available in PMC 2018 November 01.
 De La Mata et al. Page 8

In summary, we found that the CD4 response to ART is greater in those initiating in 2010–
Author Manuscript

13 and 2006–09 compared to those in 2003–05, with greater proportions of patients starting
treatment at higher CD4 counts in recent years. Patients initiating in more recent years spend
less time exposed in lower CD4 count ranges, reducing their risk for serious OIs and future
NADCs that are associated with lower CD4 count. As guidelines recommending immediate
ART are more widely implemented, it will be important to monitor their impact on
immediate and long-term clinical outcomes.

Acknowledgments
TAHOD-LITE (TREAT Asia HIV Observational Database Low-Intensity TransfEr) is an initiative of TREAT Asia,
a program of amfAR, The Foundation for AIDS Research, with support from the U.S. 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 Kirby Institute is funded by the Australian Government
Author Manuscript

Department of Health and Ageing, and is affiliated with the Faculty of Medicine, UNSW Australia (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 governments or institutions mentioned above.

References
1. Mellors JW, Munoz A, Giorgi JV, Margolick JB, Tassoni CJ, Gupta P, et al. Plasma viral load and
CD4+ lymphocytes as prognostic markers of HIV-1 infection. Annals of Internal Medicine. 1997
Jun 15; 126(12):946–54. [PubMed: 9182471]
2. Korenromp EL, Williams BG, Schmid GP, Dye C. Clinical prognostic value of RNA viral load and
CD4 cell counts during untreated HIV-1 infection--a quantitative review. PloS One. 2009;
4(6):e5950. [PubMed: 19536329]
3. Lapadula G, Chatenoud L, Gori A, Castelli F, Di Giambenedetto S, Fabbiani M, et al. Risk of Severe
Non AIDS Events Is Increased among Patients Unable to Increase their CD4+ T-Cell Counts
>200+/mul Despite Effective HAART. PloS One. 2015; 10(5):e0124741. [PubMed: 26020949]
Author Manuscript

4. Engsig FN, Zangerle R, Katsarou O, Dabis F, Reiss P, Gill J, et al. Long-term mortality in HIV-
positive individuals virally suppressed for >3 years with incomplete CD4 recovery. Clinical
Infectious Diseases. 2014 May; 58(9):1312–21. [PubMed: 24457342]
5. Bray S, Gedeon J, Hadi A, Kotb A, Rahman T, Sarwar E, et al. Predictive value of CD4 cell count
nadir on long-term mortality in HIV-positive patients in Uganda. HIV AIDS (Auckl). 2012; 4:135–
40. [PubMed: 22930645]
6. Pacheco YM, Jarrin I, Rosado I, Campins AA, Berenguer J, Iribarren JA, et al. Increased risk of
non-AIDS-related events in HIV subjects with persistent low CD4 counts despite cART in the
CoRIS cohort. Antiviral Research. 2015 May.117:69–74. [PubMed: 25766861]
7. Reniers G, Slaymaker E, Nakiyingi-Miiro J, Nyamukapa C, Crampin AC, Herbst K, et al. Mortality
trends in the era of antiretroviral therapy: evidence from the Network for Analysing Longitudinal
Population based HIV/AIDS data on Africa (ALPHA). AIDS. 2014 Nov; 28( Suppl 4):S533–42.
[PubMed: 25406756]
8. Vermund SH. Treatment as prevention for HIV in China. Lancet. 2013 Oct 5; 382(9899):1159–61.
[PubMed: 23206834]
Author Manuscript

9. van Sighem A, Danner S, Ghani AC, Gras L, Anderson RM, de Wolf F, et al. Mortality in patients
with successful initial response to highly active antiretroviral therapy is still higher than in non-HIV-
infected individuals. Journal of Acquired Immune Deficiency Syndromes. 2005 Oct 1; 40(2):212–8.
[PubMed: 16186740]
10. World Health Organisation. Scaling up Antiretroviral Therapy in resource-limited settings. Geneva,
Switzerland: 2002.
11. Booth CL, Garcia-Diaz AM, Youle MS, Johnson MA, Phillips A, Geretti AM. Prevalence and
predictors of antiretroviral drug resistance in newly diagnosed HIV-1 infection. Journal of
Antimicrobial Chemotherapy. 2007 Mar; 59(3):517–24. [PubMed: 17213262]

Int J STD AIDS. Author manuscript; available in PMC 2018 November 01.
 De La Mata et al. Page 9

12. Bangsberg DR, Moss A. When should we delay highly active antiretroviral therapy? Journal of
General Internal Medicine. 1999 Jul; 14(7):446–8. [PubMed: 10417605]
Author Manuscript

13. Hermans SM, van Leth F, Manabe YC, Hoepelman AI, Lange JM, Kambugu A. Earlier initiation
of antiretroviral therapy, increased tuberculosis case finding and reduced mortality in a setting of
improved HIV care: a retrospective cohort study. HIV Medicine. 2012 Jul; 13(6):337–44.
[PubMed: 22296211]
14. Romley JA, Juday T, Solomon MD, Seekins D, Brookmeyer R, Goldman DP. Early HIV treatment
led to life expectancy gains valued at $80 billion for people infected in 1996–2009. Health Affairs.
2014 Mar; 33(3):370–7. [PubMed: 24590933]
15. Sterling TR, Chaisson RE, Keruly J, Moore RD. Improved outcomes with earlier initiation of
highly active antiretroviral therapy among human immunodeficiency virus-infected patients who
achieve durable virologic suppression: longer follow-up of an observational cohort study. Journal
of Infectious Diseases. 2003 Dec 1; 188(11):1659–65. [PubMed: 14639536]
16. World Health Organisation. Guideline on when to start antiretroviral therapy and on pre-exposure
prophylaxis for HIV. 2015. Available online at: http://apps.who.int/iris/bitstream/
10665/186275/1/9789241509565_eng.pdf. [verified January 2016]
Author Manuscript

17. World Health Organisation. Consolidated guidelines on the use of Antiretroviral Drugs for treating
and preventing HIV infection: Recommendations for a Public Health approach. Geneva,
Switzerland: 2013. Available online at: http://apps.who.int/iris/bitstream/
10665/85321/1/9789241505727_eng.pdf. [verified October 2015]
18. Cornell M, Grimsrud A, Fairall L, Fox MP, van Cutsem G, Giddy J, et al. Temporal changes in
programme outcomes among adult patients initiating antiretroviral therapy across South Africa,
2002–2007. AIDS. 2010 Sep 10; 24(14):2263–70. [PubMed: 20683318]
19. Krentz HB, Kliewer G, Gill MJ. Changing mortality rates and causes of death for HIV-infected
individuals living in Southern Alberta, Canada from 1984 to 2003. HIV Medicine. 2005 Mar; 6(2):
99–106. [PubMed: 15807715]
20. Kee MK, Lee JH, Kim EJ, Lee J, Nam JG, Yoo BH, et al. Improvement in survival among HIV-
infected individuals in the Republic of Korea: Need for an early HIV diagnosis. BMC Infectious
Diseases. 2009 Aug.12:9.
21. De La Mata NL, Kumarasamy N, Khol V, Ng OT, Van Kinh N, Merati TP, et al. Improved survival
in HIV treatment programmes in Asia. Antiviral Therapy. 2016 Mar 10.
Author Manuscript

22. Farahani M, Price N, El-Halabi S, Mlaudzi N, Keapoletswe K, Lebelonyane R, et al. Trends and
determinants of survival for over 200 000 patients on antiretroviral treatment in the Botswana
National Program: 2002–2013. AIDS. 2016 Jan 28; 30(3):477–85. [PubMed: 26636931]
23. Smith CJ, Sabin CA, Youle MS, Kinloch-de Loes S, Lampe FC, Madge S, et al. Factors
influencing increases in CD4 cell counts of HIV-positive persons receiving long-term highly active
antiretroviral therapy. Journal of Infectious Diseases. 2004 Nov 15; 190(10):1860–8. [PubMed:
15499544]
24. Asfaw A, Ali D, Eticha T, Alemayehu A, Alemayehu M, Kindeya F. CD4 cell count trends after
commencement of antiretroviral therapy among HIV-infected patients in Tigray, Northern
Ethiopia: a retrospective cross-sectional study. PloS One. 2015; 10(3):e0122583. [PubMed:
25816222]
25. Wright ST, Petoumenos K, Boyd M, Carr A, Downing S, O’Connor CC, et al. Ageing and long-
term CD4 cell count trends in HIV-positive patients with 5 years or more combination
antiretroviral therapy experience. HIV Medicine. 2013 Apr; 14(4):208–16. [PubMed: 23036045]
Author Manuscript

26. Rajasuriar R, Gouillou M, Spelman T, Read T, Hoy J, Law M, et al. Clinical predictors of immune
reconstitution following combination antiretroviral therapy in patients from the Australian HIV
Observational Database. PloS One. 2011; 6(6):e20713. [PubMed: 21674057]
27. van Leth F, Wit FW, Reiss P, Schattenkerk JK, van der Ende ME, Schneider MM, et al. Differential
CD4 T-cell response in HIV-1-infected patients using protease inhibitor-based or nevirapine-based
highly active antiretroviral therapy. HIV Medicine. 2004 Mar; 5(2):74–81. [PubMed: 15012645]
28. Mocroft A, Phillips AN, Ledergerber B, Katlama C, Chiesi A, Goebel FD, et al. Relationship
between antiretrovirals used as part of a cART regimen and CD4 cell count increases in patients
with suppressed viremia. AIDS. 2006 May 12; 20(8):1141–50. [PubMed: 16691065]

Int J STD AIDS. Author manuscript; available in PMC 2018 November 01.
 De La Mata et al. Page 10

29. Zhou J, Kumarasamy N, Ditangco R, Kamarulzaman A, Lee CK, Li PC, et al. The TREAT Asia
HIV Observational Database: baseline and retrospective data. Journal of Acquired Immune
Author Manuscript

Deficiency Syndromes. 2005 Feb 1; 38(2):174–9. [PubMed: 15671802]

30. Avila D, Althoff KN, Mugglin C, Wools-Kaloustian K, Koller M, Dabis F, et al. Immunodeficiency
at the Start of Combination Antiretroviral Therapy in Low-, Middle-, and High-Income Countries.
Journal of Acquired Immune Deficiency Syndromes. 2014 Jan 1; 65(1):E8–E16. [PubMed:
24419071]
31. Lesko CR, Cole SR, Zinski A, Poole C, Mugavero MJ. A Systematic Review and Meta-regression
of Temporal Trends in Adult CD4(+) Cell Count at Presentation to HIV Care, 1992–2011. Clinical
Infectious Diseases. 2013 Oct 10; 57(7):1027–37. [PubMed: 23921882]
32. Duda SN, Farr AM, Lindegren ML, Blevins M, Wester CW, Wools-Kaloustian K, et al.
Characteristics and comprehensiveness of adult HIV care and treatment programmes in Asia-
Pacific, sub-Saharan Africa and the Americas: results of a site assessment conducted by the
International epidemiologic Databases to Evaluate AIDS (IeDEA) Collaboration. Journal of the
International AIDS Society. 2014; 17:19045. [PubMed: 25516092]
33. Wood E, Hogg RS, Yip B, Harrigan PR, O’Shaughnessy MV, Montaner JS. The impact of
Author Manuscript

adherence on CD4 cell count responses among HIV-infected patients. Journal of Acquired Immune
Deficiency Syndromes. 2004 Mar 1; 35(3):261–8. [PubMed: 15076240]
34. Lima VD, Harrigan R, Murray M, Moore DM, Wood E, Hogg RS, et al. Differential impact of
adherence on long-term treatment response among naive HIV-infected individuals. AIDS. 2008
Nov 12; 22(17):2371–80. [PubMed: 18981777]
35. Safren SA, Kumarasamy N, James R, Raminani S, Solomon S, Mayer KH. ART adherence,
demographic variables and CD4 outcome among HIV-positive patients on antiretroviral therapy in
Chennai, India. AIDS Care. 2005 Oct; 17(7):853–62. [PubMed: 16120502]
36. van Griensven J, Phirum L, Choun K, Thai S, De Weggheleire A, Lynen L. Hepatitis B and C Co-
Infection among HIV-Infected Adults while on Antiretroviral Treatment: Long-Term Survival,
CD4 Cell Count Recovery and Antiretroviral Toxicity in Cambodia. PloS One. 2014 Feb 12.9(2)
37. Maskew M, Brennan AT, Westreich D, McNamara L, MacPhail AP, Fox MP. Gender differences in
mortality and CD4 count response among virally suppressed HIV-positive patients. Journal of
Women’s Health. 2013 Feb; 22(2):113–20.
38. Hughes RA, Sterne JAC, Walsh J, Bansi L, Gilson R, Orkin C, et al. Long-term trends in CD4 cell
Author Manuscript

counts and impact of viral failure in individuals starting antiretroviral therapy: UK Collaborative
HIV Cohort (CHIC) study. HIV Medicine. 2011 Nov; 12(10):583–93. [PubMed: 21569188]
39. Damtie D, Yismaw G, Woldeyohannes D, Anagaw B. Common opportunistic infections and their
CD4 cell correlates among HIV-infected patients attending at antiretroviral therapy clinic of
Gondar University Hospital, Northwest Ethiopia. BMC Research Notes. 2013; 6:534. [PubMed:
24330921]
40. Franzetti M, Adorni F, Parravicini C, Vergani B, Antinori S, Milazzo L, et al. Trends and predictors
of non-AIDS-defining cancers in men and women with HIV infection: a single-institution
retrospective study before and after the introduction of HAART. Journal of Acquired Immune
Deficiency Syndromes. 2013 Apr 1; 62(4):414–20. [PubMed: 23274934]
41. Calabresi A, Ferraresi A, Festa A, Scarcella C, Donato F, Vassallo F, et al. Incidence of AIDS-
defining cancers and virus-related and non-virus-related non-AIDS-defining cancers among HIV-
infected patients compared with the general population in a large health district of Northern Italy,
1999–2009. HIV Medicine. 2013 Sep; 14(8):481–90. [PubMed: 23560682]
Author Manuscript

42. Lewden C, Chene G, Morlat P, Raffi F, Dupon M, Dellamonica P, et al. HIV-infected adults with a
CD4 cell count greater than 500 cells/mm(3) on long-term combination antiretroviral therapy
reach same mortality rates as the general population. Journal of Acquired Immune Deficiency
Syndromes. 2007 Sep 1; 46(1):72–7. [PubMed: 17621240]

Int J STD AIDS. Author manuscript; available in PMC 2018 November 01.
 De La Mata et al. Page 11

Appendix 1. The median CD4 count (cells/μL) and patient totals over the
Author Manuscript

time since ART initiation, by the year period of ART initiation and country

Appendix 2. The proportion of patients in each CD4 count (cells/μL)

Author Manuscript

category over time since ART initiation, by country and year period of ART
initiation

Appendix 3. Estimated mean CD4 count (cells/μL) change when

Author Manuscript

considering an interaction between year period of ART initiation and follow-

up time from ART initiation

Month of follow-up Year period of ART initiation Mean Diff. 95% CI p value

2003–05 ref
3 2006–09 10 (4, 15) 0.001
2010–13 18 (12, 24) <0.001

2003–05 ref
6 2006–09 5 (−1, 11) 0.093
2010–13 15 (9, 22) <0.001
Author Manuscript

2003–05 ref
9 2006–09 2 (−4, 9) 0.497
2010–13 13 (6, 19) <0.001

2003–05 ref
12 2006–09 1 (−6, 7) 0.856
2010–13 10 (4, 16) 0.002

Int J STD AIDS. Author manuscript; available in PMC 2018 November 01.
 De La Mata et al. Page 12
Author Manuscript

Month of follow-up Year period of ART initiation Mean Diff. 95% CI p value

2003–05 ref
15 2006–09 0 (−6, 6) 0.979
2010–13 8 (1, 14) 0.017

2003–05 ref
18 2006–09 0 (−6, 6) 0.888
2010–13 6 (0, 12) 0.067

2003–05 ref
21 2006–09 1 (−5, 8) 0.666
2010–13 5 (−2, 12) 0.132
Author Manuscript

2003–05 ref
24 2006–09 3 (−4, 9) 0.425
2010–13 5 (−2, 12) 0.126

2003–05 ref
27 2006–09 4 (−3, 10) 0.233
2010–13 7 (0, 14) 0.046

2003–05 ref
30 2006–09 5 (−1, 11) 0.128
2010–13 10 (3, 18) 0.004

2003–05 ref
Author Manuscript

33 2006–09 6 (−2, 13) 0.150

2010–13 16 (7, 24) <0.001

2003–05 ref
36 2006–09 5 (−6, 17) 0.354
2010–13 23 (10, 36) <0.001

Note: Multivariate model adjusts for age at ART initiation, gender, HIV mode of exposure, pre-ART HIV viral load, pre-
ART CD4 count, first ART regimen, hepatitis B co-infection, hepatitis C co-infection and clinical site.

Appendix 4. Comparison of estimated mean CD4 count (cells/μL) change

up to 36 months from ART initiation using a mixed linear model, with
random intercept and random slope for time from ART initiation
Author Manuscript

Model 1 Model 2

Mean Diff. 95% CI p value Mean Diff. 95% CI p value

Year of ART Initiation <0.001 <0.001

2003–2005 ref ref
2006–2009 6 (2, 10) 0.001 7 (3, 11) 0.001

Int J STD AIDS. Author manuscript; available in PMC 2018 November 01.
 De La Mata et al. Page 13
Author Manuscript

Model 1 Model 2

Mean Diff. 95% CI p value Mean Diff. 95% CI p value

2010–2013 9 (5, 13) <0.001 11 (7, 15) <0.001

Time from ART initiation (per 16 (15, 16) <0.001 16 (15, 16) <0.001
month)

Age at ART initiation (years) <0.001

≤30 ref ref
31–40 −5 (−8, −2) 0.003 −5 (− −9, −2) 0.001
41–50 −11 (−15, −7) <0.001 −12 (−16, −8) <0.001
51+ −7 (−12, −2) 0.004 −8 (−13, −3) 0.001
Author Manuscript

Sex
Male ref ref
Female 1 (−2, 4) 0.445 1 (−2, 5) 0.332

Mode of HIV Exposure <0.001 <0.001

Heterosexual contact ref ref
Homosexual contact 11 (5, 16) <0.001 11 (6, 16) <0.001
Injecting drug use −17 (−24, −10) <0.001 −17 (−24, −10) <0.001
Other/unknown 4 (−1, 9) 0.111 4 (−1, 9) 0.111

Pre-ART CD4 cell count (cells/μL)

≤50 ref
51–100 11 (7, 16) 0.000
Author Manuscript

101–200 9 (5, 13) 0.000

201+ −3 (−6, 1) 0.169

First ART regimen 0.009 0.013

NRTI1+NNRTI2 ref ref
NRTI1+PI3 −4 (−10, 3) 0.263 −3 (−9, 3) 0.374
Other/unknown 20 (6, 35) 0.007 20 (5, 35) 0.007

Note: Global p-values for year of ART initiation, pre-ART CD4 count and age are test for trend. Other global p-values are
test for heterogeneity.
1
NRTI = nucleoside reverse transcriptase inhibitor.
2
NNRTI = nonnucleoside reverse transcriptase inhibitor.
3
PI = protease inhibitor.
Author Manuscript

Int J STD AIDS. Author manuscript; available in PMC 2018 November 01.
 De La Mata et al. Page 14

Appendix 5. Comparison of estimated mean CD4 count (cells/μL) change

Author Manuscript

up to 36 months from ART initiation using a mixed linear model, with

random intercept and random slope for time from ART initiation

Model 1 Model 2

Mean Diff. 95% CI p value Mean Diff. 95% CI p value

Year of ART Initiation <0.001 <0.001

2003–2005 ref ref
2006–2009 6 (2, 10) 0.001 6 (2, 10) 0.002
2010–2013 9 (5, 13) <0.001 9 (5, 13) <0.001

Time from ART initiation (per month) 16 (15, 16) <0.001 16 (15, 16) <0.001
Author Manuscript

Age at ART initiation (years) <0.001 <0.001

≤30 ref ref
31–40 −5 (−8, −2) 0.003 −6 (−9, −3) <0.001
41–50 −11 (−15, −7) <0.001 −13 (−17, −9) <0.001
51+ −7 (−12, −2) 0.004 −9 (−14, −4) <0.001

Sex
Male ref ref
Female 1 (−2, 4) 0.445 2 (−1, 5) 0.319

Mode of HIV Exposure <0.001 <0.001

Heterosexual contact ref ref
Author Manuscript

Homosexual contact 11 (5, 16) <0.001 11 (6, 16) <0.001

Injecting drug use −17 (−24, −10) <0.001 −11 (−19, −2) 0.011
Other/unknown 4 (−1, 9) 0.111 3 (−1, 8) 0.168

First ART regimen 0.009 0.030

NRTI1+NNRTI2 ref ref
NRTI1+PI3 −4 (−10, 3) 0.263 −4 (−10, 3) 0.252
Other 20 (6, 35) 0.007 17 (2, 31) 0.026

Pre-ART HIV viral load (copies/mL)

≤100000 ref
>100000 30 (25, 35) <0.001
Not tested 6 (1, 10) 0.014
Author Manuscript

Pre-ART CD4 (cells/μL) 0.724

≤50 ref
51–100 12 (8, 16) <0.001
101–200 10 (6, 14) <0.001
201+ 0 (−4, 3) 0.895

Hepatitis B co-infection 0.086

Int J STD AIDS. Author manuscript; available in PMC 2018 November 01.
 De La Mata et al. Page 15
Author Manuscript

Model 1 Model 2

Mean Diff. 95% CI p value Mean Diff. 95% CI p value

Negative ref
Positive −6 (−11, −1) 0.042
Not tested −3 (−8, 2) 0.291

Hepatitis C co-infection 0.016

Negative ref
Positive −10 (−16, −3) 0.005
Not tested 1 (−4, 7) 0.601

Note: Global p-values for year of ART initiation, age at ART initiation, pre-ART HIV viral load and pre-ART CD4 count
are test for trend. Other global p-values are test for heterogeneity.
Author Manuscript

1
NRTI = nucleoside reverse transcriptase inhibitor.
2
NNRTI = nonnucleoside reverse transcriptase inhibitor.
3
PI = protease inhibitor.

TAHOD-LITE study members

PS Ly and V Khol, National Center for HIV/AIDS, Dermatology & STDs, Phnom Penh,
Cambodia;

MP Lee, PCK Li, W Lam and YT Chan, Queen Elizabeth Hospital, Hong Kong, China;

N Kumarasamy, S Saghayam and C Ezhilarasi, Chennai Antiviral Research and Treatment

Clinical Research Site (CART CRS), YRGCARE Medical Centre, VHS, Chennai, India;
Author Manuscript

TP Merati, DN Wirawan and F Yuliana, Faculty of Medicine Udayana University & Sanglah
Hospital, Bali, Indonesia;

OT Ng, PL Lim, LS Lee and R Martinez-Vega, Tan Tock Seng Hospital, Singapore;

JY Choi, Na S and JM Kim, Division of Infectious Diseases, Department of Internal

Medicine, Yonsei University College of Medicine, Seoul, South Korea;

TT Pham, DD Cuong and HL Ha, Bach Mai Hospital, Hanoi, Vietnam;

KV Nguyen, HV Bui, DTH Nguyen and DT Nguyen, National Hospital for Tropical
Diseases, Hanoi, Vietnam;
Author Manuscript

AH Sohn, JL Ross and B Petersen, TREAT Asia, amfAR - The Foundation for AIDS
Research, Bangkok, Thailand;

NL De La Mata, A Jiamsakul, DC Boettiger and MG Law, The Kirby Institute, UNSW

Australia, Sydney, Australia.

Int J STD AIDS. Author manuscript; available in PMC 2018 November 01.
 De La Mata et al. Page 16
Author Manuscript
Author Manuscript

Figure 1.
The median CD4 count (cells/μL) and patient totals over the time (months) since ART
initiation, by the year period of ART initiation.
Author Manuscript
Author Manuscript

Int J STD AIDS. Author manuscript; available in PMC 2018 November 01.
 De La Mata et al. Page 17
Author Manuscript
Author Manuscript

Figure 2.
The proportion of patients in each CD4 count (cells/μL) category over time since ART
initiation, by year period of ART initiation.
Author Manuscript
Author Manuscript

Int J STD AIDS. Author manuscript; available in PMC 2018 November 01.
 De La Mata et al. Page 18

Table 1

Summary of the patient characteristics across all countries.

Author Manuscript

2003–05 2006–09 2010–13

n (%) n (%) n (%)

Total 2874 6248 7840

Age
≤30 736 (25) 1505 (24) 2087 (26)
31–40 1440 (50) 2865 (46) 3343 (43)
41–50 508 (18) 1272 (20) 1555 (20)
51+ 190 (7) 606 (10) 855 (11)
Median [IQR] 35 [30, 40] 36 [31, 42] 36 [30, 43]
Sex
Male 2149 (75) 4312 (69) 5176 (66)
Author Manuscript

Female 721 (25) 1931 (31) 2657 (34)

Transgender 4 (<0.2) 5 (<0.1) 7 (<0.1)
Mode of HIV exposure
Heterosexual 2529 (88) 5272 (85) 5748 (73)
Homosexual 111 (4) 405 (6) 784 (10)
Injecting drug user 84 (3) 125 (2) 571 (7)
Other/Unknown 150 (5) 446 (7) 737 (10)
HCV (ever)
Negative 754 (26) 2609 (42) 4093 (52)
Positive 80 (3) 171 (3) 776 (10)
Not tested 2040 (71) 3468 (55) 2971 (38)
HBV (ever)
Author Manuscript

Negative 939 (33) 2869 (46) 4702 (60)

Positive 108 (4) 305 (5) 473 (6)
Not tested 1827 (63) 3074 (49) 2665 (34)
Pre-ART CD4 (cells/μL)
≤50 708 (25) 1295 (21) 1620 (21)
51–100 543 (19) 915 (15) 833 (10)
101–200 793 (27) 1748 (28) 1307 (17)
>200 377 (13) 1325 (21) 2990 (38)
Not tested 453 (16) 967 (15) 1094 (14)
Median [IQR] 96 [45, 171] 128 [52, 201] 172 [53, 286]
Pre-ART viral load (copies/mL)
≤105 191 (7) 458 (7) 1073 (14)
Author Manuscript

>105 204 (7) 518 (8) 1154 (15)

Not tested 2479 (86) 5272 (85) 5613 (71)
Median [IQR] 106 000 [32 000, 261 000] 114 000 [29 785, 351 500] 110 564 [30 563, 402 000]
First ART regimen
NRTI+NNRTI 2719 (95) 5930 (95) 7381 (95)

Int J STD AIDS. Author manuscript; available in PMC 2018 November 01.
 De La Mata et al. Page 19

2003–05 2006–09 2010–13

n (%) n (%) n (%)
Author Manuscript

NRTI+PI 148 (5) 290 (5) 381 (5)

Other 7 (<0.3) 28 (<0.5) 78 (1)
Previous mono/dual therapy
No 2632 (92) 6053 (97) 7680 (98)
Yes 242 (8) 195 (3) 160 (2)
Author Manuscript
Author Manuscript
Author Manuscript

Int J STD AIDS. Author manuscript; available in PMC 2018 November 01.
 Author Manuscript Author Manuscript Author Manuscript Author Manuscript

Table 2

Estimated mean CD4 count (cells/μL) change up to 36 months from ART initiation.

Univariate Multivariate

Mean Diff. 95% CI p value Mean Diff. 95% CI p value

De La Mata et al.

Year of ART Initiation <0.001 <0.001

2003–2005 ref ref
2006–2009 8 (3, 13) 0.003 8 (3, 13) 0.002
2010–2013 13 (8, 19) <0.001 15 (9, 20) <0.001

Time from ART initiation (per month) 16 (15, 16) <0.001 16 (16, 17) <0.001

Age at ART initiation (years) <0.001 <0.001

≤30 ref ref
31–40 −11 (−15, −6) <0.001 −9 (−13, −5) <0.001
41–50 −21 (−26, −16) <0.001 −20 (−25, −15) <0.001
51+ −19 (−25, −13) <0.001 −17 (−24, −11) <0.001

Sex
Male ref ref
Female 18 (14, 22) <0.001 17 (13, 21) <0.001

Mode of HIV Exposure <0.001 <0.001

Heterosexual contact ref ref
Homosexual contact 16 (9, 23) <0.001 18 (11, 25) <0.001
Injecting drug use −21 (−30, −12) <0.001 −7 (−18, 3) 0.186

Int J STD AIDS. Author manuscript; available in PMC 2018 November 01.
Other/unknown 6 (0, 13) 0.061 4 (−2, 11) 0.167

Pre-ART HIV viral load (copies/mL)

≤100000 ref ref
>100000 41 (34, 47) <0.001 38 (32, 45) <0.001
Not tested 7 (1, 13) 0.024 5 (−1, 11) 0.117

Pre-ART CD4 (cells/μL) <0.001 <0.001

≤50 ref ref
Page 20
 Author Manuscript Author Manuscript Author Manuscript Author Manuscript

Univariate Multivariate

Mean Diff. 95% CI p value Mean Diff. 95% CI p value

51–100 10 (4, 15) 0.001 9 (3, 14) 0.002
101–200 5 (1, 10) 0.040 3 (−2, 8) 0.230
201+ −8 (−13, −4) <0.001 −15 (−20, −10) <0.001
De La Mata et al.

First ART regimen 0.031 0.109

NRTI1+NNRTI2 ref ref

NRTI1+PI3 −4 (−12, 4) 0.331 −5 (−13, 3) 0.246

Other 24 (4, 44) 0.020 16 (−4, 36) 0.108

Hepatitis B co-infection
Negative ref ref
Positive −12 (−20, −5) 0.001 −11 (−18, −4) 0.003
Not tested −5 (−10, −1) 0.033 −5 (−12, 2) 0.134

Hepatitis C co-infection
Negative ref ref
Positive −20 (−28, −12) <0.001 −12 (−21, −3) 0.008
Not tested −4 (−9, 1) 0.145 3 (−5, 10) 0.478

Note: Global p-values for year of ART initiation, age and pre-ART CD4 count are test for trend. Other global p-values are test for heterogeneity.
1
NRTI = nucleoside reverse transcriptase inhibitor.
2
NNRTI = nonnucleoside reverse transcriptase inhibitor.
3

Int J STD AIDS. Author manuscript; available in PMC 2018 November 01.
PI = protease inhibitor.
Page 21