Research

Cotrimoxazole prophylaxis and antiretroviral therapy: an observational

cohort study in China
Wei Cheng,a Yasong Wu,a Yi Wen,a Ye Ma,a Decai Zhao,a Zhihui Dou,a Weiwei Zhang,a Marc Bulterysb &
Fujie Zhanga

Objective To assess if cotrimoxazole prophylaxis administered early during antiretroviral therapy (ART) reduces mortality in Chinese adults
who are infected with human immunodeficiency virus (HIV).
Methods We did a retrospective observational cohort study using data from the Chinese national free antiretroviral database. Patients
older than 14 years who started ART between 1 January 2010 and 31 December 2012 and had baseline CD4+ T-lymphocyte (CD4+ cell)
count less than 200 cells/µL were followed until death, loss to follow-up or 31 December 2013. Hazard ratios (HRs) for several variables
were calculated using multivariate analyses.
Findings The analysis involved 23 816 HIV-infected patients, 2706 of whom died during the follow-up. Mortality in patients who did and did
not start cotrimoxazole during the first 6 months of ART was 5.3 and 7.0 per 100 person–years, respectively. Cotrimoxazole was associated
with a 37% reduction in mortality (hazard ratio, HR: 0.63; 95% confidence interval, CI: 0.56–0.70). Cotrimoxazole in addition to ART reduced
mortality significantly over follow-up lasting 6 months (HR: 0.65; 95% CI: 0.59–0.73), 12 months (HR: 0.58; 95% CI: 0.49–0.70), 18 months
(HR: 0.49; 95% CI: 0.38–0.63) and 24 months (HR: 0.66; 95% CI: 0.48–0.90). The mortality reduction was evident in patients with baseline
CD4+ cell counts less than 50 cells/µL (HR: 0.60; 95% CI: 0.54–0.67), 50–99 cells/µL (HR: 0.66; 95% CI: 0.56–0.78) and 100–199 cells/µL (HR:
0.78; 95% CI: 0.62–0.98).
Conclusion Cotrimoxazole prophylaxis started early during ART reduced mortality and should be offered to HIV-infected patients in low-
and middle-income countries.

settings because of weak drug procurement systems, problems

Introduction with managing the supply chain and little awareness of this
With the rapid rise in the availability of antiretroviral therapy recommendation among medical staff.19–21 The situation is
(ART), the number of people dying from acquired immune similar in China where some health-care workers are still un-
deficiency syndrome (AIDS) has dramatically declined. aware of the substantial benefits of cotrimoxazole prophylaxis
However, in many low- and middle-income countries, ART in eligible HIV-infected patients.
coverage remains relatively poor.1 Moreover, mortality during In 2003, China officially launched its national free an-
the early stages of ART is higher in low- than high-income tiretroviral treatment programme. Two years later, the first
countries, even after adjusting for immunodeficiency at base- guidelines on cotrimoxazole in the country were released.
line.2 Reducing AIDS-related mortality in low- and middle- They recommended that cotrimoxazole prophylaxis be given
income countries will require the continued expansion of ART to all individuals older than 14 years with a CD4+ cell count
coverage, improved regimens and the adoption of additional less than 200 cells/µL, WHO stage-4 disease or a history of
cost-effective interventions. oropharyngeal candidiasis. Several studies have reported
Prophylaxis with cotrimoxazole – a combination of the substantially better survival among HIV-infected patients
antibiotics trimethoprim and sulfamethoxazole – provides since the programme was launched.22–24 However, nationally-
protection against several opportunistic infections, includ- representative data on the implementation of cotrimoxazole
ing Pneumocystis jirovecii pneumonia, malaria and cerebral prophylaxis has been available only since 2010, when these
toxoplasmosis. 3–6 Many studies have demonstrated that first started to be recorded in the programme’s information
cotrimoxazole prophylaxis increases survival among ART- system. Consequently, although the protective effect of co-
naïve patients: reductions in mortality of 19 to 46% have been trimoxazole has been documented elsewhere, evidence from
reported in low- and middle-income countries.6–12 The World China has been limited. The aim of this study was to use data
Health Organization (WHO) recommends that cotrimoxazole from the programme’s database to evaluate if cotrimoxazole
be given in such settings to adolescents (10–19 years) and prophylaxis initiated early during ART was associated with
adults with advanced human immunodeficiency virus (HIV) reduced mortality among HIV-infected patients.
clinical disease or a CD4+ T lymphocyte (CD4+) cell count
less than 350 cells/µL.13 Several studies have reported that
cotrimoxazole provides a continuous protective effect after
Methods
ART initiation.14–20 Yet, the implementation of cotrimoxazole We performed a retrospective, observational, cohort study
prophylaxis remains a challenge in low- and middle-income using data from the national free antiretroviral treatment

a
Division of Treatment and Care, National Center for AIDS/STD Control and Prevention, Chinese Center for Disease Control and Prevention, 27 Nanwei Rd, Beijing
100050, China.
b
Global AIDS Program, Centers for Disease Control and Prevention, United States Embassy, Beijing, China.
Correspondence to Fujie Zhang (e-mail: [email protected]).
(Submitted: 12 July 2014 – Revised version received: 1 December 2014 – Accepted: 3 December 2014 – Published online: 29 January 2015 )

152 Bull World Health Organ 2015;93:152–160 | doi: http://dx.doi.org/10.2471/BLT.14.142745

 Research
Wei Cheng et al. Cotrimoxazole and antiretroviral therapy in China

Fig. 1. Flowchart for the selection of HIV-infected patients who received antiretroviral more scheduled follow-up visits within
therapy, China, 2010–2013 6 months of ART initiation as belonging
to our cotrimoxazole group. Our non-
cotrimoxazole group comprised patients
127 593 HIV-infected patients registered with the national free antiretroviral treatment programme database started ART who did not report cotrimoxazole use
between 1 January 2010 and 31 December 2012 at any follow-up visit following ART
initiation.
67 551 patients were excluded from the analysis: The observation period in the
• 61 623 (91%) had a baseline CD4-cell count ≥ 200 cells/µL
• 5928 (9%) had an unreported baseline CD4-cell count non-cotrimoxazole group was the time
from the date of ART initiation to the
60 042 patients with a baseline CD4-cell count < 200 cells/µL
date of death, loss to follow-up or the
last follow-up visit before 31 December
36 226 were excluded:
2013. For the cotrimoxazole group, the
• 24 779 (68%) had data missing on cotrimoxazole use observation period was the time from
• 8753 (24%) had data missing on weight, height or history of tuberculosis the date of the last follow-up visit before
• 1654 (5%) died because of a drug overdose, suicide or an accident or stopped ART the visit during which cotrimoxazole use
• 1040 (3%) started cotrimoxazole more than 6 months after ART initiation
was first reported to the date of death,
loss to follow-up or the last follow-up
23 816 patients included in the analysis
visit before 31 December 2013. For
example, if cotrimoxazole prophylaxis
11 769 in the cotrimoxazole group
• 10 046 (85%) survived was first reported in the fourth follow-
• 1204 (10%) died up visit, the observation period started
• 519 (5%) were lost to follow-up at the date of the third follow-up visit.
Since we did not know exactly when
12 047 in the non-cotrimoxazole group cotrimoxazole was stopped, we assumed
• 10 132 (84%) survived that patients in the cotrimoxazole group
• 1502 (12%) died
• 413 (4%) were lost to follow-up continued to take the drug throughout
the observation period and, therefore,
ART: antiretroviral therapy; HIV: human immunodeficiency virus.
regarded the duration of cotrimoxazole
prophylaxis as being identical to the
observation period. The study was ap-
programme database, held at the Chi- internet-based system. Records were proved by the institutional review board
nese Center for Disease Control and screened and selected from among those of the National Center for AIDS/STD
Prevention. The database contains in the programme database on 1 January (acquired immunodeficiency syndrome/
information on the treatment of all 2014 if the following inclusion criteria sexually transmitted diseases) Control
Chinese patients with HIV infections, were satisfied: (i) ART was initiated and Prevention.
who are managed by the programme. between 1 January 2010 and 31 De-
Statistical analysis
In accordance with national policy, local cember 2012; (ii) the patient’s baseline
health-care providers recorded details CD4+ cell count was less than 200/µL; Differences between the groups in
of treatment at initiation, at follow-up (iii) there was a record of whether or baseline characteristics, such as age,
visits 0.5, 1, 2 and 3 months later and not the patient received cotrimoxazole sex, marital status, HIV infection route,
once every 3 months thereafter. If the during every scheduled follow-up visit; baseline CD4+ cell count, WHO clini-
regimen changes, the follow-up schedule and (iv) details of the patient’s weight, cal stage and tuberculosis infection in
is restarted at 0.5 months. As in previous height and history of tuberculosis at the year before ART initiation, were
analyses,24,25 patients were considered baseline were available. We excluded compared using an χ2 test for categorical
lost to follow-up if they missed four records of patients who died from either variables and a Mann–Whitney U test
consecutive follow-up appointments a drug overdose, suicide or an accident, for continuous variables since no contin-
and their records were terminated at who started cotrimoxazole more than uous variable fulfilled the Kolmogorov–
the date of the last follow-up visit. Hard 6 months after ART initiation or who Smirnov test for normality. A low body
copies of all treatment data were kept stopped ART (Fig. 1). mass index (BMI) was defined as a value
at the treatment sites and all informa- less than 18.5 kg/m2, in accordance with
Cotrimoxazole prophylaxis
tion was entered into the programme’s the threshold for malnutrition of the
database.25 By 31 December 2013, the Cotrimoxazole was administered orally Food and Agriculture Organization
database contained information on and stopped in accordance with instruc- of the United Nations.27 Patients’ data
278 085 HIV-infected patients. tions in the China free ART manual.26 were censored at 31 December 2013 or
Although cotrimoxazole has been Since the database recorded only infor- the date of loss to follow-up or death,
officially recommended in China since mation on whether or not the patient depending on which was earliest. The
2005, the national free antiretroviral was receiving cotrimoxazole at each association between mortality and
treatment programme database did follow-up visit and did not register the cotrimoxazole and baseline CD4+ cell
not record cotrimoxazole use until 1 exact date on which the drug was started count after ART initiation was analysed
January 2010, when data collection was or stopped, we regarded any patient who using Kaplan–Meier survival curves
changed from a fax-based system to an reported cotrimoxazole use at one or and crude mortality rates. In addition,

Bull World Health Organ 2015;93:152–160| doi: http://dx.doi.org/10.2471/BLT.14.142745 153

 Research
Cotrimoxazole and antiretroviral therapy in China Wei Cheng et al.

Table 1. Characteristics of HIV-infected patients receiving antiretroviral therapy, by cotrimoxazole prophylaxis, China, 2010–2013

Characteristic No. (%)a of patients No. (%)a of patients No. (%)a of patients in P
(n = 23 816) in non-cotrimoxazole groupb cotrimoxazole groupc
(n = 12 047) (n = 11 769)
Duration of follow-up in 21 (14–30) 20 (14–29) 23 (15–32) < 0.001
months, median (IQR)
Number of scheduled follow- 12 (8–16) 11 (8–15) 12 (9–17) < 0.001
up visits, median (IQR)
Age in years, median (IQR) 40 (33–49) 40 (32–49) 41 (33–50) 0.009
Age, yearsd < 0.001
< 30 3 730 (16) 2 089 (17) 1 641 (14) ND
30–39 7 621 (32) 3 774 (31) 3 847 (33) ND
40–49 6 562 (28) 3 261 (27) 3 301 (28) ND
50–59 3 251 (14) 1 594 (13) 1 657 (14) ND
≥ 60 2 652 (11) 1 329 (11) 1 323 (11) ND
Sex < 0.001
Male 16 976 (71) 8 744 (73) 8 232 (70) ND
Female 6 840 (29) 3 303 (27) 3 537 (30) ND
Marital status < 0.001
Married or living with partner 15 021 (63) 7 301 (61) 7 720 (66) ND
Single, divorced or widowed 8 795 (37) 4 746 (39) 4 049 (34) ND
Route of HIV transmissiond < 0.001
Sexual transmission 19 105 (80) 9 919 (82) 9 186 (78) ND
Injection-drug use 2 221 (9) 982 (8) 1 239 (11) ND
Blood or plasma transfusion 1 405 (6) 601 (5) 804 (7) ND
Other or unknown 1 085 (5) 545 (5) 540 (5) ND
Baseline CD4+ cell count in 84 (31–148) 103 (39–159) 69 (25–133) < 0.001
cells/µL, median (IQR)
Baseline CD4+ cell count, < 0.001
cells/µL
< 50 8 357 (35) 3 576 (30) 4 781 (41) ND
50–99 4 931 (21) 2 308 (19) 2 623 (22) ND
100–199 10 528 (44) 6 163 (51) 4 365 (37) ND
BMI in kg/m2, median (IQR) 20 (19–22) 21 (19–22) 20 (18–22) < 0.001
Body mass index, kg/m2 < 0.001
≥ 18.5 17 947 (75) 9 513 (79) 8 434 (72) ND
< 18.5 5 869 (25) 2 534 (21) 3 335 (28) ND
WHO clinical staged < 0.001
1 7 212 (30) 4 376 (36) 2 836 (24) ND
2 4 764 (20) 2 753 (23) 2 011 (17) ND
3 6 014 (25) 2 699 (22) 3 315 (28) ND
4 5 826 (24) 2 219 (18) 3 607 (31) ND
Tuberculosis before ART < 0.001
Yes 2 930 (12) 1 161 (10) 1 769 (15) ND
No 20 886 (88) 10 886 (90) 10 000 (85) ND
Cotrimoxazole started at ART NA
initiation
Yes NA NA 9 517 (81) NA
No NA NA 2 252 (19) NA
ART: antiretroviral therapy; BMI: body mass index; HIV: human immunodeficiency virus; IQR: interquartile range; NA: not applicable; ND: not determined; WHO: World
Health Organization.
a
All values in the table represent absolute numbers and percentages unless otherwise stated.
b
The non-cotrimoxazole group comprised HIV-infected patients who did not report cotrimoxazole use at any time after starting antiretroviral therapy.
c
The cotrimoxazole group comprised HIV-infected patients who reported cotrimoxazole use within 6 months of starting antiretroviral therapy.
d
Percentages may not sum to 100% due to rounding.

154 Bull World Health Organ 2015;93:152–160| doi: http://dx.doi.org/10.2471/BLT.14.142745

 Research
Wei Cheng et al. Cotrimoxazole and antiretroviral therapy in China

mortality was also evaluated using Cox mortality was 6.1 per 100 person– In addition, cotrimoxazole use was
proportional hazards regression models years: 7.0 per 100 person-years in the associated with significantly reduced
that included adjustment for covariates non-cotrimoxazole group and 5.3 per mortality over a range of ART durations:
that were predetermined to be clinically 100 person–years in the cotrimoxazole the adjusted HR for death was 0.65 for
meaningful and which had a significant group. Kaplan–Meier survival curves 6 months’ treatment, 0.58 for 12 months’
influence in the unadjusted analysis (i.e. showed that survival was lowest in indi- treatment, 0.49 for 18 months’ treatment
a P-value < 0.1). To determine whether viduals with a baseline CD4+ cell count and 0.66 for 24 months’ treatment (Ta-
cotrimoxazole’s influence on mortal- less than 50 cells/µL who did not receive ble 3). There was no significant reduction
ity varied according to the duration cotrimoxazole (Fig. 2). in mortality for ART lasting between 25
of ART and the baseline CD4+ cell Univariate analysis found that the risk and 30 months (HR: 0.80). The protective
count, the Cox models were stratified of death was significantly lower in patients association of cotrimoxazole was also
by the duration of ART (i.e. 6, 12, 18, who received cotrimoxazole (hazard ratio, significant over a range of baseline CD4+
24 and 30 months) and the baseline HR: 0.80), who were female (HR: 0.75) cell counts: the adjusted HR for death was
CD4+ cell count (i.e. < 50, 50–99 and and who had no history of tuberculosis 0.60 for patients with a count less than
100–199 cells/µL). Finally, to confirm before ART (HR: 0.69; Table 2). Factors 50 cells/µL, 0.66 for those with a count
our findings, we introduced a category significantly associated with an increased of 50–99 cells/µL and 0.78 for those with
for missing data and repeated the calcu- risk of death included age more than a count of 100–199 cells/µL (Table 3).
lations using data that included obser- 30 years, an HIV transmission route other When the analysis was repeated with
vations from individuals who had been than sexual transmission, a baseline CD4+ the inclusion of a category for missing data
excluded because of missing records. All cell count less than 100 cells/µL, a BMI less so that observations from 8753 patients
calculations were performed using SAS than 18.5 kg/m2 and WHO clinical dis- who were excluded because information
version 9.2 (SAS Institute, Cary, United ease stage 2 or higher. These associations on their weight, height or tuberculosis
States of America). remained significant in the multivariate status had not been reported, we found that
analysis after adjusting for all potential cotrimoxazole remained significantly asso-
confounders, except age in the range 30 ciated with decreased mortality (HR: 0.65;
Results to 39 years and a history of tuberculosis. 95% CI: 0.59–0.72); this value for the HR is
The analysis included data on 23 816 In particular, cotrimoxazole prophylaxis close to the 0.63 found in the multivariate
patients aged 15 years or more: 12 047 was still significantly associated with a analysis of the principle study group.
(51%) had never taken cotrimoxazole, decreased risk of death (HR: 0.63).
whereas 11 769 (49%) had taken the
drug within 6 months of ART ini-
tiation (Fig. 1). Of those who reported Fig. 2. Kaplan–Meier survival curves for HIV-infected patients receiving antiretroviral
cotrimoxazole use, 2252 (19%) did not therapy, by cotrimoxazole use and baseline CD4+ cell count, China, 2010–2013
begin taking the drug at ART initiation:
the mean starting time was 27 days
1.00
(interquartile range, IQR: 14–56) after
ART initiation. The median age of all
participants was 40 years (IQR: 33–49),
71% (16 976) were male and the most 0.95
common route of HIV infection was sex-
ual transmission, followed by injection-
Survival probability

drug use, blood or plasma transfusion 0.90

and other or unknown routes. The pa-
tients’ median baseline CD4+ cell count
was 84 cells/µL – the median was signifi- 0.85
cantly lower in the cotrimoxazole group,
at 69 cells/µL, compared with 103 cells/
µL in the non-cotrimoxazole group 0.80
(Table 1). In addition, significantly more
patients in the cotrimoxazole group than log-rank P < 0.001
the non-cotrimoxazole group had a low
0.75
BMI (28% versus 21%, respectively), a 0 3 6 9 12 15 18 21 24 27 30 33 36
history of tuberculosis before ART (15% Time after initiating ART (months)
versus 10%, respectively) and WHO CD4-cell count: 100–199 cells/µL, no cotrimoxazole
clinical stage-3 disease (28% versus 22%, CD4-cell count: 50–99 cells/µL, no cotrimoxazole
respectively) or stage-4 disease (31% CD4-cell count: < 50 cells/µL, no cotrimoxazole
versus 18%, respectively). CD4-cell count: 100–199 cells/µL, cotrimoxazole
In total, 2706 patients died during CD4-cell count: 50–99 cells/µL, cotrimoxazole
44 137 person–years of observation: 56% CD4-cell count: < 50 cells/µL, cotrimoxazole
(1515/2706) were in the non-cotrimox- ART: antiretroviral therapy; HIV: human immunodeficiency virus.
azole group and 44% (1191/2706) were Note: The P-value indicates the significant survival rate by cotrimoxazole use and baseline CD4+ cell
in the cotrimoxazole group. Overall count.

Bull World Health Organ 2015;93:152–160| doi: http://dx.doi.org/10.2471/BLT.14.142745 155

 Research
Cotrimoxazole and antiretroviral therapy in China Wei Cheng et al.

Table 2. Factors associated with death in HIV-infected patients receiving antiretroviral In high-income settings, mortality was
therapy, China, 2010–2013 reported to fall from 24 per 1000 per-
son–years during the first 6 months of
Factor Risk of death treatment to 16 per 1000 person–years
during months 7 to 12. In low-income
Unadjusted HR (95% CI) Adjusted HR (95% CI) settings, this trend was even more dra-
Cotrimoxazole use a matic: the mortality rate was reported to
No Reference Reference fall by nearly half from 51 per 1000 per-
Yes 0.80 (0.74–0.86) 0.63 (0.56–0.70) son–years during the first 6 months of
Age, years treatment to 27 per 1000 person–years
< 30 Reference Reference during months 7 to 12.2 In sub-Saharan
Africa, it has been reported that 8 to
30–39 1.22 (1.07–1.41) 1.14 (0.99–1.32)
26% of patients die during the first year
40–49 1.40 (1.21–1.62) 1.36 (1.17–1.57)
of ART, with most deaths occurring in
50–59 1.51 (1.27–1.79) 1.69 (1.42–2.01)
the first few months.28 A similar trend
≥ 60 2.03 (1.69–2.45) 2.42 (1.99–2.93) has been observed in China.23 We found
Sex that cotrimoxazole prophylaxis during
Male Reference Reference the first 6 months of ART was associated
Female 0.75 (0.69–0.82) 0.83 (0.76–0.91) with a one-third reduction in mortal-
Marital status ity. Given the high underlying risk of
Married or living with partner Reference Reference death in this period, this improvement
Single, divorced or widowed 1.05 (0.97–1.13) ND in mortality could correspond to a
Route of HIV transmission substantial reduction in the number of
Sexual transmission Reference Reference deaths. Moreover, in our study, cotri-
Blood or plasma transfusion 1.35 (1.17–1.57) 1.45 (1.25–1.70) moxazole appeared to be associated with
Injection-drug use 1.72 (1.51–1.96) 2.17 (1.88–2.50) decreased mortality with a duration of
Other or unknown 1.19 (0.97–1.47) 1.19 (0.96–1.48) ART up to about 24 months. This obser-
Baseline CD4+ cell count, cells/µL
vation still has to be explained.
In many countries, late access
100–199 Reference Reference
to ART has been identified as an im-
50–99 2.65 (2.34–3.00) 2.52 (2.22–2.86)
portant challenge for treatment pro-
< 50 4.41 (3.86–5.04) 4.23 (3.68–4.87)
grammes.23,24,31–34 One study in eastern
Body mass index, kg/m2 Africa found that the median baseline
≥ 18.5 Reference Reference CD4+ cell count in patients starting
< 18.5 2.63 (2.37–2.93) 2.13 (1.90–2.39) ART in 2008 and 2009 was 154 cells/µL32
WHO clinical stage and similar findings have been reported
1 Reference Reference in China. 23,24,31 In addition, several
2 1.44 (1.27–1.63) 1.17 (1.03–1.33) studies have demonstrated a strong as-
3 2.08 (1.82–2.36) 1.37 (1.19–1.57) sociation between a low baseline CD4+
4 2.38 (2.05–2.75) 1.43 (1.22–1.67) cell count and high mortality after ART
Tuberculosis before ART initiation, especially among people with
Yes Reference Reference a baseline count less than 50 cells/µL.23,24
No 0.69 (0.62–0.76) 0.91 (0.82–1.01) Although the baseline CD4+ cell count
at ART initiation in China improved be-
ART: antiretroviral therapy; CI: confidence interval; HIV: human immunodeficiency virus; HR: hazard ratio; ND:
not determined; WHO: World Health Organization. tween 2006 and 2009, in 2009 about one
a
Patients were regarded as having taken cotrimoxazole if they reported drug use within 6 months of third of people still started ART with a
starting antiretroviral therapy. Other patients did not report cotrimoxazole use at any time. CD4+ cell count no higher than 50 cells/
µL.31 In our study, the association be-
Our findings are consistent with those of tween co-trimoxazole-use and reduced
Discussion other studies14–20 and reinforce Chinese mortality was most marked in patients
In this large nationally-representative national guidelines, which recommend with a baseline CD4+ cell count less than
study, we found that cotrimoxazole extending the use of cotrimoxazole to 50 cells/µL: the reduction in mortality
prophylaxis started early during ART re- patients with late-disease stage. was 40%. This indicates that cotrimoxa-
duced mortality by 37% in HIV-infected Although the expanded use of ART zole has the potential to reduce mortality
adults; the decrease was greatest among has substantially improved mortality among patients with the highest risk
those with advanced disease. The reduc- and morbidity in HIV-infected patients during their most vulnerable period.
tion was evident as long as 24 months around the world, several challenges re- Worldwide, the use of cotrimoxa-
after ART initiation and was observed main that could undermine these gains, zole is suboptimal. In 2010, WHO found
in patients with severe immunosuppres- especially in low-resource countries. that, although the national policy in 38
sion. The protective effect of cotrimoxa- Higher mortality has been observed of 41 countries surveyed was to provide
zole has previously been documented during the early stages of ART in both cotrimoxazole to people living with
elsewhere in ART-naïve patients.3–12 high- and low-income settings.2,23,28–30 HIV infections, only 25 of the 38 had

156 Bull World Health Organ 2015;93:152–160| doi: http://dx.doi.org/10.2471/BLT.14.142745

 Research
Wei Cheng et al. Cotrimoxazole and antiretroviral therapy in China

characteristics between included and

Table 3. Cotrimoxazole and risk of death in HIV-infected patients receiving excluded patients and found that the
antiretroviral therapy, China, 2010–2013 distributions of key variables such as
age, sex and baseline CD4+ cell count
Characteristic         Risk of death with cotrimoxazole versus were similar in the two groups. We
no cotrimoxazolea believe, therefore, it is unlikely that our
Unadjusted HR (95% CI) Adjusted HR (95% CI)b main findings were affected by biases in
these key characteristics. Furthermore,
Duration of ART, months
our findings are consistent with those of
0–6 0.87 (0.79–0.96) 0.65 (0.59–0.73)
studies conducted in other countries.14–20
7–12 0.71 (0.60–0.85) 0.58 (0.49–0.70)
Finally, since our data reflect the realities
13–18 0.57 (0.44–0.72) 0.49 (0.38–0.63) of treatment in a middle-income setting,
19–24 0.70 (0.52–0.95) 0.66 (0.48–0.90) our conclusions may not be generaliz-
25–30 0.79 (0.50–1.24) 0.80 (0.50–1.29) able to all countries.
Baseline CD4+ cell count, In conclusion, by using data on a
cells/µL large national cohort, we found that
< 50 0.62 (0.56–0.70) 0.60 (0.54–0.67) cotrimoxazole prophylaxis adminis-
50–99 0.75 (0.64–0.87) 0.66 (0.56–0.78) tered early during ART significantly
100–199 0.83 (0.71–0.97) 0.78 (0.62–0.98) reduced mortality in HIV-infected
ART: antiretroviral therapy; CI: confidence interval; HIV: human immunodeficiency virus; HR: hazard ratio. patients, particularly in those with a
a
Patients were regarded as having used cotrimoxazole if they reported drug use within 6 months of CD4+ cell count less than 50 cells/µL.
starting antiretroviral therapy (ART). Other patients did not report cotrimoxazole use at any time. Given its beneficial effect on survival
b
Adjusted for age, sex, marital status, route of HIV transmission, body mass index, World Health
Organization clinical stage and history of tuberculosis before ART. In addition, in calculating hazard ratios
and its relatively low cost, cotrimoxa-
for different durations of ART, adjustment was also made for the baseline CD4+ cell count. zole should be offered in conjunction
with ART to HIV-infected patients
in China and other low- and middle-
fully implemented this policy.21 Major tality for different exposure times of income countries. ■
obstacles included an erratic drug sup- cotrimoxazole. Moreover, by regarding
ply, drug stocks running out and poor all patients in the cotrimoxazole group Acknowledgements
knowledge of cotrimoxazole among as having taken the drug continuously We thank local county staff of the Chi-
health-care workers and patients.19–21,35,36 from the starting date, we may have nese Center for Disease Control and
In our study, we found that only half underestimated the reduction in mor- Prevention. Wei Cheng and Yasong Wu
of participants received cotrimoxazole tality. Second, data were retrieved from contributed equally to this work.
prophylaxis. Moreover, a fifth of those an observational database and may Wei Cheng is also affiliated with
treated did not start the drug at ART therefore have inherent biases, such Zhejiang Provincial Center for Dis-
initiation. Over the past few years, China as reporting or recall bias, associated ease Control and Prevention, Hang-
has successfully established a working with the collection of nonrandom data. zhou, China. Fujie Zhang is also
system and infrastructure for the treat- However, the large number of patients affiliated with the Department of
ment of HIV-infected patients as well in the database should have helped miti- Infectious Diseases, Beijing Ditan
as institutions for monitoring treatment gate bias in any one direction. Finally, Hospital–Capital Medical University,
– these facilities could promote more the large number of patients excluded Beijing, China.
extensive use of cotrimoxazole.22–25 because of missing data may have given
This study has several limitations. rise to selection bias. However, when we Funding: This work was supported by the
First, due to the lack of detailed infor- repeated the analysis by including data National Centre for AIDS/STD Control
mation on cotrimoxazole treatment from individuals who were excluded and Prevention of the Chinese Centre for
and on adherence to treatment, we because details of their weight, height Disease Control and Prevention.
equated the duration of cotrimoxazole or history of tuberculosis were miss-
prophylaxis to the observation time in ing, our findings were consistent with Competing interests: None declared.
the cotrimoxazole group. Consequently, those of the main analysis. Moreover,
we could not assess reduction of mor- we assessed heterogeneity in baseline

‫ملخص‬
‫ دراسة أترابية قائمة عىل املالحظة يف‬:‫الوقاية باستخدام الكوتريموكسازول والعالج بمضادات الفريوسات القهقرية‬
‫الصني‬
‫الطريقة أجرينا دراسة أترابية اسرتجاعية قائمة عىل املالحظة‬ ‫الغرض تقييم ما إذا كانت الوقاية باستخدام الكوتريموكسازول‬
‫باستخدام البيانات املستمدة من قاعدة بيانات مضادات‬ ‫يف مرحلة مبكرة خالل العالج بمضادات الفريوسات القهقرية‬
‫ وتم متابعة املرىض‬.‫الفريوسات القهقرية املجانية الوطنية يف الصني‬ ‫يقلل معدل الوفيات بني البالغني املصابني بفريوس العوز املناعي‬
‫ عام ًا الذين استهلوا العالج بمضادات الفريوسات‬14 ‫األكرب من‬ .‫البرشي يف الصني‬

Bull World Health Organ 2015;93:152–160| doi: http://dx.doi.org/10.2471/BLT.14.142745 157

 Research
Cotrimoxazole and antiretroviral therapy in China Wei Cheng et al.

‫؛ فاصل‬0.58 :‫ شهر ًا (نسبة املخاطر‬12‫ و‬،)0.73 ‫ إىل‬0.59 ‫من‬ 31 ‫ حتى‬2010 ‫ يناير‬/‫ كانون الثاين‬1 ‫القهقرية يف الفرتة من‬
:‫ شهر ًا (نسبة املخاطر‬18‫ و‬،)0.70 ‫ إىل‬0.49 ‫ من‬:% 95 ‫الثقة‬ ‫ وكان إحصاء اخلاليا اللمفاوية‬2012 ‫ ديسمرب‬/‫كانون األول‬
‫ شهر ًا‬24‫ و‬،)0.63 ‫ إىل‬0.38 ‫ من‬:% 95 ‫؛ فاصل الثقة‬0.49 200 ‫) عند خط األساس لدهيم أقل من‬CD4+( ‫التائية املساعدة‬
.)0.90 ‫ إىل‬0.48 ‫ من‬:% 95 ‫؛ فاصل الثقة‬0.66 :‫(نسبة املخاطر‬ ‫ كانون‬31 ‫ أو حتى‬،‫ أو فقدان املتابعة‬،‫ميكرو لرت حتى الوفاة‬/‫خلية‬
‫واتضح االنخفاض يف معدل الوفيات بني املرىض الذين كانت‬ ‫ وتم حساب نسب املخاطر لعدة متغريات‬.2013 ‫ ديسمرب‬/‫األول‬
‫) عند خط‬CD4+( ‫إحصاءات اخلاليا اللمفاوية التائية املساعدة‬ .‫باستخدام التحليالت متعددة املتغريات‬
:‫ميكرو لرت (نسبة املخاطر‬/‫ خلية‬50 ‫األساس لدهيم أقل من‬ ً ً
‫ مريضا مصابا بفريوس‬23816 ‫النتائج اشتمل التحليل عىل‬
‫ إىل‬50 ‫ ومن‬،)0.67 ‫ إىل‬0.54 ‫ من‬:% 95 ‫؛ فاصل الثقة‬0.60 .‫ مريض ًا أثناء املتابعة‬2706 ‫ تويف منهم‬،‫العوز املناعي البرشي‬
:% 95 ‫؛ فاصل الثقة‬0.66 :‫ميكرو لرت (نسبة املخاطر‬/‫ خلية‬99 ‫وكان معدل الوفيات بني املرىض الذين استهلوا ومل يستهلوا‬
‫ميكرو لرت (نسبة‬/‫ خلية‬199 ‫ إىل‬100 ‫ ومن‬،)0.78 ‫ إىل‬0.56 ‫من‬ ‫استخدام الكوتريموكسازول خالل الستة أشهر األوىل من العالج‬
.)0.98 ‫ إىل‬0.62 ‫ من‬:% 95 ‫؛ فاصل الثقة‬0.78 :‫املخاطر‬ ‫ شخص إىل‬100 ‫ لكل‬7.0‫ و‬5.3 ‫بمضادات الفريوسات القهقرية‬
‫االستنتاج أدت الوقاية باستخدام الكوتريموكسازول التي تم‬ ‫ وارتبط الكوتريموكسازول بانخفاض‬.‫ عىل التوايل‬،‫عدد السنوات‬
‫بدأت يف مرحلة مبكرة خالل العالج بمضادات الفريوسات‬ ‫؛ فاصل‬0.63 :‫ (نسبة املخاطر‬% 37 ‫يف معدل الوفيات نسبته‬
‫القهقرية إىل تقليل معدل الوفيات ومن ثم ينبغي توفريها للمرىض‬ ‫ وأدى استخدام‬.)0.70 ‫ إىل‬0.56‫ من‬:‫ فاصل الثقة‬،% 95 ‫الثقة‬
‫املصابني بفريوس العوز املناعي البرشي يف البلدان املنخفضة‬ ‫الكوتريموكسازول باإلضافة إىل العالج بمضادات الفريوسات‬
.‫الدخل والبلدان املتوسطة الدخل‬ ‫القهقرية إىل تقليل معدل الوفيات بشكل كبري عىل مدار فرتة املتابعة‬
:% 95 ‫؛ فاصل الثقة‬0.65 :‫ أشهر (نسبة املخاطر‬6 ‫التي دامت‬

摘要
磺胺甲基异恶唑预防疗法和抗逆转录病毒治疗 ：中国的观察性队列研究
目的 评估抗逆转录病毒治疗（ART）早期进行磺胺甲 法与死亡率降低 37%（危害比、HR：0.63；95% 置信区间，
基异恶唑预防疗法是否降低中国艾滋病病毒（HIV） CI ：0.56-0.70）相关。ART 结合磺胺甲基异恶唑预防
感染成年人的死亡率。 疗法显著降低持续 6 个月（HR ：0.65 ；95% CI ：0.59-
方法 我们使用中国国家免费抗逆转录病毒数据库的数 0.73）、12 个月（HR ：0.58 ；95% CI ：0.49-0.70）、18 个
据进行回顾性观察性队列研究。对 14 岁以上在 2010 月（HR ：0.49 ；95% CI ：0.38-0.63）和 24 个月（HR ：0.66 ；
年 1 月 1 日 至 2012 年 12 月 31 日 之 间 开 始 ART 治 疗 95% CI ：0.48-0.90）随访期间的死亡率。基准 CD4+ 细
并且基准 CD4+T 淋巴细胞（CD4+ 细胞）计数小 200 胞计数小于 50 cells/µL（HR ：0.60 ；95% CI ：0.54-0.67）、
cells/µL 的患者进行跟踪随访，直至死亡、失访或到 50-99 cells/µL（HR ：0.66 ；95% CI ：0.56-0.78）和 100-
2013 年 12 月 31 日。使用多元分析计算数个变量的风 199 cells/µL（HR ：0.78 ；95% CI ：0.62-0.98） 的 患 者
险比（HR）。 的死亡率降低明显。
结果 该分析涉及 23816 名艾滋病病毒感染者，其中 结论 在 ART 治疗早期加入磺胺甲基异恶唑预防疗法
2706 人在随访期间死亡。在 ART 治疗前 6 个月开始接 可降低死亡率，应为中低收入国家艾滋病病毒感染者
受或者未接受磺胺甲基异恶唑预防疗法的病人死亡率 提供此疗法。
分别为每 100 人年 5.3 和 7.0。磺胺甲基异恶唑预防疗

Résumé
Prophylaxie par le cotrimoxazole et la thérapie antirétrovirale: une étude de cohorte par observation en Chine
Objectif Évaluer si la prophylaxie par le cotrimoxazole administrée Le cotrimoxazole était associé avec 37% de réduction de la mortalité
précocement au cours de la thérapie antirétrovirale (TAR) réduit la (rapport des risques, RR: 0,63; intervalle de confiance à 95%, IC 95%:
mortalité chez les adultes infectés par le virus de l’immunodéficience 0,56-0,70). Le cotrimoxazole ajouté à la TAR a réduit significativement
humaine (VIH) en Chine. la mortalité au cours du suivi des 6 derniers mois (RR: 0,65; IC 95%:
Méthodes Nous avons effectué une étude de cohorte par observation 0,59–0,73), des 12 derniers mois (RR: 0,58; IC 95%: 0,49-0,70), des
rétrospective en utilisant les données de la base de données 18 derniers mois (RR: 0,49; IC 95%: 0,38-0,63) et des 24 derniers mois (RR:
antirétrovirale gratuite nationale chinoise. Les patients âgés de plus 0,66; IC 95%: 0,48-0,90). La réduction de la mortalité était évidente chez
de 14 ans qui ont commencé une TAR entre le 1er janvier 2010 et le les patients avec un nombre de cellules CD4+ à la ligne de base inférieur
31 décembre 2012 et qui avaient un nombre de lymphocytes T CD4+ à 50 cellules/µL (RR: 0,60; IC 95%: 0,54–0,67), 50–99 cellules/µL (RR: 0,66;
(cellules CD4+) à la ligne de base inférieur à 200 cellules/µL, ont été IC 95%: 0,56-0,78) et 100-199 cellules/µL (RR: 0,78; IC 95%: 0,62-0,98).
suivis jusqu’à leur décès, jusqu’à ce qu’ils soient perdus de vue au suivi Conclusion La prophylaxie par le cotrimoxazole commencée
ou jusqu’au 31 décembre 2013. Les rapports des risques (RR) pour précocement au cours de la TAR a réduit la mortalité et devrait être
plusieurs variables sont calculés en utilisant des analyses multivariées. proposée aux patients infectés par le VIH dans les pays à revenu faible
Résultats Le suivi. La mortalité chez les patients qui ont et qui n’ont pas et à revenu intermédiaire.
commencé à prendre le cotrimoxazole au cours des 6 premiers mois de
la TAR était de 5,3 et 7,0 pour 100 personnes-années, respectivement.

158 Bull World Health Organ 2015;93:152–160| doi: http://dx.doi.org/10.2471/BLT.14.142745

 Research
Wei Cheng et al. Cotrimoxazole and antiretroviral therapy in China

Резюме
Профилактика котримоксазолом и антиретровирусная терапия: обсервационное когортное
исследование в Китае
Цель Определить, приводит ли профилактика котримоксазолом 5,3 и 7,0 на 100 человеко-лет соответственно. Котримоксазол
на раннем этапе антиретровирусной терапии (АРТ) к снижению ассоциировался с 37%‑ным снижением уровня смертности
смертности среди ВИЧ-инфицированных взрослых в Китае. (отношение рисков (ОР): 0,63; 95%‑ный доверительный интервал
Методы Было проведено ретроспективное обсервационное (ДИ): 0,56-0,70). Котримоксазол в дополнение к АРТ значительно
когортное исс ледование с использованием данных снижал смертность во время последующего наблюдения,
Китайской национальной базы данных свободного доступа по длившегося 6 месяцев (ОР: 0,65; 95% ДИ: 0,59–0,73), 12 месяцев
антиретровирусной терапии. Пациенты старше 14 лет, которые (ОР: 0,58; 95% ДИ: 0,49-0,70), 18 месяцев (ОР: 0,49; 95% ДИ: 0,38-
начали получать АРТ в период с 1 января 2010 года по 31 декабря 0,63) и 24 месяца (ОР: 0,66; 95% ДИ: 0,48-0,90). Заметное снижение
2012 года, и исходное число CD4+ T-лимфоцитов которых смертности наблюдалось у пациентов с исходным показателем
составляло менее 200 клеток/мкл, находились под наблюдением CD4+ менее 50 клеток/мкл (ОР: 0,60; 95% ДИ: 0,54–0,67), 50–
до летального исхода, выбытия из наблюдения или до 31 декабря 99 клеток/мкл (ОР: 0,66; 95% ДИ: 0,56–0,78) и 100–199 клеток/мкл
2013 года. С помощью многомерных анализов определялось (ОР: 0,78; 95% ДИ: 0,62-0,98).
отношение рисков (ОР) для нескольких переменных. Вывод Профилактика котримоксазолом, начатая на раннем этапе
Результаты Анализ охватывал 23 816 ВИЧ-инфицированных АРТ, приводила к снижению смертности и должна предлагаться
пациентов, 2706 из которых скончались во время наблюдения. ВИЧ-инфицированным пациентам в странах с низким и средним
Уровень смертности среди пациентов, начавших и не начавших уровнями доходов.
прием котримоксазола в течение первых 6 месяцев АРТ, составлял

Resumen
Profilaxis con cotrimoxazol y terapia antirretroviral: un estudio observacional de cohorte en China
Objetivo Evaluar si la profilaxis con cotrimoxazol en una etapa temprana con cotrimoxazol durante los primeros 6 meses de la TAR fue de 5,3 y
de la terapia antirretroviral (TAR) reduce la mortalidad en adultos de origen 7,0 por 100 personas/años, respectivamente. Cotrimoxazol se asoció
chino infectados por el virus de la inmunodeficiencia humana (VIH). a una reducción del 37 % en la mortalidad (cociente de riesgos, CR:
Métodos Se llevó a cabo un estudio de cohorte observacional 0,63, intervalo de confianza del 95 %, IC: 0,56 – 0,70). Además de la
retrospectivo utilizando los datos de la base de datos nacional de TAR, cotrimoxazol redujo la mortalidad considerablemente durante el
China de antirretrovirales de uso gratuito. Por otro lado, se realizó un seguimiento de 6 meses (CR: 0,65; IC del 95 %: 0,59 – 0,73), 12 meses
seguimiento a pacientes mayores de 14 años que iniciaron el tratamiento (CR: 0,58; IC del 95 %: 0,49 – 0,70), 18 meses (CR: 0,49; IC del 95 %: 0,38
con TAR entre el 1 de enero de 2010 y el 31 de diciembre de 2012 y con – 0,63) y 24 meses (CR: 0,66; IC del 95 %: 0,48 – 0,90). La reducción de la
recuento basal de linfocitos T CD4+ (célula CD4+) inferior a 200 células/ mortalidad resultó evidente en pacientes con recuento basal de células
µl hasta su muerte, la pérdida de seguimiento o el 31 de diciembre de CD4+ inferior a 50 células/µl (CR: 0,60; IC del 95 %: 0,54 – 0,67), 50 – 99
2013. Se calcularon los cocientes de riesgos (CR) para diversas variables células/µl (CR: 0,66; IC del 95 %: 0,56 – 0,78) y 100 – 199 células/µl (CR:
mediante análisis multivariados. 0,78; IC del 95 %: 0,62 – 0,98).
Resultados En el análisis participaron 23 816 pacientes infectados Conclusión La profilaxis con cotrimoxazol iniciada en una etapa
por el VIH, 2706 de los cuales fallecieron durante el seguimiento. La temprana de la TAR redujo la mortalidad y debe ofrecerse a los pacientes
mortalidad entre los pacientes que iniciaron y no iniciaron el tratamiento infectados por el VIH en países de ingresos bajos y medios.

References
1. Global report: UNAIDS report on the global AIDS epidemic 2012. Geneva: Joint 6. Mermin J, Lule J, Ekwaru JP, Malamba S, Downing R, Ransom R, et al.
United Nations Programme on HIV/AIDS; 2012. Available from: http://www. Effect of co-trimoxazole prophylaxis on morbidity, mortality, CD4-cell
unaids.org.cn/en/index/Document_view.asp?id=769 [cited 2013 Nov 5]. count, and viral load in HIV infection in rural Uganda. Lancet. 2004
2. Braitstein P, Brinkhof MW, Dabis F, Schechter M, Boulle A, Miotti P, et al.; ART Oct 16-22;364(9443):1428–34. doi: http://dx.doi.org/10.1016/S0140-
Cohort Collaboration (ART-CC) groups. Mortality of HIV-1-infected patients 6736(04)17225-5 PMID: 15488218
in the first year of antiretroviral therapy: comparison between low-income 7. Wiktor SZ, Sassan-Morokro M, Grant AD, Abouya L, Karon JM, Maurice C, et
and high-income countries. Lancet. 2006 Mar 11;367(9513):817–24. doi: al. Efficacy of trimethoprim-sulphamethoxazole prophylaxis to decrease
http://dx.doi.org/10.1016/S0140-6736(06)68337-2 PMID: 16530575 morbidity and mortality in HIV-1-infected patients with tuberculosis in
3. Jones JL, Hanson DL, Dworkin MS, Alderton DL, Fleming PL, Kaplan JE, et al. Abidjan, Côte d’Ivoire: a randomised controlled trial. Lancet. 1999 May
Surveillance for AIDS-defining opportunistic illnesses, 1992–1997. MMWR 1;353(9163):1469–75. doi: http://dx.doi.org/10.1016/S0140-6736(99)03465-
CDC Surveill Summ. 1999 Apr 16;48(2):1–22. PMID: 12412613 0 PMID: 10232312
4. Mermin J, Ekwaru JP, Liechty CA, Were W, Downing R, Ransom R, et al. Effect 8. Chintu C, Bhat GJ, Walker AS, Mulenga V, Sinyinza F, Lishimpi K, et al.; CHAP
of co-trimoxazole prophylaxis, antiretroviral therapy, and insecticide-treated trial team. Co-trimoxazole as prophylaxis against opportunistic infections
bednets on the frequency of malaria in HIV-1-infected adults in Uganda: in HIV-infected Zambian children (CHAP): a double-blind randomised
a prospective cohort study. Lancet. 2006 Apr 15;367(9518):1256–61. doi: placebo-controlled trial. Lancet. 2004 Nov 20-26;364(9448):1865–71. doi:
http://dx.doi.org/10.1016/S0140-6736(06)68541-3 PMID: 16631881 http://dx.doi.org/10.1016/S0140-6736(04)17442-4 PMID: 15555666
5. Béraud G, Pierre-François S, Foltzer A, Abel S, Liautaud B, Smadja D, et al. 9. Chokephaibulkit K, Chuachoowong R, Chotpitayasunondh T, Chearskul
Cotrimoxazole for treatment of cerebral toxoplasmosis: an observational S, Vanprapar N, Waranawat N, et al.; Bangkok Collaborative Perinatal HIV
cohort study during 1994–2006. Am J Trop Med Hyg. 2009 Apr;80(4):583–7. Transmission Study Group. Evaluating a new strategy for prophylaxis to
PMID: 19346380 prevent Pneumocystis carinii pneumonia in HIV-exposed infants in Thailand.
AIDS. 2000 Jul 28;14(11):1563–9. doi: http://dx.doi.org/10.1097/00002030-
200007280-00012 PMID: 10983643

Bull World Health Organ 2015;93:152–160| doi: http://dx.doi.org/10.2471/BLT.14.142745 159

 Research
Cotrimoxazole and antiretroviral therapy in China Wei Cheng et al.

10. Thuy TT, Shah NS, Anh MH, Nghia T, Thom D, Linh T, et al. HIV-associated TB 23. Zhang F, Dou Z, Ma Y, Zhao Y, Liu Z, Bulterys M, et al. Five-year outcomes of
in An Giang Province, Vietnam, 2001–2004: epidemiology and TB treatment the China National Free Antiretroviral Treatment Program. Ann Intern Med.
outcomes. PLoS ONE. 2007;2(6):e507. doi: http://dx.doi.org/10.1371/journal. 2009 Aug 18;151(4):241–51, W-52. doi: http://dx.doi.org/10.7326/0003-
pone.0000507 PMID: 17551587 4819-151-4-200908180-00006 PMID: 19687491
11. Badri M, Ehrlich R, Wood R, Maartens G. Initiating co-trimoxazole 24. Zhang F, Dou Z, Ma Y, Zhang Y, Zhao Y, Zhao D, et al. Effect of earlier
prophylaxis in HIV-infected patients in Africa: an evaluation of the initiation of antiretroviral treatment and increased treatment coverage on
provisional WHO/UNAIDS recommendations. AIDS. 2001 Jun 15;15(9):1143– HIV-related mortality in China: a national observational cohort study. Lancet
8. doi: http://dx.doi.org/10.1097/00002030-200106150-00009 PMID: Infect Dis. 2011 Jul;11(7):516–24. doi: http://dx.doi.org/10.1016/S1473-
11416716 3099(11)70097-4 PMID: 21600849
12. Zachariah R, Spielmann MP, Chinji C, Gomani P, Arendt V, Hargreaves NJ, 25. Ma Y, Zhang F, Zhao Y, Zang C, Zhao D, Dou Z, et al. Cohort profile: the
et al. Voluntary counselling, HIV testing and adjunctive cotrimoxazole Chinese national free antiretroviral treatment cohort. Int J Epidemiol. 2010
reduces mortality in tuberculosis patients in Thyolo, Malawi. AIDS. 2003 Aug;39(4):973–9. doi: http://dx.doi.org/10.1093/ije/dyp233 PMID: 19556327
May 2;17(7):1053–61. doi: http://dx.doi.org/10.1097/00002030-200305020- 26. China free ART manual. Beijing: Chinese Center for Disease Control and
00015 PMID: 12700456 Prevention; 2005.
13. Guidelines on post-exposure prophylaxis for HIV and the use of 27. Shetty PS, James WP. Body mass index. A measure of chronic energy
co-trimoxazole prophylaxis for HIV-related infections among adults, deficiency in adults. FAO Food Nutr Pap. 1994;56:1–57. PMID: 7925867
adolescents, and children. Recommendations for a public health approach. 28. Lawn SD, Harries AD, Anglaret X, Myer L, Wood R. Early mortality among
December 2014 supplement to the 2013 consolidated ARV guidelines. adults accessing antiretroviral treatment programmes in sub-Saharan
Geneva: World Health Organization; 2014. Available from: http://www.who. Africa. AIDS. 2008 Oct 1;22(15):1897–908. doi: http://dx.doi.org/10.1097/
int/hiv/pub/guidelines/arv2013/arvs2013upplement_dec2014/en/ [cited QAD.0b013e32830007cd PMID: 18784453
2015 Jan 7]. 29. Tuboi SH, Schechter M, McGowan CC, Cesar C, Krolewiecki A, Cahn P, et al.
14. Madec Y, Laureillard D, Pinoges L, Fernandez M, Prak N, Ngeth C, et Mortality during the first year of potent antiretroviral therapy in HIV-1-
al. Response to highly active antiretroviral therapy among severely infected patients in 7 sites throughout Latin America and the Caribbean. J
immuno-compromised HIV-infected patients in Cambodia. AIDS. 2007 Jan Acquir Immune Defic Syndr. 2009 Aug 15;51(5):615–23. doi: http://dx.doi.
30;21(3):351–9. doi: http://dx.doi.org/10.1097/QAD.0b013e328012c54f org/10.1097/QAI.0b013e3181a44f0a PMID: 19430306
PMID: 17255742 30. Chasombat S, McConnell MS, Siangphoe U, Yuktanont P, Jirawattanapisal
15. Lim PL, Zhou J, Ditangco RA, Law MG, Sirisanthana T, Kumarasamy N, et al.; T, Fox K, et al. National expansion of antiretroviral treatment in Thailand,
TREAT Asia HIV Observational Database. Failure to prescribe pneumocystis 2000–2007: program scale-up and patient outcomes. J Acquir Immune
prophylaxis is associated with increased mortality, even in the cART Defic Syndr. 2009 Apr 15;50(5):506–12. doi: http://dx.doi.org/10.1097/
era: results from the Treat Asia HIV observational database. J Int AIDS QAI.0b013e3181967602 PMID: 19223784
Soc. 2012;15(1):1. doi: http://dx.doi.org/10.1186/1758-2652-15-1 PMID: 31. Wen Y, Zhao D, Dou Z, Ma Y, Zhao Y, Lu L, et al. Some patient-related factors
22281054 associated with late access to ART in China’s free ART program. AIDS Care.
16. Miiro G, Todd J, Mpendo J, Watera C, Munderi P, Nakubulwa S, et al. 2011 Oct;23(10):1226–35. doi: http://dx.doi.org/10.1080/09540121.2011.55
Reduced morbidity and mortality in the first year after initiating highly 5748 PMID: 21939402
active anti-retroviral therapy (HAART) among Ugandan adults. Trop Med 32. Geng EH, Hunt PW, Diero LO, Kimaiyo S, Somi GR, Okong P, et al. Trends in
Int Health. 2009 May;14(5):556–63. doi: http://dx.doi.org/10.1111/j.1365- the clinical characteristics of HIV-infected patients initiating antiretroviral
3156.2009.02259.x PMID: 19320871 therapy in Kenya, Uganda and Tanzania between 2002 and 2009. J Int AIDS
17. Hoffmann CJ, Fielding KL, Charalambous S, Innes C, Chaisson RE, Grant AD, Soc. 2011;14(1):46. doi: http://dx.doi.org/10.1186/1758-2652-14-46 PMID:
et al. Reducing mortality with cotrimoxazole preventive therapy at initiation 21955541
of antiretroviral therapy in South Africa. AIDS. 2010 Jul 17;24(11):1709–16. 33. Nash D, Wu Y, Elul B, Hoos D, El Sadr W; International Center for AIDS Care
doi: http://dx.doi.org/10.1097/QAD.0b013e32833ac6bc PMID: 20495439 and Treatment Programs. Program-level and contextual-level determinants
18. Walker AS, Ford D, Gilks CF, Munderi P, Ssali F, Reid A, et al. Daily co- of low-median CD4+ cell count in cohorts of persons initiating ART in
trimoxazole prophylaxis in severely immunosuppressed HIV-infected adults eight sub-Saharan African countries. AIDS. 2011 Jul 31;25(12):1523–33. doi:
in Africa started on combination antiretroviral therapy: an observational http://dx.doi.org/10.1097/QAD.0b013e32834811b2 PMID: 21750418
analysis of the DART cohort. Lancet. 2010 Apr 10;375(9722):1278–86. doi: 34. Lahuerta M, Lima J, Nuwagaba-Biribonwoha H, Okamura M, Alvim MF,
http://dx.doi.org/10.1016/S0140-6736(10)60057-8 PMID: 20347483 Fernandes R, et al. Factors associated with late antiretroviral therapy
19. Suthar AB, Granich R, Mermin J, Van Rie A. Effect of cotrimoxazole on initiation among adults in Mozambique. PLoS ONE. 2012;7(5):e37125. doi:
mortality in HIV-infected adults on antiretroviral therapy: a systematic http://dx.doi.org/10.1371/journal.pone.0037125 PMID: 22615917
review and meta-analysis. Bull World Health Organ. 2012 Feb 35. Brou H, Desgrées-du-Loû A, Souville M, Moatti JP, Msellati P; Initiative
1;90(2):128C–38C. doi: http://dx.doi.org/10.2471/BLT.11.093260 PMID: Evaluation Group in Côte d’Ivoire. Prophylactic use of cotrimoxazole against
22423164 opportunistic infections in HIV-positive patients: knowledge and practices
20. Lowrance D, Makombe S, Harries A, Yu J, Aberle-Grasse J, Eiger O, et of health care providers in Côte d’Ivoire. AIDS Care. 2003 Oct;15(5):629–37.
al. Lower early mortality rates among patients receiving antiretroviral doi: http://dx.doi.org/10.1080/09540120310001595113 PMID: 12959812
treatment at clinics offering cotrimoxazole prophylaxis in Malawi. J Acquir 36. Heffelfinger JD, Voetsch AC, Nakamura GV, Sullivan PS, McNaghten AD,
Immune Defic Syndr. 2007 Sep 1;46(1):56–61. PMID: 17972365 Huang L. Nonadherence to primary prophylaxis against Pneumocystis
21. Date AA, Vitoria M, Granich R, Banda M, Fox MY, Gilks C. Implementation jirovecii pneumonia. PLoS ONE. 2009;4(3):e5002. doi: http://dx.doi.
of co-trimoxazole prophylaxis and isoniazid preventive therapy for people org/10.1371/journal.pone.0005002 PMID: 19319199
living with HIV. Bull World Health Organ. 2010 Apr;88(4):253–9. doi: http://
dx.doi.org/10.2471/BLT.09.066522 PMID: 20431788
22. Zhang F, Haberer JE, Wang Y, Zhao Y, Ma Y, Zhao D, et al. The Chinese
free antiretroviral treatment program: challenges and responses.
AIDS. 2007 Dec;21 Suppl 8:S143–8. doi: http://dx.doi.org/10.1097/01.
aids.0000304710.10036.2b PMID: 18172383

160 Bull World Health Organ 2015;93:152–160| doi: http://dx.doi.org/10.2471/BLT.14.142745