Cotrimoxazole prophylaxis and antiretroviral therapy: an observational cohort study in China

Abstract 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.


Introduction
With the rapid rise in the availability of antiretroviral therapy (ART), the number of people dying from acquired immune deficiency syndrome (AIDS) has dramatically declined. However, in many low-and middle-income countries, ART coverage remains relatively poor. 1 Moreover, mortality during the early stages of ART is higher in low-than high-income countries, even after adjusting for immunodeficiency at baseline. 2 Reducing AIDS-related mortality in low-and middleincome countries will require the continued expansion of ART coverage, improved regimens and the adoption of additional cost-effective interventions.
Prophylaxis with cotrimoxazole -a combination of the antibiotics trimethoprim and sulfamethoxazole -provides protection against several opportunistic infections, including Pneumocystis jirovecii pneumonia, malaria and cerebral toxoplasmosis. [3][4][5][6] Many studies have demonstrated that cotrimoxazole prophylaxis increases survival among ARTnaïve patients: reductions in mortality of 19 to 46% have been reported in low-and middle-income countries. [6][7][8][9][10][11][12] The World Health Organization (WHO) recommends that cotrimoxazole be given in such settings to adolescents (10-19 years) and adults with advanced human immunodeficiency virus (HIV) 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 ART initiation. [14][15][16][17][18][19][20] Yet, the implementation of cotrimoxazole prophylaxis remains a challenge in low-and middle-income settings because of weak drug procurement systems, problems with managing the supply chain and little awareness of this recommendation among medical staff. [19][20][21] The situation is similar in China where some health-care workers are still unaware of the substantial benefits of cotrimoxazole prophylaxis in eligible HIV-infected patients.
In 2003, China officially launched its national free antiretroviral treatment programme. Two years later, the first guidelines on cotrimoxazole in the country were released. They recommended that cotrimoxazole prophylaxis be given to all individuals older than 14 years with a CD4+ cell count less than 200 cells/µL, WHO stage-4 disease or a history of oropharyngeal candidiasis. Several studies have reported substantially better survival among HIV-infected patients since the programme was launched. [22][23][24] However, nationallyrepresentative data on the implementation of cotrimoxazole prophylaxis has been available only since 2010, when these first started to be recorded in the programme's information system. Consequently, although the protective effect of cotrimoxazole has been documented elsewhere, evidence from China has been limited. The aim of this study was to use data from the programme's database to evaluate if cotrimoxazole prophylaxis initiated early during ART was associated with reduced mortality among HIV-infected patients.

Methods
We performed a retrospective, observational, cohort study using data from the national free antiretroviral treatment 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  In accordance with national policy, local health-care providers recorded details of treatment at initiation, at follow-up visits 0.5, 1, 2 and 3 months later and once every 3 months thereafter. If the regimen changes, the follow-up schedule is restarted at 0.5 months. As in previous analyses, 24,25 patients were considered lost to follow-up if they missed four consecutive follow-up appointments and their records were terminated at the date of the last follow-up visit. Hard copies of all treatment data were kept at the treatment sites and all information was entered into the programme's database. 25 By 31 December 2013, the database contained information on 278 085 HIV-infected patients.
Although cotrimoxazole has been officially recommended in China since 2005, the national free antiretroviral treatment programme database did not record cotrimoxazole use until 1 January 2010, when data collection was changed from a fax-based system to an internet-based system. Records were screened and selected from among those in the programme database on 1 January 2014 if the following inclusion criteria were satisfied: (i) ART was initiated between 1 January 2010 and 31 December 2012; (ii) the patient's baseline CD4+ cell count was less than 200/µL; (iii) there was a record of whether or not the patient received cotrimoxazole during every scheduled follow-up visit; and (iv) details of the patient's weight, height and history of tuberculosis at baseline were available. We excluded records of patients who died from either a drug overdose, suicide or an accident, who started cotrimoxazole more than 6 months after ART initiation or who stopped ART (Fig. 1).

Cotrimoxazole prophylaxis
Cotrimoxazole was administered orally and stopped in accordance with instructions in the China free ART manual. 26 Since the database recorded only information on whether or not the patient was receiving cotrimoxazole at each follow-up visit and did not register the exact date on which the drug was started or stopped, we regarded any patient who reported cotrimoxazole use at one or more scheduled follow-up visits within 6 months of ART initiation as belonging to our cotrimoxazole group. Our noncotrimoxazole group comprised patients who did not report cotrimoxazole use at any follow-up visit following ART initiation.
The observation period in the non-cotrimoxazole group was the time from the date of ART initiation to the date of death, loss to follow-up or the last follow-up visit before 31 December 2013. For the cotrimoxazole group, the observation period was the time from the date of the last follow-up visit before the visit during which cotrimoxazole use was first reported to the date of death, loss to follow-up or the last follow-up visit before 31 December 2013. For example, if cotrimoxazole prophylaxis was first reported in the fourth followup visit, the observation period started at the date of the third follow-up visit. Since we did not know exactly when cotrimoxazole was stopped, we assumed that patients in the cotrimoxazole group continued to take the drug throughout the observation period and, therefore, regarded the duration of cotrimoxazole prophylaxis as being identical to the observation period. The study was approved by the institutional review board of the National Center for AIDS/STD (acquired immunodeficiency syndrome/ sexually transmitted diseases) Control and Prevention.

Statistical analysis
Differences between the groups in baseline characteristics, such as age, sex, marital status, HIV infection route, baseline CD4+ cell count, WHO clinical stage and tuberculosis infection in the year before ART initiation, were compared using an χ 2 test for categorical variables and a Mann-Whitney U test for continuous variables since no continuous variable fulfilled the Kolmogorov-Smirnov test for normality. A low body mass index (BMI) was defined as a value less than 18.5 kg/m 2 , in accordance with the threshold for malnutrition of the Food and Agriculture Organization of the United Nations. 27 Patients' data were censored at 31 December 2013 or the date of loss to follow-up or death, depending on which was earliest. The association between mortality and cotrimoxazole and baseline CD4+ cell count after ART initiation was analysed using Kaplan-Meier survival curves and crude mortality rates. In addition,

Research
Cotrimoxazole and antiretroviral therapy in China Wei Cheng et al.
mortality was also evaluated using Cox proportional hazards regression models that included adjustment for covariates that were predetermined to be clinically meaningful and which had a significant influence in the unadjusted analysis (i.e. a P-value < 0.1). To determine whether cotrimoxazole's influence on mortality varied according to the duration of ART and the baseline CD4+ cell count, the Cox models were stratified by the duration of ART (i.e. 6, 12, 18, 24 and 30 months) and the baseline CD4+ cell count (i.e. < 50, 50-99 and 100-199 cells/µL). Finally, to confirm our findings, we introduced a category for missing data and repeated the calculations using data that included observations from individuals who had been excluded because of missing records. All calculations were performed using SAS version 9.2 (SAS Institute, Cary, United States of America).

Results
The analysis included data on 23 816 patients aged 15 years or more: 12 047 (51%) had never taken cotrimoxazole, whereas 11 769 (49%) had taken the drug within 6 months of ART initiation (Fig. 1). Of those who reported cotrimoxazole use, 2252 (19%) did not begin taking the drug at ART initiation: the mean starting time was 27 days (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 common route of HIV infection was sexual transmission, followed by injectiondrug use, blood or plasma transfusion and other or unknown routes. The patients' median baseline CD4+ cell count was 84 cells/µL -the median was significantly lower in the cotrimoxazole group, at 69 cells/µL, compared with 103 cells/ µL in the non-cotrimoxazole group (Table 1). In addition, significantly more patients in the cotrimoxazole group than the non-cotrimoxazole group had a low BMI (28% versus 21%, respectively), a history of tuberculosis before ART (15% versus 10%, respectively) and WHO clinical stage-3 disease (28% versus 22%, respectively) or stage-4 disease (31% versus 18%, respectively).
In total, 2706 patients died during 44 137 person-years of observation: 56% (1515/2706) were in the non-cotrimoxazole group and 44% (1191/2706) were in the cotrimoxazole group. Overall mortality was 6.1 per 100 personyears: 7.0 per 100 person-years in the non-cotrimoxazole group and 5.3 per 100 person-years in the cotrimoxazole group. Kaplan-Meier survival curves showed that survival was lowest in individuals with a baseline CD4+ cell count less than 50 cells/µL who did not receive cotrimoxazole (Fig. 2).
Univariate analysis found that the risk of death was significantly lower in patients who received cotrimoxazole (hazard ratio, HR: 0.80), who were female (HR: 0.75) and who had no history of tuberculosis before ART (HR: 0.69; Table 2). Factors significantly associated with an increased risk of death included age more than 30 years, an HIV transmission route other than sexual transmission, a baseline CD4+ cell count less than 100 cells/µL, a BMI less than 18.5 kg/m 2 and WHO clinical disease stage 2 or higher. These associations remained significant in the multivariate analysis after adjusting for all potential confounders, except age in the range 30 to 39 years and a history of tuberculosis. In particular, cotrimoxazole prophylaxis was still significantly associated with a decreased risk of death (HR: 0.63).
In addition, cotrimoxazole use was associated with significantly reduced mortality over a range of ART durations: the adjusted HR for death was 0.65 for 6 months' treatment, 0.58 for 12 months' treatment, 0.49 for 18 months' treatment and 0.66 for 24 months' treatment ( Table 3). There was no significant reduction in mortality for ART lasting between 25 and 30 months (HR: 0.80). The protective association of cotrimoxazole was also significant over a range of baseline CD4+ cell counts: the adjusted HR for death was 0.60 for patients with a count less than 50 cells/µL, 0.66 for those with a count of 50-99 cells/µL and 0.78 for those with a count of 100-199 cells/µL (Table 3).
When the analysis was repeated with the inclusion of a category for missing data so that observations from 8753 patients who were excluded because information on their weight, height or tuberculosis status had not been reported, we found that cotrimoxazole remained significantly associated with decreased mortality (HR: 0.65; 95% CI: 0.59-0.72); this value for the HR is close to the 0.63 found in the multivariate analysis of the principle study group. Cotrimoxazole and antiretroviral therapy in China Wei Cheng et al.

Discussion
In this large nationally-representative study, we found that cotrimoxazole prophylaxis started early during ART reduced mortality by 37% in HIV-infected adults; the decrease was greatest among those with advanced disease. The reduction was evident as long as 24 months after ART initiation and was observed in patients with severe immunosuppression. The protective effect of cotrimoxazole has previously been documented elsewhere in ART-naïve patients. [3][4][5][6][7][8][9][10][11][12] Our findings are consistent with those of other studies [14][15][16][17][18][19][20] and reinforce Chinese national guidelines, which recommend extending the use of cotrimoxazole to patients with late-disease stage.
Although the expanded use of ART has substantially improved mortality and morbidity in HIV-infected patients around the world, several challenges remain that could undermine these gains, especially in low-resource countries. Higher mortality has been observed during the early stages of ART in both high-and low-income settings. 2,23,[28][29][30] In high-income settings, mortality was reported to fall from 24 per 1000 person-years during the first 6 months of treatment to 16 per 1000 person-years during months 7 to 12. In low-income settings, this trend was even more dramatic: the mortality rate was reported to fall by nearly half from 51 per 1000 person-years during the first 6 months of treatment to 27 per 1000 person-years during months 7 to 12. 2 In sub-Saharan Africa, it has been reported that 8 to 26% of patients die during the first year of ART, with most deaths occurring in the first few months. 28 A similar trend has been observed in China. 23 We found that cotrimoxazole prophylaxis during the first 6 months of ART was associated with a one-third reduction in mortality. Given the high underlying risk of death in this period, this improvement in mortality could correspond to a substantial reduction in the number of deaths. Moreover, in our study, cotrimoxazole appeared to be associated with decreased mortality with a duration of ART up to about 24 months. This observation still has to be explained.
In many countries, late access to ART has been identified as an important challenge for treatment programmes. 23,24,[31][32][33][34] One study in eastern Africa found that the median baseline CD4+ cell count in patients starting ART in 2008 and 2009 was 154 cells/µL 32 and similar findings have been reported in China. 23,24,31 In addition, several studies have demonstrated a strong association between a low baseline CD4+ cell count and high mortality after ART initiation, especially among people with a baseline count less than 50 cells/µL. 23,24 Although the baseline CD4+ cell count at ART initiation in China improved between 2006 and 2009, in 2009 about one third of people still started ART with a CD4+ cell count no higher than 50 cells/ µL. 31 In our study, the association between co-trimoxazole-use and reduced mortality was most marked in patients with a baseline CD4+ cell count less than 50 cells/µL: the reduction in mortality was 40%. This indicates that cotrimoxazole has the potential to reduce mortality among patients with the highest risk during their most vulnerable period.
Worldwide, the use of cotrimoxazole is suboptimal. In 2010, WHO found that, although the national policy in 38 of 41 countries surveyed was to provide cotrimoxazole to people living with HIV infections, only 25 of the 38 had  21 Major obstacles included an erratic drug supply, drug stocks running out and poor knowledge of cotrimoxazole among health-care workers and patients. [19][20][21]35,36 In our study, we found that only half of participants received cotrimoxazole prophylaxis. Moreover, a fifth of those treated did not start the drug at ART initiation. Over the past few years, China has successfully established a working system and infrastructure for the treatment of HIV-infected patients as well as institutions for monitoring treatment -these facilities could promote more extensive use of cotrimoxazole. [22][23][24][25] This study has several limitations. First, due to the lack of detailed information on cotrimoxazole treatment and on adherence to treatment, we equated the duration of cotrimoxazole prophylaxis to the observation time in the cotrimoxazole group. Consequently, we could not assess reduction of mor-tality for different exposure times of cotrimoxazole. Moreover, by regarding all patients in the cotrimoxazole group as having taken the drug continuously from the starting date, we may have underestimated the reduction in mortality. Second, data were retrieved from an observational database and may therefore have inherent biases, such as reporting or recall bias, associated with the collection of nonrandom data. However, the large number of patients in the database should have helped mitigate bias in any one direction. Finally, the large number of patients excluded because of missing data may have given rise to selection bias. However, when we repeated the analysis by including data from individuals who were excluded because details of their weight, height or history of tuberculosis were missing, our findings were consistent with those of the main analysis. Moreover, we assessed heterogeneity in baseline characteristics between included and excluded patients and found that the distributions of key variables such as age, sex and baseline CD4+ cell count were similar in the two groups. We believe, therefore, it is unlikely that our main findings were affected by biases in these key characteristics. Furthermore, our findings are consistent with those of studies conducted in other countries. [14][15][16][17][18][19][20] Finally, since our data reflect the realities of treatment in a middle-income setting, our conclusions may not be generalizable to all countries.
In conclusion, by using data on a large national cohort, we found that cotrimoxazole prophylaxis administered early during ART significantly reduced mortality in HIV-infected patients, particularly in those with a CD4+ cell count less than 50 cells/µL. Given its beneficial effect on survival and its relatively low cost, cotrimoxazole should be offered in conjunction with ART to HIV-infected patients in China and other low-and middleincome countries. ■ Wei Cheng is also affiliated with Zhejiang Provincial Center for Disease Control and Prevention, Hangzhou, China. Fujie Zhang is also affiliated with the Department of Infectious Diseases, Beijing Ditan Hospital-Capital Medical University, Beijing, China.

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