Bulletin of the World Health Organization

Methadone maintenance treatment and mortality in HIV-positive people who inject opioids in China

Yan Zhao a, Cynthia X Shi a, Jennifer M McGoogan a, Keming Rou a, Fujie Zhang a & Zunyou Wu a

a. National Center for AIDS/STD Control and Prevention, Chinese Center for Disease Control and Prevention, 155 Changbai Road, Changping District, Beijing, 102206, China.

Correspondence to Zunyou Wu (e-mail: wuzunyou@chinaaids.cn).

(Submitted: 15 June 2012 – Revised version received: 09 October 2012 – Accepted: 02 November 2012.)

Bulletin of the World Health Organization 2013;91:93-101. doi: 10.2471/BLT.12.108944


In Asian countries, including China, most epidemics of human immunodeficiency virus (HIV) infection and acquired immunodeficiency syndrome (AIDS) have stabilized. In 2011, the estimated prevalence of HIV infection in China was 0.058%, a rate similar to or lower than that observed in neighbouring countries. Furthermore, an estimated 780 000 people were living with HIV/AIDS and 28.4% of them were thought to have been infected by injecting illicit drugs.1 China now has the world’s largest number of people who inject illicit drugs2 and it is well known that the country’s epidemic of HIV infection originated in and spread rapidly in this group because of unsafe practices, such as sharing needles.24 The epidemic soon spread through commercial plasma donation, sexual contact and mother-to-child transmission.3

In response, since 2004 the Chinese government has implemented comprehensive strategies for treating both HIV infection and drug dependence involving, respectively, a National Free Antiretroviral Treatment Programme and a National Methadone Maintenance Treatment Programme.1,57 The National Free Antiretroviral Treatment Programme (NFATP) operates in a variety of settings, including infectious disease hospitals, general hospitals and local clinics run by the Chinese Center for Disease Control and Prevention. By April 2012, the programme had provided antiretroviral therapy (ART) to a total of 166 371 HIV-positive (HIV+) patients.8 Already, the 10-year-old programme is seeing some success: mortality rates and viral loads have decreased more among enrolled patients than among those not undergoing treatment.912 Similarly, the National Methadone Maintenance Treatment Programme has reduced drug dependence and involvement in drug-related crime among clients, increased the employment rate and improved social functioning.6 This programme has helped 357 586 opioid-dependent individuals since 2004;8 it is, in fact, the largest network for methadone maintenance treatment in the world.13

In China and globally, the introduction of ART has dramatically reduced AIDS-related mortality and the rate at which HIV+ patients progress to AIDS. However, in low- and middle-income countries, the proportion of HIV+ people injecting illicit drugs who receive ART is often disproportionally low.14 For example, people who inject illicit drugs account for an estimated 28.4% of all HIV+ individuals in China, but only 15.5% of ART patients inject illicit drugs.1,12,15 Thus, despite the increasing availability of ART, which is a life-saving therapy, HIV-related and all-cause mortality remain high among people who inject illicit drugs and are HIV+.

Studies in various countries, including Brazil,16 Denmark17 and the United States of America,18 have shown that individuals on ART who become infected with HIV by injecting illicit drugs have a higher risk of death than those infected by other transmission routes. The lower uptake of ART and the poorer outcomes observed among HIV+ people who inject illicit drugs are thought to be attributable to a variety of factors, such as delays in starting treatment, poor adherence to ART, continuing drug dependence and a high prevalence of comorbidities, such as depression and hepatitis C.19

Internationally, there is growing evidence that treatment of opioid dependence, such as methadone maintenance treatment (MMT), not only stabilizes drug dependence and its associated psychosocial conditions, but also promotes adherence to ART and improves virological responses among HIV+ people.18,2023 Although China’s national ART and MMT programmes have been shown to benefit their respective target populations when assessed independently,6,911,24,25 the advantages of coincident ART and MMT for HIV+ people who inject opioids have not yet been examined in China, though they have been observed in other settings.13,20,23,26,27 The aim of this study, therefore, was to determine whether MMT reduces mortality in HIV+ people who inject opioids and are receiving ART through China’s national ART programme.


The study involved a nationwide cohort of HIV+ people injecting opioids who received ART between 31 December 2002 and 31 December 2011. In accordance with World Health Organization recommendations,28 the current eligibility criterion for ART in China is a CD4+ T-lymphocyte (CD4+ cell) count below 350 cells/μL; before 2008, it was a count below 200 cells/μL. In China, recommended first-line ART regimens for all eligible HIV+ patients, including those injecting opioids, involve combinations of zidovudine, stavudine, lamivudine, nevirapine and efavirenz.29 Criteria for enrolment in MMT have been described previously.6 Programme guidelines suggest an initial methadone dose of no more than 50 mg per day for all patients, which is subsequently adjusted as appropriate.30

Eligibility for inclusion in this study was simply having been enrolled in China’s national ART programme at any time between 31 December 2002 and 31 December 2011. Individuals were excluded from the study if one of the following applied: (i) during medical or police registration, he or she reported acquiring an HIV infection through a transmission route other than opioid injection; (ii) ART had been received before the start of the study period; or (iii) the start date for ART was missing from the individual’s records in the ART programme data system. Participants were divided into two subgroups depending on whether or not they had received MMT at any time during ART: the ART + MMT subgroup and the ART-only subgroup, respectively. Those who initiated and discontinued MMT before receiving ART were included in the ART-only subgroup. Active drug users were not identified separately in either subgroup.

Data collection

The data for this study were obtained from the real-time, web-based data system31 of the Chinese National Center for AIDS/STD Control and Prevention, which contains three interlinked, but independently operated, data subsystems: (i) the ART-for-adults subsystem; (ii) the HIV/AIDS case reporting subsystem; and (iii) the MMT subsystem. All HIV+ individuals who met the appropriate eligibility criterion for ART described above29 and who received ART via the national ART programme were registered in the ART-for-adults data subsystem. This subsystem contains identification and demographic data for all patients and detailed information on their ongoing clinical management and ART, including the drugs used, the doses administered and HIV viral load.32 Staff at local Center for Disease Control and Prevention facilities track all individuals who have tested positive for HIV in China using the national HIV/AIDS case reporting data subsystem. Records created in this subsystem contain identification data, detailed demographic data and information on health status, health history and survival.31 All opioid-dependent individuals who take part in China’s national MMT programme are registered and tracked in the MMT data subsystem. Client records contain identification and demographic data and are updated daily with information on methadone treatment.6 Data on all individuals included in this study were initially obtained from the ART data subsystem. Then, that information was supplemented with data from the HIV/AIDS case reporting subsystem. Finally, additional data were obtained from the MMT data subsystem on all study subjects with records in that subsystem. The data sets in each subsystem were linked using unique, national, 15- or 18-digit Chinese citizen identification numbers.

An intent-to-treat approach was adopted for exposure to both ART and MMT. Thus, subjects were regarded as having received ART or MMT if they had been given at least one dose of an ART drug or methadone, respectively, regardless of their level of adherence or treatment duration. All subjects included in the final study cohort were followed through the ART programme and the observation time was the time between the date of ART initiation and one of three outcomes: (i) the end of the study (i.e. the patient was still alive and receiving ART); (ii) death while on ART; or (iii) discontinuation of ART or loss to follow-up. A subject was categorized as having discontinued ART or being lost to follow-up if no follow-up data had been recorded for more than 3 months. However, although their records in the ART data subsystem were not up to date, these patients could still be tracked through the HIV/AIDS case reporting data subsystem for events such as death. The outcome of interest in the study was all-cause mortality.

Statistical analysis

Categorical variables are presented as numbers and percentages and continuous variables are described using medians and interquartile-ranges (IQRs). Statistical tests were not used to compare the characteristics of subgroups because the large sample sizes resulted in the detection of statistically significant differences that were probably not clinically meaningful. Rather, a difference greater than 5% between subgroups was considered meaningful. The observation time was presented in person–years. The mortality rate was calculated as the total number of deaths divided by the total observation time, both overall and in different study subpopulations. Risk factors for mortality were examined using Cox proportional hazards regression analysis. Factors identified as having a significant influence on mortality on univariate analysis (i.e. P <  0.05) and factors known to be clinically meaningful were included in multivariate regression models, from which adjusted hazard ratios (HRs) and 95% confidence intervals (CIs) were obtained. Although missing data are presented in the tables, all analyses were run using only cases for which data were complete. All P-values presented are two-sided and P < 0.05 was indicative of statistical significance. Differences in mortality rates over time were calculated by dividing the total number of subjects who died in each 6-month interval by the sum of the individual observation times for subjects followed up during the corresponding interval. All statistical analyses were performed using SAS v. 9.1.3 software (SAS Institute, Cary, United States of America).

Ethics approval

This study was reviewed and approved by the institutional review board of the National Center for AIDS/STD Control and Prevention, Chinese Center for Disease Control and Prevention. As all data included here were collected during the regular administration of China’s national ART and MMT programmes and since all participants in these programmes provided signed informed consent at the time of enrolment, no further informed consent was sought.


The ART-for-adults data subsystem contained records on a total of 152 749 patients for the period from 31 December 2002 to 31 December 2011. Of these patients, 128 607 (84.2%) were excluded because their HIV infection was acquired through a route other than opioid injection. An additional 293 patients (< 0.1%) were excluded because they were not ART-naïve when they enrolled in the ART programme. Finally, 36 patients (<  0.1%) were excluded because the ART start date was missing from their records in the ART-for-adults subsystem. Thus, the final cohort consisted of 23 813 HIV+ ART patients who injected opioids.

Of the 23 813 patients, 18 652 (78.3%) were assigned to the ART-only subgroup because they did not receive MMT during the study period. It should be noted that 906 patients in this subgroup had started and discontinued MMT before ART. The ART + MMT subgroup included 5161 participants (21.7% of 23 813) who had enrolled in both ART and MMT programmes concurrently at some time during the study period. Of this subgroup, 3335 (64.6% of 5161) were enrolled in the MMT programme before ART initiation: they had been in the MMT programme for a median of 1.20 years (IQR: 0.40–2.43) and had enrolled at a median age of 33.8 years (IQR: 30.3–37.8). An additional 669 patients (13.0% of 5161) began MMT less than 6 months after initiating ART and 1157 (22.4% of 5161) began MMT more than 6 months after ART.

As shown in Table 1, the median age of participants in our study cohort was 34.0 years (IQR: 30.0–39.0 years). The majority were male (87.6%), Han Chinese (54.1%), had been educated to middle school level or higher (52.1%) and were married or cohabitating (54.1%). At the time of ART initiation (i.e. ART baseline), the median CD4+ cell count in study participants was 177.0 cells/µL (IQR: 80.0–264.0) and the median haemoglobin concentration was 13.4 g/dL (IQR: 11.5–15.0). A large proportion of patients (48.3%) were receiving a regimen that contained zidovudine, lamivudine, and nevirapine or efavirenz and 64.0% were being treated in a general hospital. By the end of the study, 16 321 patients (68.5%) were still alive and receiving ART, 3057 (12.8%) had died while still in the ART programme and 4435 (18.6%) had discontinued their ART regimen or were lost to follow-up.

Differences between subgroups

We observed several differences greater than 5% in demographic characteristics between the ART + MMT and ART-only subgroups (Table 1). Compared with members of the ART-only subgroup, fewer patients in the ART + MMT subgroup belonged to an ethnic minority (28.6 versus 35.7%), had been educated to a primary school level or lower (27.1 versus 36.2%) and were married or cohabiting (45.2 versus 56.5%). In addition, the ART + MMT subgroup contained a greater proportion of patients with a CD4+ cell count above 200 cells/µL at ART baseline (46.6 versus 40.2%) and a greater proportion who were being treated at the Center for Disease Control and Prevention clinics (19.5 versus 12.3%).

As shown in Table 2, 3057 deaths were identified over 41 959 person–years of observation time, which corresponds to an overall mortality rate of 7.3 per 100 person–years (95% CI: 7.0–7.5) in our cohort of HIV+ people injecting opioids who received ART. The highest mortality rates were observed in individuals with a haemoglobin level below 9 g/dL (17.5 per 100 person–years; 95% CI: 15.9–19.2) and those with a CD4+ cell count at ART baseline below 50 cells/µL (12.0 per 100 person–years; 95% CI: 11.2–12.7). A total of 2519 deaths were identified over 30 558 person–years of observation time in our ART-only subgroup, which corresponds to a mortality rate of 8.2 per 100 person–years (95% CI: 7.9–8.6). By contrast, only 538 deaths occurred in 11 401 person–years of observation time in our ART + MMT subgroup, which corresponds to a mortality rate of 4.7 per 100 person–years (95% CI: 4.3–5.1). Fig. 1 shows the change in mortality rate over time in the two subgroups. At 6 months after ART baseline, mortality in the ART-only subgroup was over twofold greater than in the ART + MMT subgroup: 16.9 versus 6.6 per 100 person–years, respectively. The difference becomes smaller at 12 months (7.4 versus 3.7 per 100 person–years) and disappears thereafter.

Fig.1. Mortality in people injecting opioids and infected with the human immunodeficiency virus who received antiretroviral therapy (ART), with or without methadone maintenance treatment (MMT), by time from ART initiation, China, 2002–2011
Fig. 1. Mortality in people injecting opioids and infected with the human immunodeficiency virus who received antiretroviral therapy (ART), with or without methadone maintenance treatment (MMT), by time from ART initiation, China, 2002–2011

Factors affecting survival

The results of the multivariate analysis carried out to identify risk factors for death among study participants are shown in Table 2. Factors significantly associated with an increased risk of death include: older age; male sex; being single, divorced or widowed; a CD4+ cell count of 200 cells/μL or less at ART baseline; a baseline haemoglobin level of 12 g/dL or less; being treated at a Center for Disease Control and Prevention clinic; and not participating in the MMT programme.


Our main finding is that MMT reduced the risk of death in HIV+ people injecting opioids who were receiving ART, particularly in the first 6 months of treatment: during this period, study participants receiving only ART had a mortality rate more than twice as high as those receiving ART and MMT (16.9 versus 6.6 per 100 person–years, respectively). The primary reasons for lower mortality in those receiving MMT are thought to be: (i) better adherence to the ART regimen due to regular, daily contact with health-care workers at methadone clinics; (ii) reduced active opioid injection; and (iii) the stabilizing effect of methadone on behaviour associated with opioid injection, which promotes social reintegration. These factors have already been well documented in a variety of settings.14,19,20,3335

The literature also supports our finding that a low CD4+ cell count at the start of ART is a strong independent predictor of death.10,12 A low baseline CD4+ cell count, which indicates advanced immunosuppression, is a concern for all patients in China’s national ART programme.15 It is interesting to note that the median CD4+ cell count at ART baseline was higher in our ART + MMT subgroup than in the ART-only subgroup (192 versus 172 cells/μL, respectively), though the difference was not large. This observation indicates that people who were injecting opioids and were engaged in the MMT programme often enrolled in the ART programme at an earlier stage of HIV infection. Thus, participation in China’s national MMT programme may mitigate the increased risk of death in people injecting opioids in China. In addition, China’s national MMT programme is helping these people to overcome the many barriers they face in engaging with and remaining in the ART programme.12 Regular, often daily, interactions with health-care providers in the MMT programme: (i) promote the cessation of injecting drug use; (ii) increase adherence to the ART regimen; (iii) ensure regular monitoring of the CD4+ cell count; and (iv) increase opportunities for discussing treatment options, for managing medication side-effects and comorbid conditions and for psychosocial counselling.20,34,36,37 Thus, the higher CD4+ cell counts we found at ART baseline in patients receiving ART and MMT were not unexpected.

Our finding that a low haemoglobin level at ART initiation is a strong predictor of death has also been reported previously in the literature.38,39 Anaemia is known to be an independent predictor of death and disease progression in patients with HIV/AIDS.3941

Because of the way China’s health-care system is structured, the national ART programme provides treatment for HIV infection through a variety of channels, such as hospitals specializing in infectious diseases, general hospitals and local Center for Disease Control and Prevention facilities. We found that the type of facility providing treatment was an independent predictor of death in HIV+ people injecting opioids who were receiving ART in China. This observation is consistent with the results of a previous study in China, which showed that the type of facility providing care for HIV+ patients was associated with their survival probability.11 Although this previous study involved patients living in rural areas who were not injecting illicit drugs (as opposed to ours, which only included those who were injecting opioids in urban settings), it confirms that treatment outcomes vary in accordance with the type of facility. In the present study, mortality was lowest among patients treated at infectious disease hospitals. We believe this was probably because specialized doctors and staff with better training and more experience in providing ART were available, monitoring was more effective, and the facility was located in a better-equipped urban setting.

Our study had several limitations. First, because it was a descriptive, observational study, no inferences about causality can be made. Second, although we attempted to reduce the influence of confounding variables by using multivariate analysis, there remains a risk of confounding by unmeasured variables. For example, seropositivity for hepatitis C virus and depressive symptoms are known to be poor prognostic factors for patients with HIV/AIDS and may have influenced our results.19 Third, the categorization of ART recipients as opioid injectors was based on patients’ self-reported history of opioid use at the time their HIV infection was detected. Hence, some participants may have been inappropriately excluded from our final cohort due to information bias. Fourth, the missing values for some variables in our data set could have resulted in an under- or overestimation of mortality rates and of the importance of associated risk factors. Finally, although ART and MMT providers follow guidelines published by the Chinese Ministry of Health, their clinical decisions are not standardized and may be influenced by, for example, the provider’s potentially biased opinions about poor compliance among drug users. Despite these limitations, the generalizability and reproducibility of our findings are likely to be high because the study involved a large cohort of participants drawn from a national treatment programme.

In summary, our finding that ART and MMT had a joint beneficial effect on mortality in Chinese HIV+ people injecting opioids suggests that there is an urgent need for changes in ART and MMT programmes. Cross-referrals between these programmes should be encouraged so that a greater proportion of people who inject opioids can benefit from receiving the two treatments. An even better approach may be to house ART and MMT care services at the same location – a model that has already proved to be more effective worldwide, both clinically and in terms of cost.14,23,27,4244 The next step in reducing barriers to treatment and in improving poor outcomes in HIV+ people who inject illicit drugs in China should be to study the feasibility of integrating ART and MMT services.


The authors would like to thank Enwu Liu, Mingjie Zhang and the many health-care providers across China.


This study was supported by the Chinese government HIV/AIDS Program (131-11-0001-0501) and in part by the Fogarty International Center and the National Institute on Drug Abuse at the US National Institutes of Health (China ICOHRTA2, NIH Research Grant # U2RTW06918).

Competing interests:

None declared.