As a country with more than 20% of the world's population, China has attracted international attention in the past years not only for its booming economic development but for its potential explosion of the HIV/AIDS epidemic. Thus far, all 31 provinces, autonomous regions, and municipalities in mainland China have reported HIV infections, and a total of more than 60,000 cumulative cases have been reported1; however, the estimated numbers of people living with HIV and AIDS in 2003 were 840,000 and 80,000, respectively.1 Because of a large population base, a small increase of HIV prevalence in the general population in China would mean a large increase in the number of domestic infections and a significant contribution to the total number of infections around the world.
Injection drug use has played a predominant role in the spread of HIV in China. In the past decade, drug abuse has become an increasingly more severe problem. Repeated cross-sectional surveys in 5 provinces with a high prevalence of illicit drug use found that the 1-year prevalence of drug use in the general population increased from 0.91% in 1993 to 1.17% in 1996.2,3 By the end of 2002, the number of registered injection drug users (IDUs) was approximately 1 million, with an annual increase rate of 11% in the past 4 to 5 years.4 The first HIV outbreak among IDUs in China was reported in the southwestern province of Yunnan along the border with Burma (Myanmar) in 1989.5 It then further spread among IDUs residing along the major drug-trafficking roads to Guangxi, Sichuan, Xinjiang, Guangdong, and other provinces.6 As of 2002, injection drug use contributed to 64% of the total reported HIV/AIDS cases across the nation.1
Several cohort studies have been conducted among IDUs in China, most in the provinces of Yunnan and Guangxi, which have a high HIV prevalence. These studies showed varying but generally high seroconversion rates.7,8 China has a large geographic area and a variety of cultures and customs; more cohort studies among IDUs should provide valuable data for understanding the whole picture of the HIV epidemic. In addition, evaluation of the factors associated with retention among the IDU cohort would assist in developing intervention programs among this population, considering the illegal nature of drug use in this country. This community-based study aims to estimate the seroconversion rate and to evaluate the factors associated with retention among a cohort of IDUs in Sichuan Province, China.
SUBJECTS AND METHODS
Study Design and Study Population
This prospective follow-up study was conducted in Xichang County of Sichuan Province in southwest China. Xichang is one of 142 counties in Sichuan Province and hosts a population of more than 600,000, with a majority of Han ethnic people. Approximately 10% of the population is of minority Yi ethnicity. Xichang has the highest reported prevalence of drug use and HIV infection in Sichuan Province.9
A baseline screening survey was conducted in November 2002. Study participants were primarily recruited using community-based outreach, with a supplementary method of snowball sampling as described elsewhere.10 Eligibility criteria for the cohort study included that participants must be HIV-seronegative, at least 18 years old, and have injected drugs at least once in the last 3 months. A total of 382 IDUs were screened: 2 persons did not meet eligibility criteria, 4 refused to participate (1.5% refusal rate), and 43 (11.3% prevalence rate, 95% CI: 8.2, 14.5) were HIV-positive and therefore were excluded. A total of 333 eligible and consenting IDUs were enrolled into the cohort of the prospective study. Participants were asked to come back to the study clinic for follow-up evaluation every 6 months. The study protocol was approved by the Institutional Review Board (IRB) of the Chinese Center for AIDS/STD Control and Prevention, and informed consent was obtained from the participants.
Cohort Retention Plan
A written procedure was used in the hope of keeping high retention rates at follow-up visits. Every participant was asked to provide at least 2 different contact sources. This locator information was managed by the Participant Information File System (PIFS) software program. Whenever possible, the follow-up visits were to be scheduled on the “target date.” Appointment reminder procedures at every 6-month follow-up visit included the following:
1. Study staff sent a reminder letter or made a telephone call to the participant 2 to 3 weeks before the scheduled visit. Study staff also contacted the participant on the week of the scheduled study visit to confirm that the participant planned to keep the appointment.
2. If a participant missed an appointment, staff would call the participant on the same day and reschedule an appointment for the next work day or as soon as possible.
3. If the participant was not contacted on the day of missed visit, the staff would send a letter to each address or call other contacts listed in the PIFS.
4. During the first week of the visit window (from 14 days before the target date to 30 days after the target date), in addition to the reminders by telephone calls or mail, outreach workers made home and/or street visits. If outreach workers could not contact the participant during the home visits, they left a letter with a personal note. During weeks 2 through 6 of the visit window, staff would call all the listed telephone numbers once a week at various times and outreach workers would conduct outreach visits every week.
The locator data manager continued to update the contact log and locator information throughout the study period. Outreach workers also checked drug detoxification centers and jails for reasons for loss of follow-up. Access to the PIFS was strictly protected through a database security system.
Each study participant was assigned a unique identifier code to keep all collected data confidential. An interviewer-administrated questionnaire was used during the baseline survey to collect data on demographics, drug-using behaviors, and sexual activities. A venous blood specimen was collected for testing of HIV antibodies at every 6-month follow-up visit.
HIV antibody was screened for each blood sample by enzyme-linked immunosorbent assay (ELISA; Beijing Wantai Biologic Medicine Company, China) and was confirmed by HIV-1/2 Western blot immunoassay (HIV Blot 2.2 WB; Genelabs Diagnostics, Singapore). A sample positive by both tests was considered to be HIV-positive. HIV subtype analysis was based on the sequences from the env and gag regions of the HIV RNA envelope.
Questionnaire data were double-entered and compared with EpiData software, version 2.1 (The EpiData Association, Odense, Denmark). The HIV incidence density rate was calculated based on Poisson distribution, with person-time of follow-up as the denominator. Categoric factors associated with retention were analyzed by the χ2 test, including baseline sociodemographic and behavioral characteristics. A multiple logistic regression model was constructed to select independent factors for retention after controlling for other confounding factors. All statistical analyses were done using SAS software, version 8.2 (SAS Institute, Cary, NC).
Sociodemographics and Retention of Study Cohort
Of the 333 study participants, 81.7% were male; the average age was 28.8 years (SD = 5.4), with a range from 18.3 to 44.3 years; and 66.4% participants were of majority Han ethnicity, and 28.8% were of Yi ethnicity. Twelve percent had no education, 28.2% had ever attended primary school (6 years), and 43.2% and 9.9% attended junior high school (9 years) and senior high school (12 years), respectively. The proportion of unemployed participants was 56.5%, and 24.9% were farmers. The proportion of single participants was 46.9%, 29.1% were married, 14.7% were divorced, and 7.8% were cohabiting with partners. The proportion of participants who owned their own house (apartment) was 33.3%, 12.0% rented a house, and 48.1% lived in their parents' houses. The median yearly income of the participants was US $871.70, a little bit lower than the national average.
At the 12-month follow-up visit, 70.3% (234 of 333) of the participants were retained in the cohort. Of the 99 participants lost to follow-up, 24.2% were found to be in jail or detoxification centers because of drug use and 24.2% had died. Other reasons included moving out of the area (11.1%) and wrong contact information (15.2%). Reasons for loss of follow-up were not available for 25.3% (Table 1). Among 24 deaths, 18 were a result of an overdose of illicit drugs.
Predicting Factors for Cohort Retention
In univariate analyses, associations of 12-month retention with the following sociodemographic variables were evaluated: gender, age, ethnicity, years of education, employment, marital status, owning a house (apartment), and personal yearly income. The proportion of patients who appeared at the 6-month follow-up visit was 74.8%. Ethnicity (P = 0.0030) and appearing for the 6-month follow-up visit (P < 0.0001) were associated with retention at the 12-month follow-up visit at a 0.05 significant level (Table 2). Associations were also evaluated between 12-month retention and drug-using and sexual behavioral variables, which were not found to be significantly associated with retention (Table 3). Two factors significant in univariate analyses were entered into a multiple logistic regression model. Appearing for the 6-month follow-up visit (OR = 9.03, 95% CI: 5.14, 15.89) and ethnicity (OR = 0.60, 95% CI: 0.34, 1.04) were significantly (or marginally significantly) associated with 12-month retention in the final model (not shown in Table 3).
Comparisons between majority ethnic Han and Yi and other minority ethnic participants were made for gender, age, income, needle/syringe-sharing behavior, and extramarital sex. Only needle/syringe-sharing behavior in the baseline screening survey was significantly different between the 2 groups. Yi and other minority ethnic participants reported more needle/syringe sharing than Han ethnic participants (55% vs. 33%). Because of the small number of observations in some categories, however, we have no statistical power to perform a stratified analysis by this variable to evaluate the association between ethnicity and retention (not shown in Table 3).
HIV Seroconversion and Subtype
During the 12-month follow-up period, 8 participants seroconverted to become HIV-positive, yielding an incidence density rate of 3.17 per 100 person-years (95% CI: 0.98, 5.37). Analysis of HIV-1 sequence from the env and gag regions of the virus envelope indicated that HIV-1 strains in these participants were all CRF_07BC.
In this prospective cohort study among 333 HIV-seronegative IDUs, 70.3% of the participants were successfully followed up during the 12-month study period. The main reasons for loss of retention in study cohort included a high rate of incarceration and early death; most of these deaths were the result of an overdose of illicit drugs (18 of 24 deaths [75%]). This is the first published study in China to evaluate the retention rate for a prospective cohort of drug users. Although we used similar retention strategies as other studies in Western countries,11,12 we found a lower retention rate. Some foreign studies reported that approximately 25% of the participants were lost to follow-up.13,14 Although it is generally believed that drug users are difficult to retain in a prospective cohort study, Antoine et al11 reported a 6-month retention rate of greater than 90%.
Several studies reported that HIV high-risk behaviors were not associated with cohort retention.11-13,15 Education level and moving in the last year were associated with loss of follow-up.11 In this study, drug-using and sexual behaviors were not found to be associated with retention. A high enrollment rate (98.5%) might be a reason for the relative low retention rate, but we have no confirmatory data. Appearing at the 6-month follow-up visit was associated with 12-month retention, and we believe that this finding provided valuable information for developing future long-term cohort studies among IDUs: the short-term retention rate in a preparedness pilot study should predict the long-term follow-up rate in a cohort study. Ethnicity was another predictor for 12-month retention in this study; poorer retention of Yi and other minority ethnic subjects might be a result of their lifestyle of high migration. Yi or other ethnic minority people in the study area often move temporarily from their rural residence to Xichang City or travel a longer distance for trade and business.
This study found that the HIV incidence density rate during a 12-month follow-up period was 3.17 per 100 person-years (95% CI: 0.98, 5.37). A small follow-up study among IDUs in 3 counties of Yunnan Province during 1992 through 1994 found seroconversion rates ranging from 0 to 25 per 100 person-years.7 Another study reported an HIV incidence of 2.38 per 100 person-years among drug users during the first follow-up period (February 1998-January 1999) and 6.86 per 100 person-years during the second follow-up period (January-September 1999) in Pingxiang County of Guangxi Province.8 During the 12 months of the study period, drug treatment and needle and syringe exchange programs were not available to IDUs in Xichang County. The baseline screening survey of this study showed a prevalence rate of 11.3%,10 which is much lower than prevalence rates among IDUs in some areas of Yunnan, Guangxi, and Xinjiang.5 A conversion rate of greater than 3 per 100 person-years was found in our study subjects, which implies that, without interventions, HIV is spreading rapidly among IDUs in geographic areas other than Yunnan and Guangxi, the typical epicenters of the HIV/AIDS epidemic in China.
HIV-1 subtype analysis indicated the possibility of 3 drug-trafficking routes across mainland China.6,16,17 One of these routes is from Eastern Burma to Yunnan Province, through Chengdu, the capital city of Sichuan Province, and then to Xinjiang.6 This inference was primarily based on the fact that the prevalent subtypes of HIV-1 among IDUs in Yunnan and Xinjiang were subtype C or a combination of B/C.16,17 All the subtypes of 8 HIV-1 seroconversions in our study were CRF_07BC, which further supports the existence of the overland drug-trafficking route, which goes from Kunming, the capital city of Yunnan Province, to Xichang and Chengdu in Sichuan Province, and then to Urumqi, the capital city of Xinjiang Province (Fig. 1).
Limitations of this study include a relatively small sample size and possible selection bias. Only 8 HIV seroconversions were observed during the 12-month follow-up period, which makes it impossible for us to evaluate the risk factors for HIV seroconversion. Our subject recruitment efforts through outreach contacts and snowball sampling did not guarantee that we reached all IDUs in the community; nonparticipating IDUs might have different risk behaviors and therefore have a different risk for HIV infection and a different rate of retention than the participants. In addition, those lost to follow-up have a different demographic profile (ethnicity). Selection bias might lead us to overestimate or underestimate the true HIV-1 seroconversion rate. Follow-up of the study cohort is still ongoing. The retention strategies are to be re-examined and revised to increase the retention rate, and thereby minimize potential selection bias. A relatively high HIV incidence among IDUs in our study suggested that injection drug use and needle-sharing practices are facilitating the spread of HIV to more geographic areas in China.
1. China Ministry of Health and United Nations Theme Group on HIV/AIDS in China. A joint assessment of HIV/AIDS prevention, treatment and care in China. China Ministry of Health, Beijing, China; December 1, 2003.
2. Xiao S, Hao W, Young DS. Epidemiological survey on illicit drug use in five high prevalence areas in China. Part I: demographics and prevalence rates. Zhonghua Jing Sheng Wei Sheng Za Zhi
3. Hao W, Xiao SY, Liu TQ, et al. The second national epidemiological survey on illicit drug use at six high-prevalence areas in China: prevalence rates and use patterns. Addiction
5. Ma Y, Li ZZ, Zhang KX, et al. Identification of HIV infection among drug users in China. Zhonghua Liu Xing Bing Xue Za Zhi
6. Beyrer C, Razak MH, Lisam K, et al. Overland heroin trafficking routes and HIV-1 spread in south and south-east Asia. AIDS
7. Zheng XW, Tian CQ, Choi K, et al. Injecting drug use and HIV infection in southwest China. AIDS
8. Lai S, Liu W, Chen J, et al. Changes in HIV-1 incidence in heroin users in Guangxi Province, China. J Acquir Immune Defic Syndr
9. Yang TL, Xu YC, Hu XH, et al. The prevalence of HIV, HBC and HCV among drug users in Xichang county of Sichuan province, China. Yu Fang Yi Xue He Chuan Ran Bing Xue Za Zhi
10. Ruan YH, Chen KL, Hong KX, et al. Community-based survey of HIV transmission modes among injection drug users in Sichuan, China. Sex Transm Dis
11. Antoine M, Helen N, Annet DV, et al. Sociodemographic and behavioral characteristics associated with timeliness and retention in a 6-month follow-up study of high-risk injection drug users. Am J Epidemiol
12. Cottler LB, Compton WM, Ben-Abdallah A, et al. Achieving a 96.6 percent follow-up rate in a longitudinal study of drug abusers. Drug Alcohol Depend
13. Baldwin JA, Bowen AM, Trotter RT II. Factors contributing to retention of not-in-treatment drug users in an HIV/AIDS Outreach Prevention Project. Drugs Soc
14. Des Jarlais DC, Marmor M, Friedmann P, et al. HIV incidence among injection drug users in New York City, 1992-1997: evidence for a declining epidemic. Am J Public Health
15. NIMH Multisite HIV Prevention Trial. Screening, recruiting and predicting retention of participants in a multisite HIV prevention trial. AIDS
16. Shao YM, Su L, Sun XH, et al. Molecular epidemiology of HIV infections in China. Presented at: Fourth International Congress on AIDS in Asia and the Pacific; 1997; Manila.
17. Shao YM, Su L, Zhao F, et al. Genetic recombination of HIV-1 strains identified in China [oral presentation, W. A35:429]. Presented at: 12th World Conference on AIDS; Geneva; 1998.
Keywords:© 2005 Lippincott Williams & Wilkins, Inc.
HIV-1 subtypes; prospective cohort study; incidence; retention; injection drug users; China