Introduction
In the early 1990s, the HIV/AIDS epidemic in China was mainly distributed along the southwestern border regions, particularly in Yunnan province. Injection drug users (IDU) were the most affected population, and contributed approximately 70% of reported cases in that period [1,2]. The HIV epidemic among IDU has continued to grow at a fast pace and spread inland to other parts of the country. By 2002, HIV infection among IDU had been reported in all 31 mainland provinces [3]. The virus not only spread among IDU, but also spread from IDU to others engaging in high-risk sexual behaviors and later into the general population through sexual transmission [4,5].
The main risk of HIV transmission associated with IDU is multiperson re-use or sharing of syringes and needles [6,7]. Indirect sharing of equipment and materials, such as water, cotton, ‘cookers’, and other drug preparation items may also contribute to HIV transmission [7,8]. Moreover, IDU may also put themselves and others at risk of sexual HIV transmission by exchanging sex for drugs or money [9,10]. IDU are also at high risk of infection by other blood-borne diseases such as hepatitis B virus (HBV) and hepatitis C virus (HCV) [11–13].
Sentinel surveillance and cross-sectional survey data have shown a high prevalence of HIV, HCV and HBV among IDU in southwestern and northwestern China [1,11–15]; however, there are limited data on the incidence over time [6,12,16–18]. These cohort studies have shown varying but generally high seroconversion rates. As drug injecting practices remain prevalent among drug users in many parts of China, more incidence data are needed to obtain a better understanding of the evolving epidemics of blood-borne viral infections among IDU. The aim of this paper was to report the trends in HIV, HCV and HBV seroconversion rates over 3 years among a cohort of IDU in Sichuan province, China.
Methods
Study design and study population
This study was conducted in Xichang City, Sichuan province in southwestern China. The city is located along one of China's main drug trafficking routes. A baseline screening survey was conducted in November 2002 among 379 IDU who were recruited through community outreach and snowball sampling. Those who were HIV seronegative at baseline, were 18 years or older, had injected drugs at least once in the past 3 months, and were willing to provide informed consent were eligible for enrollment in the cohort study. Participants were asked to come back to the study clinic for follow-up evaluations every 6 months. The study protocol and cohort retention have been described in detail elsewhere [16].
Data collection
An interviewer-administrated questionnaire was used to collect data on demographic characteristics, drug use and sexual behaviors. Demographic variables included sex, age, ethnicity, years of education, marital status, personal income in the past year, employment in the past 6 months and home ownership. Questions pertaining to drug use investigated the frequency of drug use and drug injection in the past 3 months and the frequency of shared injection devices in the past 3 months, including shared needles or syringes, cookers, cotton, rinsing water, and the use of ‘front or back-loading’ (i.e. transferring drug solution from one syringe into another by removing the needle or by removing the plunger before injecting). Questions about sexual behavior included having an HIV-positive sexual partner, having an IDU as a sexual partner, being a man who has sex with men, having regular and casual sex, condom use in the past month, and new sexual partners and engagement in commercial sex in the past 6 months. Unsafe sex was defined as not always using condoms with regular or casual sexual partners in the 6 months before the visit. Regular sexual partners were defined as those partners with whom the sexual relationship lasted for more than 3 months. Casual sexual partners were defined as those partners with whom the sexual relationship lasted less than 3 months. This questionnaire was developed mainly on the basis of another international cohort study, the HIV Prevention Trials Network Preparedness Study (HPTN 033) [17,18].
Laboratory analysis
A blood sample was collected from each participant for testing HIV and HCV antibodies and hepatitis B surface antigen (HBsAg). Syphilis infection was also tested but was not reported in this paper. HIV was tested by enzyme-linked immunosorbent assay (ELISA; Beijing Wantai Biological Medicine Company, China) and confirmed by HIV-1/2 Western blot immunoassay (HIV Blot 2.2 WB; Genelabs Diagnostics, Singapore). A positive result on both tests was indicative of HIV infection. HIV subtype analysis was based on the sequences from the env and gag regions of the HIV-RNA envelope. Samples were tested for antibodies to HCV by ELISA (Beijing Jinhao Biological Production Company, China). HBsAg was screened using an HBV ELISA (Beijing Wantai Biological Medicine Company).
Statistical analysis
Data were double-entered and the consistency of both databases was compared using Epi Data software (Epi Data for Windows; Epi Data Association, Odense, Denmark), and were analysed using the Statistical Analysis System version 9.1 (SAS Institute Inc., Cary, North Carolina, USA). Time zero was defined as the date of enrollment in 2002.
The events of interest in this study were seroconversions to HIV and HCV based on the detection of antibodies and infection from hepatitis B virus based on HBsAg. The date of seroconversion was estimated using the midpoint between the last negative and the first positive antibody test result. We divided the 36-month follow-up period into three observation periods: years 1, 2 and 3, based on preliminary analyses of incidence rates within each period. As some subjects might not show up in the previous year, but then show up in the following year, a weighted incidence rate was calculated for each year. For example, if one missed both appointments in year 1, and then showed up in year 2 and was tested HIV positive, he would contribute to case and partial person-year in both years, calculated assuming that HIV risk was constant over these periods. He would no longer contribute a person-year to HIV incidence in year 3, but he might continue to contribute as cases and person-years in year 3 for the HCV or HBV seroconversion rate if HCV or HBV infections were not apparent at the end of the second year.
Demographic characteristics of participants who failed to attend a follow-up visit were compared with those who attended by chi-square test. Incidence rates of HIV, HCV and HBV infections and their 95% confidence intervals (CI) were calculated using incidence density per 100 person-years of follow-up. Trends were analysed for each infection by Poisson regression. To explore risk factors associated with cumulative seroconversion of each infection, univariate Poisson regression models were used to assess the effect of both fixed covariates (e.g. sex, ethnicity, marital status, education level and unstable housing) and time-dependent covariates (e.g. frequency of injection drug use and sexual behaviors). Variables that were either marginally significantly (P < 0.10) associated with time to seroconversion in the univariate analyses were considered for inclusion in multivariate Poisson regression models.
Results
Baseline characteristics and retention of the study cohort
The prevalence rates at baseline in 2002 were 11.3% (43/379) for HIV, 71.0% (269/379) for HCV, and 16.9% (64/379) for HBV. Of 333 IDU who were HIV seronegative and eligible to enter the cohort, 107 (32.1%) and 279 (83.8%) were HCV and HBsAg negative, respectively, individuals therefore contributing to the incidence study.
Among 333 individuals in the initial HIV-seronegative cohort, 19% were recruited through community-based outreach and 81% through peer referral; 81.7% were men; the mean age was 28.8 years (range 18–44); 66.4% were of the Han ethnic group and 28.8% belonged to the Yi minority. Twelve percent had no education, 28.2% had attended a primary school (up to 6 years), and 43.2 and 9.9% attended junior (up to 9 years) and senior high (up to 12 years) schools, respectively. The proportion of participants who were unemployed was 56.5%, and 24.9% were farmers. Nearly half the participants (46.9%) were single, 29.1% were married, 14.7% were divorced, and 7.8% were cohabiting with partners. One third owned a house (or apartment), 12.0% rented a house, and 48.1% lived in their parents' houses. The median annual income per person was US$872. Heroin was the only illicit drug used by anyone in this cohort.
The demographic characteristics of IDU who attended or failed to attend one of the follow-up visits can be found in Table 1. The follow-up rates were 70.3% (234/333) in year 1, 75.7% (252/333) in year 2 and 68.8% (229/333) in year 3. Those who appeared or failed to appear in the first, second, and third years' follow-up visits were comparable in terms of ethnicity, education, length of time injecting drugs, frequency of injecting drugs in the past 3 months, and sharing needles and syringes in the past month (Table 1). Of the 104 participants lost to follow-up at the 36-month visit (average 10.4% per year), 39 (37.5%) had died; 27 (69.2%) of these deaths were attributed to a heroin overdose. Other reasons for loss to follow-up included internment in jail or detoxification centers (10.6%), migration out of the study area (22.1%), incorrect contact information (3.8%), and for the remaining 26.0%, the reason was unknown.
;)
Table 1: Demographic characteristics of injection drug users who appeared and failed to appear at the end of the first, second and third follow-up years.
Incidence rates of HIV, hepatitis C and B virus infections
Fourteen HIV seroconversions were observed over 619.0 person-years of observation, 2.3 per 100 person-years (95% CI 1.1–3.5) during 3 years of follow-up. HIV incidence was 4.1 per 100 person-years (95% CI 1.6–6.5) in year 1 (from 2002 to 2003), was 0.76 (95% CI 0.00-1.9) in year 2 (from 2003 to 2004), and was 1.4 (95% CI 0.00–3.3) in year 3 (from 2004 to 2005; Fig. 1). Co-infection of HIV with HCV was typical; among the 14 HIV seroconverting IDU, 13 (92.9%) were co-infected with HCV. Analysis of the HIV-1 sequence from the env and gag regions of the virus envelope indicated that HIV-1 strains in these seroconverting participants were all from the CRF_07BC subtypes.
Fig. 1: Trends in HIV, hepatitis C and hepatitis B virus incidence rates in a 36-month cohort of 333 injection drug users in southwestern China. Note: retention in years 1, 2 and 3 was 70.3, 75.7 and 68.8%, respectively.
HIV;
hepatitis C virus (HCV);
hepatitis B virus (HBV).
Forty-seven HCV seroconversions were observed over the study period, 33.3 per 100 person-years (95% CI 23.7–42.8). In year 1, the incidence rate was 43.9 per 100 person-years (95% CI 28.6–59.2), was 27.9 (95% CI 10.9–44.9) in year 2, and was 15.6 (95% CI 1.92–29.25) in year 3 (Fig. 1).
Fifty-one HBV seroconversions were observed out of 193 subjects with negative HBsAg at the baseline survey, 11.3 per 100 person-years (95% CI 8.2–14.4). The HBV incidence rate was 14.4 per 100 person-years (95% CI 9.1–19.6) in year 1, was 9.7 (95% CI 4.8–14.6) in year 2, and was 7.7 (95% CI 2.2–13.2) in year 3 (Fig. 1).
Pairwise comparisons of the incidence rate differences were unlikely to be the result of chance alone (P < 0.05) comparing HIV incidence in years 1 and 2 and HCV incidence comparing years 1 and 3.
Predictors of HIV, hepatitis C and B virus seroconversions
Univariate Poisson regression analyses showed that several factors were significantly associated with HIV seroconversion, including ethnicity, years of education, employment status, using drugs in the past 3 months, and sharing needles or syringes in the past 3 months (Table 2). Significant variables in univariate models were included in the multivariate model that identified two statistically significant risk factors for HIV seroconversion: being a member of an ethnic minority [risk ratio (RR) 6.1; 95% CI 1.9–19.9; P = 0.003] and more frequent sharing of needles or syringes in the past 3 months (one or more versus less than one time per week; RR 7.7; 95% CI 2.4–25.3; P < 0.001; Table 3).
Table 2: Factors associated with HIV, hepatitis C and hepatitis B virus seroconversions in a 36-month cohort of 333 injection drug users in univariate Poisson regression models.
Table 3: Factors associated with HIV, hepatitis C and hepatitis B virus seroconversions in a 36-month cohort of 333 injection drug users in multivariate Poisson regression models.
Independent predictors of HCV conversion included female sex (RR 4.0; 95% CI 1.8–8.9; P = 0.001), higher frequent drug use in the past 3 months (seven or more versus less than seven times per week; RR 3.7; 95% CI 2.0–6.9; P < 0.001) and more frequent sharing of needles or syringes in the past 3 months (one or more versus less than one time per week; RR 3.2; 95% CI 1.1–9.3; P = 0.003; Table 3).
Independent predictors for incident HBV antigenemia included higher frequent drug injection in the past 3 months (seven or more versus less than seven times per week; RR 2.9; 95% CI 1.6–5.0; P < 0.001) and co-infection with syphilis (RR 2.2; 95% CI 1.2–4.0; P = 0.01; Table 3).
Discussion
This 36-month cohort study extended our earlier research findings [16] in two ways: We reported 3 years of follow-up rather than just one year; we reported incidence rates of three blood-borne viral infections, HIV, HCV and HBV, among IDU, instead of only HIV incidence as in the previous paper. Although we observed decreased incidence rates in the second and third years of the cohort compared with year 1 incidence rates, between 2002 and 2005, these reductions were not significant by trend tests overall, perhaps because of the limited study sample size. Two pairwise comparisons were significantly different for the decline in HIV incidence from year 1 to year 2 and the decline in HCV incidence from year 1 to year 3. Several reasons may explain the decline in incidence rates. First, as this was a closed cohort, the pool of negative subjects for each infection shrank over time, particularly for HCV infection as it represented the highest seroconversion rate; second, subjects were exposed to testing-related counselling and referral to services that might have moderated their risky behaviors; third, a methadone maintenance programme was started in early 2004, and some subjects participated in this programme, further reducing their risky behaviors in study years 2 and 3.
There is limited literature about the incidence of HIV and HCV infections among Chinese IDU. A 12-month follow-up study from 2002 to 2003 among 508 and 500 IDU in Xinjiang (northwestern China) and Guangxi (southern China) found an HIV-1 incidence rate of 8.8 per 100 person-years (95% CI 6.3–12.0) and 3.1 per 100 person-years (95% CI 1.6–5.2), respectively [17,18]. An earlier study from 1993 to 1994 found that HIV incidence in Yunnan province (southwestern China) ranged from 16.3 to 25.0 per 100 person-years depending upon the location of the subgroup [19]. Another study in Guangxi province found that HIV incidence increased from 2.4 per 100 person-years in 1998 to 6.9 in 1999, and HCV incidence changed from 26.8% per 100 person-years to 28.9% in the same period [6]. Compared with these study findings, HIV incidence in our study sample was lower, and both HIV and HCV incidence rates did not show upward trends. The HIV epidemic among Chinese IDU started earliest in Yunnan province, then in Guangxi, and then in Sichuan and Xinjiang provinces [4]. Previous studies might reflect the rising waves of the epidemic in the study regions in the time periods studied [6,17,18]. In our study site, several harm-reduction programmes were available, including government-sponsored voluntary detoxification treatment (instituted in the late 1990s), needle exchange programmes managed by the China–UK AIDS Prevention and Care Project in 2002–2005, and, as stated earlier, government-sponsored methadone maintenance therapy (MMT) from 2004. An evaluation of the impact of MMT in the study site found that MMT clients reduced their frequency of risky drug use and sexual behaviors more than non-participants [20–22]. Supporting our suggestion that the reduction of risky behaviors might help explain the lack of an increase in incidence rates in our study sample, the prevalence of sharing needles and equipment in the past 3 months decreased from 22.2% in year 1 to 7.1% in year 2 and 5.7% in year 3. As seen in other international studies [23], however, interventions could not eliminate all risky behaviors, and substantial ‘residual’ risks in our cohort were likely to contribute to new infections. An incidence rate of 1.4 per 100 person-years at the end of our study is still unacceptably high from a public health perspective. Our study also showed that HCV and HBV incidence were much higher than HIV incidence despite their smaller initial pools of negative subjects. More effective HIV prevention interventions including HCV and HBV risk-reduction components are needed that cover a greater proportion of the persons in need.
Sharing needles was independently associated with HIV and HCV seroconversion; higher frequent drug use was a predictor of HCV incidence and injecting drugs was a predictor of HBV incidence. There were no risky sexual behaviors independently associated with either of the infections, and only HBV was associated with syphilis infections. These findings suggested that the major risk factor for all three viral infections among these IDU was risky drug use behaviors. Yi and other minority ethnic IDU had a higher risk of HIV infection, and female IDU had a higher risk of HBV infection. In past studies, minority ethnic and female drug users often had lower incomes and higher rates of unemployment than members of the Han ethnic group and men [24]. Yi ethnic people have been reported to be more likely to be involved in drug smuggling and drug use [25], and female IDU are more likely to sell sex for drugs or money thus putting themselves at greater risk of contracting disease [26,27]. Despite adjusting for needle sharing, individuals from ethnic minority backgrounds remain more likely to experience HIV seroconversion; this may suggest that ethnic minority status may be a surrogate for unmeasured risks. Intervention projects should give high priority to minority ethnic and female IDU.
This study has several limitations. First, a majority of participants (81%) were recruited through ‘snowball’ sampling in which a recruited individual recruits others, and this sampling method may lead to a biased estimate of incidence rates. For example, if the initial ‘seed’ drug users had extremely high-risk behaviors, the sample in our study would be more likely to have an extremely high risk and the estimates of incidence rates would be overestimated. Second, it is a one-site study, and the study findings about the changes in HIV, HCV and HBV incidence may not be generalized to other areas considering the wide geographical variations seen in China. Third, although approximately two-thirds (68.8%) of IDU were followed up at the 36-month follow-up survey, we were unable to obtain full longitudinal information on all participants. Differential loss to follow-up may have affected the study findings. If minority subjects and those with more risky drug use behavior had been more likely to drop out of the study, the incidence rates would have been underestimated. Fourth, the sample size was not large enough to provide us with adequate statistical power for some analyses. Finally, we did not test for anti-hepatitis B core antigen so some HBsAg-positive individuals may be reactivation from previousr infection rather than a new infection. This may lead to an overestimation of HBsAg incidence in our study. Cohort data from IDU in China are rare, however, and we believe that the data will help guide public health planning, preventive measures and clinical research planning, and help mobilize local officials to address the serious infectious disease hazards that accompany IDU.
Our group is seeking to continue the IDU cohort with an improved research design, and steer future research efforts towards innovative interventions for drug abuse treatment that can be implemented in China. We will also collaborate with the care and treatment sectors of the national and local health authorities and research partners to provide more comprehensive care to this population, both for drug treatment and for infectious diseases.
Acknowledgements
The authors would like to thank all staff in the National Center for AIDS/STD Control and Prevention, Sichuan Provincial Center for Disease Control and Prevention, and Xichang Center for STD and Leprosy Control who have provided assistance in many ways. They also thank the protocol team of HPTN 033 for their help in developing the parent questionnaire. The authors are also grateful to all study participants.
Sponsorship: This study was supported by grants from the Ministry of Science and Technology of China (2004BA719A01), the National Natural Science Foundation of China (30571612, 10501052) and by the China Comprehensive Integrated Programs for Research on AIDS (CIPRA U19AI51915), the National Institute of Allergy and Infectious Diseases, National Institutes of Health, USA. HZQ received supported from the Vanderbilt–UAB AIDS International Training and Research Program grant D43 TW010035 from the Fogarty International Center, National Institutes of Health, USA.
Conflicts of interest: None.
References
1. Xiao Y, Kristensen S, Sun J, Lu L, Vermund SH. Expansion of HIV/AIDS in China: lessons from Yunnan Province. Soc Sci Med 2007; 64:665–675.
2. Department of Disease Control, Minister of Health China, National Center for AIDS Prevention and Control, Group of National HIV Sentinel Surveillance.
National sentinel surveillance of HIV infection in China from 1995 to 1998 [in Chinese].
Zhonghua Liu Xing Bing Xue Za Zhi 2000;
21:7–9.
3. State Council AIDS Working Committee Office, UN Theme Group on HIV/AIDS in China.
A joint assessment of HIV/AIDS prevention, treatment and care in China [in Chinese]. Beijing: China Ministry of Health; 2004.
4. Qian HZ, Vermund SH, Wang N. Risk of HIV/AIDS in China: subpopulations of special importance. Sex Transm Infect 2005; 81:442–447.
5. Chen KT, Qian HZ. Mother to child transmission of HIV in China. BMJ 2005; 330:1282–1283.
6. Lai S, Liu W, Chen J, Yang J, Li ZJ, Li RJ,
et al. Changes in HIV-1 incidence in heroin users in Guangxi Province, China. J Acquir Immune Defic Syndr 2001; 26:365–370.
7. Strathdee SA, Galai N, Safaiean M, Celentano DD, Vlahov D, Johnson L,
et al. Sex differences in risk factors for HIV seroconversion among injection drug users: a 10-year perspective. Arch Intern Med 2001; 161:1281–1288.
8. Gaskin S, Brazil C, Pickering D. The sharing of injecting paraphernalia by intravenous drug users (IDUs) within a Worcestershire cohort, with specific reference to water and filters. Int J Drug Policy 2000; 11:423–435.
9. Kral AH, Bluthenthal RN, Lorvick J, Gee L, Bacchetti P, Edlin BR. Sexual transmission of HIV-1 among injection drug users in San Francisco, USA: risk-factor analysis. Lancet 2001; 357:1397–1401.
10. Kerr T, Stoltz JA, Strathdee S, Li K, Hogg RS, Montaner JS,
et al. The impact of sex partners' HIV status on HIV seroconversion in a prospective cohort of injection drug users. J Acquir Immune Defic Syndr 2006; 41:119–123.
11. Ruan YH, Chen KL, Hong KX, Liu SZ, He YX, Zhou F,
et al. Community-based survey of HCV and HIV co-infection in injection drug abusers in Sichuan Province of China. World J Gastroenterol 2004; 10:1589–1593.
12. Garten RJ, Lai S, Zhang J, Liu W, Chen J, Vlahov D,
et al. Rapid transmission of hepatitis C virus among young injecting heroin users in Southern China. Int J Epidemiol 2004; 33:182–188.
13. Zhao M, Du J, Lu GH, Wang QY, Xu H, Zhu M,
et al. HIV sexual risk behaviors among injection drug users in Shanghai. Drug Alcohol Depend 2006; 82(Suppl. 1):S43–S47.
14. Liu W, Chen J, Rodolph M, Beauchamp G, Masse B, Wang S,
et al. HIV prevalence among injection drug users in rural Guangxi China. Addiction 2006; 101:1493–1498.
15. Zhang Y, Shan H, Trizzino J, Ruan Y, Beauchamp G, Masse B,
et al. Demographic characteristics and risk behaviors associated with HIV positive injecting drug users in Xinjiang, China. J Infect 2007; 54:285–290.
16. Ruan Y, Qin G, Liu S, Qian H, Zhang L, Zhou F,
et al. HIV incidence and factors contributed to retention in a 12-month follow-up study of injection drug users in Sichuan Province, China. J Acquir Immune Defic Syndr 2005; 39:459–463.
17. Zhang Y, Shan H, Trizzino J, Ruan Y, Beauchamp G, Masse B,
et al. HIV incidence, retention rate, and baseline predictors of HIV incidence and retention in a prospective cohort study of injection drug users in Xinjiang, China. Int J Infect Dis 2007; 11:318–323.
18. Liu W, Chen J, Rodolph M, Beauchamp G, Li R, Wang S,
et al. HIV incidence, retention, and changes of high-risk behaviors among rural injection drug users in Guangxi, China. Subst Abuse 2006; 27:53–61.
19. Zheng X, Zhang JP, Qu SQ. A cohort study of HIV infection among intravenous drug users in Ruili and other two counties in Yunnan Province, China, 1992–1995 [in Chinese]. Zhonghua Liu Xing Bing Xue Za Zhi 1997; 18:259–262.
20. Qian HZ, Hao C, Ruan Y, Cassell HM, Chen K, Qin G,
et al.
Impact of methadone on drug use and risky sex in China.J Subst Abuse Treat 2007; September.[E-pub ahead of print].
21. Chun H, Li Z, Yuhua R, Yu Z, Zongliang F, Lu Y,
et al. On the impact of community-based Methadone Maintenance Treatment among drug addicts [in Chinese]. Yu Fang Yi Xue Qing Bao Za Zhi 2006; 22:251–255.
22. Chun H, Wei H, Yuhua R, Ze'en M, Xuegui Y, Li Z,
et al. Impact of community-based methadone maintenance treatment on high risk sexual behaviors among heroin addicts [in Chinese]. Zhong Guo Xing Bing Ai Zi Bing 2006; 12:307–310.
23. Nelson KE, Galai N, Safaeian M, Strathdee SA, Celentano DD, Vlahov D. Temporal trends in the incidence of human immunodeficiency virus infection and risk behavior among injection drug users in Baltimore, Maryland, 1988–1998. Am J Epidemiol 2002; 156:641–653.
24. He YX, Ruan YH, Teng T, Hao QN, Qin GM, Wu JL,
et al. Community-based survey of drug use and sexual behavior among female injection drug users [in Chinese]. Zhong Guo Xing Bing Ai Zi Bing 2003; 9:343–346.
25. Yin L, Qin G, Qian HZ, Zhu Y, Hu W, Zhang L,
et al.
Continued spread of HIV among injecting drug users in Southern Sichuan Province, China.Harm Reduct J 2007;
4:6. doi: 10.1186/1477-7517-4-6.
26. He YX, Teng T, Ruan YH, Zhou F, Liu SZ, Liu G,
et al. A survey of drug abuse and behaviors among injection drug users in Liangshan of Sichuan Province [in Chinese]. Zhongguo Yao Wu Lan Yong Fang Zhi Za Zhi 2004; 10:80–83.
27. Ruan Y, Cao X, Qian HZ, Zhang L, Qin G, Jiang Z,
et al. Syphilis among female sex workers in southwestern China: potential for HIV transmission. Sex Transm Dis 2006; 33:719–723.
