Secondary Logo

Journal Logo


Correlates of HIV infection among former blood/plasma donors in rural China

Ji, Guopinga,b; Detels, Rogera; Wu, Zunyouc; Yin, Yuepingd

Author Information
doi: 10.1097/01.aids.0000210613.45212.c4
  • Free



Many HIV/AIDS epidemics have been identified in rural areas of China. Plasma donation has been identified as the cause of the HIV/AIDS epidemic in these areas [1–6]. In the early 1990s, thousands of small commercial plasma collection stations were set up in rural areas in China, in addition to the government-run blood stations. People were paid to donate plasma, and some people sold their plasma several times a week. Many plasma stations did not collect plasma according to standard procedures, causing cross-infection among former plasma donors (FPDs); for example, hepatitis B and hepatitis C [7–9]. In 1995, recognizing that HIV/AIDS was being spread by commercial plasma collection stations, the government of China closed all commercial plasma collection stations and issued new regulations for blood/plasma donation [1,2].

Initially, HIV infection among FPDs was reported in rural areas in the northern part of Anhui province [1,2]. Later, other provinces also reported similar problems, including Henan, Shanxi and Shaanxi [4–6]. Anhui is located in the eastern part of China, bordering Henan province in the north. Henan has been reported to have the most HIV/AIDS cases among FPDs [10].

Nine years after the first cases of HIV were reported in northern Anhui, more people have been diagnosed and reported as having HIV infection and AIDS, including some children and adults who do not have a history of plasma donation, suggesting that vertical and sexual transmission may have already occurred. The true numbers of blood/plasma donors, HIV infections, and AIDS cases are unknown.

The objective of this study was to identify the prevalence, spread, and correlates of HIV infection among rural residents in areas with many former commercial blood/plasma stations in rural China. The study was conducted in late 2003 and early 2004.


Study site and recruitment of participants

A cross-sectional study was implemented. The target population was rural residents in areas with many FPDs in rural northern Anhui, China, where the HIV/AIDS epidemic among FPDs was first reported [1,2]. Anhui is one of the source provinces for the ‘floating population’ in urban areas (migrants from rural to urban areas). According to local officials, two-thirds or more of the young and middle-aged adults have migrated from villages in northern Anhui. There are more men than women in this ‘floating population’.

A cluster-sampling method was employed to select participants in this study. Six townships with many former commercial blood/plasma donors were selected as the study sites. A list was compiled of the villages believed to have an adult population in which over 30% had been FPDs. Forty villages were then randomly selected from the list. All residents 25–55 years of age in the selected villages were invited to participate. The age range of 25–55 years was chosen because it included most FPDs and it is also the most sexually active years of life. There were no intravenous drug users known to the health authorities and police in this area.

The questionnaire

A questionnaire was anonymously administered face-to-face at the respondents' homes or in another private place chosen by the respondents. The questionnaire included questions on demographic characteristics, history of blood/plasma donations, total number of blood/plasma donations, and sexual behavior. No identifying information was requested or recorded.

A qualitative study was carried out before the questionnaire survey, which solicited information regarding social norms, including premarital and extramarital sex, stigma towards HIV/AIDS and people living with HIV/AIDS, attitudes toward blood/plasma donation and people with a history of blood/plasma donation, etc. The qualitative study was conducted in one village that had not been selected for the main study. The questionnaire was modified according to the results of the qualitative study.

The questionnaire and the survey procedures were pre-tested after the qualitative study and before the main study. The pilot study was conducted in one village within the six townships that had not been selected as a site for the main study. The questionnaire and survey procedures were revised before the main study.

HIV testing

A venous blood specimen for HIV testing was collected by the interviewers. The blood was first tested using an enzyme-linked immunosorbent assay (ELISA) (Huamei Co., Zhenzhou, China). Another ELISA (Kehua Co., Shanghai, China) was used to retest positive samples. For the samples positive for both ELISAs, a western blot assay (Genelabs Diagnostics Pte Ltd., Singapore) was used for confirmation. About 10% of the blood samples were sent to the National Center for STDs in Nanjing for quality control testing. The results from the quality control testing were in complete accordance with the results obtained by the study laboratory.

Before blood specimen collection, pre-test counseling was provided to each participant. Post-test counseling was offered at the city and county anti-epidemic stations. A telephone number for obtaining post-test counseling was provided if the participants wanted it.

Ethical review

This study was approved by the Institutional Review Boards (IRB) of the University of California at Los Angeles (UCLA) and the Chinese National Center for AIDS Prevention & Control, Center for Disease Control and Prevention. Each participant was informed that participation was voluntary, that their responses would be anonymous, and that they could refuse to answer any question or end the interview at any time.

Data analysis

The data were input using EpiData 2.1b (The EpiData Association, Odense, Denmark) and analyzed using STATA 7.0 (STATA Corporation, College Station, Texas, USA). ‘Survey data analysis’ procedures in STATA were used to analyze the data. ‘Proportions for survey data’ were used to calculate proportions and 95% confidence intervals (CI). A ‘two-way table for survey data’ was used to explore the relationship between HIV status and other variables. ‘Logistic regression for survey data’ was used to analyze the correlates of HIV infection.

In STATA ‘survey data analysis’, the χ2 test is converted to the F test. For readers' convenience, we still used χ2 as the symbol, but used the value of F for χ2. All confidence intervals and χ2 and P values reported here were corrected for the cluster survey design using STATA ‘survey data analysis’.

Since missing values occurred mainly in ‘having multiple sexual partners’ and it is hard to impute reliable values for refusals, a category of ‘refusal’ was included in the analysis.


A total of 2120 participants were recruited, of whom 2119 (99.9%) had blood drawn for the study. Sixty-eight (3.2%) refused to answer the questionnaire. Fifty-four (2.5%) subjects were ineligible because of age. Of the 2066 eligible subjects, 1997 (96.7%) completed the questionnaire and provided a blood specimen.

HIV prevalence

Among those who completed the survey, 216 (10.8%) were found to be HIV-positive (95% CI, 7.9–13.7%) (Table 1). The prevalence was 15.1% (95% CI, 11.4–18.9%) among FPDs and 4.8% (95% CI, 2.7–7.2%) among villagers without a history of plasma donation (Table 2).

Table 1:
Demographic characteristics of HIV-positive and -negative individuals in sample.
Table 2:
HIV risk characteristics of HIV-positive and -negative participants.

Demographic characteristics

Table 1 lists the demographic characteristics of the sample stratified by HIV status. More women participated in the survey than men, with a female-to-male gender ratio of 2.0 (1331/666). Men had a higher prevalence of HIV than women (15.8 versus 8.3%). Participants who had been employed or sought employment elsewhere in the past year had a higher HIV prevalence than those who did not.

There was no significant difference in HIV prevalence according to marital status, age, having extra jobs, or education level. Different townships had different rates of HIV prevalence, ranging from 5.7 to 21.6%.

HIV risk characteristics

Comparisons of HIV risk characteristics between HIV-positive and -negative subjects were conducted (Table 2). The history of blood and/or plasma donation, total number of plasma donations, spouse's HIV status, having more than one sexual partner, and total number of blood donations were significantly associated with HIV infection. Condom use was not associated with HIV infection, but was infrequent (ever having used a condom was 5.9%, and consistent condom use was only 1.1%).

Correlates for HIV infection

Donating both blood and plasma (OR, 5.91), donating plasma only (OR, 5.74), having an HIV-infected spouse (OR, 6.56), having multiple sex partners (OR, 3.37), male gender (OR, 1.81), and place of residence (OR, 2.08, township 2 versus township 6, the reference township) (Table 3) were significantly associated with HIV infection according to multivariate logistic regression.

Table 3:
Crude and adjusted odds ratios for factors associated with HIV infectiona.

Number of plasma/blood donations

When the total number of plasma and blood donations was used in the model, multivariate logistic regression analysis showed the same variables were associated with HIV infection as in the above model (Table 4). The more plasma donations made, the higher the risk of HIV infection (OR, 8.81 for participants who donated plasma more than 10 times; OR, 3.01 for those who donated plasma 4–9 times; and OR, 2.26 for those who donated 1–3 times). Number of blood donations was not associated with HIV infection, and after controlling for other variables, blood donation history was also not associated.

Table 4:
Number of blood/plasma donations and risk for HIV infectiona,b.

Correlates of HIV infection among plasma donors

Among plasma donors, the following variables were significantly associated with HIV infection after adjusting for other potential confounders: number of plasma donations (OR 4.09 for subjects who donated plasma more than 10 times), spouse being HIV-positive (OR, 4.06), and male gender (OR, 2.04). Having multiple sexual partners was marginally associated with HIV infection (OR, 2.37; P = 0.097).

Correlates of HIV infection among non-plasma donors

Factors that were significantly associated with HIV infection among non-plasma donors after adjusting for other variables were spouse being HIV positive (OR, 11.07) and having multiple sexual partners (OR, 7.04) (Table 5). Gender was not associated with HIV infection among non-plasma donors. Condom use was rare, and none of the HIV-positive participants had used condoms in the past year, precluding analysis of correlates.

Table 5:
Crude and adjusted odds ratios of HIV infection among non-plasma donorsa.


This survey revealed that the prevalence of HIV infection was 10.8% among rural residents in villages with many FPDs in rural Anhui, China. Amongst FPDs, the prevalence was 15.2%. For residents without a history of plasma donations, the prevalence was 4.9%. Both were higher than in a previous survey done in 1996 in the same area (12.5% among FPDs and 2.1% among non-donor spouses) [2]. Yan et al. found the HIV prevalence was 25.9% among FPDs in another county in 1999 [4]. Zheng et al. found the HIV prevalence was 17.0% among FPDs and 0.2% among non-plasma donors in another county in 2000 [5]. The latter two studies had a smaller sample size of plasma donors. It is clear that HIV is now spreading beyond the FPD population, which is a cause for alarm.

HIV prevalence was associated with both a history of plasma donations and total number of plasma donations. The above-mentioned studies also showed the same relationship. This indicates that plasma donation 8–9 years ago was the major cause of the current HIV/AIDS epidemic in rural China. However, this study also clearly demonstrated that donating blood did not put donors at risk of HIV infection.

It was surprising that men had a much higher prevalence than women (15.8 versus 8.3%). The possible explanations may include but are not limited to the following factors: (1) more healthy (that is, more HIV-negative) men migrated out for jobs; and (2) more HIV-infected women died than men because the natural history of HIV infection is shorter in women than men [11–13].

Nearly 5% of non-plasma donors were found to be HIV-positive, suggesting that HIV has spread from FPDs to the general population. The multivariate logistic regression analysis conducted for the total sample, subgroups of plasma donors, and non-plasma donors all showed that a spouse being HIV-positive and having multiple sexual partners were associated with HIV infection. These indicated that HIV secondary transmission through sex is occurring in the area, both within the marriage (spouse being HIV-positive) and outside of marriage (multiple sexual partners). This spread is enhanced by low rates of condom use. Therefore, condoms should be promoted urgently to prevent further spreading of the virus through sex. Condoms must be used not only for extramarital sex, but also within marriage, especially by HIV-discordant couples. Consistent condom use should be emphasized in rural areas, because it is different from the information provided to them for the purpose of family planning.

HIV voluntary counseling and testing (VCT) should be promoted, both because free antiretroviral treatment is becoming available in these areas affected by HIV through plasma donation in China, and because knowledge of HIV status reduces risk behaviors [14–16]. Without knowledge of their HIV status, infected individuals will not seek the treatment they need and are unlikely to reduce their risk behaviors.

The difference in HIV prevalence between this and the previous study among FPDs in the same areas (15.2 versus 12.5%) and non-plasma donors (4.9 versus 2.1%) may indicate a true increase, due to sexual transmission that occurred after the first survey, nosocomial transmission that occurred after the first survey, and/or different migration rates between HIV-positive and -negative villagers. We reason that the difference probably includes all four, but sexual transmission probably accounts for the majority of the difference, given that having an HIV-positive spouse and/or having multiple sexual partners were both strongly associated with HIV infection.

There are several limitations in this study. First, many local rural residents temporarily migrate to urban areas. In this study, there were many more female than male participants, yielding a gender ratio of 2.0, so this sample was not representative of all the rural residents. Given the fact that the HIV/AIDS epidemic started in early 1990s and the average incubation of HIV/AIDS is 9 years or less, people infected with HIV were more likely to stay at home. If this were the case, the prevalence of HIV infection identified by this study might be higher than true prevalence. Second, all variables except for HIV status were self-reported; thus, some sensitive information could be under-reported, such as extramarital sex. A study using tape recorder and earphones to administer questions in the same area indicated a higher rate of extramarital sex (7.8%) [17]. If some FPDs reported being non-FPDs, the HIV prevalence would be overestimated among non-FPDs and underestimated among FPDs. Given the fact that blood/plasma donation was common in the area, this reporting bias is likely to be low. Third, nosocomial transmission of HIV infection was not assessed in this study. Given the low rate of compliance of universal precautions in rural Anhui [18], nosocomial transmission of HIV is likely to occur. A previous study found that a history of invasive medical care was not associated with HIV infection [2]. Nonetheless, the magnitude of nosocomial HIV transmission should be assessed in future studies.


The HIV/AIDS epidemic has been introduced from FPDs into the general population, which is a very disturbing observation that does not auger well for the future of the HIV epidemic in China, unless dramatic action is taken immediately.

The factors associated with HIV infection are frequency of plasma donations, spouse being HIV-positive, having multiple sexual partners, and residential location. Thus, both plasma donation and sexual transmission were responsible for the current HIV/AIDS epidemic among adults in the area.

Since condom use is low, HIV-infected villagers have multiple sexual partners, and villagers frequently migrate to urban areas to look for jobs, it is likely that HIV will continue to spread both in the local areas and in the urban areas to which villagers temporarily migrate.

Condoms should be promoted urgently for both intra- and extramarital sex to prevent further spreading of the virus through sex in rural China. HIV VCT should be promoted, especially as antiretroviral treatment is becoming available in these areas affected by HIV in China.


We thank Wendy Aft of the UCLA School of Public Health for her assistance with preparing the manuscript.

Sponsorship: This research was supported by the Fogarty AIDS International Training and Research Program (D43 TW000013).


1. Wu Z, Liu Z, Detels R. HIV-1 infection in commercial plasma donors in China. Lancet 1995; 346:61–62.
2. Wu Z, Rou K, Detels R. Prevalence of HIV infection among former commercial plasma donors in rural eastern China. Hlth Policy Plan 2001; 16:41–46.
3. Beach MV. “Blood heads” and AIDS haunt China's countryside. Lancet 2001; 357:49.
4. Yan J, Zheng X, Zhang X. The survey of prevalence of HIV infection among paid blood donors in one county in China [in Chinese]. Zhonghua Liu Xing Bing Xue Za Zhi 2000; 21:10–13.
5. Zheng X, Wang Z, Xu J. The epidemiological study of HIV infection among paid blood donors in one county of China [in Chinese]. Zhonghua Liu Xing Bing Xue Za Zhi 2000; 21:253–255.
6. Zhao S, Shang G. Shaanxi blood donor HIV continues to spread. In Southern Weekend, [Chinese newspaper], March 15, 2001.
7. Ye DQ, Hu ZP, Song HP. Serological epidemiology of blood donors in Hefei, Anhui province [in Chinese]. Chin J Public Hlth 1998; 17:366–367.
8. Zhang YX, Luo XX, Gui XE, Zhang XF, Li Y. The prevalence of HCV, HBV and HIV among paid blood donors [in Chinese]. J Clin Intern Med (Chinese) 2001; 18:308–309.
9. Yu M, Tian XL, Han HZ, Wan H. HIV infection among plasma donors who are infected with HCV [in Chinese]. Chin J AIDS/STD 2003; 9:197–198.
10. Cohen J. HIV/AIDS in China: an unsafe practice turned blood donors into victims. Science 2004; 304:1438–1439.
11. Farzadegan H, Hoover DR, Astemborski J, Lyles CM, Margolick JB, Markham RB, et al. Sex differences in HIV-1 viral load and progression to AIDS. Lancet 1998; 352:1510–1514.
12. Carpenter CC, Mayer KH, Fisher A, Desai MB, Durand L. Natural history of acquired immunodeficiency syndrome in women in Rhode Island. Am J Med 1989; 86:771–775.
13. Carpenter CC, Mayer KH, Stein MD, Leibman BD, Fisher A, Fiore TC. Human immunodeficiency virus infection in North American women: experience with 200 cases and a review of the literature. Medicine 1991; 70:307–325.
14. MacNeil JM, Mberesero F, Kilonzo G. Is care and support associated with preventive behaviour among people with HIV? AIDS Care 1999; 11:537–546.
15. Louria DB, Skurnick JH, Palumbo P, Bogden JD, Rohowsky-Kochan C, Denny TN, et al. HIV heterosexual transmission: a hypothesis about an additional potential determinant. Int J Infect Dis 2000; 4:110–116.
16. Meursing K, Sibindi F. HIV counselling––a luxury or necessity? Hlth Policy Plan 2000; 15:17–23.
17. Liu H, Xie J, Yu W, Song W, Gao Z, Ma Z, et al. A study of sexual behavior among rural residents of China. J Acquir Immune Defic Syndr Hum Retrovirol 1998; 19:80–88.
18. Ji G, He L, Liu Y, Chu H. Report on universal precaution compliance and occupational exposure among rural grassroots health workers who provide obstetric and gynecologic services. Chin J AIDS STDs 2003; 9:235–238.

HIV/AIDS; former plasma donors; China; HIV transmission

© 2006 Lippincott Williams & Wilkins, Inc.