Prevalence and trend of HIV infection among voluntary blood donors in China since implementation of the Blood Donation Law: a systematic review and meta-analysis

Authors


Corresponding Author Yan Hong, Department of Social and Behavioral Health, School of Rural Public Health, Texas A&M Health Science Center, 1266 TAMU, College Station, TX 77843, USA. Tel.: +1 979 862 1700; Fax: +1 979 458 4264; E-mail: yhong@srph.tamhsc.edu

Abstract

Objective  In 1998, the Blood Donation Law was enacted in China in response to the outbreak of HIV endemic in central rural China as a result of unhygienic and commercial blood collection. This study aims to provide a first comprehensive review of available data on the prevalence, trend and other epidemiological characteristics of HIV infection among voluntary blood donors since 1998.

Methods  Major English and Chinese databases were searched and a systematic review was constructed. Pooled infection rates by province and year were calculated using random-effect or fixed-effect models.

Results  A total of 87 studies met our inclusion criteria. A total of 2573 HIV-positive cases were identified among voluntary blood donors in the past 10 years; the pooled prevalence was 13.22/100 000, with a range of 0.74–125.97 per 100 000. Among the 24 provinces that reported yearly data, the prevalence of HIV increased from 5.62/100 000 to 28.90/100 000. The male-to-female ratio was 2.8; about 60% were below the age of 30 years.

Conclusions  The prevalence of HIV in voluntary blood donors has risen steadily and fast. Efficient measures need to be taken urgently to prevent HIV test-seeking through blood donor programmes, to promote voluntary blood donation in low-risk groups and to enforce the Blood Donation Law strictly.

Abstract

Objectif:  En 1998, la loi sur le don de sang a été promulguée en Chine en réponse à l’épidémie du VIH endémique dans le centre de la Chine rurale en raison de la collecte non hygiénique et commerciale du sang. La présente étude vise à fournir un premier examen détaillé des données disponibles sur la prévalence, la tendance et autres caractéristiques épidémiologiques de l’infection VIH parmi les donneurs volontaires de sang depuis 1998.

Méthodes:  Les principales bases de données en anglais et en chinois ont été recherchées et une analyse systématique a été menée. Les taux d’infection poolés par province et par année ont été calculés en utilisant des modèles à effet aléatoire ou à effet fixe.

Résultats:  87 études répondaient à nos critères d’inclusion. Au total 2.573 cas de VIH positifs ont été identifiés parmi les donneurs volontaires de sang au cours des 10 dernières années; la prévalence poolée étant de 13,22/100.000, avec un intervalle de 0,74 à 125,97 pour 100.000 habitants. Dans les 24 provinces qui ont rapporté des données annuelles, la prévalence du VIH est passée de 4,9/100.000 à 125,97/100.000. Le ratio homme/femme était de 2,8; environ 60% d’eux étaient âgés de moins de 30 ans.

Conclusions:  La prévalence du VIH chez les donneurs volontaires de sang a augmenté de façon constante et rapide. Des mesures efficaces devraient être prises d’urgence pour prévenir le recours au test VIH dans les programmes de donneurs de sang, pour promouvoir le don de sang volontaire chez les groupes à faible risque et pour appliquer strictement la loi sur le don de sang.

Abstract

Objetivo:  En 1998 se promulgó la Ley de Donación de Sangre en China, como una respuesta al brote de VIH, endémico en el área central y rural de la China por una recolección de sangre comercial y poco higiénica. Este estudio busca realizar una primera revisión integral de los datos disponibles sobre la prevalencia, la tendencia y otras características epidemiológicas de la infección por VIH entre donantes voluntarios desde 1998.

Métodos:  Se realizó una búsqueda en bases de datos inglesas y chinas y se construyó una revisión sistemática. Se calcularon las tasas de infección agrupadas por provincia y año, utilizando modelos de efectos aleatorios o de efectos fijos.

Resultados:  87 estudios cumplían los criterios de inclusión. En total se identificaron 2,573 casos VIH positivos entre donantes voluntarios durante los últimos 10 años; la prevalencia acumulada era de 13.22/100,000, con un rango de 0.74 a 125.97 por 100,000. Entre las 24 provincias que reportaron datos tempranos, la prevalencia del VIH aumentó de 4.9/100,000 a 125.97/100,000. La razón hombre/mujer era de 2.8; aproximadamente un 60% eran menores de 30 años.

Conclusiones:  La prevalencia del VIH entre donantes voluntarios de sangre ha aumentado de forma constante y rápida. Es urgente tomar medidas eficientes para prevenir la búsqueda del testaje para VIH mediante los programas de donación de sangre, para promover la donación voluntaria de sangre entre grupos de bajo riesgo, y para cumplir estrictamente la Ley de Donación de Sangre.

Introduction

As a result of traditional Chinese beliefs on the detrimental effects associated with blood loss, recruitment of voluntary blood donors had always been difficult in China (Tison et al. 2007). In the 1980s to early 1990s, a huge shortage of blood donation had created a big market for commercial blood collection, which targeted the poorest farmers in rural central China. From the 1980s to mid-1990s, governmental and commercial blood stations/centres pooled and centrifuged the blood, retained the plasma and reinfused donors with red blood cells from the pool, which enabled people to sell their plasma more frequently without developing anaemia. Donors were paid $5 for each plasma collection. But this practice turned out to be an effective way of transmitting HIV and other blood-borne diseases (Cohen 2004) which may have infected hundreds of thousands of people in seven provinces (Grusky et al. 2002). The catastrophic outbreak of HIV/AIDS in poor rural China drew wide international media attention and forced the government to rethink its blood collection practices (Gill et al. 2002). In 1998, with a commitment to control infectious diseases and restructure blood donation and collection practices, the Chinese government passed the first Blood Donation Law, signalling that blood donation has since entered an era of scientific management and quality control that adhered to international protocol of blood bank (Wu et al. 2004; Qian et al. 2005).

The Blood Donation Law bans all paid or commercial blood donation or collection. In the first 7 years since the enforcement of the law, because of the persistence of traditional belief and distrust of blood donation system in the Chinese public, the need for voluntary blood donation still could not be met (Shan et al. 2002). The responsibility of donor recruitment was placed on local governments, who were often assigned with donation quotas. The local governments subsequently assigned the quotas to ‘work units’ such as factories and universities. Because of the general reluctance to donate blood, employers often pressured employees into donation, using paid leave or monetary or spiritual incentives (Shan et al. 2002). Although not truly voluntary, such practice was fundamentally different from commercial blood collection and greatly reduced the HIV infection in blood transfusion. Thanks to effective public campaigns (Adams et al. 2009), such ‘employer-organized blood donation’ has been gradually replaced by individual voluntary blood donation. From 2000 to 2005, the proportion of individual voluntary blood donation in total blood supply increased from 21% to 84%. In 2005, the government stopped employer-organised blood donation (Gao, 2010). In addition, to prioritise blood donation from voluntary non-remunerated donors, policies and regulations are enforced to screen donated blood, promote rational use of blood and implement effective quality control systems (WHO 2002, 2008, 2011a, 2011b). According to the national policies, HIV test is performed using two different quality assays, positive on either assay would lead to discarding the blood. The positive sample is then sent to local Center for Disease Control and Prevention (CDC) confirmation using Western blot (WB) test; local CDC will also contact the positive donors for counseling and treatment. Such testing procedure has been in effect since 1997 (China Ministry of Health 1997).

The past decade witnessed not only the structuring of blood bank safety, but also a rapid growth of HIV/AIDS epidemic in China. The number of HIV infections has grown 30-fold since 1985; the current official estimate of people living with HIV/AIDS reached 740 000 (Wang et al. 2009). Since the late 1990s, the Chinese government has gradually shifted from denial to a more pragmatic attitude towards HIV/AIDS epidemic (Wu et al. 2004, 2007). In 1998, the Principles for STI/HIV Education and Prevention Messages was issued jointly by the Ministry of Health and eight other ministries, acknowledging the growing epidemic (Wu et al. 2004). Meanwhile, drug use and commercial sex remained two of the ‘social evils’ that were subject to regular ‘strike hard’ policies. With increasing pressure from the growing HIV epidemic and criticism from the international community, since 2001, a series of harm reduction programmes was initiated and gradually scaled up (Wu et al. 2007, 2011; Hong & Li 2009). However, stigma against people living with HIV/AIDS and people at risk of HIV such as men having sex with men and female sex workers is still widespread, even among health-care providers (Li et al. 2007; Hong & Li 2009). Prevalence of self-initiated HIV testing is very low, although voluntary counselling and testing (VCT) is available in most cities and is free (Hesketh et al. 2005; Wu et al. 2006). Most HIV positives are unaware of their status and most people shun HIV-related communication (Wang et al. 2009). Although the HIV epidemic in China is still classified as ‘concentrated’ (because its prevalence has not reached 1%) (UNAIDS 2011), the rapid increase of sexual transmission of HIV in recent years signalled a potential spread of the virus to the general population (Gill & Okie 2007).

Hence it is important to examine the prevalence of HIV infections in voluntary blood donors in China. Such data would be critical to evaluate the effectiveness of enforcement of the Blood Donation Law and to make recommendations on policies and practices to further improve blood bank safety in the most populated country.

Methods

Literature search

Major English and Chinese electronic databases were searched: PubMed, Wanfang, China National Knowledge Infrastructure (CNKI) and VIP (WeiPu). The keywords used for searching the databases were blood donation, blood donors, HIV, Chinese and China. To maximise outputs, each keyword was searched individually or in combination. We conducted the literature review process from two directions: firstly, an inductive electronic search using keywords for all potential articles; secondly, a deductive approach through searching and retrieving articles from reference sections of identified publications as well as review articles related to blood donation in China.

Selection criteria and data abstraction

Inclusion criteria for identified articles were as follows: (i) studies were conducted among voluntary blood donors; (ii) studies were conducted after 1999 (after implementation of Blood Donation Law); (iii) studies reported HIV infection rates in voluntary blood donors (including individual voluntary donors and employer-organised voluntary donors); (iv) HIV infection was confirmed by WB (the protocol issued by China CDC) and (v) studies were conducted in mainland China.

Exclusion criteria were as follows: (i) study samples included commercial donors; (ii) studies did not report actual HIV infection rates, i.e. those not reporting the number of blood donors (so that the rate could not be calculated) and those employing mathematical modelling only; (iii) HIV infection was not confirmed by WB and (iv) studies conducted outside mainland China such as Hong Kong and Taiwan. The literature search and selection process are depicted in Figure 1.

Figure 1.

 Flow chart of systematic literature search.

Statistical analysis

Firstly, we extracted data on HIV prevalence in each province and each year and entered the longitudinal data into Stata. Secondly, we transformed the raw data into sin data to stabilise the variance following the steps detailed by Freeman and Tukey (1950). Thirdly, we tested the heterogeneity between studies using Q test. If heterogeneity existed (when P < 0.05), random-effect models were used for meta-analysis; if heterogeneity was not significant, fixed-effect models were used for meta analysis (Deeks et al. 2001). Such approaches of variance stability and heterogeneity testing were used in other meta-analyses of prevalence studies (Higgins & Thompson 2002; Gao et al. 2011). Then, ‘Metan’ functions in Stata were used for the analysis (Sterne et al. 2001; Harris et al. 2008). After the meta-analysis, we transformed the summary estimates and 95% confidence intervals (Freeman & Tukey 1950). Such steps were taken for meta-analysis in each year and each province; because of data limitation, simple calculation was performed for the demographic characteristics of the HIV-positive donors. All data manipulation and statistical analyses were performed using Stata 10.1; ArcGIS 9.3 was used for mapping the provincial prevalence.

Results

General information of identified studies

Of 87 studies that reported HIV infection rates in voluntary blood donors, all reported data from 2000 to 2009. Most data were retrieved from studies conducted in southwest China and northwest China; some data were from North China and Central China; no data were available from Heilongjiang and Shaanxi Provinces. Only 11 provinces provided consecutive yearly data from 2000 to 2009; the rest had data only for certain years; some only reported cumulative prevalence for several years, for example Shanghai, Ningxia and Jilin each had one report for several years of HIV prevalence. Among the studies that did report yearly data, none provided provincial data; in other words, most studies included in the current review were municipal- or city-level data. Data from several cities were pooled to calculate the prevalence in the province. Similarly, yearly prevalence was calculated from the available yearly city-level data.

Provincial HIV prevalence and overall trend

As depicted in Table 1, of the 29 provinces that reported any data on HIV prevalence from 2000 to 2009, the pooled HIV prevalence was 13.22/100 000 (12.10, 14.40), representing 2573 confirmed HIV-positive cases. Huge regional differences existed in the HIV prevalence, with a range of 0.74/100 000 (Ningxia) to 125.97/100 000 (Yunnan). Most of the provinces (23 of 29) had a prevalence lower than the national average, and the remaining six provinces had infection rates higher than the national average: Yunnan (125.97/100 000), Guangxi (32.40/100 000), Guizhou (19.60/100 000), Xinjiang (44.09/100 000), Chongqing (18.22/100 000) and Tibet (24.02/100 000). All these six provinces are located in Southwest and Northwest. Jiangxi (12.10/100 000), Tianjin (13.22/100 000) and Sichuan (12.10/100 000), and Henan (13.22/100 000) reported prevalence rates close to the national average: the rates have been increasing rapidly in recent years. Figure 2 depicts the provincial heterogeneity in HIV prevalence.

Table 1.   Prevalence of HIV by province from 2000 to 2009
RegionProvinceNo. of studies*Years of data availableNo. of HIV infection†Prevalence per 100 000 (95% CI)Hetero-geneity (X2) P-value I 2‡ (%)
  1. *Number of studies refers to the number of times the assessment was taken and reported. For example if the HIV prevalence was reported in each year from 2000 to 2008, the number of studies was 9; if the cumulative HIV prevalence from 2000 to 2008 was reported once, the number of studies was 1. The number of studies is also the degree of freedom (plus 1) in calculation of heterogeneity.

  2. †The number of HIV infection refers to the total absolute HIV infection cases reported in the study.

  3. ‡I square (I2) refers the proportion of variance explained by heterogeneity.

East ChinaAnhui102000–200932.50 (0.09,5.62)40.9110.00
Fujian52004–200834.22 (1.22,10.00)3.880.420.00
Jiangsu192003–2008263.60 (2.02, 5.62)30.650.03241.30
Shandong182000–2009313.02 (1.60, 4.90)46.06063.00
Zhejiang212000–20091305.62 (4.22, 8.10)89.03077.50
Jiangxi32005–2007612.10 (4.90, 22.50)10140.560.00
Shanghai12000–2008984.24 (4.24, 4.24)
North ChinaBeijing102000–20081759.02 (7.22, 12.10)24.750.00267.70
Shanxi92000–2009274.90 (3.60, 7.22)15.570.07642.20
Hebei12000–200831.77 (1.76, 1.78)
Inner Mongolia152004–200894.90 (2.50, 9.02)11.740.6280.00
Tianjin42004–2007613.22 (4.90, 25.60)3.880.27522.70
South ChinaGuangdong622000–200516610.00 (8.10, 12.10)103.960.00141.30
Guangxi402000–200941632.40 (24.02, 39.99)236.81085.20
Hainan102000–2009274.90 (2.02, 9.02)24.640.00363.50
Central ChinaHenan102002–20097013.22 (10.00, 15.62)5.910.7490.00
Hubei282000–2009384.22 (3.02, 5.62)33.110.19418.40
Hunan72005–200831.22 (0.22, 3.60)11.140.08446.10
NorthwestGansu102000–2009224.90 (1.60, 9.02)24.770.00363.70
Ningxia12002–200810.74 (0.74, 0.74)0.74
Qinghai12000–200415.09 (5.07, 5.11)5.09
Xinjiang62002–20032744.09 (7.22, 112.18)36.15086.20
ShaanxiNo data
SouthwestGuizhou262000–20097919.60 (13.22, 27.22)59057.60
Sichuan82001–200727012.10 (8.10, 16.90)62.18088.90
Yunnan342000–2009636125.97 (102.37, 155.94)241.19086.30
Chongqing102000–200921218.22 (5.62, 36.10)215.69095.80
Xizang/Tibet52005–2009624.02 (0.40, 87.00)13.260.0169.80
NortheastLiaoning302000–2009806.40 (4.22, 9.02)80.78064.10
Jilin12006–200726.26 (6.26, 6.26)
Hei-long-jiangNo data   No data
Total 4052000–2009257313.22 (12.10, 14.40)3693.08 88.90
Figure 2.

 Pooled HIV prevalence by province from 2000 to 2009.

Of the 29 provinces that reported any HIV prevalence data, five did not report any yearly data; Table 2 presents the yearly trend of HIV prevalence in 24 provinces. We found that the number of HIV infection cases identified in the voluntary blood donors increased from 73 in 2000 to 400 in 2005, and then dropped to 250 in 2009; but the prevalence remained high. The overall yearly prevalence has been increasing, from 5.62/100 000 in 2000 to 28.90/100 000 in 2009. It should be noted that the total number of cases fell (perhaps because the number of studies shrank), although the prevalence continued to rise, as shown in Figure 3. Taking a closer look at the yearly trend of the six provinces with highest HIV prevalence (Table 3) reveals that the prevalence in Yunnan peaked in 2005 and then slowly declined and maintained at the level of 130 s per 100 000 since 2007. In Guangxi, it peaked in 2003 and then slowly declined and maintained at the level of 40 s per 100 000 since 2006. In Guizhou and Chongqing, the prevalence has been increasing steadily from 2000 to 2009 and reached 53 and 73 per 100 000 respectively. Because of data limitation, the trends in Xinjiang and Tibet are difficult to describe.

Table 2.   Yearly HIV prevalence in voluntary blood donors in 24 provinces*
YearNo. of studiesNo of HIV infectionsPrevalence (per 100 000) (95% CI)Heterogeneity (X2) P I 2 (%)
  1. *A total of 24 provinces were included in the analysis of yearly trend: Of 31 provinces in China, no data were available from Shaanxi and Hei-long-jiang; only one-time data were reported from Shanghai, Hebei, Ningxia, Qinghai and Jilin. The total number of HIV infection cases in 24 provinces was 2468.

200033735.62 (3.60, 8.10)83.25061.6
200134965.62 (3.60, 9.02)142.18076.8
2002421779.02 (5.62, 12.10)248.80083.5
2003431789.02 (6.40, 12.10)220.88081.0
20044731314.40 (10.0, 21.02)484.07090.5
20055140015.62 (10.0, 21.02)713.66093.0
20065139119.60 (13.22, 25.60)459.73089.1
20074632816.90 (12.10, 22.50)321.47086.0
20083726218.22 (13.22, 25.60)364.57090.1
20092125028.90 (19.60, 42.02)179.31088.8
Figure 3.

 Trend of HIV infection rates in 24 provinces from 2000 to 2009. See Table 2 for confidence intervals for yearly prevalence.

Table 3.   Yearly HIV prevalence in voluntary blood donors in six provinces*,
YearYunnanXinjiangGuangxiTibetGuizhouChongqing
  1. *These six provinces reported a prevalence higher than the national average.

  2. †All numbers are prevalence by 100 000.

2000 28.90 55.21 5.62 1.60 3.02
2001 65.01140.5613.22 1.22 0.40
2002 60.01140.5616.9022.50 3.60
2003 75.61140.5689.9713.22 4.90
2004168.0142.02 18.2230.62 6.40
2005206.8828.90172.1325.6024.02
2006176.30  3.0242.02  5.6232.4024.02
2007129.54  2.5042.02  6.4022.5048.39
2008133.17 42.0252.89  5.6214.4055.21
2009136.8444.0952.8972.88

Demographic characteristics and transmission routes

Only a small number of studies reported the demographic characteristics of the HIV-positive blood donors. Table 4, pooled from available data, revealed that the male–to-female ratio among the HIV positives was 2.79; close to 60% of HIV positives were below the age of 30 years, and sexual transmission comprised 85% of HIV transmission, including 50% of homosexual transmission. From the seven studies that examined the mode of blood donation, most HIV-positive cases (78%) were individual voluntary donors; 22% were employer-organised voluntary donors. These studies also reported that some blood donors repeatedly returned for blood donation even after being confirmed HIV positive.

Table 4.   Demographic characteristics and transmission modes among HIV-positive voluntary blood donors
Gender*Age†Transmission mode‡Donation type§
Male
N (%)
Female
N (%)
18–30
N (%)
31–40
N (%)
41–55
N (%)
IDU
N (%)
Sexual
N (%)
Other
N (%)
Individual-voluntary
N (%)
Employer-organised
N (%)
HomoHeteroUn-reported
  1. IDU, intravenous drug use.

  2. *Pooled from 49 studies that report gender of HIV-positive blood donors.

  3. †Pooled from 39 studies that report age of HIV-positive blood donors.

  4. ‡Pooled from 16 studies that report transmission modes of HIV-positive blood donors. Of the 16 studies, only five differentiated homosexual transmission from heterosexual transmission; the remaining 11 only differentiated sexual transmission from IDU.

  5. §Pooled from seven studies that report donation type.

1391 (73.6)498 (26.4)969 (59.6)461 (28.4)195 (12.0)40 (13.3)66 (21.9)48 (16.0)143 (47.5)4 (1.3)90 (73.8)32 (26.2)

Discussion

Human Immunodeficiency Virus infection is one of the major transfusion transmission infections (WHO 2002). Reports of HIV transmission through blood transfusion have mostly been published about developing countries (Volkow & Del Rio 2005; Stokx et al. 2011). The prevalence of HIV infection among voluntary blood donors in China is 13.22/100 000, much lower than that in African countries, but higher than that in some developed countries (Nanu et al. 1997; Glynn et al. 2000; Heyns et al. 2006).

Several limitations should be addressed before we interpret the findings. Firstly, data reporting of HIV infection varied considerably across provinces. Some provinces reported yearly data from 2000 to 2009 with city-level data and province-level data; but most provinces only had inconsecutive data from some years and some cities. Although statistical procedures were adjusted for heterogeneity, biases exist in the pooled provincial prevalence or yearly prevalence. As a result of uneven data reporting from each province, especially limited data or no data from some provinces, the pooled prevalence reported in this study could not be viewed as true HIV prevalence in voluntary blood donors in China. Secondly, all studies were observational and blood donors were not randomly chosen, which might lead to selection bias. Thirdly, most studies included in this review were written in Chinese (some with English abstracts and available in PubMed), which makes it difficult for non-Chinese readers to trace back to the original materials. And finally, our ability to assess study quality was limited by the fact that many studies did not offer detailed information of selected subjects or valid data on important factors such as demographic characteristics and donation types. Despite these limitations, our study is the first meta-analysis to report HIV prevalence in voluntary blood donors in China.

Since the enforcement of the Blood Donation Law in 1998, blood donation and collection practices in China have undergone substantial changes. As commercial or paid blood donors were replaced by voluntary blood donors, a sharp decline was observed in HCV prevalence (Gao et al. 2011). By contrast, our review showed a rapid increase in HIV prevalence among the blood donors. This increase might be attributed to the increasing HIV/AIDS epidemic in China as well as the ever more stringent blood quality control policies and practices. The considerable regional differences in HIV prevalence in blood donors we observed reflect the wide heterogeneity of the HIV epidemic in the country. For example Yunnan, Guangxi and Xinjiang are the three provinces with highest general HIV prevalence (Wang et al. 2009). But Tibet, Guizhou and Chongqing, which do not have a high general HIV prevalence, did report high HIV prevalence rates in blood donors. Such discrepancies might be a result of the imbalanced and potentially biased data reported in existing studies, or to the especially high rates of HIV prevalence in blood donors in these provinces.

The demographic profiles of HIV-positive donors were similar to the general profile of HIV-positive individuals in the nation, with most infected among the young population under the age of 30 years. Of the HIV-positive donors, 85% were infected through sexual transmission, compared to about 50% in total HIV infection cases. This discrepancy suggests that many people with high-risk sexual behaviours were unaware of their HIV-positive status, or that some HIV-positive individuals who were infected through sexual contacts sought HIV testing through blood donation. Literature from other countries suggests that as many as 10% blood donors were HIV test-seeking individuals (Lau et al. 2002; Goncalez et al. 2010). Recent studies in China found that some people repeatedly sought HIV testing through blood donation to confirm their HIV status after high-risk sexual behaviours. Such test-seeking behaviours would increase the cost of blood collection and testing and the risk of blood contamination, and deserves more attention. The fact that most HIV-positive donors were among the young population with high-risk sexual behaviours suggests the urgent need for targeted intervention.

Several policy and intervention recommendations can be derived from our findings. Firstly, data reporting of HIV incidence and prevalence among blood donors needs to be strengthened. The rising trend of HIV prevalence in voluntary blood donors suggests potential risks, but more data are needed to monitor the epidemic and to design appropriate testing and screening practices. We also need data on demographic characteristics, transmission routes and donation types to better identify the at-risk population and design relevant risk-reduction strategies to safeguard our blood bank.

Secondly, culturally appropriate public campaigns are needed to recruit and maintain a population of low-risk voluntary blood donors. The HIV epidemic in China is still concentrated in high-risk populations, although the rapid increase of heterosexual transmission in recent years suggests the potential risk of spreading to the general population (Gill & Okie 2007). At this critical moment of epidemiologic transition, blood bank safety might be endangered if high-risk groups are not excluded from the blood donors. In some cities, the nucleic acid testing (NAT) has been implemented to identify HIV-positive cases in a shorter time window (El Ekiaby et al. 2010). But NAT cannot eliminate the window period; it is also costly and requires high-level technical and staff support which preclude its scale-up to all blood stations in China, given the prevailing limited financial and personnel resources (Liu et al. 2010). Instead, we propose culturally appropriate public campaigns to recruit low-risk voluntary blood donors and effective pre-donation behavioural screenings to reduce the number of at-risk blood donors.

Thirdly, HIV test-seeking behaviours deserve more attention. Such behaviours may be a result of the stigma associated with HIV and HIV testing; people might perceive HIV testing through blood donation as less stigmatizing than HIV testing in VCT clinics, STD clinics or primary care clinics. We propose campaigns to reduce stigma associated HIV testing so that at-risk individuals could seek appropriate testing and counselling at VCT clinics. Effective campaigns are also needed to educate the public about the devastating effect of a contaminated blood bank and to discourage HIV test-seeking through blood donation programmes.

Conclusions

This meta-analysis provides a relatively comprehensive profile on the prevalence and trend of HIV prevalence in voluntary blood donors since the enforcement of Blood Donation Law in China. The increasing prevalence suggests a potential danger in blood safety; the high heterogeneity across regions and provinces underscore the need for culturally appropriate interventions to reduce risks, especially to exclude individuals at risk of HIV infection from blood donation.

Acknowledgement

We appreciate statistical advice from Drs. Hongwei Zhao and Luohua Jiang. We also thank Chia-Yuan Yu for assistance in GIS mapping.

*Articles included in the meta-analysis but are not cited in the main text.

Ancillary