Chapter 8

How much HIV infection comes from blood exposures?


What proportion of HIV infections in countries with generalized epidemics comes from blood exposures (except IDU and blood transfusions)? When infection control in health care is not routine and reliable, the answer to this question is both important and not immediately obvious.


Two preliminary issues


HIV survival outside the body


Laboratory studies through 1988 showed that HIV survives outside the body at room temperature for a few hours to a few days when dry, and for several weeks when wet (see Chapter 5). Several studies after 1988 show that 10 percent of an initial deposit of HIV can survive in dry conditions – such as on a glass slide – for several hours to more than a day.[i] Research published in 1999 showed that HIV can survive for weeks at room temperature in wet conditions, such as in a used syringe or needle.[ii] Wiping or rinsing does not reliably remove or kill (inactivate) HIV. Except in carefully controlled conditions, even soaking contaminated instruments in bleach or alcohol does not reliably kill HIV, because fluids might not penetrate clots or small spaces in syringes and needles, solutions might weaken over time, and presence of organic matter interferes with bleach.[iii] Heating to boiling reliably kills HIV. Soaking in glutaraldehyde or other special chemicals is also effective.

Many public health experts understate HIV survival. For example, the British Medical Journal in 2007 quoted a ‘senior research scientist’ at Harvard’s School of Public Health to say, ‘The HIV virus is extremely fragile, dying easily and quickly once exposed to air.’[iv] Similarly, web-based training materials for nurses posted by Johns Hopkins aver that ‘HIV cannot live outside body fluids more than a few seconds,’ and that ‘HIV can live between 30 seconds to one minute when exposed to the air.’[v] These statements are not only dead wrong, they are deadly.


Risk to transmit HIV through trace amounts of blood              


In a 1991 article, senior officials in WHO’s Global Programme on AIDS estimated that the risk to transmit HIV from an HIV-positive patient to a subsequent patient through reused ‘equipment/needles’ was less than 0.5 percent.[vi] Because this is an important estimate, it is useful to look at relevant evidence.

A 1991 review of risks to transmit HIV through health care averred that needlestick injuries to healthcare workers ‘give an idea of the risk of transmission from non-intravenous injection of blood and suggest that this is not an efficient mode of transmission.’[vii] Through 1996, evidence compiled from multiple studies that followed healthcare workers exposed to HIV through needlestick accidents showed that only 20 (0.3 percent) of 6,202 acquired HIV infections.[viii]

However, the parallel does not fit. In the mid-1990s, CDC looked at what was different about needlestick accidents that led to HIV infections vs. those that did not. Not surprisingly, deep needlesticks – deep enough for the hole of the needle to be within the skin, as in an injection – were much more dangerous than shallow scratches. And only 7 percent of needlestick accidents were deep. With data from this study, the average risk for a healthcare worker to acquire HIV infection from a deep needlestick can be estimated at 2.3 percent.[ix]

Thus, based on information from needlestick accidents, a good first estimate of the risk to transmit HIV from patient to patient through syringes and needles reused without sterilization is 2.3 percent. However, many other factors – presence of visible blood, whether the needle came from a vein, the size of the needle – raise or lower this risk. If HIV contaminates multidose vials or rinsing pans, HIV may spread from one patient to infect more than one subsequent patient. Rinsing or washing syringes or needles without sterilization may lower the risk.

Other estimates of the risk to transmit HIV through trace amounts of blood come from studies among IDUs. Two studies in New Haven, US, and in Bangkok, Thailand, used information from IDUs to estimate the risk to transmit HIV through a single intravenous injection with shared and contaminated equipment. These estimates are 0.67 percent and 0.8 percent in the US and Thailand, respectively.[x] However, both studies rejected IDUs’ statements about how often they shared injection equipment, and assumed instead that many if not most shared more often than they reported. The researchers also assumed that all sharing was random, with an ‘average’ IDU. If the researchers had used reported rather than assumed rates of sharing, and had considered that IDUs who share often do so with regular partners, their estimated risks to transmit HIV through injections would have been much higher.

The best evidence on risks to pass HIV from patient to patient through invasive healthcare procedures comes from hospital-based outbreaks where that has happened on a large scale. What one needs to know to calculate the risk is the average numbers of invasive procedures (and details about the procedures) administered to HIV-positive persons that result in one HIV transmission to another patient. Unfortunately, no one has reported sufficiently detailed information from any large nosocomial outbreak to calculate the risk. However, there is enough information to make some rough estimates.

For example, in southern Russia in 1988-89, invasive procedures at 13 hospitals passed HIV from one child directly and indirectly (through other children) to more than 260 children over 15 months (Chapter 9 provides more information on this outbreak).[xi] Most transmissions during this outbreak came from inpatient children who themselves had been infected not more than six weeks earlier.[xii] If we assume that the risk to transmit HIV through healthcare was constant throughout the outbreak, and that each infected child infected an average of 1-2 children over 1-2 months, this would produce the observed nosocomial infections.

With this model, how many procedures did HIV-positive children have over one month to infect one other child? No one has reported this information, but a rough estimate is possible. If HIV-positive children had an average of 10-50 procedures per month, after which instruments were subsequently reused without sterilization, the calculated average risk to transmit HIV per procedure would be 2 percent to 10 percent. Similar rough estimates can be derived from several other documented nosocomial outbreaks.[xiii]


What proportion of HIV infections comes from blood exposures?       


Some people propose that blood exposures (except transfusions and IDU) account for as little as 0.01 percent of HIV infections in the world, while others have estimated that blood exposures account for more than 50 percent of infections in generalized epidemics. One way to approach the debate is to distinguish between weaker and stronger evidence, and to consider what additional evidence could be collected to settle the matter.


Estimates based on no evidence and/or indirect evidence      


As discussed in Chapter 5, experts consulted by WHO in 1988 estimated that ‘inadequately sterilized skin-piercing instruments (health sector and outside)’ accounted for 1.6 percent of HIV infections in Africa.[xiv] The paper that reports this estimate provides no calculations or argument linking it to any evidence. In 1991, senior officials in WHO’s Global Programme on AIDS proposed that ‘equipment/needles’ accounted for less than 0.1 percent of HIV infections in Africa, again without presenting any supporting evidence.[xv]

A 1991 review acknowledged that ‘the proportion of HIV infection transmitted parenterally [though skin-piercing procedures] has not been well studied,’ but estimated nonetheless that medical injections accounted for not more than ‘occasional transmission of HIV’ in Africa. To reach this conclusion, the review relied heavily on indirect evidence, arguing that ‘injections in the 5-14-year age group are very common, yet cases of HIV infection and AIDS in this age group are relatively rare.’[xvi] This common argument not only discounts studies that report appreciable numbers of unexplained HIV infections in children, but also relies on the invalid assumption that adults’ risks for nosocomial HIV infection would be not much greater than children’s risks (see Chapter 5).


Estimates from models using numbers of unsafe injections


In 1999 and 2004, two WHO teams published estimates of the number of HIV infections from medical injections, calculated from survey-based estimates of the number of unsafe injections. According to these estimates, 2-5 percent of new HIV infections in the world were from unsafe injections.

To see how these estimates are derived, consider the calculations that produced the first such estimate in 1999.[xvii] (This explanation follows the logic of the model, but simplifies the calculations.) Briefly, the WHO team estimated that 20 million Africans were HIV-positive. They also estimated (from surveys in several African countries) that Africans received 1-2 injections per year, and they assumed that HIV-positive Africans received the same number of injections as did other Africans. Hence, each year, doctors and others administered 20-40 million injections to HIV-positive patients. Next, from surveys of injection practices, the modelers estimated that doctors reused syringes without sterilization after 50 percent of all injections, and thereby after 10-20 million injections administered to HIV-positive patients.

But here comes the crucial assumption: the WHO team assumed that only 0.5 percent of patients who received injections with HIV-contaminated syringes or needles (that is, with equipment reused without sterilization after injecting an HIV-positive patient) contracted HIV infections. Thus, the model calculated that 0.5 percent of 10-20 million Africans – approximately 50,000-100,000 people – contracted HIV each year from injections. If the WHO team had used a higher rate of HIV transmission through reused equipment – such as the 2.3 percent average risk of transmission to healthcare workers through deep needlestick accidents – their calculations would have ‘found’ many more HIV infections.

As of 2006, UNAIDS[xviii] accepted the estimate from WHO’s later model-based calculation that medical injections account for 5 percent of HIV infections in the world.[xix] These estimates focus on injections. Other blood exposures taken together – drawing blood, tattooing, minor surgeries, etc – might be more common and more damaging than injections. But even with good information about the numbers of unsafe procedures, model-based estimates are unreliable. Their weakness is that they depend on an assumed – not an observed – rate of HIV transmission per event.


Estimates from studies of risks for incident HIV infection      


Compared to models, studies of risks for incident HIV infection – studies that follow HIV-negative people, retest them to see who becomes HIV-positive, and ask about sex and blood exposures – are able to provide more reliable estimates of the proportion of HIV infections coming from healthcare.

In 2001, UNAIDS commissioned Nicole Seguy to review studies of risks for HIV infection that asked about medical injections. She found four studies in Africa that followed and tested people in the general population, and asked about injections as risks for incident infections. Her September 2002 draft concluded that these studies showed that ‘contaminated injections may cause between 12% and 33% of new HIV infections’ in Africa.[xx] UNAIDS did not publish the paper.

During 2004-08, other studies that tested and followed people in the general population (such as villagers or women at antental clinics, but not sex workers or patients) reported additional information on injections as risks for incident HIV infection in Africa. Taking all of these studies together, 13 of 14 results show that men and/or women who reported injections were more likely to show up with a new HIV infection as compared to those who reported no injections (Table 8.1). Reporting an injection increased their risk to acquire an HIV infection by as much as 8.5 times, with a median increase of 1.5 times. Using standard analyses to link HIV infections to injections (see notes to Table 8.1), these studies attribute as much as 54 percent (median 19 percent) of new HIV infections to injections. (These analyses measure the ‘extra’ infections in people who report injections vs. people who do not, but do not prove that injections caused these infections.)

A handful of studies have examined other blood exposures as risks for incident HIV infections. For example, in Pune, India, during 1993-2000 new infections were more common in people who had received a tattoo.[xxi] In addition, some studies of risks for prevalent HIV infection find higher prevalence in people who report blood exposures that would not be related to HIV treatment (so there is no reason to suspect reverse causation). For example, in several African countries where men are circumcised during puberty or later, circumcised vs. uncircumcised adolescent men were more likely to be HIV-positive, and this was true both for virgin men and for non-virgins.)[xxii] And a study in Zanzibar, Tanzania, found that HIV prevalence was 1.9 times greater in those who reported having had an operation vs. those who reported no operations.[xxiii]


Table 8.1: Estimates of the proportion of incident HIV infections from injections

Country, year of study


Relative risk for HIV in persons reporting any vs no injections

% of HIV infections “attributable” to injections*

Studies included in Seguy’s review

    DRC, Kinshasa, 1984-86[xxiv]

Healthcare workers



    Uganda, Rakai, 1989-90[xxv]




    Rwanda, Butare, 1989-93[xxvi]




    Uganda, Masaka, 1990-97[xxvii]







Other results

    Rwanda, Butare, 1989-93[xxviii]




    Uganda, Masaka, 1990-2005[xxix]




    Tanzania, Mwanza, 1991-94[xxx]







    Uganda, Rakai, 1997-99[xxxi]




    Zimbabwe, Manicaland,








    Malawi, Blantyre, 2003-05[xxxiii]




    Uganda, 2004-05[xxxiv]




* This column reports the crude population attributable fractions (PAFs) of incident infections associated with injections. These PAFs are calculated from the ratio of the rates (RR) of HIV incidence in persons who reported vs. persons who did not report injections, and the proportion (ρ) of people with incident infections who reported injections, using the formula: PAF = ρ(RR-1)/RR. For more information about PAFs see standard epidemiology texts.

† This result appears to be based on a subset of adults selected from a larger study.[xxxv]

‡ This result is based on reported hormone injections for birth control in the study clinic.


Studies not done         


The best way to determine the proportion of HIV infections from various risks is to trace infections to their source. In countries with generalized epidemics, some studies have traced a minority of infections to spouses. But only rarely has an infection been traced to any other source.


Initial vs. total impact on HIV epidemic expansion


All the above estimates measure only the direct contribution of blood exposures to HIV epidemics, which is less than the total – direct plus indirect – contribution. Suppose, for example, that 10 percent of people with new HIV infections in Zambia during 1989-2009 acquired their infections directly from blood exposures. Many of these 10 percent would subsequently transmit HIV to others through sex and/or blood, and these others would in turn transmit through multiple modes, and so on. Thus, the direct plus indirect contribution of HIV transmission through blood exposures during 1989-2009 on the number of infections in Zambia in 2009 would be far greater than 10 percent.

Moreover, HIV transmission through blood exposures accelerates sexual transmission by allowing HIV to jump from one sexual network to others. When husband and wives are both HIV-positive and have no other sex partners, their HIV has no chance to reach anyone through sex. But if, for example, a husband transmits HIV to someone through a dental clinic, the HIV can get into another sexual network, where sexual transmission again becomes possible.


WHO, UNAIDS, and allies repeat low estimates        


Seguy’s unpublished paper along with several articles published in 2002-03[xxxvi] challenged WHO’s and UNAIDS’s lack of attention to HIV transmission through healthcare. In response, WHO and UNAIDS staff led a group of 15 authors in a review defending WHO’s low estimates of HIV infections from healthcare.[xxxvii]

Their review, published in 2004, did not mention Seguy’s unpublished draft. Moreover, at least some of the authors were aware of – and did not report – unpublished data from Masaka, Uganda, and from Mwanza, Tanzania, which linked more than 20 percent of incident HIV infections to injections. The Mwanza team did not report these data until 2006, and the Masaka team did not do so until 2007 (see Table 8.1). While not reporting conflicting evidence from studies of risks for incident HIV infection, the authors relied on indirect evidence – including low HIV prevalence in children in selected studies – to defend WHO’s low estimates of HIV infections from injections.

Many organizations involved in HIV prevention in generalized epidemics have continued to use unsupported low estimates for the proportion of HIV infection from blood exposures. For example, USAID, in a 2002 document, attributed only 0.01 percent of global HIV infections to healthcare (except blood transfusions).[xxxviii] And WHO’s 2002 World Health Report ignored WHO’s own model-based estimates to say, ‘Current estimates suggest that more than 99% of the HIV infections prevalent in Africa in 2001 are attributable to unsafe sex.’[xxxix]


Not only HIV: Other bloodborne viruses


Hepatitis B virus


In much of Africa and Asia, 70-95 percent of adults have been infected with hepatitis B virus at some time in their lives, and 8-20 percent have active (chronic or new) infections. Childhood acquisition of hepatitis B, which often leads to lifetime chronic infection, accounts for most active hepatitis B infections in Africa.

For several decades after tests for hepatitis B were available, no researcher looking at risks for hepatitis B infection in African children reported information on injections.[xl] The first study to do so tested children aged 6 months to 5 years from seven villages in Gambia in 1988, and then followed and retested them in 1990. The baseline survey in 1988 reported no association between hepatitis B infection and ‘number of injections for immunization.’[xli] The study, unfortunately, did not report the relevant data (which might have shown more immunizations for infected children, but with more than a 5 percent chance that the observation was a statistical accident). The study also did not report any information about curative injections.

The follow-up study found that 30 percent of susceptible children (setting aside children who had past or current infections in the baseline survey) had contracted hepatitis B infections over two years. The study reported a ‘slight excess of injections recorded by infected children: 0.82…versus 0.66…in uninfected children.’[xlii] Assuming that each child had received 0 or 1 injection, and using standard analyses (as explained in notes to Table 8.1), these reported data are consistent with injections accounting for 2/5ths of new infections. The study also reported large unexplained differences in hepatitis B incidence between villages – with incidence over two years ranging from 21 percent to 54 percent. The study did not consider that such differences might have been due to differences in infection control practices on the part of the clinics and healthcare workers serving the different villages, and the study reported nothing about infection control practices.

With few exceptions, researchers in Africa continued to avoid, ignore, or deny evidence that healthcare is a risk for hepatitis B infection. They continued, at the same time, to come up with no explanation for high rates of infection in African children or adults. For example, the Gambian study mentioned in the previous paragraph found that ‘insecticide spraying of the child’s dwelling…[reduced] exposure to bedbugs but there was no effect on hepatitis B infection.’[xliii] A 1993 review of hepatitis B in Africa acknowledged, ‘The phenomenon of horizontal [child-to-child] transmission of HBV [hepatitis B virus] has never been adequately explained.’[xliv]

In contrast, researchers in Asia and Eastern Europe looked at healthcare risks for hepatitis B and were more successful in both explaining and curbing high rates of infection. For example, when researchers in China found that 21 percent of children aged 1 year had active hepatitis B infections, they noted that healthcare workers changed needles but reused syringes without sterilization. Healthcare managers responded with ‘stricter measures…to prevent HBV [hepatitis B virus] infection in these clinics, including training of medical personnel…and the use of sterilized syringes and needles for single injections.’[xlv] With these changes, the prevalence of active hepatitis B infection among children aged 1 year whose mothers were not infected (to exclude mother-to-child transmission) fell from 15 percent to 3.3 percent.

The same WHO teams that modeled the contribution of medical injections to HIV infections (see above) also modeled the contribution of injections to hepatitis B infections. According to the latest estimate from 2004, injections account for only 10 percent of hepatitis B infections in Africa.[xlvi] Considering that IDUs and MSMs contribute little to hepatitis B transmission in Africa – and certainly not among children – this estimate is very likely low.

Hepatitis B vaccine was introduced in 1981. Most countries have begun to vaccinate infants against hepatitis B. Preventing hepatitis B infection in infants and young children prevents most new chronic infections.


Hepatitis C virus        


WHO estimates much higher prevalence of hepatitis C virus infection in Africa than in developed countries.[xlvii] Although hepatitis C transmits almost exclusively through blood exposures, few studies in Africa have identified the blood exposures that transmit hepatitis C. A large majority of studies in Africa that have tested people for both hepatitis C and HIV find that HIV infection is more common in people with hepatitis C virus or antibodies.[xlviii] Because IDUs have been rare in Africa, this points to other blood exposures transmitting both HIV and hepatitis C.

Although many countries in Africa have relatively high prevalence of hepatitis C virus or antibodies, low prevalence in Southern Africa – lower even than in Western Europe or the US – presents a puzzle. This low prevalence has been presented as evidence that blood exposures are not common in Southern Africa, and therefore that HIV acquisition among adults in the region must be almost entirely through sex.[xlix]

However, much remains unknown about how to interpret negative tests for hepatitis C antibodies, and more specifically about how to do so in Africa. Recent studies suggest that many people who have been infected with the virus defeat it within several months, and do not subsequently carry antibodies. For example, 8 of 116 US health care workers exposed to hepatitis C through needlestick accidents became at least transiently infected – determined by finding the virus in their blood – but only 2 of 8 developed antibodies, and only 1 of 8 (1 of the 2 with antibodies) developed a chronic infection.[l] Also, some people defeat chronic infections over time, and some who do subsequently lose antibodies.

The ability to resist or to defeat hepatitis C may vary by age, virus type, and other factors. Notably, the hepatitis C virus type that circulates in Southern Africa – type 5 – is rare outside the region.[li] Furthermore, people infected with hepatitis B, which is common in Africa, are more likely to defeat hepatitis C.[lii] Thus, low prevalence of hepatitis C infection and antibodies in Southern Africa does not reliably show the proportion of people who have been exposed to hepatitis C or to other bloodborne viruses – and does not, therefore, show that sex accounts for most HIV infections in Southern Africa.


Ebola, Marburg, and Lassa virus outbreaks  


Ebola, Marburg, and Lassa viruses – all of which cause viral hemorrhagic fever – circulate over large areas of Africa among non-human hosts. All three occasionally infect humans. Some infections may be unnoticed or mild, while others lead to pain, fever, bleeding, and often death. Person-to-person transmission outside hospitals appears to be too inefficient – too infrequent – to sustain an epidemic.

In the last several decades, hospitals in Africa have from time to time spread rare infections of these bloodborne viruses in short but deadly outbreaks. The viruses appear to be most deadly when spread through blood, such as through syringes and needles reused for injections.[liii] After the first three documented Ebola outbreaks during 1976-79 (see Chapter 3), no one recognized any new infections in Africa for more than a decade. Ebola returned in 1994-96, causing several outbreaks in Gabon with a total of almost 100 deaths, and a large outbreak with more than 250 deaths in Kikwit, DRC.[liv] Another large outbreak in Uganda in 2000-01 caused several hundred deaths.[lv] During 2001-03, five Ebola outbreaks in Gabon and Congo led to a combined total of more than 260 deaths,[lvi] and other outbreaks emerged in south-central DRC and western Uganda in 2007.[lvii]

Two hospitals in Nigeria transmitted Lassa virus during an outbreak in 1995. According to an international team, ‘Compelling, indirect evidence revealed that parenteral [skin-piercing] drug rounds with sharing of syringes, conducted by minimally educated and supervised staff, fuelled the epidemic among patients.’[lviii]

The Marburg virus caused an outbreak of hemorrhagic fever in DRC in 1998-2000, and another large outbreak in Angola with more than 300 deaths in 2004-05. More than 75 percent of the early cases in Angola were among children, mostly infants. Infections in children were ‘most probably related to needle use, perhaps for vaccines, perhaps from multidose vials.’[lix]

A 2005 review of recurrent epidemics of Ebola, Lassa, and Marburg hemorrhagic fever in Africa concluded that[lx]


transmission of blood-borne viruses in medical facilities of all kinds is probably common [across much of Africa]...Indeed, hepatitis C virus (HCV) and human immunodeficiency virus (HIV) may be the viruses most commonly spread by this method. The difference with the haemorrhagic fever viruses is that the consequences…are immediately noticeable, whereas with HCV and HIV it takes years, even decades, for the transmission to be appreciated.


Why don’t we know?  


During hospital-based outbreaks of Ebola, Lassa, and Marburg infections, people in the community do not need an investigation to see a connection between the hospital and the illness. But because most HIV infections cause no or only minor symptoms for years, people cannot so easily see the link between HIV infection and medical procedures. In the absence of investigations, an antenatal clinic could, for example, infect hundreds of women with HIV over many years without coming under suspicion.

Finding out how HIV infects so many women is crucial to understanding and stopping generalized epidemics. Most studies that have asked women about medical injections as risks for incident HIV infection have found that women reporting injections were more likely to acquire HIV infection (Table 8.1). What is more remarkable, however, is not the evidence that links HIV infection in women to healthcare, but the rarity of such research.

For example, more than 40 percent of women in Botswana aged 25 to 39 years are HIV-positive. Harvard University, CDC, and other foreign organizations fund and advise HIV research in Botswana. Through early 2009, no one has reported research that has asked any Botswanan woman (or man for that matter) about injections or other blood exposures as risks for HIV infection. Sexual acquisition of HIV infection is assumed.

[i] van Bueren J, Simpson RA, Jacobs P, et al. ‘Survival of human immunodeficiency virus in suspension and dried onto surfaces’, J Clin Microbiol, 1994, 32: 571-4; Tjotta E, Humgnes O, Grinde B. ‘Survival of HIV-1 activity after disinfection, temperature and pH changes, or drying’, J Med Virol 1991; 35: 223-7; Kramer A, Schwebke I, Kampf G. ‘How long do nosocomial pathogens persist on inanimate surfaces? A systematic review’, BMC Infect Dis, 2006, 6: 130.

[ii] Abdala N, Stephens PC, Griffith BP, et al. Survival of HIV-1 in syringes’, J Acquir Immune Defic Syndr, 1999, 20: 73-80.

[iii] Sattar SA, Springthorpe VS. ‘Survival and disinfectant inactivation of the human immunodeficiency virus: A critical review. Rev Infect Dis, 1991, 13: 430-47; van Bueren J, Simpson RA, Salman H, et al. ‘Inactivation of HIV-1 by chemical disinfection: Sodium hypochlorite’, Epidemiol Infect, 1995, 115: 567-79.

[iv] Mosczynski P. ‘Unhygienic circumcisions may increase risk of HIV in Africa’, BMJ, 2007, 334: 498.

[v] Johns Hopkins Center for Clinical Global Health Education (CCGHE). ‘Nursing Training Curriculum, HIV Practice and Reducing Stigma, Module I: HIV transmission.’ Baltimore: CCGHE, no date. Available at: (accessed 9 September 2007). pp. 6, 24.

[vi] Heymann DL, Edstrom K. ‘Strategies for HIV prevention and control in sub-Saharan Africa’, AIDS, 1991, 5 (suppl 1): S197-208. p. S197.

[vii] Berkley S. ‘Parenteral transmission of HIV in Africa’, AIDS, 1991, 5 (supp 1): S87-92. p. S90.

[viii] Bell DM. ‘Occupational risk of human immunodeficiency virus infection in healthcare workers: An overview’, Am J Med, 1997, 102 (suppl 5B): 9-15.

[ix] Cardo DM, Culver DH, Ciesielski CA, et al. ‘A case-control study of HIV seroconversion in health care workers after percutaneous exposure’, N Eng J Med, 1997, 337: 1485-90; Gisselquist D, Upham G, Potterat JJ. ‘Efficiency of human immunodeficiency virus transmission through injections and other medical procedures: Evidence, estimates, and unfinished business’, Infect Control Hosp Epidemiol, 2006, 27: 944-52.

[x] Kaplan EH, Heimer R. ‘A model-based estimate of HIV infectivity via needle sharing’, J Acquir Immune Defic Syndr, 1992, 5: 1116-18; Hudgens MG, Longini IM Jr, Halloran ME, et al. ‘Estimating the transmission probability of human immunodeficiency virus in injecting drug users in Thailand’, Appl Statist, 2001, 50: 1-14.

[xi] Bobkov A, Garaev MM, Rzhaninova A et al. ‘Molecular epidemiology of HIV-1 in the former Soviet Union: Analysis of env V3 sequences and their correlation with epidemiologic data’, AIDS, 1994, 8: 619-24.

[xii] Bobkov A, Cheingsong-Popov R, Garaev M, et al. ‘Identification of an env G subtype and heterogeneity of HIV-1 strains in the Russian Federation and Belarus’, AIDS, 1994, 8: 1649-55.

[xiii] Gisselquist D et al. ‘Efficiency of human immunodeficiency virus transmission’.

[xiv] Chin J, Sato PA, Mann JM. ‘Projections of HIV infections and AIDS cases to the year 2000’, Bull WHO, 1990, 68: 1-11. p. 3.

[xv] Heymann DL, Edstrom K. ‘Strategies for HIV prevention’.

[xvi] Berkley S. ‘Parenteral transmission in Africa’. pp. S87, S90.

[xvii] Kane A, Lloyd M, Zaffran M, et al. ‘Transmission of hepatitis B, hepatitis C and human immunodeficiency viruses through unsafe injections in the developing world: Model-based regional estimates’, Bull WHO, 1999, 77: 801-7.

[xviii] UNAIDS. 2006 Report on the Global AIDS Epidemic. Geneva: UNAIDS, 2006.

[xix] Hauri AM, Armstrong GL, Hutin YJF. ‘The global burden of disease attributable to contaminated injections given in health care settings’, Int J STD AIDS. 2004, 15: 7-16.

[xx] Randerson J. ‘WHO accused of huge HIV blunder’, New Scientist, 6 December 2003, 180 (2424): 8-9.

[xxi] Reynolds SJ, Risbud AR, Shepherd ME, et al. ‘Recent herpes simplex virus type 2 infection and the risk of human immunodeficiency virus type 1 acquisition in India’, J Infect Dis, 2003, 187: 1513-21.

[xxii] Brewer D, Potterat JJ, Roberts JM, Brody S. ‘Male and female circumcision associated with prevalent HIV infection in virgins and adolescents in Kenya, Lesotho, and Tanzania’, Ann Epidemiol, 2007, 17: 217-26.

[xxiii] Croce F, Fedeli P, Dahoma M, et al. ‘Risk factors for HIV/AIDS in a low HIV prevalence site of sub-Saharan Africa’, Trop Med Int Health, 2007, 12: 1011-17.

[xxiv] N’Galy B, Ryder RW, Bila K, et al. ‘Human immunodeficiency virus infection among employees in an African hospital’, N Eng J Med, 1988, 319: 1123-7.

[xxv] Wawer MJ, Sewankambo NK, Berkley S, et al. ‘Incidence of HIV-1 infection in a rural region of Uganda’, BMJ, 1994, 308: 171-3.

[xxvi] Bulterys M, Chao A, Habimana P, et al. ‘Incident HIV-1 infection in a cohort of young women in Butare, Rwanda’, AIDS, 1994, 8: 1585-91.

[xxvii] Quigley MA, Morgan D, Malamba SS, et al. ‘Case-control study of risk factors for incident HIV infection in rural Uganda’, J Acquir Immune Defic Syndrome, 2000, 23: 418-25.

[xxviii] Bulterys M, Chao A, Dushimimana A, et al. ‘HIV transmission through health care in sub-Saharan Africa, authors’ replies [letter]’, Lancet, 2004, 364: 1665-6.

[xxix] Whitworth JA, Birao S, Shafer LA, et al. ‘HIV incidence and recent injections among adults in rural southwestern Uganda’, AIDS, 2007, 21: 1056-8.

[xxx] Todd J, Grosskurth H, Changalucha J, et al. ‘Risk factors influencing HIV infection incidence in a rural African population: A nested case-control study’, J Infect Dis, 2006, 193: 458-66; Gisselquist D. ‘New information on the risks of HIV transmission in Mwanza, Tanzania [letter]’, J Infect Dis, 2006, 194: 536-7.

[xxxi] Kiwanuka N, Gray RH, Serwadda D, et al. ‘The incidence of HIV-1 associated with injections and transfusions in a prospective cohort, Rakai, Uganda’, AIDS, 2004, 18: 342-4.

[xxxii] Lopman BA, Garnett GP, Mason PR, et al. ‘Individual level injection history: A lack of association with HIV incidence in rural Zimbabwe’, PLoS Med, 2005, 2: 142-6.

[xxxiii] Kumwenda NI, Kumwenda J, Kafuafula G, et al. ‘HIV-1 incidence among women of reorductive age in Malawi’, Int J STD AIDS, 2008, 19: 339-341. Kumwenda JJ, Makanani B, Taulo F, et al. ‘Natural history and risk factors associated with early and established HIV type 1 infection among reproductive-age women in Malawi’, Clin Infect Dis, 2008, 46: 1913-1920.

[xxxiv] Mermin J, Musinguzi J, Opio A, et al. ‘Risk factors for recent HIV infection in Uganda’, JAMA, 2008, 300: 540-549.

[xxxv] Gisselquist D. ‘HIV transmission through health care in sub-Saharan Africa [letter]’, Lancet, 2004, 364: 1665.

[xxxvi] Gisselquist D, Potterat JJ, Brody S, et al. ‘Let it be sexual: How health care transmission of AIDS in Africa was ignored’, Int J STD AIDS, 2003, 14: 148-161; Gisselquist D, Rothenberg R, Potterat JJ, Drucker E. ‘HIV infections in sub-Sahara Africa not explained by sexual or vertical transmission’, Int J STD AIDS, 2002, 13: 657-66.

[xxxvii] Schmid GP, Buve A, Mugyenyi P, et al. ‘Transmission of HIV-1 infection in sub-Saharan Africa and effect of elimination of unsafe injections’, Lancet, 2004, 363: 482-8.

[xxxviii] USAID. USAID’s Expanded Response to HIV/AIDS. Washington, DC: USAID, 2002.

[xxxix] WHO. The World Health Report 2002. Geneva: WHO, 2002. p. 9.

[xl] Hudson C. ‘How AIDS forces reappraisal of hepatitis B virus control in sub-Saharan Africa’, Lancet, 1990, 336: 1364-7.

[xli] Mayans MV, Hall AJ, Inskip HM, et al. ‘Risk factors for transmission of hepatitis B virus to Gambian children’, Lancet, 1990, 336: 1107-9. p. 1108.

[xlii] Mayans MV, Hall AJ, Inskip HM, et al. ‘Do bedbugs transmit hepatitis B?’, Lancet, 1994, 343: 761-3. p. 762.

[xliii] Ibid. p. 761.

[xliv] Kiire CF. ‘The epidemiology and control of hepatitis B in sub-Saharan Africa’, Prog Med Virol, 1993, 40: 141-56. p. 146.

[xlv] Yao GB. ‘Importance of perinatal versus horizontal transmission of hepatitis B virus infection in China’, Gut, 1996; 38 (suppl 2): S39-42. p. S40.

[xlvi] Hauri AM et al. ‘The global burden of disease’.

[xlvii] ‘Hepatitis C – global prevalence (update)’, Wkly Epidemiol Rec, 1999, 74: 425-7.

[xlviii] Gisselquist D, Perrin L, Minkin SF. ‘Parallel and overlapping HIV and bloodborne hepatitis epidemics in Africa’, Int J STD AIDS, 2004, 15: 145-52.

[xlix] Walker PR, Worobey M, Rambaut A, et al. ‘Sexual transmission of HIV in Africa’, Nature, 2003, 422: 679.

[l] Personal communication from Jonathan Schwartz, 23 August 2007.

[li] Davis GL. ‘Hepatitis C virus genotypes and quasispecies’, Am J Med, 1999, 106 (suppl 6B): 21S-26S.

[lii] Reid S. ‘High prevalence of chronic hepatitis B limits hepatitis C prevalence in Africa,’ unpublished, 2009.

[liii] Peters CJ. ‘Marburg and Ebola – Arming ourselves against the deadly filoviruses’, N Eng J Med, 2005, 352: 2571-3.

[liv] Colebunders R, Borchert M. ‘Ebola haemorrhagic fever – A review’, J Infect, 2000, 40: 16-20; Khan AS, Tshioko FK, Heymann DL, et al. ‘The reemergence of Ebola hemorrhagic fever, Democratic Republic of the Congo, 1995’, J Infect Dis, 1999, 179 (suppl 1): S76-86.

[lv] Okware SI, Omaswa FG, Zaramba S, et al. ‘An outbreak of Ebola in Uganda’, Trop Med Int Health, 2002, 7: 1068-75.

[lvi] Leroy EM, Telfer P, Kumulungui B, et al. ‘A serological survey of Ebola virus infection in Central African nonhuman primates’, J Infect Dis, 2004, 190: 1895-9.

[lvii] Towner JS, Sealy TK, Khristova ML, et al. ‘Newly Discovered Ebola Virus Associated with Hemorrhagic Fever Outbreak in Uganda.’ PLoS Pathog, 2008, 4: e1000212. doi:10.1371/journal.ppat.1000212

[lviii] Fisher-Hoch SP, Tomori O, Nasidi A, et al. ‘Review of cases of nosocomial Lassa fever in Nigeria: the high price of poor medical practice’, BMJ, 1995, 311: 857-9. p. 857.

[lix] Fisher-Hoch SP. ‘Lessons from nosocomial viral haemorrhagic fever outbreaks’, Brit Med Bull, 2005, 73 and 74: 123-37. p. 134.

[lx] Ibid. pp. 133-134.