Chapter 5

From first tests to a consensus on risks, 1984-88


During 1984-88, researchers tested millions of people all over the world for HIV infection. Testing was sufficient to find where HIV had reached and to make ballpark estimates of the numbers infected. In Africa and in Haiti, public health experts faced the additional challenge of finding and closing the paths through which HIV transmitted among the general population.    


First estimates: Pattern 1 or concentrated epidemics 


In 1988, WHO described AIDS epidemics in the Americas (except parts of the Caribbean), Western Europe, Australia, and New Zealand as Pattern 1 epidemics, where ‘most cases occur among homosexuals or bisexual males and intravenous drug users,’ and ‘the male-to-female sex ratio of infection ranges from 10:1 to 15:1.’[i] In countries with Pattern 1 or concentrated epidemics, surveys during 1984-88 sampled members of high-risk groups, such as MSMs attending a specific clinic. Then, to estimate the total number of people infected, information from these surveys was combined with estimates of the percentages of the population in high-risk groups.

Testing found more HIV infections than most people had expected. As of 1988, WHO estimated 2.3 to 2.8 million HIV infections in Pattern 1 epidemics.[ii] This agreed with CDC’s estimate of 1 million infections in the US in 1989, equivalent to 0.6 percent adult HIV prevalence.

Through December 1988, CDC attributed 89 percent of cumulative AIDS cases in the US to MSMs and IDUs, 3 percent to transfusions of blood and blood products, 3 percent to heterosexual exposures (setting aside persons born in countries with generalized epidemics), and a few percent to other and unknown risks.[iii] Similar concentrations of AIDS cases in MSMs and IDUs appeared in Western Europe, Australia, and Canada.


First estimates: Pattern 2 or generalized epidemics in Africa and Haiti          


In Africa and parts of the Caribbean, WHO described Pattern 2 epidemics, in which ‘The male-to-female ratio of infection is approximately 1:1,’ and IDU and MSM account for few infections.[iv] In Africa and Haiti, most HIV infections spread across adults with no recognized unusual behaviors or risks. Small groups with high HIV prevalence, such as sex workers, accounted for small minorities of infections. Thus, it was necessary to survey the general population to determine the number of infections.

During 1985-90, governments of four African countries arranged national surveys. In some other countries, surveys targeted the general population in one or more cities or rural communities. Despite some questions about survey design (whether communities and individuals were chosen at random), and HIV tests (number of HIV-positive results that were false positives), these surveys provide good information about the numbers and distribution of HIV infections in Africa in 1988.

Other important data come from repeated (sentinel) surveys of HIV prevalence among pregnant women attending selected urban clinics for antenatal care. Although HIV prevalence among women at urban antenatal clinics has generally been far greater than in all urban and rural adults, data from urban antenatal clinics is useful to compare HIV prevalence across countries and to track changes over time.


Table 5.1: Countries with the worst HIV epidemics in 1988

Region, country

WHO’s 1988 estimated HIV prevalence (%), all ages[v]

HIV prevalence (%) in adults:

General population surveys, 1985-90[vi]

Urban women at antenatal clinics[vii]





West Africa






    Cote d’Ivoire*






    Guinea Bissau*






Central Africa
























East Africa








11§, 16






















Southern Africa




































* Some data for these countries include HIV-2 infections.

† Survey in 1980

‡ Combined results from two surveys in Bangui in 1986-87

§ Survey includes children

¶ Survey in semi-urban areas

Sources: See references by column.


After ‘a detailed meta-analysis of available seroprevalence data for each country in sub-Saharan Africa,’ WHO estimated that as of 1988, 2.5 million people were living with HIV infection in generalized epidemics in Africa and the Caribbean.[viii] At the same time, WHO estimated that HIV prevalence in the total population – including children – was 1.0 percent or more in 10 African countries, and 0.5 percent or more in two others (Table 5.1). Because African populations are heavily weighted towards younger ages (due to high rates of population growth, 45 percent of the population was aged 0-14 years in the late 1980s),[ix] and because adults account for most HIV infections, this was roughly equivalent to estimating adult HIV prevalence of at least 2 percent and 1 percent in 10 and 2 countries respectively.

In 1988, the combined population of the 12 countries that WHO considered to have the worst HIV epidemics in Africa was 130 million. Using WHO’s lower bound estimate for the percentage of the population infected in each country, these countries had at least 1.1 million infections. Thus, WHO’s estimate of 2.5 million infections for countries with generalized epidemics assumed higher prevalence in some of these 12 countries, as well as some infections in other countries in Africa and in the Caribbean. But WHO did not publish these details.


Central Africa


As of 1988, WHO estimated that at least 1 percent of the population (2 percent of adults) was HIV-positive in three countries in Central Africa: CAR, Congo, and DRC. In all three, median HIV prevalence among women attending antenatal clinics in urban areas was around 4-6 percent in 1987-88. Considering that most people lived in rural areas with less HIV, adult HIV prevalence may have been around 2 percent for these countries (consistent with WHO’s estimate).

There is some evidence for relatively slow epidemic expansion in DRC during the 1980s. In Kinshasa, 3 percent of mothers attending well-baby clinics were HIV-positive in 1980 (from stored blood).[x] In eight years – i.e., by 1988 – HIV prevalence in Kinshasa more than doubled to 7 percent among pregnant women at antenatal clinics. Near Yambuku in north central DRC, 5 (0.8 percent) of 659 rural adults and children were HIV-positive in 1976 (from tests of stored blood collected after the community’s 1976 Ebola outbreak). In 1986 researchers went back to see what had happened. Two of the five people who had been HIV-positive in 1976 were alive and HIV-positive in 1986, while three had died with symptoms suggesting AIDS. At the same time, a new survey found that 3 (0.8 percent) of 388 rural adults were HIV-positive, which suggests stable HIV prevalence in the area from 1976 to 1986.[xi]

In Cameroon, where the HIV-1 M group began, surveys in 11 rural and urban communities in 1985-87 found from 0 percent to 1 percent (with a median of 0.5 percent) of adults to be HIV-positive. From this, the team that organized the surveys estimated adult HIV prevalence in Cameroon to be not more than 0.5 percent.[xii]


West Africa     


Among the 12 African countries with the worst HIV epidemics in 1988, WHO listed two in West Africa: Cote d’Ivoire and Guinea-Bissau. During the 1980s, HIV-1 expanded into West Africa, where HIV-2 had circulated at low levels in previous decades. By 1988, HIV-1 accounted for most HIV infections in Cote d’Ivoire and Burkina Faso. HIV-2 continued to be more common in Guinea-Bissau, Senegal, and Gambia.

Cote d’Ivoire appears to have had one of the fastest growing HIV epidemics in the 1980s. Surveys in randomly selected communities in February 1989 found 4.9 percent prevalence of HIV-1 and/or HIV-2 infection in rural adults and 7.3 percent in urban adults, excluding Abidjan (see Table 5.1). If these results are reliable, adult HIV prevalence in Cote d’Ivoire was around 6 percent in 1988 (two-fifths of the population was urban), and Cote d’Ivoire along with Uganda (see below) had the two worst HIV epidemics.

For Guinea-Bissau, tests of stored blood from a 1980 rural survey found 0.9 percent HIV-2 prevalence in children and adults (as reported in Chapter 3), including 1.8 percent HIV-2 prevalence in adults (9 infected of 510 tested), and there were no HIV-1 infections. Through the 1980s, HIV-1 remained rare in Guinea-Bissau. Although some studies during the 1980s reported higher HIV-2 prevalence in some communities, the persistence of higher prevalence in older people suggests that Guinea-Bissau’s HIV-2 epidemic expanded little, if at all, during the 1980s.


East Africa     


At the end of the 1980s, several countries east of DRC appeared to have some of the worst epidemics in Africa. As of 1987-88, median HIV prevalence in selected urban antenatal clinics ranged from 18 percent to 32 percent in Burundi, Rwanda, and Uganda. WHO estimated at least 1 percent HIV prevalence in the total population in all three countries in 1988, and at least 0.5 percent in Tanzania.

A national survey in Rwanda in 1986 found that 21 percent of adults were HIV-positive in urban areas (with 5.8 percent of the population), but only 1.5 percent in rural areas (see Table 5.1). Combining rural and urban data, estimated adult HIV prevalence across the country was 2.7 percent. This was, notably, less than one-tenth the 32 percent HIV prevalence reported from an urban antenatal clinic in Kigali in 1988!

Somewhat similar data are available for Burundi. A national survey in 1989-90 reported 11 percent HIV prevalence among adults and children in urban areas, 16 percent among adults in semi-urban areas, and 0.7 percent among adults in rural areas (see Table 5.1). The information I have from this survey does not show HIV prevalence among urban adults, or the percentage of the population that lived in semi-urban areas. But because most of the population was rural, if the survey’s results are at all accurate for rural areas, Burundi’s adult HIV prevalence could not have been much more than about 2 percent in 1988 – matching WHO’s lower-bound estimate.

Uganda very likely had the highest HIV prevalence in 1988. The government of Uganda arranged a national survey in 1987-88. Several accounts of this survey report HIV prevalence in various regions ranging from 0 percent to 12 percent in rural adults and from 7.7 percent to 29 percent in urban adults (see Table 5.1). Two second-hand accounts of survey results report estimates of 600,000-800,000 infections in Uganda,[xiii] which would be equivalent to 7-10 percent adult HIV prevalence. Considering that HIV prevalence may have been lower in several regions not surveyed, that false positives were a problem,[xiv] and that only 11 percent of Uganda’s population was urban, HIV prevalence may have been near or even below the low end of that range.

Kagera Region, west of Lake Victoria, had the worst HIV epidemic in Tanzania in the 1980s. Surveys of the general population in Kagera Region in 1989 found 24 percent of adults to be HIV-positive in Bukoba town, 10 percent in adjacent rural communities, 4.5 percent in more remote districts, and only 0.4 percent in the most remote districts in the Region.[xv] Through 1988, HIV prevalence reached 7.8 percent among women attending an antenatal clinic in Dar es Salaam. Considering that 20 percent of Tanzania’s population was urban, even though HIV had not reached much of the country, adult HIV prevalence in 1988 may well have exceeded WHO’s lower-bound estimate of 1 percent.


Southern Africa 


WHO listed three countries in Southern Africa – Malawi, Zambia, and Zimbabwe – among the 12 African countries with the worst HIV epidemics in 1988. Through 1988, infections in these three countries appear to have concentrated in large cities and in Zambia’s Copperbelt, adjacent to DRC. Even so, HIV prevalence in urban antenatal women was much lower than in Rwanda and Burundi. However, the epidemic was growing fast. During 1985-87, HIV prevalence among women at urban antenatal clinics in Malawi and Zambia more than doubled to 8-12 percent (see Table 5.1); comparable data are not available from Zimbabwe.


The Caribbean                       


WHO provided no estimates for HIV prevalence in 1988 in Haiti and in other Caribbean countries considered to have generalized epidemics. Based on surveys among blood donors, pregnant women, and other groups in urban areas (there was little information from rural areas), HIV prevalence in Haiti was comparable to what was found in DRC.[xvi] Similar surveys in other Caribbean countries generally found much less HIV through 1988.


First estimates: Pattern 3 HIV epidemics in the rest of the world


In 1988, WHO characterized epidemics in North Africa, the Middle East, Central and Eastern Europe, Asia, and most of the Pacific as Pattern 3 epidemics, with ‘recent onset of the HIV/AIDS pandemic (mid to late 1980s),’ and with few AIDS cases to date. In this vast region with most of the world’s population, WHO estimated 100,000 HIV infections in 1988 – less than 1 infection in 20,000 adults. Although most countries with Pattern 3 epidemics ‘have not as yet shown predominant modes of HIV transmission,’[xvii] some countries were already showing signs of Pattern 1 or 2 epidemics.

Information from south India through 1988 suggested the emergence of a Pattern 2 or generalized epidemic. In 1986, tests on 102 women in sex work in Tamil Nadu found 10 to be HIV-positive.[xviii] Very likely HIV had been circulating in India for some time when it was discovered, because expanded testing in 1987-88 in multiple sites across thousands of kilometers found infections in pregnant women, sex workers, patients with tuberculosis or sexually transmitted disease, and MSMs. Most HIV in India belongs to the HIV-1 C clade, and is related to HIV that circulates in much of Southern and East Africa. Family ties between India and Africa provided many opportunities for HIV to reach India. Analyses of HIV sequences from India might reveal the dates that several C clade viruses reached India and then multiplied to contribute to India’s epidemic.

Through 1988, Thailand’s HIV epidemic looked like a Pattern 1 or concentrated epidemic. The first known HIV infection in Thailand was in an MSM returning from the US with AIDS in 1984. He most likely acquired his infection in the US, and did not pass it to anyone in Thailand. A year later, in 1985, doctors in Bangkok found an HIV-positive male sex worker who had not lived outside Thailand.[xix] Through 1987, doctors identified approximately 200 HIV infections. Then, in nine months from late 1987 to September 1988, HIV prevalence in IDUs attending drug treatment clinics in Bangkok soared from 1 percent to 43 percent.[xx] A royal amnesty in December 1987 likely contributed to this epidemic surge by releasing HIV-positive IDUs from prison into the community.


Assessing generalized epidemics: Who was infected? 


Researchers looking for high-risk groups in Africa and the Caribbean found few MSMs and IDUs (except some MSMs in Haiti).[xxi] However, researchers found other groups that characteristically had relatively high HIV prevalence, including women, urban residents, people with high socio-economic status, patients, women in sex work, and blood donors.


Table 5.2: HIV prevalence (percentages) in selected populations in generalized epidemics, 1984-88

Region, country

Women in sex work*

Patients at clinics treating sexually transmitted disease*

Inpatients: Median [range]

Blood donors: Median [range]

West Africa





    Cote d’Ivoire



34 [6.8-70]

12 [2.8-19]




14 [9.8-16]

8.5 [5.1-9.8]

Central Africa








6.5 [1.1-9.7]





32 [7.7-38]

7.1 [3.4-14]




31 [4.1-87]

6.0 [1.0-9.0]

East Africa









8.8 [8.3-9.2]





18 [11-20]




10 [2.3-34]

5.1 [0.7-11]




37 [27-38]

16 [9.2-23]

Southern Africa








21 [4.5-29]





19 [15-27]

14 [6.0-19]





3.0 [2.3-37]









49 [35-63]

4.0 [2.2-16]

* These columns show medians from sentinel surveys in urban areas reported by WHO[xxii]; for each country, the columns show the last available median from 1985-88, except that for patients with sexually transmitted disease in Congo, CAR, Malawi, Uganda, and Zimbabwe, the column shows the first reported medians from 1989-90.

† These columns summarize data compiled by the US Census Bureau[xxiii] from studies that tested at least 50 inpatients or 100 blood donors (excluding data identified by the US Census Bureau as ‘data of unknown quality’).

‡ Includes HIV-2 infections.

Sources: See references in notes for each column.


Whereas women were much less likely than men to be HIV-positive in countries with concentrated epidemics, women in Africa were more often infected than men. For example, among thousands of patients found to be HIV-positive in Kinshasa in 1986, the ratio of women to men was 1.3 to 1.[xxiv]

Within each African country, urban adults had higher HIV prevalence than rural adults. The difference varied from country to country.

During 1985-88, many studies across Africa found that high socioeconomic status was a risk for HIV infection. Studies in Kinshasa reported higher HIV prevalence in administrators or higher-paid staff vs. other staff at a hospital,[xxv] bank, and textile factory.[xxvi] Among inpatients and outpatients in Zambia in 1985, HIV prevalence increased progressively from 8 percent among those with 0-4 years of education up to 33 percent for those with more than 14 years of education.[xxvii]

During 1984-88, hundreds of studies found high HIV prevalence among various categories of patients (Table 5.2). There were many reasons for patients to be HIV-positive, so the sources of their infections were not clear. People with HIV infection and weakened immune systems are more likely to get tuberculosis, genital ulcers, diarrhea, and other conditions that bring them to clinics. On the other hand, sexually transmitted diseases may have been markers for sexual behaviors that are risks for HIV infection. In addition, people who attended clinics for long-term or repeat treatment (such as for tuberculosis) may have been infected during treatment.

High rates of HIV infection among women in sex work are characteristic of generalized epidemics (see Table 5.2). In one of the first studies that tested African blood for HIV, Van de Perre found 88 percent of a sample of sex workers in Rwanda in 1984 to be HIV-positive.[xxviii] As later studies showed, this was unusually high, not only for sex workers in Africa but for all high-risk groups in all countries.

Although HIV screening of transfused blood was absent or erratic in most of Africa during 1985-88, many hospitals tested at least some blood intended for transfusion. HIV prevalence in blood donors often exceeded prevalence in urban adults. Many donors, including replacement donors recruited by patients’ families, were paid, repeat donors.


Assessing generalized epidemics: Sexual transmission


From 1984, the CDC began to list ‘heterosexual contact…with a person with AIDS or at risk for AIDS’ as a high-risk category in its reports of AIDS cases in the US.[xxix] Studies in the US and Europe during 1984-88 found that as many as 25 percent or more of the spouses of HIV-positive IDUs, MSMs, or hemophiliacs were also infected.[xxx] This was on the one hand reassuring, because HIV seemed to transmit inefficiently through penile-vaginal sex. But on the other hand, it was disturbing, because it meant that people were at risk from heterosexual partners. Where HIV had invaded the general population – as in Africa and Haiti – this called for programs to warn people to be aware of risk, and to use condoms with heterosexual partners who might be HIV-positive.

However, the finding that HIV transmitted (inefficiently) through penile-vaginal sex did not by itself explain generalized HIV epidemics. If, as many people believed, most HIV-infected adults in Africa and Haiti acquired their infections from heterosexual partners, then it was necessary to explain how heterosexual sex could transmit HIV so much more extensively in Africa and Haiti than in the US and Europe.


Conflicting information about African sexual behavior and HIV


An early, influential study from Belgium and Rwanda reported scarcely believable rates of partner change among African men with AIDS or pre-AIDS. According to the study, the men had from 12 to 60 sexual partners per year, and 81 percent had visited sex workers at least once per month for the past two years.[xxxii] One 1987 analysis of Africa’s AIDS epidemic generalized that ‘most traditional African societies are promiscuous by Western standards. Promiscuity occurs both premaritally and postmaritally.’[xxxiii]

Other studies during 1984-88 found that such behavior was not characteristic of Africans with HIV infections. For example, a study in Kinshasa in 1987-88 found that 71 percent of 7,058 men working at a textile factory or bank and 99 percent of their 4,548 wives reported no non-marital partners in the past year. Only 2.3 percent of men reported more than five non-marital partners in the past year. Notably, men and women who reported no non-marital partners were found to have most of the HIV infections. Men who reported no non-marital partners in the past year had 143 (61 percent) of the 236 HIV infections in men, while men who reported more than five non-marital partners had only 13 (5.4 percent) of the infections in men. Women who reported no non-marital partners had 170 (97 percent) of the 175 HIV infections in women. Furthermore, more than 70 percent of HIV-positive wives had HIV-negative husbands. The study did not report the numbers of injections received by workers or wives, but noted that ‘in Zaire patients who are ill always have high rates of receipt of injection with needles which may often have been re-used and not properly sterilized.’[xxxiv]

Another early study in Kigali, Rwanda, in 1988 found 25 women who were HIV-positive with HIV-negative husbands. Fifteen of these 25 women reported that their HIV-negative husband was their only lifetime sex partner.[xxxv] The study team suggested that women had underreported their sexual activity.


Unresolved risks for HIV among African sex workers


During 1984-88, some of the most influential research on risks for HIV infection in Africa was conducted through clinics treating sexually transmitted disease. During 1981-85, before anyone knew that AIDS was circulating in Kenya, researchers working through a special study clinic in Nairobi studied genital ulcers among sex workers. Tests of stored blood from this study found 4 percent of sex workers to be HIV-positive in 1981, increasing precipitously to 82 percent in 1983, before falling back to 61 percent in 1985.[xxxvi] During this period, HIV prevalence in pregnant women attending antenatal clinics in Nairobi increased from 0 percent to 2 percent. Sex workers appeared to be leading Kenya’s HIV epidemic.

Two subsequent studies that followed sex workers and male clients through Nairobi clinics that treated sexually transmitted disease observed high rates of HIV incidence.[xxxvii] During 1985-87, 67 percent of initially HIV-negative sex workers followed for an average of 18 months became HIV-positive. Similarly, 8 percent of male clients followed for an average of less than 4 months after seeking treatment for sexually transmitted disease became HIV-positive. Neither of these studies considered the possibility that blood tests and injections in study clinics transmitted HIV. Notably, sex workers who visited the study clinic more frequently, as well as sex workers who had genital ulcers (which were routinely treated with injections), were more likely to acquire HIV infections.

During the mid-1980s, Vachon,[xxxviii] Wyatt,[xxxix] and others proposed that unsterile medical injections in Africa – especially injections to treat sexually transmitted disease – worked synergistically with sexual transmission to spread HIV infection among sex workers and clients. In 1985, Wycoff criticized one study that attributed high HIV prevalence among sex workers’ clients to sexual transmission:[xl]


the reader is left to wonder if the risk factor for HTLV-III/LAV [HIV] spread in central Africa is prostitute exposures (e.g., heterosexual activity) or STD [sexually transmitted disease] clinic attendance (e.g., infected needles exposures). The implications of this uncertainty – economic, medical, and social – are very important to any evaluation of the AIDS epidemic in central Africa.


Observed high rates of HIV prevalence and incidence among sex workers in Africa during 1984-88 contrasted sharply with findings from contemporaneous studies among sex workers in countries with concentrated epidemics. For example, in seven studies in Western Europe through 1987, less than 2 percent of a cumulative total of more than 2,700 sex workers were HIV-positive, and more than 60 percent of those who were HIV-positive were IDUs.[xli] Few sex workers in Africa were IDUs, so that could not explain their high rates of HIV prevalence.

The hypothesis that clinics treating sexually transmitted disease infected African sex workers and clients was not disproved – it was not even tested. Nevertheless, clinic-based studies reporting high HIV prevalence and incidence among African sex workers and clients were widely accepted to show a dominant role for heterosexual transmission in Africa’s HIV epidemics. Conflict of interest could explain why studies failed to ask the right questions to trace HIV infections either to sex or to blood exposures. Researchers working through clinics treating sexually transmitted disease could have been hurt by the finding that their clinics were spreading HIV.


The key argument that sex accounts for Africa’s HIV epidemics


In October 1985, WHO organized a workshop on AIDS in Central Africa, in Bangui, CAR. The memorandum from that workshop estimated that 80-90 percent of AIDS in African adults was from sexual transmission.[xxxi] Through 1988, this position swept the field. It is useful to consider the evidence and arguments available through 1988 that supported this hypothesis.

As noted in previous paragraphs, some studies reported high HIV prevalence in prostitues and their customers. But most HIV infections were found in African men and women with average or even conservative sexual behavior. Even if Africans were promiscuous relative to Americans or Europeans (which was soon disproved; see Chapter 7), and even if there were other factors that accelerated HIV transmission through heterosexual contact in Africa, these differences would only support arguments that heterosexual exposures accounted for a higher proportion of HIV infections in Africa than in the US and Europe. However, that meant only an indeterminate percentage more than 3 percent – which was the proportion of cumulative AIDS cases among US adults attributed to heterosexual transmission through 1988. Such arguments did not support estimates that sex accounted for most infections in Africa. Similarly, the common argument that the equal sex ratio among HIV-positive adults mean that most had contracted their infections through sex was specious (see Chapter 4).

In the 1980s – and onwards – the key argument supporting the hypothesis that sex accounted for most HIV infections in African adults had nothing to do with sex. Instead, the key argument was that blood exposures in health care did not account for many HIV infections, so that by elimination of other risks, infections in African adults must have come from heterosexual exposures. The next section considers evidence available through 1988 to assess the importance of HIV transmission through blood exposures.


Assessing generalized epidemics: Transmission through trace blood exposures


HIV survival outside the body


In 1985, several French scientists who had helped to discover HIV reported its ‘unusual stability’ at room temperatures. For HIV kept in wet conditions, there was only a ‘slight decrease’ in its ability to grow in cell culture after seven days. Even when dried, some HIV remained infectious for seven days. The scientists advised that ‘more safety precautions should be taken…in hospitals and by dentists…’[xlii]

Another study published in 1986 reported that some HIV (beginning with a concentrated solution with more HIV than would normally be found in blood) survived for more than 15 days in wet conditions and three or more days after drying.[xliii] In 1987, CDC reported results from its own studies showing that drying HIV for ‘several hours’ leaves 1 percent to 10 percent of HIV alive and infectious.[xliv]


Exposures to HIV-contaminated equipment   


Hundreds of studies during 1984-88 reported high HIV prevalence in patient populations in Africa and Haiti (Table 5.2). Whatever the reasons for these infections, they presented a threat to other patients when clinics reused instruments without sterilization.

During the 1980s, virtually everyone who was aware of healthcare practices in Africa reported that medical instruments were often reused without sterilization. For example, a 1987 study in Uganda observed that injections ‘are often not given under aseptic conditions and themselves could theoretically be a vehicle of transmission.’[xlv] In Haiti, ‘injections may be given by either medical personnel or piqurists (untrained injection doctors). Disposable needles and syringes are not readily available in Haiti, so needles and syringes may be reused without sterilization.’[xlvi]

Moreover, studies during the 1980s showed that injections and other invasive procedures were common.


Transmission risk through contaminated instruments            


Studies during 1986-88 showed that almost everyone who received a transfusion of blood from an HIV-positive donor became infected.[xlvii] However, this said little about the lower risk associated with exposures to very small amounts of blood, such as what might be found on medical instruments reused without sterilization.

Similarly, studies through 1988 found high HIV prevalence among IDUs. But because IDUs get so many injections, this finding did not provide much information about the risk to transmit HIV through less frequent blood exposures during healthcare.

Another view of patients’ risks to acquire HIV from contaminated instruments came from studies of healthcare workers after needlestick accidents. At least seven studies published in 1985-86 followed healthcare workers exposed to HIV through accidents, such as scratches or jabs with instruments that had just been used on HIV-infected patients. Combining information from these studies, only 3 (0.4 percent) of 818 workers developed HIV infections.[xlviii] However, because most needlestick accidents are scratches that do not puncture the skin, this was not a reliable estimate of the risk to transmit HIV from one patient to others through injections and other skin-piercing procedures. Thus, other studies were required to measure the risks that people in Africa and in the Caribbean faced of acquiring HIV infection in healthcare settings, and to determine how much such risks contributed to HIV epidemics.


Early research linking HIV infections to health care in Africa           


During 1984-88, the best research on blood exposures as risks for HIV infection in Africa came from Project SIDA (syndrome d’immuno deficience acquise), which was headquartered at Mama Yemo Hospital in Kinshasa. The governments of DRC, the US, and Belgium jointly sponsored Project SIDA. Researchers in Rwanda, Tanzania, Uganda, and other countries also asked about blood exposures as risks for HIV infection and reported useful information.

In 1985, Project SIDA tested inpatient and outpatient children aged 1-24 months and their mothers for HIV infection. Among 449 child-mother pairs, 44 children were HIV-positive. Surprisingly, 17 (39 percent) of the 44 HIV-positive children had mothers who were HIV-negative. Sixteen of the HIV-positive children with HIV-negative mothers were inpatients. These 16 children were an average of 11 months old and had received an average of 44 injections. Among children with HIV-negative mothers, ‘medical injections seemed to be the most important risk factor for HIV seropositivity, followed by previous blood transfusions or hospital admission.’ The study team noted, ‘Injections are often administered in dispensaries which reuse needles and syringes yet may not adequately sterilize them.’[xlix]

In 1984-86, Lepage and colleagues conducted a similar study in Kigali, Rwanda, that produced similar results.[l] They tested the mothers of 76 children with AIDS aged 1-48 months. Eighteen (24 percent) of the mothers were HIV-negative. The 18 children with HIV-negative mothers had received an average of 23 injections, seven had been transfused, and eight had been previously hospitalized. Lepage and colleagues accepted that transfusions had infected some children, but did not want to accept that injections had done so as well. To see if there was some mistake, they retested 12 mothers who had previously tested HIV-negative. On this second test, three of 12 tested HIV-positive, leaving 15 (20 percent) of 76 HIV-positive children with HIV-negative mothers.

From this, the study team speculated that the HIV-negative test results for the other mothers may have been false-negatives as well. (Another possible – even likely – explanation for several mothers first testing negative and then positive is that they had contracted HIV infection from their children, possibly through breastfeeding, as later reported from Russia and Libya. See Chapter 9.) After raising doubts about evidence they did not like, Lepage and colleagues did not, either then or later, do the further research required to validate their doubts. The lack of resolution meant that Rwandan children were dying of unexplained AIDS, while medical researchers were satisfied to impugn evidence pointing to healthcare mistakes.

Thus, studies among children suggested that healthcare in Africa might be responsible for a lot of HIV infections among children. However, studies among children could not show how much HIV in adults was due to blood exposures. That could only be determined by studies among adults.

In mid-1984, Project SIDA tested 2,384 staff of Mama Yemo Hospital in Kinshasa.[li] One hundred and fifty-two (6.4 percent) were HIV-positive. Staff who reported injections in the past three years were 1.8 times more likely to be HIV-positive than those who reported no injections. During 1985-88, other studies in Rwanda, Tanzania, Uganda, and Zimbabwe asked about injections as risks for prevalent (existing) HIV infections in adults.[lii] These studies found that adults who reported injections in recent years were 1.7 to 3.6 times more likely to be HIV-positive than those who reported no injections. Because most of the people in these studies reported injections, injections seemed to be linked to a lot of HIV infections.

However, the repeated finding that adults who had received injections in recent years were more likely to be HIV-positive did not necessarily show that injections transmitted HIV. Compared to people without HIV infections, people who were HIV-positive were more likely to be sick and therefore more likely to go for injections. This ‘reverse causation’ could explain some or all of the observed link between injections and prevalent HIV infection. Other study designs were needed to see how many HIV infections among adults were coming from injections and other invasive healthcare.

One way to minimize the problem of reverse causation is to find a group of HIV-negative people, and then follow and re-test them later to see who becomes HIV-positive, and ask everyone about risks. In 1986, Project SIDA re-tested 1,905 hospital employees who had been HIV-negative in 1984. In two years, 62 staff had acquired new HIV infections. Staff who reported one or more injections during those two years vs. staff who reported no injections were 1.53 times more likely to show up with an incident (new) infection.[liii] However, the study team avoided the conclusion that injections were a risk for HIV infection. They reported that 6 of 62 people with incident infections had unspecified ‘HIV-associated signs and symptoms’ over the two years, and speculated that these six persons might have sought injections to treat HIV-related symptoms. After cutting these six persons out of the analysis, staff who reported injections were still 1.35 times more likely to show up with a new HIV infection. However, with the revised data, the observed greater risk to acquire HIV infection among those who reported injections was no longer ‘significant.’ To say the risk was no longer significant meant that the more frequent HIV infections among staff who reported injections may have been a statistical accident. In layman’s terms, there was at least a 5 percent chance that injections did not infect any of the hospital staff.

Thus, during 1984-86, researchers had found HIV in children with HIV-negative mothers, and had followed HIV-negative adults to find that those who reported injections were more likely to acquire HIV infections. One interpretation of these findings is that injections and possibly other blood exposures caused a lot of HIV infection. Nevertheless, many researchers resisted this evidence with arguments that tests may have been bad, and that unwelcome findings may have been a statistical accident.

Considering the importance of the issue, and the suggestive evidence from these three early studies, one could have expected new studies of injections and other blood exposures as risks for incident HIV infection. These new studies could have been carefully designed to meet all doubts, so that health experts – and the public – would soon know if blood exposures during healthcare and cosmetic procedures were major contributors to Africa’s HIV epidemics.

This did not happen. During 1985-88, only two other studies in Africa[liv] and none in Haiti asked about injections as risks for incident HIV infections. These two studies (which have been mentioned earlier in this chapter) worked through clinics treating sexually transmitted disease in Nairobi, Kenya, to study HIV incidence in sex workers and clients. In these studies, the few sex workers and clients who reported injections or scarification outside the study clinics were not more likely to show up with new HIV infections. However, neither study considered risks to transmit HIV through the much more common invasive procedures – injections and blood tests – in the study clinics. With their incomplete data, these two studies provide weak support for the position that blood exposures only rarely transmit HIV in Africa.


What about low HIV prevalence in children?


In the 1980s – and onwards – one of the most influential arguments that blood exposures account for few HIV infections in Africa has been that low HIV prevalence in children (except young children with HIV-positive mothers) shows that blood exposures seldom transmit HIV. Those who present this argument can cite many studies during 1984-88 (and later) that found few HIV infections in African children that could not be attributed to mother-to-child transmission. However, many other studies through 1988 (and later) have reported unexplained infections in children. For example, Rwanda’s 1986 national survey found 4.2 percent of urban children aged 6-15 years to be HIV-positive.[lv] And, as noted above, several hospital-based studies in 1984-86 found high percentages of HIV-positive children to have HIV-negative mothers.

Moreover, even in communities where children have few unexplained and possibly iatrogenic infections, this finding cannot be extrapolated to adults. Packard and Epstein have argued that the age-sex distribution of AIDS is not reliable evidence for sexual transmission:[lvi]


…if one looks at the age-sex distribution of tuberculosis [TB], one sees a similar pattern. Surely no one would argue that TB is a sexually transmitted disease.…[I]t is generally noted that children in the age range from 4-14, for reasons which are not altogether clear, have a higher level of resistance to a number of diseases including TB. This period is often referred to as the ‘golden years.’ This phenomenon may in fact occur in AIDS.


Extending their argument, children vs. adults may be more able to resist HIV infection when exposed. Furthermore, children vs. adults may be treated in clinics where providers more often use sterile instruments, and where other patients are less often HIV-positive. Finally, adults might be exposed to blood during dental care, tattooing, shaving, and other procedures not common among children. Thus, low HIV prevalence in children does not show that adults are not contracting HIV from blood exposures during healthcare and cosmetic services.


Assessing generalized epidemics: Transmission through blood transfusions


During 1984-88, at least 12 studies of risks for HIV infection in African adults and children (excluding studies among patients) asked about transfusions. Averaging results from these studies, transfusions appeared to have been responsible for about 5 percent of HIV infections.[lvii] In the late 1980s, AIDS experts generally accepted these results as an accurate measure of HIV infections from transfusions.

The contribution of transfusions to Africa’s HIV epidemic was determined by the percentage of blood donors who were HIV-positive and the frequency of transfusions. In 1984-88, in the 12 African countries with the worst HIV epidemics and in Haiti, the median HIV prevalence among studies that tested blood donors for HIV ranged from a low of 3.0 percent in Zimbabwe to 18 percent in Rwanda and 20 percent from one study in Malawi (Table 5.2). Most African governments that answered a 1984 WHO questionnaire reported heavy reliance on paid and replacement donors.[lviii]

Some studies in Africa reported frequent transfusions, suggesting that doctors often transfused when other and safer treatments would suffice. For example, a 1984-85 study in Kinshasa found that 36 percent of inpatient children aged 2-14 years had been transfused before their current hospitalization, and so had 14 percent of their healthy siblings.[lix] People in sub-Saharan Africa (except South Africa) received an estimated 20 million transfusions in the 1980s[lx] – equivalent to about 0.5 percent of the population receiving a transfusion each year. This shows a large increase from an estimated 12 million transfusions in the 1970s.


Risk for HIV among repeat blood donors 


I have found no studies in Africa or Haiti during 1984-88 that asked about blood or plasma donation as a risk for HIV infection. However, reports from Spain[lxi] and Mexico in 1987-88 showed that donating plasma was a risk for HIV infection.

In 1986, shortly after the Mexican government ordered that all blood and plasma donations should be tested for HIV, tests found that 7 percent of paid donors were HIV-positive. Subsequently, tests of stored plasma from one private plasma-buying center identified 281 HIV-positive donors, including 62 repeat donors with new infections between June and October 1986.[lxii] Characteristically,[lxiii]


a paid donor would be a young man from a rural area who had migrated to one of the shanty towns that surround large cities…After hearing about the opportunity from a friend or family member, he would become a regular customer at one of the local blood banks or plasmapheresis centers, being paid to donate…as often as every 2 or 3 days. The more times he donated, the higher his risk became for becoming infected with HIV during the blood collection process.


After this study, Mexico banned blood sales from 1987. The Mexican Government’s aggressive response to investigate and stop nosocomial transmission has been characteristic of countries, such as Mexico, with concentrated HIV epidemics. Subsequently, Mexico and the Panamerican Health Organization listed professional blood donors as a high-risk category for HIV infection in official statistics. WHO and African governments did not adopt this category.


AIDS control programs accept HIV transmission through healthcare in Africa      


During the mid-1980s, many health experts in Europe and the US expected that healthcare practices in Africa would change to prevent HIV transmission to patients. For example, a Lancet editorial in 1987 identified ‘the three main control activities’ for HIV in Africa to be ‘screening of donated blood, improvements in the use of sterile procedures by health workers, and health education designed to modify sexual behavior.’ The editorial supposed that ‘the first two measures are…at least in principle, moderately easy to achieve.’[lxiv]

However, managers of the world’s response to AIDS in Africa showed only a limited commitment to stop HIV transmission through healthcare. As already described, almost as soon as doctors in DRC and Rwanda started to test for HIV, they found HIV-positive children with HIV-negative mothers. Without investigations – asking where and when children had received care, and then testing others treated at the same facilities around the same time – it was not possible to determine the extent of the damage from infection control lapses, or to target interventions to the responsible facilities and procedures. There is no indication that international agencies or foreign health aid programs urged or offered to help with investigations, and there is no report that public health agencies in the affected countries investigated these or any other unexplained HIV infections. Lack of investigations implicitly accepted continuing unknown numbers of nosocomial infections.

An important 1986 article on AIDS in Africa projected continuing unsafe practices: ‘one cannot hope to prevent reuse of disposable injection equipment when many hospital budgets are insufficient for the purchase of antibiotics,’ and ‘one cannot expect public health officials to upgrade blood transfusion services to prevent HIV infection...’[lxv] The authors of that article include the heads of WHO’s Global Programme on AIDS and later UNAIDS for most of the next 22 years. They expected – and subsequent HIV prevention programs accepted – that African clinics and hospitals would not reliably sterilize skin-piercing instruments, and would not test all transfused blood.


Initial but declining efforts to protect patients              


The 1985 Bangui Workshop on AIDS in Central Africa discussed blood exposures as well as sex. The memorandum from the meeting noted, ‘Epidemiological data show that non-sterile injections are responsible for a large proportion of AIDS cases in children,’ and recommended, ‘The public must be warned of the risk of AIDS transmission, especially through injections or scarifications carried out with equipment that is not disinfected and is re-used, both in modern medicine and in traditional medicine.’[lxvi]

From 1986, WHO established the Global Programme on AIDS (initially named the Control Programme on AIDS). Along with an overwhelming emphasis on sexual behavior, early Global Programme documents discussed risks to transmit HIV in healthcare, and what to do about them. In 1986, the Global Programme’s proposed outline for a national AIDS control program advised:[lxvii]


Prevention of [HIV] transmission through injections may require programmes to prevent reuse of single-use syringes/needles or methods to ensure reliable sterilization of reusable injection material. Programmes may also be directed towards sterilization of other instruments which pierce the skin or contact mucous membranes (e.g., scarifications, circumcision knives, ear-piercing).


In 1986, the Global Programme’s proposed model budget for national AIDS prevention programs allocated 16 percent to a ‘syringe/needle pilot project,’ and 24 percent to test transfused blood for HIV. With this money, the Global Programme proposed that AIDS prevention programs provide sterilizers and self-destructing syringes, which could not be reused. Corresponding to this advice, Uganda in 1987 formulated a 5-year AIDS control plan, including, along with other components, infection control and blood safety. The program provided sterilizers, syringes, and needles for health centers and hospitals, trained health staff in sterile techniques, [lxviii] and warned the public to avoid injections.[lxix]

The Global Programme’s interest in infection control to protect patients was at its high point in 1986. After 1986, the Programme’s documents show continuing awareness of widespread infection control lapses, but declining commitment to address them. For example, the Programme’s 1988 guidelines for national AIDS programs urged governments to review infection control practices in formal and informal healthcare, but proposed no inputs or activities – aside from preparation of guidelines and professional training – to address infection control lapses.[lxx] Hence, for countries that recognized their epidemics early and asked for advice in 1986, the Global Programme urged them to allocate substantial efforts and resources to improve infection control. For countries that asked for advice after 1986, the Programme was much less intent on promoting infection control.


Limited commitment to blood safety (i.e., safe blood transfusions)    


Through early 1986, only two African countries – South Africa and Zimbabwe – reported routine screening of transfused blood for HIV.[lxxi] Limited donor support for safe blood transfusions is illustrated by Project SIDA’s activities in Kinshasa.

Project SIDA, headquartered at Mama Yemo Hospital, began using HIV tests in research in 1984. Not until two years later, in November 1986, did the project help the hospital to begin screening transfused blood for HIV. Even then, ‘due to the lack of trained staff, ELISAs [testing] could be done only during the day.’[lxxii] During the night, doctors continued to give emergency transfusions with untested fresh blood from someone who was available. This situation – testing during the day, but not at night – continued for 6 months. (Project SIDA took advantage of this situation to see what happened to people transfused with HIV-positive blood. By testing blood samples after blood had already been transfused, the Project identified 90 people who were HIV-negative when they entered the hospital, but who received HIV-positive blood. All or almost all patients – records were unreliable – given HIV-contaminated blood acquired HIV infections, and they were more likely to die over the following year than patients given HIV-negative blood.)

Like Project SIDA, WHO accepted transfusion of HIV-contaminated blood in Africa. In 1986, an expert committee organized by WHO recommended that countries should consider whether or not to test blood for HIV ‘in the context of their overall health programmes and the availability of human and material resources.’[lxxiii] In 1987, the Programme Committee of WHO’s Executive Board recommended transfusion of untested blood for an indefinite period:[lxxiv]


…in many developing countries…if one excluded all those who suffered from various tropical and parasitic diseases and hepatitis, as well as those who harboured HIV, the reservoir of potential safe donors would be substantially depleted. The cost of the necessary screening procedures to ensure the safety of the blood being donated was prohibitive…a pragmatic approach had to be followed in these countries to improve the situation progressively…


Initially, large budgets were required for new laboratories to test for HIV. In 1987, relatively low-cost rapid tests became available. From 1987, the Commission of European Communities (which later became the European Union) provided aid for blood safety in nine African countries.[lxxv] Despite some improvements, progress was constrained by lack of money and commitment. In 1988, the Global Progamme on AIDS together with the League of Red Cross and Red Crescent Societies began the Global Blood Safety Initiative.[lxxvi] This initiative soon produced good plans to stop HIV transmission through blood transfusions, but these plans subsequently did not receive the support required to do so (Chapter 9).


Increasing but limited attention to protect healthcare workers          


Whereas the Global Programme’s attention to patients’ risks fell over time, its attention to healthcare workers’ risks increased. The Programme’s first proposal for national AIDS control programs, prepared in 1986, does not mention healthcare workers’ risks to acquire HIV from patients.[lxxvii] Two years later, the Global Programme’s 1988 ‘Guidelines for nursing management of people infected with human immunodeficiency virus (HIV)’ recommended infection control to protect both nurses and patients.[lxxviii] From 1987, the government of Uganda addressed healthcare-worker safety through short-term training and through the provision of gloves and other protective wear, ‘especially for midwives and traditional birth attendants.’ However, ‘the gap between demand and supply remains wide…’[lxxix] The situation for healthcare workers in Uganda was better than in most African countries.


Competing priorities to extend invasive healthcare   


During the 1980s, WHO, UNICEF, and foreign donors were heavily committed to programs to extend immunizations for children and prenatal and delivery care for women. Such vertical programs (that focused on specific interventions) generated incentives to overlook systemic issues such as infection control. For managers and staff of vertical programs, concerns about infection control threatened their efforts to extend interventions as fast as possible at the lowest possible cost.

Among vertical programs, the Expanded Programme on Immunization (EPI) paid the most attention to infection control. As EPI’s managers struggled to expand coverage in Africa, the surfacing of the AIDS epidemic forced them to reconsider injection practices. EPI decisions during 1985-88 provide a good example of the tension between vertical programs and infection control.

In 1986, an EPI policy document observed, ‘Countries which for many years have tolerated the use of unsterile techniques for immunization and other injections are now faced with rising concerns about the risks which such practices entail.’[lxxx] The comment applies as well to EPI, WHO, and UNICEF managers.

A 1985 meeting of WHO and UNICEF Regional Directors noted only 20 percent coverage of EPI immunizations in the Africa region, and asked for new initiatives to reach full coverage by 1990.[lxxxi] The report from the meeting did not mention HIV. Subsequently, WHO’s Regional Committee for Africa declared 1986 to be African Immunization Year. During 1986, EPI delivered 62 million immunizations in Africa.[lxxxii]

A 1986 EPI report noted that ‘the practice of multiple immunizations with a single needle and syringe (or with multiple needles and a single syringe) still occurs in immunization programmes,’ and exhorted ‘These practices must be stopped [emphasis in original] and the “each child – one sterile syringe – one sterile needle” rule meticulously followed.’[lxxxiii] However, another 1986 WHO and UNICEF document left a loophole for BCG immunization: ‘The same syringe and needle are often used to administer BCG, but the only acceptable practice here is to flame the needle between injections.’[lxxxiv]

Although EPI managers asked for safe injections, they did not insist. In 1987, a ‘Joint WHO/UNICEF statement on immunization and AIDS’ envisioned a choice between unsafe injections for immunization and no immunizations: ‘The potential for spread of HIV infection in childhood immunization sessions is low even where sterilization practices are below standard,’ while ‘Immunization programmes in developing countries are now preventing almost a million deaths a year…’ Having posed this choice, without considering safe injections as a third option, WHO and UNICEF opted for unsafe immunizations: ‘Halting immunization efforts because of the fear of AIDS would increase deaths among children, while doing little to stop HIV transmission.’[lxxxv]

Mann et al. in Kinshasa attributed many HIV infections in children to medical injections, but they pulled their punches when it came to immunizations. They argued from their data that ‘childhood vaccination…was not associated with HIV seropositivity.’[lxxxvi] This was a weak argument. Because only about 1 in 10 injections were immunizations, and because most children had received injections for immunization, it would have been difficult from their data to see a link between immunizations and HIV infection even if immunizations had infected several of the 44 HIV-positive children in their study. In 1988, Lepage and colleagues similarly asserted: ‘The risk of HIV contamination [through immunization injections] is low if any, and should not compromise the immense benefit that wide-spread immunization campaigns have on children’s health.’[lxxxvii]

By October 1988, EPI had increased immunization coverage among African children to 56 percent for BCG, 40 percent for measles, and 38 percent for three doses of DPT (diphtheria, pertussis, tetanus), and 23 percent of pregnant women received two doses of tetanus vaccine.[lxxxviii] On the other hand, limited commitment to safe injections brought limited results. In 1988, EPI’s Global Advisory Group reported: ‘The 1986 recommendation of a single sterile needle and syringe is now widely adopted as national EPI policy but is proving hard to realize in the field.’[lxxxix]


African public awareness and response to HIV risks in blood exposures       


Some early AIDS prevention programs in Africa warned the public about blood exposures. These warnings likely had the biggest impact on public awareness of risks in countries with the worst epidemics, where people recognized AIDS deaths. In Uganda, for example, where early AIDS control messages warned the public to avoid injections, 2/3rds of mothers interviewed in 1989 said that injections were a risk to transmit HIV.

However, by the late 1980s, public health managers in Africa pulled back on efforts to warn the public about risks to contract HIV from healthcare. In Uganda, for example, public health managers linked a slowdown in the expansion of the EPI program in 1988 to mothers’ worries about the safety of reused glass syringes. At the time, health experts recognized that ‘syringes and needles are routinely reused’ in Uganda. Even so, ‘radio messages about AIDS in Uganda have been modified to ensure that parents are aware of the safety of immunizations.’[xc] Ironically, EPI managers agreed with Ugandan mothers that injections with reused glass syringes were not safe, and had already initiated efforts to shift EPI support for injections worldwide to auto-destruct syringes, that would break after one use (see Chapter 9).

As illustrated in Uganda, public health managers have considered too much public awareness of risks in healthcare to be a problem. Immunization programs that accepted common reuse of syringes and needles without sterilization, and public messages that such injections were safe, no doubt “educated” both healthcare providers and the public that such dangerous practices were safe enough. The consequences of this mis-education go far beyond whatever infections might have been transmitted through unsafe immunizations. Even so, some people aware of risks avoided blood exposures. In Uganda, for example, many people kept syringes at home, which they brought to clinics when they wanted an injection. Africans alert to risks no doubt found other ways to reduce their blood exposures during healthcare and cosmetic services.


1988 consensus on risks for HIV in generalized epidemics  


From 1984, as soon as tests for HIV infection came available, researchers in Africa began to study the possibility that blood exposures in healthcare transmitted HIV. This was a far cry from what had happened after tests for hepatitis B infection were developed at the end of the 1960s – through 1988, no study had reported information on medical injections as risks for hepatitis B infection in African children.

Even so, research on risks for HIV infection did not resolve key questions. After several early studies published unwelcome results suggesting that injections accounted for a lot of HIV infections in African children and adults, there were no follow-up studies to confirm or refute these results. Instead, researchers avoided the issue. In 1988, leaders of the world’s response to Africa’s AIDS epidemic acknowledged that ‘The role of certain practices in traditional and Western medicine, such as the reuse of needles, razor blades, or other instruments, in the spread of HIV infection has not been well documented…’[xci]

In 1988, WHO asked a panel of experts to estimate the proportions of HIV infections in Africa and the Caribbean from all of the various routes of transmission. The panel estimated that heterosexual partners accounted for 80 percent of infections, mother-to-child transmission for 10.8 percent, blood tranfusions for 6 percent, MSMs and IDUs for 1.6 percent, and medical injections and other blood exposures for 1.6 percent.[xcii] If these and other similar estimates had been recognized as hypotheses – which they were – they could have guided and stimulated research to prove or to disprove them. But this did not happen. Instead, the scientific community treated such estimates as established facts. For example, an editorial in the Lancet in 1988 misrepresented available evidence to claim that[xciii]


…epidemiological studies have shown that transfusion of HIV-infected blood or the use of unsterile needles or other skin-piercing instruments within the health system or as part of traditional healing (or other) practices can account for only a small proportion of HIV infections [in Africa].


From 1988, funders, journals, and colleagues accepted that researchers looking at risks for HIV infection in Africa did not have to ask about medical injections or other blood exposures in healthcare or cosmetic services. For example, a 1988 review of AIDS in sex workers did not mention risks of contracting HIV from unsterile medical instruments in Africa or elsewhere – but did note that sex workers who were IDUs in the US and Europe might contract HIV infections from needles and syringes reused to inject illegal drugs.[xciv]

The consensus to ignore HIV transmission during healthcare released HIV prevention programs in countries with generalized epidemics from having to do anything about common lapses in standard precautions during healthcare. As far as AIDS experts were concerned, it was none of their business.

[i] Chin J, Mann JM. ‘The global patterns and prevalence of AIDS and HIV infection’, AIDS, 1988, 2 (suppl 1): S247-52. p. S250.

[ii] Sato PA, Chin J, Mann JM. ‘Review of AIDS and HIV infection: Global epidemiology and statistics’, AIDS, 1989, 3 (suppl 1): S301-7.

[iii] CDC. HIV/AIDS Surveillance Report: AIDS cases reported through December 1988. Atlanta: CDC, 1989.

[iv] Chin J, Mann JM. ‘The global patterns’. p. S250.

[v] Table 5.1, WHO’s estimates for HIV prevalence are from: Sato PA et al. ‘Review of AIDS and HIV infection’.

[vi] Table 5.1, reports from surveys, by country, are as follows: Cote d’Ivoire – Benoit SN, Gershy-Damet GM, Coulibaly A, et al. ‘Seroprevalence of HIV infection in the general population of the Cote d’Ivoire, West Africa’, J Acquir Immune Defic Syndr, 1990, 2: 1193-6; Guinea-Bissau – Piedade J, Venenno T, Prieto E, et al. ‘Longstanding presence of HIV-2 infection in Guinea-Bissau (West Africa)’, Acta Trop, 2000, 76: 119-24; CAR and Congo – Merlin M, Josse R, Trebucq A, et al. ‘Surveillance epidemiologique du syndrome d’immunodepression acquise dans six etats d’Afrique Centrale’, Med Trop, 1988, 48: 381-9; Burundi – Ministry of Health (Burundi). ‘Surveillance Epidemiologique du VIH/SIDA/MST’, unpublished document, Ministry of Health, 1992, reported in: US Census Bureau. HIV/AIDS Surveillance Data Base, June 2003 release. Washington, DC: US Census Bureau, 2003; Rwanda: Bizimungu C, Ntilivamunda A, Tahimana M, et al. ‘Nationwide community-based serological survey of HIV-1 and other human retrovirus infections in a Central African country’, Lancet, 1989, i: 941-3; Uganda – Kangeya-Kayondo JF, Amaana A, Naamara W. ‘Anti-HIV seroprevalence in adult rural populations of Uganda and its implications for preventive strategies’, 5th Int Conf AIDS, Montreal 4-6 June 1989, poster TAP 11 (the US Census Bureau’s HIV/AIDS Surveillance Data Base, June 2003 release, reports data from this abstract that are not in the published abstract); Naamara W. ‘Official Release of the National Serosurvey for Human Immunodeficiency Virus (HIV) in Uganda’, unpublished letter to Minister of Health (Uganda), 1990, reported in: US Census Bureau. HIV/AIDS Surveillance Data Base, June 2003 release.

[vii] Table 5.1, HIV prevalence in women attending urban antenatal clinics is the average of medians for 1985-86 and 1987-88 from: UNAIDS, WHO. Cote d'Ivoire: Epidemiological Fact Sheet on HIV/AIDS and Sexually Transmitted Infections, 2000 update. Geneva: WHO, 2000; and similar Fact Sheets for other countries; except for Guinea-Bissau, for which data are from: US Census Bureau. HIV/AIDS Surveillance Data Base, June 2003 release.

[viii] Sato PA et al. ‘Review of AIDS and HIV infection’. p. S304; Chin J, Sato PA, Mann JM. ‘Projections of HIV infections and AIDS cases to the year 2000’, Bull WHO, 1990, 68: 1-11.

[ix] World Bank. African  Development Indicators 1998/99. Washington DC: World Bank, 1998.

[x] Desmyter J, Goubau P, Chamaret S, et al. ‘Anti-LAV/HTLV-III in Kinshasa mothers in 1970 and 1980 [abstract]’, 2nd Int Conf AIDS, Paris 23-25 June 1986, abstract s17g.

[xi] Nzilambi N, De Cock KM, Forthal DN, et al. ‘The prevalence of infection with human immunodeficiency virus over a 10-year period in rural Zaire’, N Eng J Med, 1988, 318: 276-9.

[xii] Merlin M et al. ‘Surveillance epidemiologique’.

[xiii] Musinguzi J. ‘Women and AIDS: Are women at increased risk and what are the implications?’ East Afr Med J, 1993, 70: 245-8; Mugerwa RD, Marum LH, Serwadda D. ‘Human immunodeficiency virus and AIDS in Uganda’, East Afr Med J, 1996, 73: 20-6.

[xiv] Berkley S, Naamara W, Okware S, et al. ‘AIDS and HIV infection in Uganda – Are more women infected than men?’, AIDS, 1990, 4: 1237-42.

[xv] Killewo J, Nyamuryekunge K, Sandstrom A, et al. ‘Prevalence of HIV-1 infection in the Kagera Region of Tanzania: A population-based study’, AIDS, 1990; 4: 1081-5.

[xvi] US Census Bureau. HIV/AIDS Surveillance Data Base, June 2003 release.

[xvii] Sato PA et al. ‘Review of AIDS and HIV infection’. p. S304.

[xviii] Simoes EAF, Babu PG, John TJ, et al. ‘Evidence for HTLV-III infection in prostitutes in Tamil Nadu (India)’, Indian J Med Res, 1987, 85: 335-8.

[xix] Beyrer C. War in the Blood: Sex, politics and AIDS in Southeast Asia. London: Zed Books, 1998.

[xx] Wright NH, Vanichseni S, Akarasewi P, et al. ‘Was the 1988 HIV epidemic among Bangkok’s injecting drug users a common source outbreak?’, AIDS, 1994, 8: 529-32.

[xxi] Pape JW, Liautaud B, Thomas F, et al. ‘The acquired immunodeficiency syndrome in Haiti’, Ann Intern Med, 1985, 103: 674-8.

[xxii] UNAIDS, WHO. Cote d’Ivoire: Epidemiological Fact Sheet on HIV/AIDS and Sexually Transmitted Infections, 2000 update; similar UNAIDS’ Fact Sheets for other countries, except that for Haiti the data are from the 2002 update.

[xxiii] US Census Bureau. HIV/AIDS Surveillance Data Base, June 2003 release.

[xxiv] N’Galy B, Ryder RW. ‘Epidemiology of HIV infection in Africa’, J Acquir Immune Defic Syndr, 1988, 1: 551-8.

[xxv] Mann JM, Francis H, Quinn TC, et al. ‘HIV seroprevalence among hospital workers in Kinshasa, Zaire’, JAMA, 1986, 256: 3099-102.

[xxvi] Ryder RW, Ndilu M, Hassig SE, et al. ‘Heterosexual transmission of HIV-1 among employees and their spouses at two large businesses in Zaire’, AIDS, 1990, 4: 725-32.

[xxvii] Melbye M, Njelesani EK, Bayley A, et al. ‘Evidence for heterosexual transmission and clinical manifestations of human immunodeficiency virus infection and related conditions in Lusaka, Zambia’, Lancet, 1986, ii: 1113-15.

[xxviii] Van de Perre P, Clumeck N, Carael M, et al. ‘Female prostitutes: A risk group for infection with human T-cell lymphotropic virus type III’, Lancet, 1985, ii: 524-7.

[xxix] CDC. ‘Acquired immunodeficiency syndrome (AIDS) weekly surveillance report – United States, 31 December 1984’. Atlanta: CDC, 1984.

[xxx] Johnson AM, Laga M. ‘Heterosexual transmission of HIV’, AIDS, 1988, 2 (suppl 1): S49-56.

[xxxi] WHO. ‘Workshop on AIDS in Central Africa, Bangui, Central African Republic, 22-25 October 1985.’ Geneva: WHO, 1986. Doc. no. WHO.CDS.AIDS/85.1.

[xxxii] Clumeck N, Carael M, Rouvroy D, et al. ‘Heterosexual promiscuity among African patients with AIDS [letter]’, N Eng J Med, 1985, 313: 182.

[xxxiii] Hrdy DB. ‘Cultural practices contributing to the transmission of human immunodeficiency virus in Africa’, Rev Infect Dis, 1987, 9: 1109-19. p. 1112.

[xxxiv] Ryder RW et al. ’Heterosexual transmission of HIV-1’. p. 729.

[xxxv] Allen S, Tice J, Van de Perre P, et al. ‘Effect of serotesting with counselling on condom use and seroconversion among HIV discordant couples in Africa’, BMJ, 1992, 304: 1605-9.

[xxxvi] Piot P, Plummer FA, Rey M-A, et al. ‘Retrospective seroepidemiology of AIDS virus infection in Nairobi populations’, J Infect Dis, 1987, 155: 1108-12.

[xxxvii] Plummer FA, Simonsen JN, Cameron DW, et al. ‘Cofactors in male-female sexual transmission of human immunodeficiency virus type 1’, J Infect Dis, 1991, 163: 233-9; Cameron DW, Simonsen JN, D’Costa LJ, et al. ‘Female to male transmission of human immunodeficiency virus type 1: Risk factors for seroconversion in men’, Lancet, 1989; ii: 403-7.

[xxxviii] Vachon F, Coulaud JP, Katlama C. ‘Epidemiologie actuelle du syndrome d’immunodeficit acquis en dehors des groupes a risque’, Presse Med, 1985, 14: 1949-50.

[xxxix] Wyatt HV. ’Injections and AIDS’, Trop Doctor, 1986, 16: 97-8.

[xl] Wykoff RF. ‘Female-to-male transmission of AIDS agent [letter]’, Lancet, 1985; ii: 1017-18.

[xli] Cohen JB, Wofsy C, Gill P, et al. ‘Antibody to human immunodeficiency virus in female prostitutes’, MMWR, 1987, 36: 157-61.

[xlii] Barre-Sinoussi F, Nugeyre MT, Chermann JC. ‘Resistance of AIDS virus at room temperature [letter]’, Lancet, 1985; ii: 721-2.

[xliii] Resnick L, Veren K, Salahuddin Z, et al. ‘Stability and inactivation of HTLV-III/LAV under clinical and laboratory conditions’, JAMA, 1986, 255: 1887-91.

[xliv] CDC. ‘Recommendations for prevention of HIV transmission in health-care settings’, MMWR, 1987, 36 (suppl 2S): 3S-18S.

[xlv] Konde-Lule JK, Berkley SF, Downing R. ‘Knowledge, attitudes and practices concerning AIDS in Ugandans’, AIDS, 1989, 3: 513-18. p. 517.

[xlvi] Pape JW, Liautaud B, Thomas F, et al. ‘The acquired immunodeficiency syndrome in Haiti’, Ann Intern Med, 1985, 103: 674-8. p. 677.

[xlvii] Berglund O, Beckman S, Grillner L, et al. ‘HIV transmission by blood transfusion in Stockholm 1979-85: nearly uniform transmission from infected donors’, AIDS, 1988, 2: 51-4; Colebunders R, Ryder R, Francis H, et al. ‘Seroconversion rate, mortality, and clinical manifestations associated with the receipt of a human immunodeficiency virus-infected blood transfusion in Kinshasa, Zaire’, J Infect Dis, 1991, 164: 450-6.

[xlviii] Wormser GP, Joline C, Sivak SL, et al. ‘Human immunodeficiency virus infections: considerations for health care workers’, Bull New York Acad Med, 1988, 64: 203-15.

[xlix] Mann JM, Francis H, Davachi F, et al. ‘Risk factors for human immunodeficiency virus seropositivity among children 1-24 months old in Kinshasa, Zaire’, Lancet, 1986, ii: 654-7. p. 656.

[l] Lepage P, Van de Perre P, Carael M, et al. ‘Are medical injections a risk factor for HIV in children?’, Lancet, 1986, ii: 1103-4; Lepage P, Van de Perre P. ‘Nosocomial transmission of HIV in Africa: What tribute is paid to contaminated blood transfusions and medical injections?’, Infect Control Hosp Epidemiol, 1988, 9: 200-3.

[li] Mann JM et al. ‘HIV seroprevalence among hospital workers’.

[lii] Van de Perre P, Carael M, Nzaramba D, et al. ‘Risk factors for HIV seropositivity in selected urban-based Rwandese adults’, AIDS, 1987, 1: 207-11; Killewo J et al. ‘Prevalence of HIV-1 infection in the Kagera Region of Tanzania’; Konde-Lule JK et al. ‘Knowledge, attitudes and practices’; Bassett MT, Latif AS, Katzenstein DA, et al. ‘Sexual behavior and risk factors for HIV infections in a group of male factory workers who donated blood in Harare, Zimbabwe’, J Acquir Immune Defic Syndr, 1992, 5: 556-9.

[liii] 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.

[liv] Plummer FA et al. ‘Cofactors in male-female sexual transmission’; Cameron DW et al. ‘Female to male transmission’, Lancet, 1989, ii: 403-7.

[lv] Bizimungu C et al. ‘Nationwide community-based serological survey’.

[lvi] Packard RM, Epstein P. ‘Epidemiologists, social scientists, and the structure of medical research on AIDS in Africa’, Soc Sci Med, 1991, 33: 771-83. p. 780.

[lvii] 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: 148-61.

[lviii] WHO. ‘Blood and Blood Products: Report by the Director-General, 14 January 1987.’ Geneva: WHO, 1987. Doc. no. EB79/7, Add. 1.

[lix] Mann JM, Francis H, Davachi F, et al. ‘Human immunodeficiency virus seroprevalence in pediatric patients 2 to 14 years of age at Mama Yemo Hospital, Kinshasa, Zaire’, Pediatrics, 1986, 78: 673-7.

[lx] Schneider WH, Drucker E. ‘Blood transfusions in the early years of AIDS in sub-Saharan Africa’, Am J Pub Health, 2006, 96: 984-94.

[lxi] Navarro V, Roig P, Nieto A, et al. ‘A small outbreak of HIV infection among commercial plasma donors [letter]’, Lancet, 1988, ii: 42.

[lxii] Avila C, Stetler HC, Sepulveda J, et al. ‘The epidemiology of HIV transmission among paid plasma donors, Mexico City, Mexico’, AIDS, 1989, 3: 631-3.

[lxiii] del Rio C, Sepulveda J. ‘AIDS in Mexico: Lessons learned and implications for developing countries’, AIDS, 2002, 16: 1445-57. pp. 1446-7.

[lxiv] ‘AIDS in Africa’, Lancet, 1987, ii: 192-4. p. 193.

[lxv] Quinn TC, Mann JM, Curran JW, Piot P. ‘AIDS in Africa: An epidemiologic paradigm’, Science, 1986, 234: 955-63. p. 962.

[lxvi] WHO. ‘Workshop on AIDS in Central Africa, Bangui, Central African Republic, 22-25 October 1985.’ Geneva: WHO, 1986. Doc. no. WHO.CDS.AIDS/85.1. pp. 11, 12.

[lxvii] WHO. ‘Global WHO Strategy for the Prevention and Control of Acquired Immunodeficiency Syndrome: Projected Needs for 1986-1987.’ Geneva: WHO, 1986. Doc. no. AIDS/CPA86.2. pp. 22, 27.

[lxviii] Okware SI. ‘Towards a national AIDS-control programme in Uganda’, West J Med, 1987, 147: 726-9.

[lxix] Berkley S, Weeks M, Barenzi J. ‘Immunization and fear of AIDS [letter]’, Lancet, 1990, i: 47-48.

[lxx] WHO. ‘Guidelines for the development of a national AIDS prevention and control program’, WHO AIDS series 1. Geneva: WHO, 1988. Available at: (accessed 4 September 2007).

[lxxi] WHO. ‘Blood and Blood Products’.

[lxxii] Colebunders R et al. ‘Seroconversion rate, mortality, and clinical manifestations’. p. 450.

[lxxiii] ‘Acquired Immune Deficiency Syndrome (AIDS): WHO meeting and consultation on the safety of blood products’, in: Petricciani JC, Gust ID, Hoppe PA, et al. (eds), The Safety of Blood and Blood Products. Geneva: WHO, 1987. pp. 355-359. The quote is from page 356.

[lxxiv] WHO. ‘Blood and Blood Products: Report by the Programme Committee of the Executive Board.’ Geneva: WHO, 1987. Doc. no. EB79/7. pp. 4-5.

[lxxv] Fleming AF. ‘HIV and blood transfusion in sub-Saharan Africa’, Transfu Sci, 1997, 18: 167-79.

[lxxvi] WHO. ‘Report of the Global Blood Safety Initiative Meeting, Geneva, 16-17 May 1988.’ Geneva: WHO, 1988. Doc. no. WHO/GPA/DIR/88.9.

[lxxvii] WHO. ‘Global WHO Strategy’.

[lxxviii] WHO. ‘Guidelines for nursing management of people infected with human immunodeficiency virus (HIV)’, WHO AIDS series 3. Geneva: WHO, 1988. Available at: (accessed 4 September 2007).

[lxxix] Okware SI. ‘Towards a national AIDS-control programme in Uganda’. p. 729.

[lxxx] WHO. ‘Immunization Policy.’ Doc. no. WHO/EPI/GEN/86/7 Rev 1. Geneva: WHO, 1986. p. 8.

[lxxxi] WHO. ‘Final Report, WHO/UNICEF Regional Directors’ Consultation, Brazzaville 3-4 September 1985.’ Geneva: WHO, 1985. Doc. no. AFR/EXM/10.

[lxxxii] WHO. ‘Report of the Expanded Programme on Immunization, Global Advisory Group meeting, Washington DC, 9-13 November 1987.’ Geneva: WHO, 1988. Doc. no. WHO/EPI.GEN/88.1.

[lxxxiii] LaForce FM. ‘Immunization of children infected with human immunodeficiency virus’, Geneva: WHO, 1986. Doc. no. WHO/EPI/GEN/86/6 Rev. 1. p. 4.

[lxxxiv] WHO. ‘Selection of Injection Equipment for the Expanded Programme on Immunization’, EPI Technical Series. Geneva: WHO, 1986. Doc. no. WHO/UNICEF/EPI.TS/86.2. p. 4.

[lxxxv] ‘Expanded Programme on Immunization: Joint WHO/UNICEF statement on HIV and immunization’, Wkly Epidemiol Rec, 1987, 62: 53-4. p. 53.

[lxxxvi] Mann JM et al. ‘Risk factors for human immunodeficiency virus seropositivity’. p. 656.

[lxxxvii] Lepage P, Van de Perre P. ‘Nosocomial transmission of HIV in Africa’. p. 203.

[lxxxviii] WHO. ‘Report of the Expanded Programme on Immunization Global Advisory Group Meeting, Abidjan, Cote d’Ivoire, 17-21 October 1988’, Geneva: WHO, 1989. Doc. no. WHO/EPI/GEN/89.1.

[lxxxix] Ibid. p. 71.

[xc] Berkley S et al. ‘Immunization and fear of AIDS [letter].’

[xci] Piot P, Plummer FA, Mhalu FS, et al. ‘AIDS: An international perspective’, Science, 1988, 239: 573-9. p. 578.

[xcii] Chin J et al. ‘Projections of HIV infections and AIDS cases’.

[xciii] ‘AIDS in sub-Saharan Africa’, Lancet, 1988, i: 1260-1. p. 1260.

[xciv] Day S. ‘Prostitute women and AIDS: Anthropology’, AIDS, 1988, 2: 421-8.