The African AIDS Epidemic. John Iliffe

of childbearing age were raped each year. These were only the most blatant forms of coercion. More took the form of steady pressure rather than violence. Many men and women believed that a man who had given bridewealth for a woman, spent money on her, or received encouragement from her had a right to sex regardless of her wishes: ‘Once you have kissed each other that means you are preparing for sex. If she refuses at that point you must just force her.’⁶⁵ Not all needed to be forced. For some poor young women, their sexuality might be their only means of survival or of acquiring coveted goods and other benefits. A study in Cape Town found that about 20 per cent of teenage women reported sex for money or presents. ‘If he wants a woman like me, a man must pay,’ one said. ‘Forget about marriage . . . that was something for our mothers and grannies, it’s not for us.’⁶⁶ Yet even young women eager to be regte, steady girlfriends with hope of marriage, were equally at risk of infection, for it commonly implied unprotected sex. In the mid 1990s one-third of South Africa’s teenage women bore a child. Ten per cent of these women had HIV. By the age of 25, one-quarter would be infected.⁶⁷

      In its silent origins, its rapid expansion, its association with mobility, its exploitation of gender inequality, and its growing concentration among the poor, South Africa’s epidemic was an extreme version of a continental pattern, much as Apartheid had been an extreme version of a wider colonial order. The epidemic that had begun two decades earlier close to the equatorial forest had culminated at the southern extremity of the continent. From that extremity the counter-attack would eventually begin.

      6

       The Penetration of the West

      The penetration of HIV-1 from the equatorial region into West Africa differed markedly from its expansion to the east and south. Except in Côte d’Ivoire, it was more gradual and less complete, reaching in the early 2000s prevalences only one-fifth or one-sixth of the highest elsewhere. The reasons for this are unclear but probably include obstacles to overland mobility from east to west, the wider economic opportunities open to West African women in towns, widespread male circumcision, relatively low HSV-2 prevalences, and the barriers to infection presented by Islamic moral and marital patterns. Another difference, of less certain relevance, was that when the HIV-1 virus entered West Africa, it found HIV-2 already established.

      As a human disease, HIV-2 was probably older than HIV-1. It was closely related to the simian immunodeficiency virus found in sooty mangabey monkeys (SIVsm) living only in the West African forest region between the Casamance River in Senegal and the Sassandra River in Côte d’Ivoire, which was also the endemic location of the human virus. HIV-2 shared some 70 per cent of its genome with SIVsm but only about 42 per cent with HIV-1. Indeed, some of the eight groups of HIV-2 known in 2004 were more like SIVsm than they were like one another. This was because SIVsm was very widespread and diverse (although completely harmless) in sooty mangabey monkeys and because each HIV-2 group was probably the result of a separate transmission from a monkey.¹ Of the eight groups, six had failed to establish themselves in human beings, having infected only seven known cases between them. Of the two more successful, group A was the more common throughout the coastal region west of Côte d’Ivoire, while group B was found chiefly in Côte d’Ivoire and Ghana, although scattered cases of both existed elsewhere.² A study using molecular clock techniques estimated that the most recent common ancestor of group A existed in 1940±16 and of group B in 1945±14.³ Yet, given the high prevalence of SIV among sooty mangabeys, their close interaction with human beings, and the frequency of twentieth-century transmissions, similar transmissions had probably taken place in earlier centuries.

      This was more likely with HIV-2 than HIV-1 because HIV-2 was a less virulent and visible disease in human beings, possibly because its progenitor was so fully adapted to monkeys. HIV-2 was about three times more difficult than HIV-1 to transmit through sexual intercourse and at least ten times more difficult to pass from mother to child.⁴ Mortality from HIV-2 may have been only about one-third of that from HIV-1, for viral loads were generally lower, those infected were often older, and progression to Aids might take on average as much as 25 years, so that many of those infected never reached that stage, although if they did the final illness was similar.⁵ Given that the opportunistic infections fatal to Aids patients were often those common to the local disease environment, it was understandable that HIV-2 passed unnoticed until 1985, when researchers investigating the existence of HIV-1 in Senegal discovered the other virus almost by chance.⁶ This probably explains why HIV-1 and HIV-2 appear to have emerged virtually simultaneously: the appearance is an optical illusion.

      Once the discovery was made, retrospective testing of the earliest stored blood for HIV-2 antibodies revealed an intriguing pattern.⁷ Apart from one obscure reference to an alleged case in Mali in 1957, the earliest may have been a Portuguese man who had lived in Guinea-Bissau between 1956 and 1966. Other infections there during the 1960s are also recorded. Five cases were found in Côte d’Ivoire during the 1960s. Stored blood taken in 1967 also revealed two cases each in Nigeria and Gabon, both outside the range of sooty mangabeys and presumably infected through travel. They were followed in the 1970s by infections from Mali, Senegal, and Angola, the last probably transmitted through the movement of Portuguese troops from Guinea-Bissau. By the 1980s scattered cases were reported from many parts of western Africa, often from the countryside, suggesting a low-intensity disease much like HIV-1 in its pre-epidemic days in western equatorial Africa. In Guinea-Bissau, however, the liberation war of 1960–74, the presence of Portuguese troops, the movement of refugees, and perhaps especially the widespread use of injections by Portuguese military doctors appear to have bred a localised and probably unique epidemic.⁸ Hospitals in Portugal later treated many cases contracted in Guinea-Bissau at this time. A study in Bissau town in the late 1990s showed that levels of infection peaked among men in their sixties and women in their fifties who would have been sexually most active during the 1960s. Prevalence there among men who had served in the Portuguese army was 23 per cent; among the nineteen women who had had sex with white men it was 37 per cent. This wartime legacy gave Guinea-Bissau much the highest prevalence of HIV-2. In the mid 1980s, 26 per cent of paid blood donors there tested positive, as did 8.6 per cent of Bissau’s pregnant women and 36.7 per cent of its sex workers in 1987.⁹ Ten years later HIV-2 infected 13.5 per cent of people over 35 living on the outskirts of the town. High levels were also reported in rural areas and spilled over (largely through migrant sex workers) to southern Senegal and The Gambia.¹⁰ Yet the epidemic never spread beyond this region. That would presumably have required a virus more infectious than HIV-2.

      HIV-1 was such a virus. Its arrival in West Africa (as distinct from western equatorial Africa) is difficult to trace but possibly took place in about 1980, slightly after its appearance in East and Central Africa. A claim to have discovered one case in stored blood taken in Burkina in 1963 can almost certainly be dismissed. A Malian migrant who had never visited equatorial Africa died in Paris in 1983 with Aids-like symptoms, although this could as well have been HIV-2 as HIV-1. Ghanaian doctors came to believe that they had seen Aids cases as early as 1981, but no details are available and HIV-2 would again have been possible.¹¹ Otherwise, the earliest evidence comes from Côte d’Ivoire. Retrospective tests on stored blood taken there between 1970 and 1983 all proved negative. Adult mortality in Abidjan declined until 1985, the year when its first Aids cases were diagnosed, and then began to increase rapidly. In 1985, 38 of 79 sex workers were found to be infected there, together with 10 of 71 in the northern Ivoirian town of Korhogo. A year later HIV-1 prevalence was 3.0 per cent among pregnant women and 4.9 per cent among hospital staff in Abidjan. French researchers concluded that the first HIV infections there probably took place in about 1980.¹² Observers suggested at the time that the city’s sex workers might have been infected by European tourists, but this is unlikely because the B subtype of HIV-1 prevalent in Europe did not become established in Abidjan or elsewhere in West Africa. Rather, the dominant strain came to be CRF02_AG, the circulating recombinant form rare in the DR Congo but common in Cameroun and Gabon, implying a northward diffusion comparable to the eastward diffusion of subtypes A and D into East Africa – a diffusion that in West Africa could have been carried in the first instance along the coast by sex workers and their clients moving between Libreville, Douala, and Abidjan. CRF02_AG became dominant among