The beginning

During the last decade, much progress has been made in understanding the events that triggered the transformation of HIV-1 from a rare zoonosis of hunters in the forests of Central Africa into the worst pandemic of modern times. The source of HIV-1 was the Pan troglodytes troglodytes chimpanzee, whose simian immunodeficiency virus (SIV_cpz) virus is phylogenetically related to its human counterpart.1 2 The first person to be infected must have lived somewhere within the habitat of this ape, in Cameroon, the Central African Republic, Gabon, Equatorial Guinea, Congo-Brazzaville, or small parts of the Democratic Republic of Congo and Angola that lie to the north of the Congo River. The cross-species transmission event occurred during the first three decades of the 20th century.3

After some local transmission between humans, the virus eventually managed to move out of the bush and enter the binational conurbation of Léopoldville (currently Kinshasa) and Brazzaville, the capital cities of Belgian Congo and French Equatorial Africa, where it flourished and differentiated. These two cities are the areas of the world with the broadest genetic diversity of HIV-1 group M, harbouring all the subtypes as well as many sub-subtypes and recombinants.4−7 The diversity of HIV-1 in a given place is influenced by the duration of its presence and by the efficacy of its transmission, and HIV-1 diversity in Kinshasa 25 years ago was already far more complex than that found today anywhere else in the world.8 Molecular clocks have shown that in Léopoldville, the number of infected persons remained stable for several decades but started increasing exponentially in the early 1950s.3 This paper presents a personal view of what could have happened for the dynamics of HIV-1 to change so drastically.

Les Femmes Libres

By the early 1930s, Léopoldville (‘Léo’; population: 40 000) had become the economic hub of Central Africa, the capital of its most prosperous colony, a booming town which benefited from its geographical location: the terminus of all navigation on the Congo basin and the departure point of the railway that carried goods to the Atlantic coast. Brazzaville, a few kilometres away, on the other side of the river, remained a sleepy colonial town with 18.000 inhabitants, the capital of a poor and under-populated federation. World War II accelerated this contrast between Brazzaville and Léopoldville, whose population now doubled every 5 years. Because of colonial policies, a substantial gender imbalance developed in Léo, with two adult men for each woman, a ratio which was much higher among the unmarried fraction of the population and in some ethnic groups.9 The consequence was predictable: large-scale prostitution. In 1928, out of 6000 adult women in Léo, only 358 were living with a husband, and a tax was levied on the ‘healthy women theoretically living alone’, a pragmatic approach to sex work. Aware that these women spread sexually transmitted pathogens, a governor's edict compelled them to be examined monthly. A health card had to be stamped each time.10

This fiscal category of ‘femmes libres’ or ‘free women’ (not financially dependent on a man) encompassed genuine sex workers, petty traders and the grey areas in between. The pattern of sex work radically differed from the one that was to appear decades later. Free women had a few regular clients to whom they provided diversified services: a good meal, laundry, hair care, conversation and sex. Their clients gave them some regular income in return for sexual favours that were granted regularly over a period of time rather than a fee for each act of intercourse. These relationships were understood by both parties not to be exclusive and there was no intent of having children (figure 1).11–13

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Entertainment in a bar of Léopoldville (early 1950s).28

This soft prostitution, in essence a form of concurrent sexual partnerships, sufficed for the persistence of HIV-1 in Léo but probably not for the dramatic, exponential expansion that began in the 1950s, a decade during which the pattern of sex work did not change. Researchers have hypothesised that an epidemic of genital ulcer diseases (GUD), especially syphilis, fuelled the early (pre-1929) spread of HIV-1 in Belgian Congo's capital through their co-factor effect on transmission.14 Could the same mechanisms have also caused the expansion of the 1950s?

For their own good

In 1926, the Belgian Red Cross established a branch, Croix-Rouge du Congo. Three years later, the charity opened a Dispensaire Antivénérien (sexually transmitted disease (STD) clinic), in the Barumbu district of Léopoldville-Est. Later, a smaller satellite clinic was opened in Léo-Ouest.15 16 The STD clinics provided free care to male and female patients who presented spontaneously with a genital complaint or were referred by other healthcare facilities. Contact tracing generated additional workload as did, until 1952, the screening of migrants to Léo, who had to show up on arrival in order to comply with the health regulations. And there were thousands of asymptomatic free women (as defined by the tax department) who were compelled to come for screening at least a few times per year. At its peak, 32 000 such visits took place annually.

Tables 1 and 2 and figures 2 and 3 present the statistics for the two clinics together, but the Léo-Ouest clinic remained embryonic and 95% of the total number of cases and number of injections correspond to the work done in Léo-Est. Data on the incidence of syphilis or gonorrhoea, the number of women and men seen at least once during the year, the number of serological assays for syphilis and the number of injections administered were provided in the annual reports of Croix-Rouge du Congo and in the health services reports.16 17 The number of new cases of syphilis or gonorrhoea fluctuated substantially, reflecting the expansion of the capital's population and changes not just in the diagnostic strategies but also in the efforts to screen free women.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Total number of new cases of gonorrhoea and syphilis in men and women at the Dispensaires Antivénériens, Léopoldville, Belgian Congo.29

• Download figure
• Open in new tab
• Download powerpoint

Figure 3

Number of injections of various drugs (to men and women) and number of visits for free women at the Dispensaires Antivénériens, Léopoldville, Belgian Congo.29

Chancroid, donovanosis and lymphogranuloma venereum (LGV) remained uncommon.16 17 Perhaps because the clinical diagnosis of the various etiologies of GUDs is unreliable, no precise information was given on these other ulcerative diseases, but the 1953 Croix-Rouge report noted that ‘in Léopoldville, while there were a few cases of chancroid, the Nicolas-Favre disease, so common in the Equateur province, was absent’.16 Later on, 24 and 38 cases of chancroid were reported in 1957 and 1958, compared to 1 and 11 cases of LGV. For the whole colony (population 10.5 million), public and private healthcare institutions reported <1500 cases of chancroid and fewer than 3000 cases of LGV each year while the number of cases of ‘syphilis’ peaked at 103 458 in 1948.17

Given that individuals treated for syphilis or gonorrhoea had to attend repeatedly for prolonged courses of injectable drugs and for a lengthy follow-up, by the early 1950s, up to 1000 patients attended the four-room Léo-Est clinic every day. The medical officers estimated that they were providing more or less regular check-ups to 3500 free women. The clinic opened at 04:30 so that men could receive their treatments before going to work. Blood samples were sent to a reference laboratory; in 1954, 85 654 Bordet–Wasserman non-treponemal tests were performed.16

In retrospect, the treatment received by a substantial fraction of patients, apart from symptomatic men, was not very useful: debatable diagnoses and rather ineffective drugs. By the late 1940s, the medical officers began to wonder whether some of their efforts were futile. In the 1949–1954 reports, tables show that of the 7622 new diagnoses of syphilis, only 115, 111, and 52 patients had signs compatible with primary, secondary, and tertiary syphilis, respectively (table 1). All the others, 96% of the patients with so-called syphilis, were given injectable drugs merely because of a positive serology. Even today, the treponemal and non-treponemal serological assays do not discriminate between syphilis and yaws, the non-venereal cutaneous disease caused by Treponema pallidum subsp. pertenue. For both infections, the non-treponemal serology can remain positive at a low titre for a year or more. At that time, the medical orthodoxy was that patients needed to be treated repeatedly until their serology became completely negative. Given the high incidence of yaws in the rural areas where many of these individuals originated from, most cases with a positive serology probably corresponded to a past episode of yaws, which had often occurred during childhood. However, if that person was a free woman or a migrant and happened to be tested at the STD clinic, she/he was always considered to be syphilitic and received long courses of the intravenous arsenical Neo-Salvarsan (once a week) plus an intramuscular bismuthic or mercurial drug. On average, combining initial and subsequent treatments, each patient with ‘syphilis’ received 24 intravenous arsenical injections and 42 intramuscular injections of bismuthic/mercurial products.16

The precision of diagnoses among the free women was no better for gonorrhoea. The clinics could not cultivate gonococci and did simple staining of the vaginal secretions in women who were mostly asymptomatic. The existence of non-pathogenic Neisseria species was not yet discussed. Patients thought to have gonorrhoea were treated for up to 2 months with products that were supposed to combat the infection by triggering high fever, such as injections of milk, typhoid vaccine or an emulsion of dead gonococci. Starting in 1951, effective antibiotics (penicillin, sulphonamides or streptomycin) were given but only after this ‘preparatory’ course.

As a result of these diagnostic and therapeutic approaches, during the 1930s and 1940s, the STD clinics administered around 50 000 injections per year, about 60% of which were given intravenously. In the 1950s, during the post-war demographic boom, the number of injections fluctuated around 100 000 per year, peaking at an extraordinary level of 154 572 by 1953 (table 2). This decreased rapidly thereafter for several reasons, including the introduction of penicillin, a course of which required fewer injections.

It is generally difficult to gather information about the procedures used for sterilisation of syringes and needles during the colonial era. Here, however, it is instructive to read a paper about hepatitis in Léopoldville written by Dr Paul Beheyt, an internal medicine specialist at Hôpital des Congolais, the sole hospital for Congolese in Léo.18 The author distinguished ‘epidemic hepatitis’ (presumably, hepatitis A or E) from ‘inoculation hepatitis’; the latter was diagnosed when a patient developed jaundice 45–150 days after having received intravenous injections or transfusions. It is plausible that many of the cases that Beheyt referred to as ‘inoculation hepatitis’ were caused by the parenteral transmission of the hepatitis B virus, because acute hepatitis C is rarely severe enough for a patient to develop jaundice. Of the 69 cases of inoculation hepatitis diagnosed during 1951–1952, 32 had received intravenous arsenical drugs for the treatment of syphilis at the Léo-Est STD clinic, which corresponded to 1% of the syphilis cases treated during the same period by this institution.18 This measure of risk was greatly underestimated because the reference hospital presumably diagnosed only a fraction of all the cases of inoculation hepatitis in Léo, and the iatrogenic infection could only occur among patients who had not been infected with the hepatitis B virus earlier in their lives, 5% of the adults in Central Africa at the most.

As Beheyt wrote: ‘The Congo contains various health institutions where every day local nurses give dozens, even hundreds, of injections in conditions such that sterilisation of the needle or the syringe is impossible. At the Dispensaire Antivénérien in Léopoldville, on an average 300 injections are administered each day. The large number of patients and the small quantity of syringes available to the nursing staff preclude sterilisation by autoclave after each use. Used syringes are simply rinsed, first with water, then with alcohol and ether, and are ready for a new patient. The same type of procedure exists in all health institutions where a small number of nurses have to provide care to a large number of patients, with very scarce supplies. The syringe is used from one patient to the next, occasionally retaining small quantities of infectious blood, which are large enough to transmit the disease.’18

In 1955, the Croix-Rouge withdrew from running the STD clinics, and the colony's Department of Hygiene took over its clients and nurses. By the end of 1958, 4384 free women were registered, a substantial percentage, given that the medical officers had estimated that there were 5000–6000 of these in the city. The number of injections was drastically reduced. Long-acting penicillin replaced arsenical drugs for patients with syphilis, and the indications for treatment were tightened as physicians acknowledged that it was unnecessary to treat those who carried only a ‘serological scar’ due to a past, adequately treated, episode of yaws or syphilis. Cross-reactivity between these two treponemal infections was recognised, as was the inadequate specificity of stains to detect gonococci in women.19−21

The Department of Hygiene ran an annual medical census of the whole population of Léopoldville, during which every inhabitant was examined summarily to look out for tropical diseases, a substantial effort. Between 148 584 (1949) and 322 198 (1958) individuals were examined to detect a few dozen cases of sleeping sickness or leprosy. During the last years of the colonial rule, the health agents required every adult man to drop his pants; of the 99 446 men seen in 1958, 163 were found to have gonorrhoea, 335 had non-gonococcal urethritis while 44 had GUD.22 This rather energetic attitude had been driven by a long-standing preoccupation with the undesirable effects of STDs on fertility and population growth, which were prerequisites for the success of the colonial enterprise.

The incidence of tropical diseases, for which injectable drugs were massively used in the rural areas around Léo, remained minimal in the capital. After 1930, there were generally <100 cases of sleeping sickness or yaws diagnosed each year, either actively during the census or passively in health facilities. Through the systematic screening of migrants, sleeping sickness cases were identified and treated early on, which reduced the risk of transmission of the lethal parasite within Léo. This was very good news for its large (>30 000) Belgian population, as the risk of being bitten by an infectious tsetse fly was reduced accordingly. Yaws remained uncommon, reflecting easier access to healthcare and hygienic conditions which were better than in rural areas. Schistosomiasis was not endemic, filariasis was left untreated, and intravenous quinine for severe malaria was used mostly in children. Consequently, in colonial Léopoldville, the treatment of STDs, especially ‘syphilis’, provided the best opportunity for the iatrogenic transmission of bloodborne viruses among adults.

Independence at last

In June 1960, the Congo became independent. Chaos and civil wars rapidly followed. Hundreds of thousands sought refuge permanently in the capital, whose population expanded dramatically. An unexpected outcome was a change in the pattern of sex work. In studies on Léo's sex workers in 1961–1965, social scientists noted a diversification of the trade: there were now sex workers who could have a quick session without much conversation being exchanged, the first description of high-risk sex workers for whom sex was purely a commercial transaction. This had emerged around brothels called ‘flamingos’ or the downmarket ‘Londones’: small shacks with one or two rooms quickly set up on some unoccupied piece of land near the street markets. Outside, men could be seen patiently queuing for their short moment of pleasure.11 23−25

This change was a consequence of the social changes that occurred after 1960 with a staggering rise in unemployment and massive pauperisation. Clients did not have enough money to provide regular support to a free woman whom they would visit time and again; they barely had a few pennies for a brief session. And poverty among women had grown so much that some had no option but to accept those few pennies.

Assembling the puzzle

Less than 0.1% of the total of the Central African chimpanzee population inhabited the Belgian Congo; therefore, it is unlikely that ‘patient zero’, the one who started the pandemic, lived there. However, Léo was the most dynamic city in Central Africa, attracting many migrants. A SIV_cpz-infected hunter moving into the city or an HIV-infected trader wishing to spend some time in the capital would have to present himself at the STD clinic on arrival, where he would receive treatment for syphilis if his serological assay was positive, generally because of a prior yaws infection. Alternatively, the first HIV-infected free woman, sexually infected by one of her patrons, may have been treated with intravenous drugs because of presumed syphilis.

Once the virus was introduced within the Léopoldville-Brazzaville conurbation, it may have found an opportunity for amplification through non-sterile syringes and needles at the Dispensaire Antivénérien of Léo-Est, where the turnover of patients was extreme due to the obsession of local physicians for treating anybody with a positive ‘syphilis’ serology. Iatrogenic transmission of another bloodborne virus, presumably hepatitis B, was well documented as occurring in 1951–1952 because of inadequate sterilisation of injection equipment, at exactly the same time as HIV-1 started increasing exponentially. If the hepatitis B virus was transmitted iatrogenically, the same may have occurred with SIV_cpz/HIV-1 once it was introduced into the cohort of patients treated for a presumed STD. Given the caseload, the interval between two injections with the same needle/syringe was probably less than an hour, and the only obstacle to HIV-1 transmission was the short contact with ether. Patients who developed primary HIV infection experienced high viraemia; even an inactivation that was 99% effective would have left viable virions.

Many of these iatrogenically infected cases would have been free women who had concomitant sexual relationships with several men—arguably a perfect storm. Those free women who were infected parenterally could then transmit the virus sexually to some of their regular clients, who, in turn, infected other sex workers or, later, other women, eventually allowing the virus to move out of the core group. This second, sexual part of the amplification process could have developed during the early 1960s, when the face of Léopoldville changed abruptly with the emergence of a different type of sex work in which some women might entertain up to 1000 clients per year, a pattern similar to the one in Nairobi which, 20 years later, exponentially amplified HIV-1.26 The conditions for an urban sexual epidemic of HIV-1 were now ripe, several decades after the virus had managed to travel down the Congo River from its original crucible somewhere in the forests of Central Africa.

The alternative scenario, that GUD once more enhanced the transmission of HIV-1, seems unlikely for several reasons. First, in the early 1950s, the vast majority of patients reported to have had ‘syphilis’ never experienced a chancre. Second, chancroid and other bacterial causes of GUD were uncommon in the city. Third, as noted elsewhere,14 the apparent incidence of GUD was substantially lower at the time of the expansion of the virus than in the long period during which the number of HIV-infected persons remained fairly stable.3

The cofactor effect of GUD in Léo may have been more important in the 1920s, as suggested elsewhere.14 However, that particular mathematical model included parameters that might have heavily influenced its conclusion: a GUD-related RR of HIV-1 transmission which varied between 9 and 430, and a mean ulcer duration of 10 weeks.14 Instead, most studies of the co-factor effect of GUD estimated that this RR varies between 3 and 5,27 and with syphilis, herpes and LGV, it seems unlikely that the ulcer would persist for as long as 10 weeks.

In conclusion, there is no historical evidence that GUD or changes in the pattern of sex work could explain the exponential amplification of HIV-1 that occurred in Léopoldville from the early 1950s. Unless this chronology is mistaken, iatrogenic transmission may have played a more important role during that pivotal decade.