14
Assembling the puzzle
After reviewing the many elements of the puzzle piece by piece throughout this book, it is now time to assemble them into a coherent summary of the events that led to the transformation of SIV
cpz
into HIV-1, triggering the worst pandemic of modern times. Several pieces of this puzzle are irrefutable, while others remain the most plausible hypotheses explaining parts of the story, given the currently available circumstantial evidence. However, as the years go by, it becomes less and less likely that researchers will uncover novel information that could substantially alter this narrative.
We
have seen in
Chapter 2
that, for at least several hundred years, the
Pan troglodytes troglodytes
chimpanzee of central Africa has been infected with a simian immunodeficiency virus, SIV
cpz
, which is genetically identical to HIV-1. The distribution of SIV
cpz
among chimpanzees in the pre-colonial era was probably not much different from what it is today. Apart from the higher level of threat from humans, the social and sexual behaviour of chimps has not changed over time. SIV
cpz
is mainly transmitted within well-defined troops of chimpanzees, presumably through sexual intercourse, but only sporadically to other communities, with which contacts are infrequent. This resulted in a heterogeneous distribution of SIV
cpz
, absent from some communities while infecting a third of the members of other troops. Overall, around 6% of
P.t. troglodytes
chimps are infected with SIV
cpz
. Some naturally infected chimps develop a disease reminiscent of AIDS, but only after several years during which their intense sexual promiscuity allowed them to spread the virus. It is clear that the other three subspecies of
Pan troglodytes
are not the source of HIV-1. The other chimpanzee species, the
Pan paniscus
bonobo, has been less investigated but there is so far no evidence that it is infected with SIV
.
1
Human populations of central Africa have been in contact with
P.t. troglodytes
for as long as they have lived there, around 2,000 years for the
Bantus and longer for the
pygmies, and in contact with SIV
cpz
-infected
chimps for as long as the virus has been present among apes. Why did not SIV
cpz
emerge into HIV-1 sooner? First, human contacts with chimpanzee blood may have been less common during the pre-colonial era as the lack of
firearms made it more difficult to hunt for apes, and the dearth of even rudimentary roads in densely forested areas reduced the interactions between humans and chimps (
Chapter 4
). Secondly, the conditions that would later facilitate the large-scale sexual and/or parenteral amplification of SIV
cpz
/HIV-1 did not yet exist. At the time there was little opportunity for the parenteral transmission of blood-borne viruses, apart from traditional scarifications and ritual
circumcisions. And there were no cities where high-risk prostitutes would have sex with more than 1,000 men each year. Thus, when pre-colonial hunters or cooks acquired SIV
cpz
, such infections remained epidemiological dead ends: the hunter infected his wife or wives, the cook infected her husband, both would die of AIDS ten years later, and that would be the end of it. The number of cases would be too low for the virus to be transported during the slave
trade on a scale that would have made it recognisable 300 years later, epidemiologically or phylogenetically, or for the disease to be identified among many others by the pioneers of tropical medicine in the early twentieth century. Then, around 1921, the date of the most recent common ancestor of pandemic HIV-1 strains, the situation changed. Not so much that many more people had contacts with chimpanzee blood, although
small firearms were by then readily available, but the mechanisms that would allow the exponential amplification of the infection appeared.
In the following pages, I will try to estimate the probability that a number of events did or did not occur. There are many sources of error and many assumptions had to be made. Therefore, these figures should be seen as generally indicative of what may have happened rather than as precise mathematical measures. Even if one doubles or halves this or that number, it does not materially alter the conclusions.
We have seen in
Chapter 4
that, around 1921, 1.35 million adults lived in areas inhabited by
P.t. troglodytes
: Cameroun Français, Gabon, Moyen-Congo, Oubangui-Chari, Guinea Espanola, the Cabinda enclave of Angola and the small part of the Belgian Congo that lies north of the Congo River. Presuming that the frequency of exposure to chimpanzee blood was the same as in recent times, 0.1% of these adults had been exposed to blood potentially containing SIV
cpz
at least once in their lives. If we multiply the total number of adults by 0.1% and multiply this by the 5.9% SIV
cpz
prevalence among
P.t. troglodytes
chimps, we can calculate that eighty adults were exposed to the virus while
hunting or handling chimpanzee carcasses. With a 1% risk of transmission during occupational exposure through the skin, we end up with one human infected from chimps, or three if transmission per exposure was closer to 3% (these estimates of risk of transmission are based on extrapolations from
healthcare workers exposed to HIV-infected blood). To facilitate the subsequent calculations, let us use the median number and say that in 1921 there were two SIV
cpz
-infected humans, probably both men, living in one of the countries inhabited by
P.t. troglodytes
.
1
–
2
Then, if we were to assume that urbanisation and urban prostitution were the only amplifying mechanisms (the current standard theory on the emergence of HIV-1), we need one of these SIV
cpz
-infected hunters to infect a first prostitute in a colonial city, perhaps Yaoundé, Bangui, Libreville, Brazzaville or Léopoldville, for a chain of sexual transmission to be initiated. A number of factors made such a process a bit unlikely. First, the proportion of the population of central Africa that lived in urban areas around 1921 was at most 5%. Second, those who were more likely to be SIV
cpz
-infected, the illiterate villagers who had occupational contacts with chimpanzees, must have been less prone to move into urban areas than individuals with at least a few years of primary school education, who spoke some French and could be hired by the colonial administration, private companies or expatriates. Third, not all men living in the cities had sex with prostitutes. So if we start with two SIV
cpz
-infected men, assume a 2.5% probability of moving to a colonial city, and also assume that half of the city dwellers bought sex from free women, the probability that at least one of the two might have had sex with a free woman would be something like one in forty. However, the probability of male-to-female transmission of HIV-1 is not 100%. There are many serodiscordant couples in Africa (and elsewhere), in which one remains seronegative despite hundreds of unprotected intercourses with the seropositive spouse. In general, the risk of HIV-1 transmission is estimated at about one per 1,000 intercourses. Transmission is more effective, however, in the presence of an STD, especially those causing genital ulcers, and if the man is not circumcised. If we assume that the SIV
cpz
-infected man had repeated intercourses over many years with free women, the cumulative probability of at least one forward transmission of HIV-1 would be around 50%. So the 1:40 probability becomes 1:80. It is entirely possible that this did occur and
that the pandemic was in essence caused by an unpredictable factor: bad luck.
Would the odds have been different if the first chains of sexual transmission had not occurred in the cities but in the camps housing the unfortunate men conscripted into forced labour for building the Congo–Océan railway, as described in
Chapter 3
? If we exclude those from Tchad (no
P.t.
troglodytes
), about 110,000 men from
Moyen-Congo and
Oubangui-Chari were forced to work on the railway between 1921 and 1932, which was roughly 16% of all adult men living within the
P.t. troglodytes
habitat. If we do the same maths as above, we end up with a 1:7 probability that one of them was infected with SIV
cpz
. Let us say that half had sex with local free women, but for a shorter period (unlike the urban migrants, most CFCO workers would not spend much more than a year or two
in situ
), thus with a lower risk of forward transmission. So we would probably end up with the same type of odds, something like 1:40 or 1:80
.
3
However, this chain of events would have been infinitely more likely, even unavoidable, if there had been, somewhere in one of these countries, an initial phase of parenteral amplification of SIV
cpz
/HIV-1 through re-usable syringes and needles for the treatment of tropical diseases. For this to happen, we would need one of the two SIV
cpz
-infected men to be diagnosed with a tropical disease and to receive an IV drug. For villagers, this was much more probable than migrating to a city and infecting a prostitute. They did not have to move at all: the mobile disease control teams came to their villages with their microscopes, syringes, needles and drugs. For the Congo–Océan workers, rudimentary hospitals provided care along the railway, and many workers would need their services, however primitive these were.
Opportunities for the parenteral transmission of blood-borne viruses first arose with the campaigns against sleeping sickness, then with the early treatments of
leprosy, at exactly the right place and time (
Chapter 8
). A few years later, massive iatrogenic transmission of
HCV occurred in south
Cameroon through the parenteral treatment of
malaria when fixed health centres and hospitals, which could administer intravenous quinine, were established in the 1930s. In some regions, transmission of HCV may have been enhanced by large-scale campaigns for the control of
yaws and syphilis, which used mostly parenteral drugs. The areas highly endemic for malaria and yaws happened to correspond to the forested areas inhabited by
P.t.
troglodytes
.
For instance in the
Ntem,
Kribi and
Sanaga-Maritime regions of
Cameroun, as well as in most of
Gabon and
Moyen-Congo, population incidence of yaws was often greater than 200 per 1,000 per year between 1930 and 1950, and most humans would be bitten by a mosquito carrying the malaria parasite every other day. Over a period of a few years, almost the entire population, including our two SIV
cpz
-infected individuals, would be treated with injectable drugs.
The efficacy of transmission of SIV
cpz
from one patient to another would have been comparable to what has been described in many parts of the world, mostly after 1981, among drug
addicts
. As reviewed in
Chapter 7
, once the virus is introduced into such groups who share syringes and needles, it spreads quickly, and up to 50% of addicts can acquire HIV-1 within a few years.
4
Transmission of HIV-1 is ten times more effective through the sharing of needles and syringes than via sexual intercourse because the minute quantity of blood from the first user which remained in the syringe or needle is then injected IV, directly into the next user’s blood, where HIV-1 can easily spread to the latter’s lymphocytes. When a second person, an addict or, in our case, another individual treated for the same tropical diseases, developed primary HIV-1 infection, a very high degree of viraemia ensued, reaching 10
5
to 10
7
viral copies per ml. After two or three weeks at this level, viraemia slowly decreased over four to six months until it reached a steady state, generally around 10
4
viral copies per ml. During this brief peak, the second person’s blood was highly infectious so that the risk of transmission to a third person must have been higher than 1% per episode of needle/syringe sharing. Further injections of antitreponemal, antitrypanosomiasis or antimalarial drugs with the syringe or needle used for this second patient would expose to HIV-1 other patients who were treated on the same or following day. Inevitably, a third person would become infected, further increasing the risk for the other patients treated at the same health facility or disease control mobile clinic. The tragic iatrogenic epidemics of HIV-1 in Romania and Libya, which, long after HBV, HCV and HIV were identified, occurred in countries with far more resources than those available in central Africa during the early twentieth century, demonstrated the potentially devastating efficacy of health care in spreading HIV-1 parenterally.
The fact that in parts of
Cameroun and
Gabon up to 50% of some birth cohorts were infected iatrogenically with
HCV suggests that if SIV
cpz
/HIV-1 was ever introduced into the group of patients receiving
injections in a given healthcare facility, the amplification must have been substantial. So from one SIV
cpz
/HIV-1-infected patient, there could have been hundreds after a few years. Such amplification was probably limited geographically to the patients receiving medical care in the same hospital, dispensary or disease control mobile clinic as the first SIV
cpz
-infected patient. However, once the number of HIV-1-infected individuals had increased to a few hundred, the probability of at least one of them moving to a city, infecting a free woman and initiating a chain of sexual transmission increased proportionally. From a long shot this became an unavoidable certainty.