by Pepin
The other pathogen that can be used for such proof of concept studies is HCV. HCV is compatible with prolonged survival, as less than 25% of infected individuals go on to develop cirrhosis or cancer of the liver. Furthermore, its sexual transmission and mother-to-child transmission are rather ineffective, so that the parenteral route accounts for the overwhelming majority of cases. In industrialised societies, HCV transmission during blood transfusions has been eliminated through screening, so that most new cases of HCV infection are acquired during intravenous injections of recreational drugs. In sub-Saharan Africa, some transmission still arises during the administration of unscreened transfusions, but most cases are thought to be acquired via injections during health or dental care.5–6
In contrast, HBV is not a good marker for the parenteral transmission of viruses in sub-Saharan Africa because it is more infectious and transmitted effectively through several other mechanisms. There, a large number of children get infected with HBV: some from their mothers but the majority from other children, possibly through chronic skin wounds which are extremely common in underprivileged populations. Very few of these children develop acute hepatitis with jaundice when they acquire HBV, but 10–15% remain chronic (and infectious) carriers and can develop long-term complications, while the remaining 85–90% becomes immune. Those who managed to avoid HBV during childhood acquire it sexually at adolescence, and in central Africa around 95% of adults have antibodies against HBV. There is little room for the parenteral transmission of HBV in adults, and for that reason HCV is a much better marker of the parenteral/iatrogenic transmission of viruses. In industrialised countries there was never much transmission of HBV during childhood with the result that, until the vaccine became widely used, most adults were immunologically naive with regard to HBV and could be infected parenterally, either during health care or while using injectable recreational drugs.
HIV among drug addicts
The high efficacy of the parenteral transmission of HIV-1 has been documented time and again in studies of IDUs conducted during the early part of the epidemic, before this risk was understood and before partially effective prevention measures could be implemented in this hard-to-reach target group. Along with outbreaks that occurred during health care, this constitutes the best proof of concept for the potential transmission of HIV-1 during medical injections with improperly sterilised syringes and needles, because the quantity of blood (and thus the number of viruses) left in a syringe after an IV injection of cocaine or heroin is approximately the same as after similar equipment has been used for the IV administration of a pharmaceutical agent. This is in stark contrast to transfusion-related HIV infection, where the amount of infectious blood is much larger. Among IDUs the transmission of HIV-1 occurred from a first to a second addict, from this second to a third one, and so on, unlike what happened to haemophiliacs, many of whom could be infected via the same lot of a coagulation factor, which had been prepared through a pooling process from thousands of donors.
Collections of samples obtained before HIV assays were marketed allowed a retrospective description of the spread of the virus among IDUs. In New York, where the virus was introduced in the mid-1970s, more than half of addicts were infected by 1982. In Edinburgh, HIV prevalence rose rapidly to 51% in the early 1980s, with most seroconversions occurring within a two-year period. In Bari, Italy, HIV seroprevalence among IDUs increased from 0% in 1979 to 76% in 1985 while in Valencia, Spain, it increased from 11% in 1983 to 48% two years later. In Geneva, prevalence increased from 6% in 1981 to 38% in 1983. Among the drug addicts of Milan, HIV prevalence was 8% in 1981, 35% in 1983 and 60% in 1985. In Bangkok, HIV prevalence among IDUs increased from 1% at the end of 1987 to 43% by the end of the following year.7–13
Transmission among IDUs continued even where comprehensive interventions, including needle exchange programmes, were put in place. In Vancouver, at the very time and place of an international conference heralding the development of effective antiretroviral therapies, 23% and 88% of IDUs tested in 1996 were infected with HIV and HCV respectively. And such outbreaks continue to occur today, nearly thirty years after the identification of HIV as the cause of AIDS. In Sargodha, in the Punjab province of Pakistan, HIV prevalence among IDUs increased from 9% in 2005–6 to 51% the following year.14,15
These are all good examples of the exponential transmission of HIV from blood to blood through the sharing of syringes and needles. The probability of transmission per act when an HIV-negative addict uses a syringe or needle recently used by an HIV-infected drug user is somewhere between 0.7% and 1.1%, an order of magnitude higher than when a seronegative individual has sex with an HIV-infected partner (the risk of transmission per intercourse is 0.1%). These are summary estimates and the risk of parenteral (or heterosexual) transmission must be substantially higher when the index HIV-infected person is experiencing the high viraemia that characterises primary HIV infection. This creates a vicious circle in which the HIV-negative addicts are, month after month, increasingly likely to share their injection paraphenalia with someone who is HIV-infected. The result is that today HIV prevalence nationwide among drug addicts is 37% in Russia, 40% in Spain, 43% in Indonesia and Thailand and 72% in Estonia. Worldwide, out of 16 million intravenous drug users, about 3 million are HIV-infected.16–18
Iatrogenic epidemics of HIV
Several small outbreaks of HIV transmission through health care have been reported. As these generally involved just a few patients, it is difficult to use this information to infer what could have occurred on a much larger scale in sub-Saharan Africa. However, more to the point of this discussion, two epidemics (by definition, an epidemic corresponds to a much larger number of patients than an outbreak), both of which involved children, can be used as a further proof of concept for a large-scale transmission of HIV-1 during health care.
Shortly after the fall of Ceausescu in 1989, a medical journal reported that 367 children in Romanian hospitals and orphanages had tested positive for HIV. Of the 1,168 cases of AIDS reported to the Romanian ministry of health by December 1990, 94% had been documented among children less than four years of age. This age distribution was extraordinary. Two-thirds of paediatric cases occurred among children abandoned by their parents and living in orphanages or long-term care facilities. Less than 10% of the mothers who could be located were found to be HIV-infected, implying an unusual mode of infection of the children.19,20
At the time in Romania, transfusions were given to malnourished children as a way to provide them with many nutrients. Because the recipients needed small volumes of blood, a unit was generally split and administered to two or three children. One third of cases occurred in children who received transfusions of unscreened blood, and the others were attributed to multiple therapeutic injections with syringes and needles used over and over again without proper sterilisation. Many children with AIDS had received 300 or more lifetime IM injections, an extraordinary number.20
Cases were distributed unevenly throughout the country. The largest numbers, about half of the total, were seen in the Constanta district, east of Bucharest. A nationwide survey of 12,000 institutionalised children under four years of age revealed that 10% were HIV-infected. Among institutionalised children in Constanta, HIV prevalence was 48% in those aged less than four years, but 0% in those aged four or more. This reflected a much more intensive use of medical injections and transfusions in the first years of life, as well as a cohort effect: nosocomial transmission occurred only after HIV-1 had been introduced into Romania in the 1980s.
Further studies suggested that injections rather than transfusions were the mode of infection for most children, even those who had received transfusions. The testing of more than 400,000 units of blood in 1990–1 documented that the HIV prevalence among donors was only 0.007% for the whole country and 0.025% in Constanta. This implied a low risk for the recipients, even if the children were transfused ten times during their lives. A case-control study documented an association between
HIV and multiple injections: the HIV-infected children, only a few years old, had received 280 injections on average, compared to 187 for the HIV-negative controls. These injections consisted of vitamins, antibiotics and vaccines.21,22
More than 5,000 children had been iatrogenically infected with HIV before this came to an end. When typing became available, it was shown that almost all HIV-infected children born of a seronegative mother were infected with HIV-1 subtype F. Among HIV-infected adults, subtype F accounted for 68% of infections, but subtypes A, B, C, D and CRF02_AG were also detected. Thus, multiple subtypes were introduced by Romanians who travelled to Africa or by foreigners who lived in Romania. Then subtype F was exponentially amplified, from child to child, by unsafe medical practices within the vulnerable subgroup of institutionalised infants and children. Subtype F isolates from Romania are similar to those from Angola, another communist country at the time, where the source of this tragic epidemic may lie. The other subtypes disseminated slowly in the adult Romanian population as they did in other European countries, mostly through sexual intercourse.21,23–25
A second dramatic epidemic occurred among children attending the Al-Fateh Pediatric Hospital in Benghazi, Libya, and attracted media attention because five unlucky Bulgarian nurses and one Palestinian doctor were sentenced to death for their alleged role in this tragedy. They were released after spending nine years in Libyan prisons. The problem was first recognised in 1998, when serological analyses showed evidence of a recent HIV infection in several children. Eventually, about 450 children (mean age: four years) were found to be HIV-infected. Many of them subsequently received health care in Swiss and Italian institutions, from where some of the details emerged. By 2007, fifty-two children had died.26–28
Two-thirds of the HIV-infected children had been hospitalised at the Benghazi hospital, while the others had received outpatient care; 46% of the HIV-infected children were also infected with HCV. Of twenty-five mothers initially tested, only one was HIV-infected, implying that the other children had certainly not been infected from mother-to-child. Later, eighteen additional mothers were found to be HIV-infected, but it seems that most were infected from their child (presumably during breast feeding) rather than the other way around, because none of the fathers was HIV-infected. All children were infected with the same recombinant CRF02_AG strain, with little genetic variation between isolates, implying a common source for the HIV-1 epidemic and rapid transmission of the virus.26–27
Unfortunately, foreign investigators had little direct access to epidemiological information from Libya, and of course Libyan investigators could not talk since colonel al-Gadhafi’s dictatorial regime had already decided who to blame: the foreigners, who they said had deliberately infected the children as part of a Central Intelligence Agency (CIA)- or Mossad-inspired plot. So we do not know for sure exactly which procedure(s) led to nosocomial transmission of HIV-1 and HCV. All the victims had had blood drawn at the hospital, most had received some form of IV therapy, but few had received transfusions. Many children had spent some time in the ‘semi-intensive care unit’ of the hospital. Infection control practices were so deplorable that it seems likely that transmission occurred through more than one type of procedure.26–27
Molecular clock analyses showed that the most recent common ancestors for the HIV-1 and HCV isolates predated 1998 by a few years, implying that the two viruses were already spreading in the hospital before the expatriate healthcare workers arrived in Benghazi, a conclusion confirmed by the finding that some of the infected children had not received care at the hospital after 1997. Given that the CRF02_AG strain is found mainly in central and West Africa, it seems likely that the original case, the Benghazi patient zero, was one of the 1.5 million African migrant workers who lived in Libya.29
It is of course a scandal and a tragedy that so many children got infected with HIV through the re-use of needles and other medical devices in a country so rich with oil money, and that scapegoats were jailed for so long as a public relations diversion from the real culprit: the Libyan regime. But this epidemic, the previous one in Romania and two smaller outbreaks in the former Soviet Union illustrate how HIV-1 can be exponentially amplified through unsafe injections, much like what happens among IDUs. If such epidemics could arise after the existence of HIV-1 was known, they must certainly have occurred before the new disease and its aetiological agent were identified.30
The largest ever iatrogenic epidemic
At the turn of the century, studies in Egypt demonstrated that the iatrogenic transmission of another blood-borne virus, HCV, could reach a massive scale. Of course Egypt is not inhabited by P.t. troglodytes, so HIV-1 could not possibly have emerged in the Middle East. But as an example of the potential for well-intentioned disease control interventions to transmit blood-borne viruses, the Egyptian experience is hard to beat.31
Schistosomiasis is caused by a parasite with a complex life cycle, which is acquired during exposure to water inhabited by the snails that constitute its intermediate host. In the most common form of schistosomiasis, adult worms live in the blood vessels around the rectum, causing bloody diarrhoea, or liver fibrosis when their eggs manage to travel up to the liver. There is also a variety in which the adult worms live around the bladder, with the eggs excreted into the urine rather than the stools; such patients pass blood in their urine.
Schistosomiasis used to be highly endemic in the Nile delta, where mass treatment campaigns with parenteral drugs led to the HCV infection of millions of individuals. The use of tartar emetic, a very old drug against schistosomiasis, had begun in 1921, and the ministry of health started organising large-scale campaigns in the early 1950s. Between 1964 and 1982, more than two million injections of tartar emetic were administered each year to 250,000 patients. On average they received ten to twelve weekly IV injections through hastily sterilised syringes and needles, boiled for only one or two minutes or not at all. In specialised centres, more than 500 patients could be treated in an hour. Tartar emetic was not very effective: many patients relapsed or were re-infected, and had to receive further cycles of the drug. The historical information is congruent with what was calculated using the same molecular approaches as we reviewed for HIV-1, which indicated an exponential growth in the number of HCV-infected individuals between 1940 and 1980.32–33
For the whole country, 22% of individuals aged ten to fifty years became infected with HCV. HCV prevalence is lower (6–8%) in residents of Cairo and Alexandria, but reaches 19% to 28% in Upper, Middle and Lower Egypt. Prevalence is higher among the older age groups, who were exposed repeatedly to schistosomiasis treatments. In Lower and Middle Egypt, more than 50% of individuals aged forty or more are HCV-seropositive. In all these regions, there was a correlation between exposure to schistosomiasis treatment and HCV infection. The same interventions also led to the iatrogenic transmission of HBV, but in this case the relationship was blurred by the other modes of transmission of HBV.31
HCV infection in central Africa
Now let us get geographically closer to the heart of the story. After Egypt, central Africa has the highest HCV prevalence in the world: 6.0% of adults overall and 13.8% in Cameroon. In several areas of Cameroon, more than 40% of elderly individuals became HCV-seropositive. What epidemiologists call a cohort effect (the exposure to a given pathogen varying according to the year of birth) was demonstrated, with HCV prevalence reaching 40–50% among people born before 1945, about 15% for those born in 1960, and 3–4% for younger individuals born after 1970. In several studies, HCV prevalence plateaued at the same point, corresponding to a year of birth around 1930–5 (Figure 9). So most of the parenteral transmission of HCV must have occurred between 1930 and 1970.34–42
Molecular clock analyses confirmed this conclusion and revealed that in Cameroon, Gabon and the Central African Republic, the number of HCV-infected individuals started increasing exponentially between 1920 and 1940, continuing for two or three decades. Since the heterosexual transmission of HCV
is relatively ineffective, this indicates massive parenteral transmission of at least one blood-borne virus in the very areas inhabited by the P.t. troglodytes source of HIV-1. This parenteral transmission took place during the colonial era, starting at the same time, give or take a few years, as SIVcpz successfully emerged into human populations to become HIV-1. It is hard to believe that this represents merely a strange coincidence.42,43
In Cameroon, populations with a high HCV prevalence come from Yaoundé or villages in the southern rain forest (Figure 9 and Map 5), and prevalence is much lower in the north. A similar north–south gradient in HCV prevalence was observed in territories that used to be part of AEF, with a low prevalence in Tchad and the Central African Republic and a high prevalence in Gabon. This means that the diseases during the treatment of which transmission of HCV occurred were more common in the southern rain forests than in the arid north.42–45
Figure 9 Prevalence of HCV infection at various sites in Cameroon by year of birth.