Corresponding AuthorBrian Greenwood, Department of Infectious and Tropical Diseases, London School of Hygiene and Tropical Medicine, London, UK. Tel.: +44 (0) 207 299 4707; Fax: +44 (0) 207 299 4720; E-mail: firstname.lastname@example.org
Invasive meningococcal infection is probably a relatively new disease. The first recorded outbreak occurred in Geneva in 1805 (Vieusseux 1806). The following year, typical cases of meningococcal meningitis (cerebro-spinal meningitis) were seen in New England (Danielson & Mann 1806) and epidemics occurred across Europe and North America throughout the nineteenth century. In 1840, the first outbreak in Africa was reported among French troops based in Algiers (Chalmers & O'Farrell 1916) and during the latter half of the 19th century several outbreaks occurred in Egypt and the Sudan (Ensor & Balfour 1904; Chalmers & O'Farrell 1916). However, it was not until 1905 that the first major epidemic in West Africa was reported (McGahey 1905a).
Epidemic meningitis in West Africa – 1905
One hundred years ago, an outbreak of 32 cases of epidemic meningitis in Zungeru, northern Nigeria was reported in the Journal of Tropical Medicine (McGahey 1905a), a progenitor journal of Tropical Medicine and International Health. At this time, the northern part of Nigeria was only just coming under British colonial rule. The ancient Hausa city states of Kano and Sokoto had been subjugated by Lord Lugard and his army only 2 years previously and the area was still unsettled. McGahey had no doubt that he was observing an outbreak of epidemic meningococcal meningitis, presumably because he had had previous experience of such outbreaks in Europe. Review of the 32 cases presented in his paper suggests that he was right as they showed all the features of epidemic meningococcal meningitis as it is seen in Africa today – a wide age range, characteristic symptoms and signs and complicating herpes infection. The outbreak was widespread. Further cases were soon reported from Yola, nearly 500 miles east of Zungeru (McGahey 1905b; Figure 1) and soon much of northern Nigeria was affected. There are no detailed records of the extent of this epidemic but it is estimated that there were at least 20 000 cases and many thousands of deaths. The epidemic spread rapidly westwards and, in 1906, it reached northern Ghana where it was described in detail by Horn (1908). His report describes all the main epidemiological features of epidemic meningococcal disease as it is seen in West Africa today – a high attack rate, seasonality and spread from area to area over a period of 2 or 3 years. Horn managed to culture the meningococcus from cerebrospinal fluid so it is certain that the 1906–1908 epidemic in Ghana was caused by a meningococcus and highly likely that the earlier Nigerian epidemic was caused by this bacterium also.
Was the 1905 outbreak the first epidemic of meningococcal meningitis in West Africa?
It will never be known for certain whether the 1905 outbreak was the first epidemic of meningococcal meningitis to affect West Africa but the limited evidence available suggests that it was (Greenwood 1999). This comes from two main sources:
•Discussions with traditional rulers. There are a number of reports from colonial medical officers involved in the management of meningitis epidemics in the early part of the 20th century of discussions with traditional rulers on the origins of these outbreaks. Most suggested that this was a new disease that had not occurred in their community previously.
•Early European travellers to West Africa. As far as I am aware, there is no record of an epidemic of meningitis in the writings of the early European travellers to West Africa. The journals of Mungo Park (Figure 2), a surgeon, are especially valuable as he travelled twice across a large part of the African meningitis belt at the end of the 18th century. Park had a professional interest in local diseases and their treatment and mentions several tropical infections but there is no mention in his journals of epidemic meningitis. Had epidemics been occurring in countries of the Sahel at that time, it is very unlikely that he would not have commented upon this, as their impact on the community is so dramatic.
How did epidemic meningitis reach West Africa?
It is possible that the strain of meningococcus responsible for the first West African epidemic emerged within Nigeria, perhaps as a result of the acquisition of a virulence gene by a previously non-pathogenic neisseria, as may have happened in Europe 100 years previously. However, it seems more likely that the strain responsible for the first epidemics in West Africa came from the Sudan. By the latter half of the 20th century, epidemic meningococcal disease was well established in Egypt and Sudan and in 1899, a large epidemic was recorded in Omdurman (Ensor & Balfour 1904). There are at least two ways in which the epidemic strain could have reached Nigeria from the Sudan – with pilgrims to the Holy Places or with followers of the Mahdi:
•Pilgrims. By the beginning of the 20th century, many West African pilgrims to the Holy Places travelled across present day Chad and the Sudan before crossing the Red Sea to the Arabian Peninsula and returned by the same route.
•Soldiers. Volunteers in the armies of the Mahdi who fought against Anglo-Egyptian forces in the Sudan came from across West Africa and from as far away as Senegal. On the defeat of the Mahdi and his successor, many returned home along the route of pilgrims.
Travelling by horse and/or by camel it would have been possible for travellers to complete the journey from the Sudan to northern Nigeria within a few months, well within the period in which a meningococcus can persist in the nasopharynx. The idea that the epidemic strain responsible for the 1905 Nigerian epidemic might have come from Mecca was recognised by McGahey (1905b), who noted that the epidemic commenced shortly after the arrival in Yola of Arab traders from Mecca.
Characteristics of the early epidemics
Description of the cases seen by McGahey (McGahey 1905a) and of the epidemiological characteristics of the 1906/1908 epidemic in northern Ghana described by Horn (1908) leave little doubt that this was the same disease as seen in West Africa today. The major difference was the high mortality of the infection in the absence of any effective treatment or means of prevention.
Epidemic meningitis in West Africa in 2005 – what has changed?
The ability to determine the cause of an African epidemic of meningitis has advanced a long way since the struggles of Horn to culture the organism in a grass hut in northern Ghana. The development of effective transport media makes it possible to send samples for culture from a peripheral clinic to a central laboratory. Antigen detection assays have been developed which can be used under field conditions, although the high costs of commercial latex agglutination tests has limited their availability. Efforts are now being made to develop diagnostic dip-sticks that would be well suited for use in peripheral clinics (Chanteau et al. unpublished data). Multiplex PCR assays that can be done on stored samples (Parent du Châtelet et al. 2005) make identification of the cause of outbreaks much easier than was the case in the past.
It is unlikely that the serogroup of the meningococcus responsible for the first African epidemics will ever be known, although it might be possible to determine this using PCR, if burial sites from the time of the first epidemics could be identified and exhumation was culturally acceptable. Serogrouping employing current nomenclature first came into widespread use only in the 1950’s and there is a possibility of confusion over the identity of organisms characterised using older typing systems. However, serotyping of isolates obtained during African epidemics during the 1940–1960s showed that, with the possible exception of an outbreak in the Sudan in the 1930s (Davis 1931), epidemics were caused by meningococci belong to serogroup A (Lapeyssonnie 1963). In the 1970s, epidemics caused by group C meningococci were recorded in Nigeria and Niger (Whittle et al. 1975; Broome et al. 1983) and, in 2002, a W135 strain caused a major epidemic in Burkina Faso (Decosas & Koama 2002). Whether prior, widespread immunisation with group A + C polysaccharide vaccines contributed to the W135 epidemic in Burkina is uncertain but vaccination is unlikely to have contributed to the earlier group C outbreaks, as vaccines used at that time contained both groups A and C polysaccharides. It is likely that natural variations in the pre-dominance of different serotypes take place over time independently of vaccination, a view supported by the results of longitudinal carriage studies which have shown changes in the serotype of nasopharyngeal isolates in a defined community over time (Gagneux et al. 2002). However, in spite of these natural fluctuations, group A meningocccus has remained the dominant cause of African epidemics throughout the past 70 years. It has been recognised for many years that during epidemics some cases are caused by Streptococcus pneumoniae (Tugwell et al. 1976) and not the meningococcus. However, during recent epidemics in the meningitis belt this proportion appears to have increased, perhaps because of the circulation of a virulent strain of serotype 1 pneumococcus (Leimkugel et al. 2005).
There are some indications that epidemics of meningococcal disease have become more frequent in Africa in recent years, although it is difficult to be sure that this is a true increase and not a consequence of better reporting. Because the population of the countries of the meningitis belt has increased substantially during the past 50 years, the total number of cases occurring each decade is higher than it was 50 years previously. The advent of HIV infection in Africa does not appear to have had a major impact on the incidence of meningococcal disease, as has been the case for invasive pneumococcal disease, perhaps because the prevalence of HIV is relatively low in countries of the meningitis belt.
The characteristic epidemiological features of the first epidemics – periodicity, geographical restriction, massive size and marked seasonality – have persisted.
The concept of a geographically constrained African ‘meningitis belt’, extending across the Sahel and sub-Sahel from The Gambia in the west to Ethiopia and Sudan in the east, first defined by Lapeyssonnie (Lapeyssonnie 1963; Figure 3), has generally held up to more recent analyses (Molesworth et al. 2003). However, during the past decade, outbreaks of meningococcal disease have been reported in areas adjacent to the classical meningitis belt and also in several countries in East Africa and southern Africa. These outbreaks have shown some of the epidemiological characteristics of meningitis belt epidemics such as seasonality, although they have rarely been as large. These observations suggest that the meningitis belt may be expanding, perhaps in association with a reduction in rainfall and absolute humidity in sub-humid areas adjacent to the meningitis belt and in southern Africa (Molesworth et al. 2002). In countries of the belt which bridge forest and savanna, the tendency for outbreaks to be restricted to the northern, more arid regions, noted by early observers such as Horn (1908), has persisted with very few exceptions.
An explanation for the existence of the African meningitis belt is little clearer today than it was 100 years ago. Molesworth et al. (2003) have correlated the distribution of recent and earlier outbreaks with a number of environmental characteristics. Correlations were found with absolute humidity, type of land-cover, dust concentration and population size. Using type of land-cover and absolute humidity as the main variables, a model has been constructed which allows prediction of areas at high risk of epidemics (Molesworth et al. 2003) and these match well with areas in which epidemics have been recorded. The model has also allowed the identification of areas of moderate, and perhaps increasing, risk including substantial areas in eastern and southern Africa. Absolute humidity, identified by earlier workers such as Waddy (1957) as a determinant of epidemics emerges, as the most important variable in the new model but how absolute humidity exerts its impact is still unclear.
A characteristic feature of the first epidemics of meningococcal disease seen in West Africa was the way in which they almost always started around the middle of the dry season, rapidly built up to a peak at the end of the dry season and then subsided abruptly with the coming of the rains, only to start again in neighbouring areas in the dry season of the following year. This pattern has persisted with almost no exceptions but there is still no explanation for this remarkable seasonality. Outbreaks are seen at the time of the year when absolute humidity is low and when a hot, dusty wind (the harmattan) blows from the Sahara. Following on suggestions by previous authors I have postulated previously that the seasonality observed in the incidence of cases of meningococcal disease reflects a change in the ratio of cases of disease to nasopharyngeal carriers, normally in the range of 1–100, rather than a change in the overall incidence of infection (Greenwood et al. 1985) perhaps reflecting an impairment of local defences at the height of the dry season. How this might be brought about is uncertain but study of the nature of the particles present in harmattan dust and of whether these might impact on host protective defences deserves further study.
Why African epidemics of meningococcal disease start in a particular place at a particular time is still poorly understood. There is evidence that epidemics can follow the introduction of a new, epidemic strain, for example the sub-group III-1 strain of group A meningococcus that was responsible for many African epidemics in the 1980s and which is believed to have been introduced into Africa by pilgrims returning from Mecca. This strain, of multilocus sequence type ST5, is now being replaced by another serogroup A strain belonging to ST7 (Nicolas et al. 2001). However, introduction of a strain with a new antigenic variant of an outer membrane protein would only be expected to initiate an epidemic in a population with a previously high level of exposure to group A meningococci if the major component of naturally acquired immunity was directed at this protein antigen rather than at the capsular polysaccharide. This does not appear to be the case in areas with a high level of exposure to the group A meningococcus (Amir et al. 2005). The cause of the W135 epidemic in Burkina Faso is a further mystery. Although, this was the first large epidemic to be caused by a W135 meningococcus in Africa, strains of W135 meningococci had been circulating in West Africa for many years (Denis et al. 1982; Kwara et al. 1998) without causing epidemics. Furthermore, W135 meningococci similar to the strain that caused the epidemic in Burkina Faso have appeared in several other countries in the African meningitis belt but no further major W135 epidemics have occurred. Environmental factors such as the severe overcrowding and deprivation that may occur in refugee situations can pre-dispose to epidemics but these were not present in Burkina Faso and there were no obvious environmental factors that could account for this epidemic. Perhaps African epidemics require several events involving both the organism and the environment occurring concurrently but what grouping of circumstances is necessary to set off an epidemic remains a mystery and no system has yet been devised that allows an epidemic to be accurately predicted by time and place. It is 10 years since the last major epidemic spread across the meningitis belt so another major epidemic may not be far away.
The clinical features of the patients described by early observers (McGahey 1905a; Horn 1908) are identical to those seen in African epidemics today and there are no indications that the clinical pattern of the disease has changed during the past 100 years. There is some evidence that the proportions of cases of meningococcaemia and meningitis vary between epidemics and during epidemics (Greenwood et al. 1979) but differences in reporting procedures and access of patients to treatment may account for some of these differences. Overall, there is no evidence that the disease has become more or less severe over time.
In 1905, there was no effective treatment for meningococcal meningitis, although a number of heroic measures such as antiseptic throat washes were tried. The introduction of sulphonamides was almost miraculous and had a major impact on those who saw the transition of an infection, which was almost always fatal to one with a 90% cure rate over a period of 2 or 3 years (Kirk 1950). Fortunately, cheap and highly effective antibiotics in the form of penicillin and chloramphenicol were available to take the place of sulphonamides when resistance to the latter emerged. In West Africa, the meningococcus remains sensitive to both penicillin and chloramphenicol so that single injection treatment with oily chloramphenicol is still an effective treatment. However, the efficacy of ceftriaxone in the treatment of African meningococcal meningitis has recently been confirmed (Nathan et al. 2005). As this antibiotic is easier to use than oily chloramphenicol, and is becoming cheaper, it may soon become the treatment of choice. There has been little progress in the management of cases of meningococcal septicaemia who account for a substantial proportion of the 10% of patients who still die from the infection despite receiving an appropriate antibiotic.
In 1905, there were no effective methods for containing epidemics of meningococcal disease. Measures such as closing of markets and the prevention of social gatherings such as funerals were sometimes adopted in an attempt to prevent spread of the epidemic strain. These may have had some effects, but they were very unpopular and sometimes led to cases being hidden from the authorities. The advent of sulphonamides provided a means of prevention as well as cure and sulphonamides were used widely for this purpose in the 1950s and 1960s. It was generally believed that they were highly effective but this was never proved in a well-controlled trial (Greenwood 1999). With the emergence of resistance to sulphonamides and the need to use more expensive drugs for chemoprophylaxis, such as rifampicin or cephalosporins, interest in this form of control has waned and chemoprophylaxis is now rarely used in Africa except in the management of localised outbreaks.
For 30 years, group A + C polysaccharide vaccines have been used to control epidemics and recently a trivalent vaccine, including W135 polysaccharide, has been developed and deployed in Burkina Faso. These vaccines have saved many lives but, even when used early in an epidemic, they probably reduce the number of cases by only around a half (Woods et al. 2000). Ensuring that vaccination takes place early in the course of an outbreak is difficult, even with the help of an organisation set up to facilitate this – the International Co-ordinating Group (ICG). There is no evidence that use of polysaccharide vaccines for epidemic control has reduced the incidence as opposed to the size of African epidemics.
Meningococcal conjugate vaccines, which induce immunological memory and, in contrast to polysaccharide vaccines, have a substantial effect on nasopharyngeal carriage (Snape & Pollard 2005) provide a possible means of finally eliminating African meningococcal epidemics. The technology to develop group A meningococcal conjugate vaccine has been available for at least 15 years and the first trial of such a vaccine took place in Africa in 1993 (Twumasi et al. 1995). However, the development of these vaccines has, until recently, made little progress as no major pharmaceutical company was prepared to take on the challenge of developing a vaccine which would have a very limited commercial market. This situation has recently changed and now a heptavalent vaccine containing DPT, hepatitis B, Hib polysaccharide and group A + C meningococcal conjugates is being developed by GlaxoSmithKline (GSK). In addition, a monovalent group A conjugate vaccine is being developed in India, specifically for use in Africa, through a public private partnership, the Meningitis Vaccine Programme (MVP), which is supported by the Bill and Melinda Gates Foundation (Jodar et al. 2003). This vaccine should be ready for deployment in 2009. It is proposed to use this vaccine for mass immunisation of children and young adults in countries at high risk of epidemics with the aim of interrupting transmission as well as providing personal protection, a strategy that has proved very successful in the UK in the prevention of group C meningococcal disease (Snape & Pollard 2005). Protection will be sustained either by incorporating a meningococcal conjugate vaccine into the EPI programme or through periodic mass vaccination campaigns of unvaccinated children. Provided that there are no problems in the development process and that the international community is prepared to help with the costs of mass immunisation, conjugate vaccines provide a means by which it should finally be possible to make epidemics of meningococcal meningitis in Africa a matter or interest only to historians.
Although epidemic meningococcal disease is one of the conditions most feared by those who live in the African meningitis belt, and responsible for thousands of deaths each year, it has received little attention from the international research community and little more is known about the reasons for the peculiar epidemiology of this disease in Africa than was the case 100 years ago. However, work on the development of conjugate vaccines has awakened more interest in this condition in the international research community and among donors and this may lead to more epidemiological and immunological studies that will help to elucidate the reasons for the peculiar epidemiology of this infection in Africa. It is likely that the new group A conjugate vaccines will achieve much of their impact through their effect on nasopharyngeal carriage. However, little is known about carriage of the group A meningococcus in the African meningitis belt and the factors that influence it, for example the effects of age and season and the impact of the host's immune response. Without this information it will not be possible to identify the key targets for mass vaccination campaigns. More research on epidemiology of nasopharyngeal infection with the meningococcus in Africa is needed urgently.