Chapter 8: Tuberculosis caused by Other Members of the M. tuberculosis Complex
by Angel Cataldi and Maria Isabel Romano
8.1. Mycobacterium bovis disease in humans
Bovine tuberculosis (TB) is caused by Mycobacterium bovis, a mycobacterium highly similar to Mycobacterium tuberculosis and belonging to the M. tuberculosis complex. The main host of M. bovis is cattle (Bos taurus) but it affects many other mammalians including man. In man, it is the most frequent cause of zoonotic TB, i.e. TB transmitted from animals to humans, which is clinically indistinguishable from TB caused by M. tuberculosis. Before milk pasteurization, M. bovis was an important cause of human TB, especially intestinal TB in children. After the generalized adoption of pasteurization of milk and other dairy products, the occurrence of zoonotic TB dropped sharply.
Very few studies are published on zoonotic TB. In the last 50 years, research on zoonotic TB was influenced by scientific trends, societal worries such as human immunodeficiency virus/acquired immunodeficiency syndrome (HIV/AIDS) and contaminated food, as well as by the availability of tools for the identification of the bovine TB bacillus. For example, the development of the polymerase chain reaction (PCR) and other molecular tools to identify M. bovis and differentiate it from other members of the M. tuberculosis complex have allowed the discovery of more cases in retrospective studies and have suggested new forms of transmission. The medical literature on the incidence of zoonotic TB is marked by numerous clinical descriptions of cases of M. bovis at the regional or nosocomial level, but there are very few systematic surveys of M. bovis diagnosis on a national level (Anon 2003, Barrera 1987, Cousins 1999, Pavlik 1998). There are three main explanations for the absence of an accurate and methodical estimation of the contribution of M. bovis to the global TB burden. First, at the clinical or radiological level, there is no difference between TB caused by M. tuberculosis or that of M. bovis. Second, most laboratories use Löwenstein-Jensen culture medium with glycerol, which does not promote M. bovis growth. Furthermore, cultivation is always an expensive option for many low-income countries compared to the cheaper and faster acid-fast staining. Third, and perhaps most important, in most cases the treatment of TB caused by M. tuberculosis or M. bovis was the same; therefore, there was no clinical interest in differentiating the causative agent.
There is a direct correlation between the prevalence of human TB of bovine origin and that of TB in livestock (Cosivi 1998). At the global level the situation of bovine TB is disparate. In many developed countries bovine TB was eradicated 30-40 years ago by strong campaigns based on tuberculin skin testing (TST) and mandatory sacrifice of animals at the slaughterhouse. In these countries, human TB caused by M. bovis accounts for around 1 % of all TB cases, and sporadic cases occur in elderly people by reactivation of ancient infections or in immigrants from countries where bovine TB has not been eradicated. Importantly, some developed countries, including England or New Zealand, could not completely eliminate bovine TB, or worse, there is a re-emergence of the disease (Thoen 2006). The persistence of M. bovis in wildlife is frequently indicated as the main cause of this re-emergence. On the other hand, in many low-income countries, bovine TB continues to be an important animal health problem. Different epidemiological scenarios can be observed. Meat and bovine products are important resources in some countries, such as Argentina, Brazil, Mexico, and Venezuela, where the number of cattle equals or exceeds that of the human population, and the risk of zoonotic TB could be higher. Less clear is the situation in countries where livestock industry is less developed and intensive, and cattle farming is a family affair for milk consumption or retail commercialization; on the other hand, if the total number of cattle is highly reduced, people live near the animal folds and sometimes consume milk raw. In Central American and African countries, as well as in China, cows are preserved for milk production and meat is consumed from other species such as sheep and swine that are less susceptible to M. bovis. Finally, in India, a high proportion of people do not eat cattle meat but do consume milk and are in close contact with cattle, increasing the risk. Some low-income countries, including Cuba, Mongolia, and Costa Rica are remarkable exceptions, because they have eradicated bovine TB, probably because the cattle population is relatively small.
The epidemiology of zoonotic TB, was recently examined by Thoen et al. (Thoen 2006), who reviewed publications from 1966 on. In this chapter we will concentrate on the most recent findings.
In Africa, there are several reports about the incidence of zoonotic TB. Many of these studies involved research on pre-existing mycobacterial collections or in limited clinical settings, such as in Egypt, Nigeria, Madagascar, Zaire, and Tanzania (Table 8-1). Other authors looked for M. bovis in cattle, as in Ghana and Zambia, where high incidences were described. Another study demonstrated the presence of M. bovis in milk in Tanzania (Kazwala 1998). An excellent review about bovine TB was published by Ayele (Ayele 2004). Genotyping analysis demonstrates different clonal populations depending on the geographical region under study. Due to the consumption of raw milk in regions where AIDS is highly prevalent, many studies concentrated on patients having lymphadenitis. One study in Djibouti showed a low prevalence of M. bovis in those patients (Koeck 2002), while others in Ethiopia (Kidane 2002) and Tanzania (Mfinanga 2004) demonstrated 17 % and 10 % prevalence respectively (Table 8-1).
In Asia, where a policy of bovine TB control was adopted in few countries, there are very few publications on zoonotic TB (Table 8-1). Clearly, a more active search for M. bovis is needed on the Asian continent, where the burden of TB is high.
In Latin America, most of the studies were published in Argentina describing incidence ranging from 0.7 % to 6.2 % in a main milk region, with a much lower national prevalence. In Brazil and Mexico, only one publication was available per country (Table 8-1). In a small collection of human isolates from Chile, M. bovis was not found (Mancilla 2006), a finding that merits a more extended survey in that country (Table 8-1). Genotyping of M. bovis from humans in Argentina showed that the predominating genotype in cattle also predominates in humans, strongly indicating that infection is transmitted from cattle. Furthermore, the majority of patients are related to the farming or meat industry (Zumárraga 1999). In contrast, Romano et al. recently selected isolates with scarce growth in Löwenstein-Jensen media from different hospitals in Argentina, and many of them were M. bovis. Genotyping of these unsuspected M. bovis isolates showed that they belong to a spoligotype that is not predominant in cattle, suggesting that a clone circulates among humans (M. Romano, unpublished observations). The situation of zoonotic TB in Latin America has recently been reviewed by Ritacco et al. (Ritacco 2006).
In the United States, the description of an outbreak of M. bovis TB cases in San Diego was of special interest (Table 8-1). The ingestion of raw milk products by immigrant children was suspected as the source of the infection (Dankner 1993, Dankner 2000). No recent cases were informed from Canada. Other reports also describe a higher incidence of M. bovis among third world immigrants residing in industrialized countries (Cousins 1999, Jalava 2007).
In Europe, in the last decades, M. bovis in humans was reported sporadically (Thoen 2006). The prevalence of TB cases caused by M. bovis is around 1.0 % of all TB cases in the United Kingdom. A similar figure is found in Germany (where Mycobacterium caprae is relevant) and in Spain (Table 8-1). A recent paper described that in Lyon, France, there was no genetic relatedness among nine M. bovis isolates collected from patients over five years, strongly indicating that there is no active transmission (Mignard 2006). In contrast, another typing study in Italy described spoligotyping and Mycobacterial Interspersed Repetitive Units (MIRU) patterns in a collection of 42 isolates, with one genotype accounted for 32 % of all isolates, while the others were unique (Lari 2006). Importantly, a study in the United Kingdom showed that there is no increase in M. bovis disease in humans in spite of an important increase in the incidence of bovine TB (Jalava 2006).
In New Zealand, M. bovis accounts for 2.7 % of laboratory-confirmed human TB cases. Many of the genotypes were identical to patterns from farmed and wild animals (Baker 2006). In contrast, there has been no publications on zoonotic TB in Australia during the last six years.
Download of the entire textbook
(Tuberculosis 2007, 687 pages, PDF, 8.3 MB)
Table 8-1: Bovine TB in humans Country Target of study Main findings Reference Egypt Identification of M. bovis in humans using cultures. 6 % of the total TB cases were caused by M. bovis Elsabban 1992 Egypt Identification of M. bovis in humans using cultures High incidence of M. bovis Nafeh 1992 Nigeria Identification of M. bovis in humans using cultures 3.9 % of TB cases caused by M. bovis Idigbe 1986 Madagascar Identification of M. bovis in humans using cultures 1.25 % of TB cases caused by M. bovis Rasolofo-Razanamparany 1999 Madagascar Genotyping of M. bovis collection A genotype is preva-lent in humans and cattle Rasolofo-Razanamparany 2006 Zaire Identification of M. bovis in humans using cultures High incidence of M. bovis Mposhy 1983 Zambia Large field diagnos-tic test of cattle 33 % of positive herds Cook 1996 Tanzania Detection of M. bovis in milk 6 % of samples posi-tive for M. bovis Kazwala 1998 Burundi Culture of myco-bacteria from hu-man and cattle samples. No M. bovis in hu-mans, 38 % in cattle. Rigouts 1996 Ghana Large field diagnos-tic test of cattle 13.8 % of positive animals. Bonsu 2000 Tanzania Screening of TB patients from rural communities 16 % of TB cases caused by M. bovis Kazwala 2001 Tanzania Genotyping of M. bovis collection Low clustering of cases in humans and cattle Kazwala 2005 Djibouti Biopsies of lymph nodes from TB patients Low prevalence of M. bovis Koeck 2002 Ethiopia Biopsies of lymph nodes from TB patients 17.1 % of samples positive for M. bovis Kidane 2002 Tanzania Biopsies of lymph nodes from TB patients 10 % of samples positive for M. bovis Mfinanga 2004 Nigeria Genotyping of M. bovis collection No common patterns of M. bovis from cattle and humans Cadmus 2006 China Clinical report Case of disseminated TB due to M. bovis Wei 2004 India Screening of CSF from patients Molecular evidence of M. bovis Prasad, 2005 Argentina Identification of M. bovis in humans using cultures 8 % of M. bovis in extrapulmonary TB Peluffo 1982 Argentina Identification of M. bovis in humans using cultures M. bovis identified in 0.47 % of sputum samples Barrera 1987 Argentina Identification of M. bovis in humans using cultures Annual variations of M. bovis going from 0.7 % to 6.2 % of human TB Sequeira 1990 Argentina Identification of M. bovis in humans using cultures 7 % of extrapulmonary TB due to M. bovis Solda 2005 Brazil Mycobacterial cul-tures from children M. bovis provoked 3.5 % of cases of pediatric TB Correa 1974 Chile Identification of M. bovis in humans using cultures No M. bovis isolation Mancilla 2006 Mexico Identification of M. bovis in humans using cultures 3/19 isolates were M. bovis Toledo Ordoñez 1999 Argentina Genotyping of M. bovis collection A genotype is preva-lent in humans and cattle Zumarraga 1999 USA Diagnostic of TB in workers from a dairy farm Risk factor for zoo-notic TB, but no cases demonstrated. Winthrop 2005 USA Diagnostic of TB in immigrant children Pediatric TB due to M. bovis Dankner 1993, 2000 Australia Identification of M. bovis in humans using cultures M. bovis present in immigrant workers United King-dom Identification of M. bovis in humans using cultures No increase of zoo-notic TB in spite of increase in cattle. Jalava 2007 United King-dom Identification of M. bovis in humans using cultures M. bovis provoked 1.0% of TB cases Yates 1988 France Genotyping of M. bovis collection No genetic related-ness among M. bovis isolates collected from patients Mignard 2006 Italy Genotyping of M. bovis collection A genotype accounts for 32 % of human isolates Lari 2006 Germany Identification of M. bovis in humans using cultures M. bovis represents 1 % of all TB cases. 31 % of the isolates were M. caprae Kubica 2003 United King-dom Clinical report An intrafamilial spread of M. bovis Smith 2004 United King-dom National survey for M. bovis M. bovis represents between 0.5 % and 1.5 % of TB cases de la Rua-Domenech 2006 Spain Identification of M. bovis in humans using cultures 9 M. bovis cases in patients in 4 years Remacha 2006 Spain Identification of M. bovis in humans using cultures 0.95 % of all cases of tuberculosis due to M. bovis Esteban 2005 New Zealand Identification of M. bovis in humans using cultures 2.7 % of TB cases due to M. bovis. Many of the genotypes were identical to patterns from animals Baker 2006 In the last 10 years, human disease due to drug-resistant M. bovis has been de-scribed (Blazquez 1997, Hughes 2003, Sechi 2001). One case was of special con-cern because it affected many patients, most of them HIV-positive (Blazquez 1997). This strain spread over Europe and into Canada, and affected 141 patients (S. Samper, personal communication). This fact highlights the high risk of spread of MDR M. bovis, especially in parts of Africa where M. bovis animal disease and HIV human infection co-exist. In humans, the disease caused by M. bovis or M. tuberculosis is clinically indistin-guishable. However, if the physiopathology of TB in humans and cattle are com-pared, some differences are observed. In humans, the apical lobes of the lungs are most affected. In cattle, lesions are most frequently observed in the dorsal caudal lung regions (Cassidy 2006). This part is the most distant from the mouth and nos-trils, meaning that the droplets must travel the longest possible route. In bovines, on the other hand, the lesions are frequently located in lymph nodes associated with the respiratory tract, and not in the lung parenchyma. This observation may be related to the fact that the detection of infected cattle is made in the early stages of disease progression, before the presentation of advanced cavitary lesions. At the histological level, the differences are related to the cell types intervening in the immune response and granuloma formation. For example, the content of gd T cells is much higher in cattle and these cells, as well as neutrophils, participate in granuloma and lesion formation (Cassidy 2001). Nowadays, and due to control campaigns, large liquefied lesions are less frequently observed in cattle in contrast with findings in wildlife where it is possible to observe advanced lesions (Cassidy 2006). Sheep and horses are rarely infected. The infection in goats shows extreme varia-tion according to the geographic location. There are less reports of TB in domestic than in feral pigs. The direct transmission of M. bovis from wildlife animals to humans is much less frequent. Transmission from deer to humans has been re-ported in Canada (Long 1999). Cats, but not dogs, have been reported in several countries as the source of human TB (Fernandez 1999, Underwood 1999, Monies 2000). Importantly, there are no reports of human infection by M. bovis coming from a direct environmental source (Biet 2005). The fact that in situations where the prevalence of M. bovis in cattle is high but does not seem to be associated with a higher incidence in humans may suggest that humans are less susceptible to M. bovis than to M. tuberculosis (de la Rua-Domenech 2006). After the introduction of milk pasteurization, there was a clear impact on the death rate of children under five years of age (Thoen 2006). A recent review (de la Rua-Domenech 2006) described the survival of M. bovis in different foods. M. bovis survives well in cows’ milk. Viable bacilli can be found in yogurt and cream cheese made from unpasteurized milk for up to 14 days after preparation, and in butter for up to 100 days. The consumption of unpasteurized raw milk or milk products is still allowed in many European countries. In low-income countries, consumption of raw milk or dairy products is common in rural areas. The detection of M. bovis in milk from infected cattle is problematic because M. bovis is usually present in low amounts. As contaminating microbiota exist in raw milk, other bacteria and fungi overgrow M. bovis. Decontamination methods ap-plied to other clinical samples with higher bacillary loads, such as sputum or ne-cropsy samples, kill the few M. bovis that may exist in tested milk. This has led to a worrying situation in which there are no validated methods for its detection in milk or milk products. The main problem is the failure of culture as a gold standard. PCR methods for the detection of members of the M. tuberculosis complex in clini-cal specimens were developed in the mid-90s. Although the first developed PCR method used primers directed at a M. bovis specific sequence (Rodriguez 1995), most PCR protocols use primers derived from the insertion sequence IS6110 inser-tion sequence, present in all members of the M. tuberculosis complex. The detec-tion limit in artificially contaminated milk is generally low: 10-1,000 colony form-ing units (cfu) (Zanini 1998, Zumárraga 2005, Antognoli 2001). The sensitivity among tuberculin skin test (TST)-positive cows also varies in different studies, from 11-50 % (Cornejo 1998, Romero 1999, Sreevatsan 2000, Zumárraga 2005). One study in Argentina did not find M. bovis in cattle milk (Perez 2002). This variation is expected, as excretion of M. bovis in milk is sporadic and not all in-fected animals excrete bacilli. There are no published studies on the detection of M. bovis by PCR in cheese. In summary, PCR is powerful in detecting M. bovis in milk but there is an urgent need to validate this technique on a wider level. 8.2. The BCG vaccine: adverse reactions The bacille Calmette-Guérin (BCG) is a live, attenuated vaccine derived from M. bovis. BCG is known to cause local reactions consistent with primary infection with an attenuated strain (i.e. a small localized ulcer and possible regional lymph-adenopathy), and more severe reactions are thought to be rare. Deep ulcers, pro-longed drainage, lymphadenitis (1 %), abscess (2 %) (Turnbull 2002), osteitis (0.04 %) (Kroger 1995), and rarely disseminated infection have all been reported (Albot 1997). The age of the recipient and the dose of vaccine affect the incidence of local com-plications. Disseminated disease is thought to be rare, in the order of 1/1,000,000 doses and directly related to immune dysfunction (Turnbull, 2002). The major worldwide concern about the risk of disseminated infection has been connected to the risk of HIV-related immunosuppression in the recipient. BCG is given routinely to newborns in many countries. However, this practice is under active review be-cause of concerns that the vaccine’s problems may outweigh its efficacy. Some authors recommend that BCG vaccination should be confined to groups of infants with a high risk of TB infection, and should be given at six months of age, in order to reduce severe disease and deaths among infants with immunodeficiency disor-ders (Romanus 1993). From 1993 to 2001, 20 adverse events of BCG vaccination were reported in mem-bers of TB-endemic Aboriginal communities in Canada. Six of these were dissemi-nated disease and five were in children from Aboriginal communities (the sixth one was vaccinated as an infant outside Canada). All of these cases were confirmed as being caused by BCG. One of these children was HIV-infected and the other four had congenital immunodeficiencies, which presented for the first time as dissemi-nated BCG infection. All of these children died as a result of their underlying im-munodeficiency. This rate of disseminated infection indicates a higher rate of un-derlying congenital immunodeficiency in this population and an unanticipated serious risk in this population of BCG recipients (Hutmacher 2002). Health Canada recommends administration of BCG vaccine to all newborn infants who are mem-bers of TB-endemic Aboriginal communities because of the high rate of TB infec-tion and the high risk of serious disease in young children after primary infection. The debate about BCG vaccine in Canada has accelerated as a result of concerns about adverse events (Clark 2006). Several studies have shown clear benefits of BCG vaccination when the risk of tuberculous infection is higher than 1 % per year (Rouillon 1965, Immunization Practices Advisory Committee 1988, Health Canada 2002). These benefits become less clear when the risk of infection is lower than 0.1 %, as rates of severe TB dis-ease and deaths are quite low, regardless of the BCG vaccination policy. Results of these studies are therefore consistent with recommendations of the IUATLD and World Health Organization (WHO), which state that discontinuation of BCG can be considered in populations with an annual risk of tuberculous infection lower than 0.1 % (IUATLD 1994; World Health Organization 2001). The vaccine may be considered in select situations where exposure to TB infection cannot be readily controlled with anti-tuberculous chemotherapy, particularly where multidrug resis-tance is documented. The recipients in this situation may include household con-tacts as well as laboratory personnel and travelers (National Advisory Committee on Immunization 2002). A study in which the vaccine was administered to high risk newborn infants before environmental exposure to mycobacteria could have occurred, showed an overall efficacy of 73 % (range 59 % to 80 %) for disease and 87 % for death (Rosenthal 1961, Fordham von Reyn 2002). The overall trends are that newborns are better protected and primary disease, miliary TB, and TB menin-gitis are better prevented. Another concern about the administration of BCG is its effect on the TST. Because administration of BCG induces a positive skin test of variable size in a large pro-portion of vaccinated individuals, this reaction will affect the interpretation of TST results in contact tracing, thus, jeopardizing the use of a valuable tool in the control of TB transmission in the community. The place of BCG vaccination in TB control programs is being carefully reassessed because of the significant risk of dissemination in immunocompromised patients. BCG has to be administered to newborns from endemic countries in Africa, Asia, and Latin America, because the vaccine is effective for the prevention of dissemi-nated TB and meningitis. However, a careful review and identification of underly-ing risks for immunodeficiency should also be performed. This should include a careful family history for immunodeficiency and prenatal HIV screening. On the other hand, in non-endemic countries, BCG could be discontinued, but if the vaccine is no longer to be given routinely to newborns from endemic communities in these countries, then possible consequences must be anticipated. The rates of miliary TB and meningitis in these infants will increase if ongoing exposure of young infants continues. In the meantime, if the routine infant BCG vaccine pro-gram is abandoned in these communities, this must be compensated for by support of enhanced TB detection and treatment programs. An effective TB prevention and control program requires effective assessment of active disease, effective therapy including DOTS, finding and screening of contacts of infectious cases, and identi-fication and management of latently infected individuals. Otherwise, there is little doubt that infants from these endemic communities will be at increased risk of disseminated primary TB. For example, Sweden moved from the mass vaccination of newborns with BCG to a selective vaccination program for high risk groups. This strategy met with some success, measured at 82 % effectiveness (Romanus 1992). This was accompanied by a higher rate of atypical mycobacterial infection in the non-BCG-vaccinated population (Romanus 1995). In some countries, such as Canada, the high risk population is already being vaccinated, but higher selectivity may be required given the identified risk of the vaccine; possibly limiting newborn BCG use to communities with active cases until the outbreak can be brought under control. Safer and more effective vaccines for TB prevention may soon be available (Do-herty 2005). Alternative vaccines to BCG are on the horizon and it is hoped that they will have better efficacy, be more standardized, and have fewer side effects, especially in immunocompromised individuals, including the HIV-infected popu-lation worldwide, which is at a high risk of TB co-infection. (see chapter 10). One alternative intervention already exists in the form of the early detection and treat-ment of tuberculous infection. The administration of isoniazid is highly effective in reducing the risk of disease (International Union Against Tuberculosis Committee on Prophylaxis 1982) and protection may last for up to 30 years (Hsu 1984). Treatment of infection is generally well tolerated by children (Kopanoff 1978), and compliance is usually much higher than in adults (Wobeser 1989, McNab 2000). Considering the safety issues outlined in this report, the best course of action may be the removal of the BCG vaccine combined with improvements in TB program-ming in non-endemic countries (Vaudry 2003). Such improvements must include early case finding in adults to prevent transmission, and early detection and treat-ment of infection in children through contact tracing and screening in high-risk communities. 8.3. Mycobacterium africanum subtypes M. africanum is predominantly isolated in Africa and, in certain areas of the conti-nent, it is thought to produce a significant proportion of the cases of pulmonary TB (Frothingham 1999, Haas 1997). Reports on the sporadic isolation of M. africanum in Europe and the United States (Desmond 2004) have also been made, including one outbreak of multidrug-resistant (MDR) M. africanum (Schilke 1999). Based on biochemical characteristics, two major subgroups of M. africanum have been described, corresponding to their geographic origin in Western (subtype I) or Eastern (subtype II) Africa. Numerical analyses of biochemical characteristics revealed that M. africanum subtype I is more closely related to M. bovis, whereas subtype II more closely resembles M. tuberculosis (Niemann 2002, Sola 2003). M. africanum subtype II was classified by its resistance to thiophen-2-carboxylic acid hydrazide (TCH). It is the main cause of human TB in Kampala, Uganda (East Africa). Spoligotyping does not lead to a clear differentiation of M. tuberculosis and M. africanum, but all M. africanum subtype II isolates lack spacers 33 to 36, differen-tiating them from M. africanum subtype I. In an IS6110 restriction fragment length polymorphism (RFLP)-based dendrogram, M. africanum subtype II isolates were clustered into two closely related strain families (Uganda I and II) and clearly sepa-rated from M. tuberculosis isolates. An additional characteristic of both M. africa-num subtype II families is the absence of spoligotype spacer 40. In addition, all strains of the M. africanum subtype II family Uganda I also lack spacer 43 (Nie-mann 2002, Viana-Niero 2001). Lack of spacers 40 and 43 are not exclusive mark-ers for the M. africanum subtype II family Uganda I, but might represent a useful additional criterion for M. africanum subtype identification in combination with biochemical test results (Brudey 2004, Mostowy 2004) (see figure 8-1). The gyrB desoxyribonucleic acid (DNA) sequence allows the differentiation of M. africanum subtype I strains from M. bovis, M. caprae, and M. microti. M. africa-num subtype I and M. pinnipedii, however, display identical gyrB DNA sequences and the same occurs with M. africanum subtype II and M. tuberculosis (Niemann 2000). Thus, differentiation of M. africanum subtype II from M. tuberculosis con-tinues to be based on phenotypic characteristics such as growth on bromocresol purple medium. In recent studies, based on the regions of difference (RD) to distinguish M. africa-num, three groups were identified: M. africanum subtype II isolates that have dele-tion of TbD1 and have retained RD9, RD7, RD8 and RD10 intact. Some have sug-gested that these organisms should be included in the species M. tuberculosis. A second group consists of M. africanum subtype I with deletion of RD9, RD7, RD8, and RD10, called 1a. Finally, M. africanum subtype I with deletion of RD9 but not RD7, RD8, and RD10, and called 1b, forms the third group. Both subtype I branches of M. africanum have indistinguishable gyrB sequences. In addition, Mostowy et al. (Mostowy 2004) have recently described several novel RD loci within M. africanum organisms. Among these were RD711 and RD713, deleted in M. africanum subtype Ib, and RD701and RD702, found deleted in a larger study of strains M. africanum subtype Ia (Mostowy 2004). Figure 8-1: Spoligotypes of Mycobacterium tuberculosis complex strains. 1M. africanum sub-type II Uganda 1; 2M. africanum subtype II Uganda 2; 3M. microti vole type; 4M. microti llama type 8.4. Mycobacterium microti disease M. microti is a member of the M. tuberculosis complex and was first isolated in 1937 as the causative agent of pulmonary TB in the wild vole (Microtus agrestis) (Wells 1937). It was considered to be avirulent for humans, cattle and laboratory animals and was therefore proposed as a live vaccine against TB. The efficacy of vaccination with M. microti was assessed in clinical trials in the United Kingdom (Hart 1977) and the Czech Republic (Sula 1976), and indeed the strain was used as a vaccine in Africa for more than 15 years (Fine 1995). In all cases, M. microti proved to be safe and effective in preventing disease, showing a protective efficacy similar to that of BCG. However, M. microti has been recently identified as the causative agent of pulmo-nary TB in both immunocompromised and immunocompetent humans (van Sool-ingen 1998, Horstkotte 2001). Genotypic analysis of M. microti showed the exis-tence of two different variants of M. microti: vole and llama types. The vole type, isolated from voles, ferrets, and pigs, shows hybridization with only two of the 43 spacers in spoligotyping; whereas in the llama-type, the spoligotype-PCR product hybridizes with nine spacer sequences (figure 8-1). The first four M. microti iso-lates from humans in the Netherlands showed spoligotype patterns of the vole-type. Three of these four human M. microti isolates were obtained from immunocom-promised patients (two had undergone kidney transplantation; one was HIV-infected). Two of the patients with M. microti infection had a history of contact with mice, which was found to be suggestive of zoonotic transmission (van Soolin-gen 1998, Brodin 2002). The first case of human infection with M. microti of the llama-type was reported in Germany: the patient was HIV-infected and presented with pulmonary TB (Horstkotte 2001). The time span required for cultivation of vole-type strains (3 and 4 months) is significantly longer than that required for growth of M. microti llama-type strains. The patient with M. microti llama-type infection was successfully treated with isoniazid, rifampin, and pyrazinamide, which indicates that the standard TB therapy is sufficient for treatment of patients with M. microti infection. A possible source of infection could not be identified in this patient. Recent data demonstrated that M. microti can cause severe pulmonary TB in im-munocompetent patients (Niemann 2000a). M. microti has been isolated in Ger-many from two HIV-negative immunocompetent patients with pulmonary TB. According to spoligotype patterns, one of the isolates belonged to the llama type and the other to the vole type. These findings emphasize the