19b Other bacterial diseasesAnaplasmosis, ehrlichiosis and neorickettsiosis

In 2001, taxonomic reorganization of the bacterial genera Anaplasma, Ehrlichia, Cowdria and Neorickettsia resulted in the transfer of numerous species between these taxa, and the renaming of the transferred species to reflect their new taxonomic position (Dumler et al. 2001). Among the members of these genera, there are four species of established zoonotic importance (Table 19b.1), which are therefore the subject of this chapter. Two of these species were affected by the changes outlined above.

Although these four species possess markedly different ecologies, they share the fundamental biological character of being obligate intracellular bacteria that reside within vacuoles of eukaryotic cells. This lifestyle underlies their fastidious nature in the laboratory and hence our limited knowledge of their biology and pathogenicity. Nonetheless, despite this shortfall, all four are associated with diseases of established or emerging importance:

The first three pathogens are transmitted by hard (ixodid) ticks and are encountered across the temperate zones of the northern hemisphere (and maybe beyond), although the vast majority of human infections caused by them are currently reported in the USA. There, HME and HGA are second only to Lyme disease (caused by Borrelia burgdorferi) in terms of public health significance. Furthermore, given that there is evidence of increasing

population sizes and changing distributions for ixodid species (Scharlemann et al. 2008), it is not unreasonable to predict that the infections they transmit will present an increased medical burden in the future. N. sennetsu remains an enigmatic pathogen; case reports remain scarce, but serological surveys suggest high levels of exposure. The widespread consumption of raw fish across east Asia presents specific infection risks to this region, and an increased awareness that sennetsu neorickettsiosis is among the infections that can be acquired from this source is required before its public health importance can be accurately assessed.

E. chaffeensis and E. ewingii can still be considered as recent discoveries, having both been first described in the last 20 years. HME was first recognized in the USA in the mid 1980s (Maeda et al. 1987), and E. chaffeensis was demonstrated to be its aetiological agent a few years later after being isolated from a HME patient (Anderson et al. 1991). E. ewingii was first described in 1992, when it was implicated in canine granulocytic ehrlichiosis (Anderson et al. 1992), but it was not recognized as a zoonotic pathogen until 1999, when PCR was used to detect DNA from the species in the blood of four patients from Missouri, USA with suspected ehrlichiosis (Buller et al. 1999). In the USA, national collation of reports of E. chaffeensis infections has been ongoing for several years. These figures reveal that there was an almost four-fold rise in the number of cases between 2004 to 2008 culminating in almost 1,200 reports. In 2009, this number declined, although about 800 cases were still reported (Centers for Disease Control and Prevention (CDC) 2010). Conversely, reports of E. ewingii remain rare; only 15 cases were collated during 2008 and 2009 (CDC 2010).

A. phagocytophilum was first implicated as a human pathogen in 1994 (Bakken et al. 1994; Chen et al. 1994). However, prior to this date, it had long been recognized as a veterinary pathogen, associated with pasture fever and tick-borne pyaemia in livestock primarily in Northern Europe (Woldehiwet 2010). The first reported cases of HGA were in patients living in Midwestern USA, and, despite A. phagocytophilum having a distribution stretching across most of the temperate regions of the northern hemisphere, the vast majority of subsequent cases have also been from the USA. Like Ehrlichia infections, those caused by A. phagocytophilum are reportable in the USA, and the collation of data has shown that between 2004 and 2009, over 4,500 HGA cases were reported nationally (CDC 2010). Elsewhere in the world, HGA was first reported in Slovenia in 1997 (Petrovec et al. 1997), and a handful of further cases have been encountered in that country (Lotric-Furlan et al. 2004). Occasional case reports have demonstrated the presence of HGA in numerous other countries.

The earliest reports of an illness compatible with sennetsu neorickettsiosis, associated with the consumption of raw fish on Kyushu, the most south-westerly of the four main islands that comprise Japan, date from the 1800s (Misao and Katsuta 1956; Rikihisa 1991). However, demonstration of the pathogenic role N. sennetsu was not obtained until far more recently, with its recovery in mice from the samples collected from a patient with a typical clinical history and mononucleosis in the 1950s (Misao and Kobayashi 1954). This achievement provoked a flurry of activity in which the isolate was further characterized and its physiology explored, but further cases were not reported and the disease apparently vanished. However, more recently, serological evidence of exposure to N. sennetsu among febrile Malaysian patients was reported (Ristic 1990) together with the acquisition of an isolate from the blood of one patient (Weiss et al. 1990). Very recently, PCR-based demonstration of N. sennetsu infection in a febrile patient from Laos was reported, together with evidence of a high prevalence of anti-N. sennetsu antibodies in a sample of over 1,000 blood donors and febrile patients from the country (Newton et al. 2009).

The primary natural reservoir of E. chaffeensis is the white-tailed deer (Odocoileus virginianus). Experimental inoculations of white-tailed deer have shown them to be susceptible to infection and capable of transmitting E. chaffeensis to its natural vector, the lone star tick (Amblyomma americanum). In deer, infections are chronic and characterized by bacteraemia and sequestration of bacteria in lymph nodes and bone marrow for at least several months (Davidson et al. 2001). The range of the white-tailed deer extends across most of the USA, northwards into Canada and south, through Central America as far as Peru. Serological surveys for E. chaffeensis antibodies in cervids in numerous states in the USA have revealed a generally high seroprevalence, with typically at least 25% (and sometimes a far higher proportion) of animals having evidence of exposure. PCR-based surveys for E. chaffeensis DNA in deer tissues are generally concordant with serosurveys (Yabsley 2010). Low infection prevalences in deer tend only to be encountered on the periphery of the range of Am. americanum. Other mammals have also been implicated in the natural maintenance of E. chaffeensis in the USA, and, given the broad host range of Am. americanum, the exploitation of diverse vertebrates by the pathogen is feasible. Raccoons (Procyon lotor) and possums (Didelphis virginiana) feed all three life stages of Am. americanum, and serosurveys of these species have indicated that both are naturally exposed to E. chaffeensis. However, experimental infections of raccoons resulted in only short-lived infections (Yabsley et al. 2008). Serological or PCR-based evidence of natural infections have also been reported for coyotes (Canis latrans) and captive and free-ranging lemur species. Rodents do not appear to play a role as a reservoir for E. chaffeensis (Yabsley 2010).

Am. americanum is a three-host tick with a catholic feeding behaviour, although white-tailed deer are the most important hosts for all three life stages. Larvae and nymphs will, however, also feed on other small and medium sized mammals and birds, and, within its natural range, Am. americanum is the tick most commonly found on humans (Merten and Durden 2000). The distribution of Am. americanum is considerably more limited than that of white-tailed deer, and most cases of HME occur where the population densities of Am. americanum are highest, in the south-central, south-eastern and, increasingly, eastern USA. Although E. chaffeensis cases have been reported in most US states, not all are thought to be autochthonous and doubts have been raised about the accuracy of diagnosis in some cases (Paddock and Childs 2003). Survey of ticks has demonstrated the widespread presence of E. chaffeensis at varying prevalences, typically under 10%, but rising to almost 30% in some locations (Yabsley 2010). Although some circumstantial evidence for the role of ticks other than Am. americanum in the transmission of E. chaffeensis, further studies are required to confirm their involvement.

There is now clear evidence that E. chaffeensis, or at least closely related organisms, are present outside of the USA. A study in China (Gao et al. 2001) reported the presence of DNA from an E. chaffeensis-like organism in the blood of febrile, tick-bitten patients, and the same diagnostic approach was used to identify infection in a Venezuelan child (Martinez et al. 2008). In Cameroon, DNA from an organism indistinguishable from E. chaffeensis was detected in the blood of 12/118 patients with fever of unknown origin (Ndip et al. 2009). Numerous reports have presented serological evidence of infection in most regions of the world, but in the absence of other diagnostic evidence, these must be treated with a degree of caution. Serological and PCR-based surveys of wildlife and livestock have indicated the existence of an E. chaffeensis-like agent in many parts of the world including East Asia and South America, and PCR-based surveys of various ticks in Asia, Africa and South America have also demonstrated the presence of such organisms (Yabsley 2010).

Both dogs and deer have been implicated as important reservoir hosts for E. ewingii. A survey of 88 dogs, most with suspected acute ehrlichiosis, but some apparently healthy, revealed the presence of E. ewingii DNA in the blood of 20 dogs (18 of which were sick and two of which were healthy), and provoked the suggestion that dogs serve as a reservoir for the pathogen (Liddell et al. 2003). The presence of E. ewingii DNA has been demonstrated in the blood of white-tailed deer, and naïve fawns inoculated with blood drawn from infected deer were susceptible to the pathogen and remained bacteraemic for 68 days (Yabsley et al. 2002). Like E. chaffeensis, E. ewingii is transmitted by Am. americanum. An early study demonstrated that this tick, but not other species, was capable of experimental transmission of E. ewingii from infected to naïve dogs (Anziani et al. 1990). E. ewingii DNA has been detected in host-seeking Am. americanum ticks in several states in the USA at a prevalence of infection of up to 5% (in adults), but more often somewhat lower (Wolf et al. 2000).

A. phagocytophilum is a generalist species, capable of exploiting numerous mammal species as reservoir hosts and multiple ixodid tick species as vectors, and is encountered throughout the temperate regions of the northern hemisphere, from the western coast of north America, across Europe, North Africa and Asia, and as far east as Siberia and Japan. Early interest in A. phagocytophilum related to its aetiological role in infections of livestock in northern Europe, including sheep and cattle (MacLeod and Gordon 1933; Woldehiwet 2010). It was first recognized in North America as a parasite of voles (Tyzzer 1938), then as a veterinary pathogen in horses (Gribble 1969) and subsequently in dogs (Madewell and Grimble 1982). A. phagocytophilum infections in deer in the UK were first described in 1965 (McDiarmid 1965). The demonstration of A. phagocytophilum as a human pathogen in the early 1990s, resulted in renewed efforts to define the transmission network of the bacterium. In the USA, the importance of rodents in this network is well established (Telford et al. 1996; Foley et al. 2008), although other species including deer (Massung et al. 2005), foxes (Gabriel et al. 2009), and maybe even reptiles (Nieto et al. 2009) are also thought to be involved. In Europe, a similarly wide range of wildlife species have been implicated as reservoir hosts including rodents, insectivores, cervids, and even bears (Vichova et al. 2010; Woldehiwet 2010). Rodents and cervids also appear to be key reservoir hosts for A. phagocytophilum in Asia (e.g. Kawahara et al. 2006; Zhan et al. 2009). There is currently no evidence to indicate that birds are involved in the natural maintenance of A. phagocytophilum.

Several Ixodes species have been implicated in the transmission of A. phagocytophilum including, in North America, Ixodes scapularis (Pancholi et al. 1995), Ixodes pacificus (Barlough et al. 1997), and Ixodes spinipalpis (Holden et al. 2003, Zeidner et al. 2000), in Europe and North Africa, Ixodes ricinus (MacLeod and Gordon 1933), Ixodes trianguliceps (Bown et al. 2003), and Ixodes ventalloi (Santos et al. 2003), and in Asia, Ixodes persulcatus (Cao et al. 2000), and Ixodes ovatus (Ohashi et al. 2005). It is likely that other Ixodes species are also competent vectors for the pathogen, and there is some evidence than ticks from other genera may also fulfil this role; A. phagocytophilum DNA has been detected in questing members of the genera Haemaphysalis (Barandika et al. 2008; Kim et al. 2003; Yoshimoto et al. 2010), Dermacentor (Baldridge et al. 2009; Cao et al. 2006; Holden et al. 2003), Hyalomma and Rhipicephalus (Barandika et al. 2008; Sarih et al. 2005; Toledo et al. 2009). Furthermore, the vector competence of Dermacentor albipictus has been demonstrated in an experimental study (Baldridge et al. 2009). However, A. phagocytophilum does not appear to be transmitted by Am. americanum, the vector of E. chaffeensis and E. ewingii (Ewing et al. 1997).

The extent to which ticks are able to maintain A. phagocytophilum trans-ovarially, and the relative importance of this transmission mode for the natural persistence of the species, remains uncertain. Early studies on I. ricinus found no evidence for trans-ovarial transmission (MacLeod and Gordon 1933) and subsequently, consensus agreed with these findings. However, conflicting evidence has occasionally been published; for example, a report of the natural transmission of infection to sentinel mice by an I. spinipalpis larvum (Burkot et al. 2001), and, perhaps more compelling, experimental evidence for infection in a small proportion of experimentally-reared D. albipictus larvae (Baldridge et al. 2009).

There is increasing evidence that although A. phagocytophilum can be considered a generalist, within the species are subpopulations, or ecotypes, adapted to specific host and vectors. An ecotype specifically adapted to exploiting white-tailed deer rather than other mammals has been encountered in the USA (Massung et al. 2002; Massung et al. 2005) and studied in some depth (Massung et al. 2003; Massung et al. 2007). Evidence for other host-restricted A. phagocytophilum ecotypes in the USA has also emerged (Foley et al. 2009). Furthermore, the existence of different ecotypes circulating in distinct yet coexisting transmission cycles has been described in the UK (Bown et al. 2009). Interestingly, the basis of this distinction appears to be the adaptation of A. phagocytophilum ecotypes to specific vector species rather than their reservoirs (Bown et al. 2009).

Very little is known about the ecology of Neorickettsia sennetsu. Its natural cycle is unknown but is thought to involve infection of fish-associated trematodes. Human infection results from the consumption of uncooked fish that are parasitized by infected trematodes. An organism referred to as the SF agent has been identified in Stellantchasmus falcatus flukes parasitizing gray mullet (Mugil cephalus)(Wen et al. 1996), but this organism does not appear to be the same as the human-infecting N. sennetsu. The recent rediscovery of sennetsu neorickettsiosis in Laos (Newton et al. 2009) provoked exploration of possible sources of infection, with N. sennetsu DNA being detected in a gill tissue samples from a climbing perch (Anabas testudineus).

All these pathogens are obligate intracellular Gram-negative bacteria belonging to the order Rickettsiales within the α subclass of the proteobacteria. E. chaffeensis and N. sennetsu have a tropism for monocytes, whereas E. ewingii and A. phagocytophilum parasitize neutrophils. Given the role of these cell types at the frontline of innate antimicrobial defence, it is remarkable that bacteria have evolved to exploit such potentially toxic targets. The unique molecular and cellular mechanisms that underlie this exploitation have begun to be unravelled, and themes common to Ehrlichia and Anaplasma have emerged. Both genera of bacteria possess cell walls that lack peptidoglycan and lipopolysaccharide (LPS), and the absence of these molecules strips the bacteria of two of the key triggers of host immunity. Without these two pathogen-associated molecular patterns (PAMPs), the bacteria cannot be effectively recognized by Toll-like receptors and other monocytic and granulocytic cell surface molecules, thereby facilitating their survival inside the host (Rikihisa 2010). Arthropod innate immunity is also responsive to PAMPs, hence the lack of peptidoglycan and LPS in Ehrlichia and Anaplasma species is likely to also facilitate exploitation of their tick vectors.

Differential recognition of cell surface receptors is likely to underlie the different cell tropisms of E. chaffeensis and A. phagocytophilum. Receptors for A. phagocytophilum on the neutrophil cell surface have been identified, and binding to these, and maybe other as yet unrecognized molecules, is thought to provoke endocytosis (Carlyon and Fikrig 2006; Rikihisa 2010). Both Ehrlichia and Anaplasma species remain within the phagosome but inhibit its fusion with lysosomes or NAPDH oxidase components, thereby preventing its acidification and creating an intracellular niche in which to replicate (Carlyon and Fikrig 2006; Rikihisa 2010). Exploration of the influence of intracellular infection on host cell physiology has revealed significant changes in iron metabolism and inhibition of interferon gamma signalling pathways (Rikihisa 2010). The exploitation of neutrophils and monocytes is extended by prolonging their lifespan through the inhibition/delay of apoptosis (Carlyon and Fikrig 2006; Rikihisa 2010). Both E. chaffeensis and A. phagocytophilum upregulate production of several proinflammatory cytokines including IL-1β and IL-8, which may result in the recruitment of additional neutrophils and monocytes to the site of infection (Carlyon and Fikrig 2006; Rikihisa 2010). A. phagocytophilum and E. chaffeensis possess type 4 secretion systems (T4SS) that bring about host cell subversion via the export of bacterial effector molecules, and, in both species, upregulation of T4SS is associated with intracellular growth (Rikihisa 2010). T4SS effector molecules are beginning to be identified; A. phagocytophilum delivers at least two proteins via its T4SS, AnkA and Ats1 (Niu et al. 2010). Whereas AnkA migrates to the neutrophil nucleus, Ats1 travels to, and enters mitochondria, where it is thought to act by reducing the sensitivity of mitochondria to respond to apoptosis-inducing factors, leading to the inhibition of host cell apoptosis (Niu et al. 2010).

Exploitation of reservoir hosts by A. phagocytophilum is characterized by chronic infection. During this time the bacteria must circumvent host immunity. Work outlined above has explored how A. phagocytophilum subverts innate immunity during the acute phase of infection, but, in order to chronically persist, it must also counter acquired humoral responses; antibodies are produced in abundance by infected hosts, but appear to be ineffective in curtailing infection. The basis for this failure appears to be the ability of A. phagocytophilum to express a multitude of antigenic variants during the course of infection, thus, although variants dominating during the first weeks of infection will eventually be countered by specific antibodies, other variants will replace those lost, thereby facilitating persistence. A. phagocytophilum possesses a number of immunogenic surface proteins, but one, termed P44 (MSP2), is immunodominant (Dumler et al. 1995). Exploration of the A. phagocytophilum genome has revealed that it contains more than 100 well dispersed msp2 pseudogenes characterized by conserved sequences flanking a hypervariable region (Caspersen et al. 2002; Zhi et al. 1999), and these pseudogenes are inserted into a unique msp2 expression site by gene combinatorial conversion mechanisms (Zhi et al. 1999; Barbet et al. 2003). Some experimental evidence for the sequential expression of different pseudogenes during the course of prolonged infection has been obtained (Barbet et al. 2003; Granquist et al. 2008; Wang et al. 2004).

As well as possessing a wide diversity of mechanisms for interaction with their reservoir hosts, Ehrlichia and Anaplasma species must also exploit ticks to complete their natural cycles. Very little is known about the mechanisms that underlie this exploitation. However, studies have revealed that A. phagocytophilum persists within the secretory salivary acini of tick salivary glands (Telford et al. 1996), and that tick feeding stimulates the replication and migration of the bacteria from the salivary glands to the mammalian host (Hodzic et al. 2001). Studies suggest that transmission of A. phagocytophilum occurs between 24 and 48 hours after tick attachment (Katavolos et al. 1998), and acquisition of A. phagocytophilum by uninfected I. scapularis larva begins within two days of tick attachment on A. phagocytophilum-infected mice (Hodzic et al. 1998). Once in the tick, A. phagocytophilum moves across the gut wall and infects the salivary glands, a migration that occurs as early as 24 hours after engorgement (Sukumaran et al. 2006). During tick engorgement, A. phagocytophilum is thought to induce expression of salivary gland proteins to facilitate its transmission. One such protein, Salp16, has been characterized and its essential role in A. phagocytophilum survival demonstrated (Sukumaran et al. 2006), although the mechanism underlying this interaction remains to be elucidated. Very recently, it has been shown that A. phagocytophilum also interacts with α1,3-fucose to facilitate its acquisition by ticks (Pedra et al. 2010).

Comparative genomics of A. phagocytophilum, E. chaffeensis and N. sennetsu has been reported following the sequencing of complete genomes for each and for a number of close relatives (Dunning Hotopp et al. 2006). All three species have single, circular genomes, but differ quite markedly in size. The N. sennetsu genome is only 860 kilobases, whereas that of E. chaffeensis is 1.18 megabases and that of A. phagocytophilum is almost 1.5 megabases. These differences are, in part, due to a remarkably large expansion of immunodominant outer membrane proteins that is thought to facilitate antigen variation in Anaplasma and Ehrlichia species. Unlike their near-neighbours within the Rickettsiales, these two genera have also retained the ability to make vitamins, enzyme co-factors and nucleotides for themselves. That N. sennetsu possesses a far less expansive inventory of genes suggests that the complexity of its natural cycle is less than those of A. phagocytophilum and E. chaffeensis, which involve mammalian hosts and arthropod vectors (Dunning Hotopp et al. 2006).

E. chaffeensis, E. ewingii and A. phagocytophilum infections can be transmitted to humans through the bite of a single tick. Onset of symptoms occurs between one week and three weeks after being bitten. All infections range dramatically in severity, from asymptomatic seroconversion to a severe febrile illness and even organ failure and death. Most patients present with non-specific symptoms such as headache, fever, malaise, and myalgia, eliminating the possibility of a reliable clinical diagnosis. However, identifying risk factors can help steer a clinician towards a diagnosis; thus, it is important to establish if a patient has had a recent tick bite, or even if he/she has been pursuing outdoor activity in tick-infested areas at the time of the year when questing ticks are most prevalent. Another potentially useful indicator is the presentation of a similar illness in other family members or even pet dogs (Thomas et al. 2009); temporal and geographical clusters of HME have been reported (Standaert et al. 1995; Yevich et al. 1995). However, as many people do not realise they have been bitten by ticks, a patient’s ignorance of tick bite should not exclude diagnosis of HME or HGA.

A recent meta-analysis of HME and HGA symptoms, signs and clinical laboratory findings indicated that fever was the most common symptom, occurring in over 90% of patients, and that over 75% of patients reported headaches and malaise. Myalgia occurred in 77% of HGA patients, but only 57% of those with HME. Other relatively common symptoms included nausea, vomiting, diarrhoea, cough, and arthralgias. A rash appeared on 31% of HME patients, but only 6% of HGA patients (Dumler et al. 2007). Common clinical laboratory findings include leukopenia (62% HME, 49% HGA), thrombocytopenia (71% HME and HGA), and elevated serum aspartate aminotransferase or alanine aminotransferase levels (83% HME, 71% HGA) (Dumler et al. 2007). Thus, the key to diagnosing HME or HGA is the identification of fever and thrombocytopenia, leukopenia and elevated serum aminotransferase levels in a patient with a history of tick bite or who is likely to have been exposed to tick bite in areas where these diseases are extant. 42% of HME patients and between 33 and 50% of HGA patients are hospitalized, primarily to rule out more sinister differentials (Dumler et al. 2007). Complications are not very common, although between 7–17% of patients may develop serious, life threatening syndromes; HME patients can develop fulminant toxic shock, particularly those with underlying immunocompromise, and HGA infection can progress to involve the central nervous system, with peripheral neuropathies, including brachial plexopathy, demyelinating polyneuropathy and facial palsy being the most common presentations (Dumler et al. 2007). Deaths are extremely rare.

Most commonly, patients with sennetsu neorickettsiosis experience sudden-onset chills and a fever of 38–39°C, which lasts for approximately two weeks. Other manifestations include headache, malaise, myalgias, arthromyalgia, pharyngitis and generalized lymphadenopathy (Tachibana 1986). The most recent case report relates to a patient from Laos (Newton et al. 2009), who presented with a two week history of fever, headache and weakness, and who, on admission, was pale, jaundiced and febrile, with palpable inguinal and cervical lymph nodes and hepato-splenomegaly. He was anaemic but had normal peripheral white cell count except relative and absolute lymphocytosis and raised liver enzymes. Treatment with ofloxacin led to rapid resolution of all symptoms and the patient was discharged five days after admission.

Given the non-specific clinical presentation of Anaplasma and Ehrlichia infections, laboratory confirmation of aetiology is essential. The simplest and cheapest means of providing this confirmation is the microscopic examination of Giemsa or Wright-stained peripheral blood smears. Inclusions of bacteria, termed morulae, stain purple within neutrophils (A. phagocytophilum or E. ewingii) or monocytes (E. chaffeensis). This approach is most sensitive during the first week of infection, but is considerably better for diagnosing A. phagocytophilum infections than those caused by E. chaffeensis; only about 10% of HME patients have visible intra-monocytic morulae, compared to between 25–75% of HGA patients (Dumler et al. 2007). PCR-based methods, performed using blood held on anticoagulant, are now the diagnostic test of choice, particularly when samples are taken soon after the patient has first presented. A variety of assays have been described and their sensitivity is relatively high, ranging from 60–85% for E. chaffeensis infection and 65–90% for A. phagocytophilum infection (Dumler et al. 2007). This difference may be due to A. phagocytophilum provoking more intense bacteraemia, although it may also reflect more sensitive assays; for example, PCR targeting conserved sections of p44 pseudogenes, which exist in excess of 100 copies per genome, is likely to significantly enhance the sensitivity of A. phagocytophilum-specific assays (Courtney et al. 2004). PCR is currently the only definitive test for E. ewingii infection. The sensitivity of both blood smear analysis and PCR-based assays for all three pathogens is adversely affected by antecedent antibiotic treatment, thus, if possible, blood samples should be collected before initiation of therapy (Dumler et al. 2007).

Proof of ongoing infection is unequivocally obtained by in vitro cultivation of bacteria. This is possible for E. chaffeensis and A. phagocytophilum, but not for E. ewingii, which, to date, has resisted all isolation attempts. However, the obligate intracellular nature of these pathogens requires that they are co-cultured with appropriate eukaryotic cell lines and thus the routine use of this approach is limited to a relatively small number of reference laboratories that have the relevant facilities and expertise. E. chaffeensis and A. phagocytophilum grow in a variety of cell lines, but E. chaffeensis is most often isolated by inoculation of monocytes into the DH82 canine histiocytic cell line (Standaert et al. 2000). Typically, co-cultures must be incubated for up to six weeks before infected cells are detected (Dumler et al. 2007). A. phagocytophilum is most frequently isolated by inoculation of leukocytes into HL60, a human promyelocytic leukaemia cell line (Goodman et al. 1996). Again, patience is required as infected cells may not become apparent for several weeks after inoculation (Dumler et al. 2007). Alternative cell lines, specifically those derived from Ixodes ticks, are now being increasingly used for the cultivation of A. phagocytophilum strains (Munderloh et al. 1994).

Serodiagnosis is the most frequently used approach to laboratory confirmation of E. chaffeensis and A. phagocytophilum infections (but not for E. ewingii as, in the absence of isolates, suitable antigen cannot be produced). Although of no use for identifying acute-phase infections, serology has proven an extremely sensitive means of confirming aetiology at two weeks or more after onset of symptoms (Walls et al. 1999). Infections are confirmed using the classical criteria of a four-fold increase in antibody titre or a seroconversion to a titre of 128 or higher. However, a task force aimed at developing a consensus approach to diagnosis of HME has also proposed that a single titre of 64 or greater is highly suggestive of infection, and a single titre of greater than 256 is indicative of infection (Walker et al. 2000). Nonetheless, as IgG antibodies can persist for many months, antibody titres must be considered in the context of the patient’s clinical history (Thomas et al. 2009). Immunoflorescence assays, incorporating whole cell antigens, are the most commonly used format for serodiagnosis, and have a sensitivity rate of about 90% for HME and between 80–100% for HGA (Dumler et al. 2007). The specificity of serodiagnosis of HGA has been estimated to be between 83–100% (Dumler et al. 2007), although cross-reactivity between E. chaffeensis, E. ewingii and A. phagocytophilum antibodies can occur (Buller et al. 1999; Dumler et al. 2007). Furthermore, sera from patients with Rocky Mountain spotted fever, Q fever, brucellosis, Lyme disease and Epstein Barr virus infection, together with those with autoimmune conditions, may provoke false positive serological test results (Dumler et al. 2007), re-emphasizing the importance of a polyphasic diagnostic approach, combining, if possible, a variety of laboratory methods and clinical examination.

The recent case of sennetsu neorickettsiosis was diagnosed using PCR-based assays to amplify N. sennetsu DNA fragments from buffy coat (Newton et al. 2009). This study also used a micro-immunofluoresence assay and Western blotting to screen patients with fever of unknown origin and/or jaundice for N. sennetsu antibodies. The antigen for both assays was whole cell N. sennetsu Miyayama strain (ATCC VR367). Western blotting revealed an apparent N. sennetsu-specific antigen of about 25 kilodaltons size (Newton et al. 2009).

The two antibiotics most commonly prescribed for the treatment of Ehrlichia   and Anaplasma infections are doxycycline and tetracycline. Doxycycline is probably considered the drug of choice as it has a less frequent dosage. Although prolonged or repeated exposure to this antibiotic increases the risk of dental staining, its limited use has negligible effect. Indeed, the Committee on Infectious Diseases of the American Academy of Pediatrics promotes the use of doxycycline for the treatment of ehrlichiosis in children of all ages (Dumler et al. 2007). The recommended dosage for doxycycline is 100 milligrams for adults, and 2.2 milligrams per kilogram for children, given orally every 12 hours for between 5 and 14 days. Tetracycline is orally administered every six hours at a dosage of 500 milligrams per day for adults or 25 milligrams to 50 milligrams per kilogram per day for children. Clinical response to treatment is usually dramatic, with marked resolution of symptoms with 48 hours and full recovery within a few days. Prophylactic antibiotic treatment for those bitten by ticks is not recommended because most people who get bitten by ticks do not develop an Ehrlichia or Anaplasma infection.

The best way for preventing HME or HGA is to avoid tick bites. However, this does not preclude outdoor activities; wearing protective clothing such as boots or solid shoes, long trousers tucked into socks and long-sleeves shirts, or using repellent sprays will markedly reduce the risk of tick bite. Although the rapid removal of ticks from the skin does not necessarily exclude the chance of infection, being aware of tick bite, and informing the clinician of this if symptoms should develop, will greatly facilitate accurate diagnosis of Ehrlichia and Anaplasma infections, hence the prescription of an appropriate treatment regimen.

In cell culture, N. sennetsu is sensitive to doxycycline, ciprofloxacin and rifampicin (Brouqui and Raoult 1990) although the relative efficacies of these antibiotics for the treatment of infections remains untested. A recent case report described apparently successful treatment with ofloxacin (Newton et al. 2009).

N. sennetsu infections are closely linked with the consumption of uncooked fish, and thus are considered preventable by appropriate cooking. However, raw fish form part of the staple diet of many in Eastern and South Eastern Asia, as demonstrated by the widespread human Opisthorchis and Clonorchis liver fluke infections, which are also acquired from this source (Sripa 2008).