Rickettsia felis: from a rare disease in the USA to a common cause of fever in sub-Saharan Africa


  • P. Parola

    1. Unité de Recherche en Maladies Infectieuses et Tropicales Emergentes (URMITE), UMR CNRS 6236 – IRD, WHO Collaborative Center for Rickettsioses and Other Arthropod Borne Bacterial Diseases, Faculté de Médecine, Université de la Méditerranée, Marseille Cedex 5, France
    Search for more papers by this author

Corresponding author: P. Parola, Unité de Recherche en Maladies Infectieuses et Tropicales Emergentes (URMITE), UMR CNRS-IRD 6236-198, Faculté de Médecine, 27 Boulevard Jean Moulin, 13385 Marseille Cedex 5, France
E-mail: philippe.parola@univmed.fr


Clin Microbiol Infect 2011; 17: 996–1000


Rickettsia felis is a spotted fever group rickettsia that has been definitely described in 2002. Within the last 20 years, there have been a growing number of reports implicating R. felis as a human pathogen, parallel to the fast-growing reports of the worldwide detection of R. felis in arthropod hosts, mainly the cat flea Ctenocephalides felis felis. R. felis is now known as the agent of the so-called flea-borne spotted fever, with more than 70 cases documented in the literature. Recently, two studies respectively conducted in Senegal and Kenya, have challenged the importance of R. felis infection in patients with unexplained fever in sub-Saharan Africa. We focus here on the epidemiological and clinical aspects of R. felis infection. More studies are needed, including the study of other arthropod vectors, but it can be speculated that R. felis infection might be an important neglected agent of fever in sub-Saharan Africa.

Rickettsia felis is an obligate intracellular Gram-negative bacterium belonging to the spotted fever group (SFG) of Rickettsia [1,2]. It was probably first detected in European cat fleas (Ctenocephalides felis felis) in 1918 and tentatively named ‘Rickettsia ctenocephali’ [3]. This work was, however, overlooked until 1990, when a Rickettsia-like organism (the ‘ELB agent’, for the Elward Laboratory, Soquel, CA, USA) was found in C. felis fleas by electron microscopy [4]. The species R. felis was formally validated by molecular criteria in 2001, and the reference strain was isolated and definitely characterized in 2002 [5]. Particularly, it was shown that that this rickettsia can be cultivated at low temperature only. Recent outcomes include the study of the genome of R. felis [6]. The presence of up to two plasmid forms has been shown, as well as strain variation of plasmid content [7,8].

A Worldwide-distributed Rickettsia

Since its definitive description, the interest in R. felis and its association with fleas has increased [1,9]. The use of molecular assays, including regular and quantitative real-time polymerase chain reaction (qPCR) has provided rapid and reproducible tools to detect R. felis in arthropods. R. felis has now been described in infected arthropods from more than 20 countries on five continents (Fig. 1). In America, after the USA, Brazil and Mexico, it has been found in fleas from Peru, Uruguay, Chile, Argentina [1,9] and recently Hawaii [10], Canada [11], Panama [12], and the Caribbean Island St Kitts [13]. In Europe, R. felis has been detected in fleas from Spain, France, the UK, Portugal, Cyprus [1,9] and recently Germany [14] and Italy [15]. In Asia, it has been detected in Japan, Indonesia, Thailand, Afghanistan, Israel [1,9] and recently Laos [16], Taiwan [17] and Lebanon [18]. It has also been detected in Australia, New Zealand [1,9] and New Caledonia [19]. Finally, in Africa, R. felis has been detected in fleas from Algeria, Ethiopia, Gabon [1,9], and recently the Ivory Coast [20] and Morocco (Parola P, unpublished data).

Figure 1.

 Distribution by 2011 of clinical findings of Rickettsia felis infections (yellow stars) and of arthropods infected with R. felis (red circles). The numbers in the stars indicate the number of clinical cases. In these countries (but not yet Tunisia, Senegal and Kenya) R. felis has also been detected in arthropods. The red circles indicate locations where R. felis has been detected in arthropods (mainly fleas) only. Updated from [29].

Beside C. felis, new arthropods have been found to be infected with R. felis, including: other flea species (C. canis, C. orientis, Anomiopsyllus nudata, Archaeopsylla erinacei, Ctenophthalmus sp., Xenopsylla cheopis X. brasilliensis, Tunga penetrans, Ceratophyllus gallinae, Spilospsyllus cuniculi and Echidnophaga gallinacean); ticks (Haemaphysalis flava, Rhipicephalus sanguineus, Ixodes ovatus, and Carios capensis); and chigger (South Korea) and mesostigmata mites (Taiwan) [1,9,20]. The reported hosts for these vectors were mainly cats, dogs and rodents, and more rarely opossums, hedgehogs, horses, sheep, goats, gerbils and monkeys [1,9].

However, most of the reports state that the cat flea C. felis is the most recurrent arthropod in which R. felis has been detected. Furthermore, C. felis is currently the only known biological vector of R. felis [9]. Within the flea, R. felis infection is disseminated, having been identified in the midgut, ovaries and salivary glands [21]. Studies examining the transmission of R. felis using colonized cat fleas have shown stable vertical transmission (transovarial and trans-stadial) [9,22]. Vertical transmission of R. felis persists in C. felis for at least 12 generations without the aid of an R. felis-infected bloodmeal; however, over successive generations prevalence wanes to low levels (10%) [22]. It has also been recently shown that cat fleas are able to acquire R. felis infection from an infectious blood meal [23]. These data further support the wide distribution of R. felis in relation to the worldwide distribution of C. felis. The cat flea is extremely common on cats and dogs in many temperate and tropical regions, but it also infests opossums, raccoons and rats [24]. It represents the great majority of fleas in human homes. If associated with an infectious agent, the worldwide distribution of C. felis represents therefore a threat to the human population because of lack of host specificity of the cat flea, and its ability to bite people.

The Emerging Flea-borne Spotted Fever

In 1994, ‘ELB agent’ DNA fragments were detected in blood samples obtained from a patient from Texas in 1991 [25]. This became the first evidence of R. felis’ potential as a human pathogen. More arguments for the pathogenicity of R. felis for humans were provided in 2000 in Mexico, when three patients with fever, exanthem, headache and central-nervous system involvement were diagnosed with R. felis infection by specific PCR of blood or skin and seroconversion to rickettsial antigens [26]. All the patients had had contact with fleas or animals known to carry fleas. In 2001, high antibody titres to R. felis were found in two French patients with clinical signs of rickettsioses and two of 16 Brazilian patients with febrile rash. Moreover, specific sequences of R. felis were identified in the serum of one Brazilian patient [27]. In 2002, two cases of typical spotted fever were reported in an adult couple in Germany. Clinical features included fever, marked fatigue, headache, generalized maculopapular rash and a single black, crusted, cutaneous lesion surrounded by a halo, and enlarged, painful lymph nodes in the inguinal region for the man. Serological techniques discriminated R. felis infection among several rickettsiae for the woman and this was confirmed by detection of R. felis DNA in the woman’s sera [28]. Thereafter, within the last 20 years, there have been a growing number of reports implicating R. felis as a human pathogen, parallel to the fast-growing reports of the worldwide detection of R. felis in arthropod hosts [1]. In 2009, together with the report of a French case, a PubMed search found reports (case reports and seroprevalence studies) of 68 R. felis infections [29]. Cases of the so-called flea-borne spotted fever (also referred to as cat flea typhus) had been reported in the Americas (Brazil and Mexico), Asia (Thailand, South Korea and Laos), Northern Africa (Tunisia and Egypt) and Europe (France, Germany and Spain) [29]. The available clinical findings for 34 persons infected with R. felis included fever (32 cases), cutaneous rash (24 cases, mostly maculopapular), cutaneous eschar (four cases), neurological signs (five cases), digestive symptoms, cough without pneumonia (three cases), and pneumonia (two cases) [29,30]. However, clinical findings for R. felis are often confused with those found for patients with murine typhus caused by R. typhi, or other febrile illnesses, and they appear to be more complex and more severe than initially thought. Since 2008, a case has been diagnosed with acute polyneuropathy-like symptoms [30]. Recently, two cases of subacute meningitis caused by R. felis have been documented in Sweden [31]. More interestingly, a case of what has been documented by serology as murine typhus was reported in 2008 in Israel [32]. At the end of 2010, the authors revised the diagnosis, which was retrospectively documented by molecular tools as a R. felis infection, the first in the Middle-East region [33].

Like other rickettsiose, R. felis infections can be diagnosed by serological testing, microimmunofluorescence (MIF) being the method of reference [34]. However, most commercially available MIF assays offer a very limited selection of antigens (e.g. R. rickettsii in the United States or R. conorii and R. rickettsii in France for the spotted fever group; R. typhi for the typhus group). Even national reference centres may routinely test for only a few other SFG rickettsiae (e.g. R. akari or R. africae). Serological profiles for R. felis infections differ; cross-reactions with SFG rickettsiae as well as with SFG rickettsiae and R. typhi have been observed, but a lack of cross-reactions has also been observed [29,34]. In this context, it is important to remind the practising physician that MIF may be adequate to diagnose the class of infection (e.g. a spotted fever rickettsiosis) but is likely to be insufficient to definitively identify R. felis unless other, more-sophisticated, serological assays (Western blot and cross-absorption studies) are performed, or blood or tissue samples can be evaluated by culture or PCR-based methods [34]. However, although real-time PCR assays are becoming increasingly used as sensitive and rapid tools to detect and identify rickettsiae in blood and skin biopsy specimens throughout the world, these facilities are not available everywhere, particularly in developing countries. As for many rickettsioses, the reference treatment is based on doxycycline [34].

An Common Agent for ‘Fever of Unknown Origin’ in Subsaharan Africa?

Until recently, almost nothing was known about R. felis in sub-Saharan Africa. The bacterium had been detected in fleas from Ethiopia in 2001 [27], in Gabon in 2005 [35], and in the Ivory Coast in 2009 [20], but no case had been documented. In July 2010, two studies were published in the same issue of Emerging Diseases, challenging the importance of R. felis infection in patients with unexplained fever [36,37] (Table 1).

Table 1. R. felis as an emerging cause of fever of unknown origin in sub-Saharan Africa in non-malaria patients
CountryLocation(inhabitants)Period ofthe studyPatientsPrevalence of R. felis infectionOverall incidenceAverage age of infected patients in years (range)Incidence in children<10 years oldSeasonality trends
Senegal (36)Dielmo (391)November 2008 to July 2009103 1.7% (7/391) 3.5% (5/143) 
Ndiop (313) 31 1.3% (1/313) 0.9% (1/109) 
Total (704) 134 (204 samples)4.4% (9/204) 15 (2–57)6/252 (vs. 2/452; p 0.02)Monthly incidence from 2.38% (1/42) in July to up to 16.7% (3/18) in April
Kenya (37)Garissa, NorthEasternProvince23 months from 1August 2006through to 30June 2008.1633.7% (6/163)29 (7–47)5/6 from September through to February (rainy season)

In Senegal, our team investigated the origin of unknown fever in Senegalese patients who had a negative test result for malaria [36]. We focused on potential Rickettsia spp. infection as a cause of fever. Sampling took place during November 2008 to July 2009 in two rural Senegalese villages in the Sine-Saloum region: Dielmo (13°43′N, 16°24′W) and Ndiop (13°41′N, 16°23′W). A total of 204 samples from 134 patients were tested by two quantitative real-time PCRs (qPCR) for all spotted fever group rickettsiae. We identified nine samples from eight patients (6%) that were positive by both genus-specific qPCR systems; seven from Dielmo and one from Ndiop. The prevalence of flea-borne spotted fever in all tested samples was 4.4%. The overall incidence was 1.7% in Dielmo and 0.3% in Ndiop (7/391 vs. 1/313; p 0.06). Reasons for the significantly different prevalence of these infectious diseases in the two geographically close villages remained unexplained. A higher attack rate of flea-borne spotted fever was identified in children <10 years of age, with an attack rate of 3.5% in Dielmo during a 9-month period. Monthly incidence for positive samples was up to 16.7% in April. Beside fever and unspecific symptoms (weakness, headache with sleep disorders, and digestive and respiratory signs), no rashes and no eschars were found in infected patients. However, another rickettsial study in an indigenous African population reported that cutaneous rash might be imperceptible in patients with pigmented skin. Overall, this study provided definitive molecular evidence for R. felis infection in West Africa in the initial days of infection in febrile Senegalese patients who did not have malaria [36].

In Kenya, a similar study has been conducted by another team aiming to determine the aetiologic agents of non-malaria acute febrile illnesses in the semi-arid North Eastern Province, Kenya, during inter-epidemic periods of Rift Valley fever, for 23 months during 2006–2008 [37]. A total of 163 patients with fever were included. Two quantitative real-time PCR (qPCR) assays were used to screen for rickettisal nucleic acid in the patients’ serum samples. A R. felis-specific qPCR was also used. At the end, nucleic acid preparations of serum from 6 (3.7%) of 163 patients were positive for rickettsial DNA as determined by qPCR and were subsequently confirmed by molecular sequencing to be positive for R. felis. The six febrile patients’ symptoms included headache, nausea, and muscle, back and joint pain. No patients reported skin rash, and no information about possible eschars was available. Most of the cases were diagnosed during August 2007 to June 2008. All patients reported contact with livestock animals in a place where all livestock owners in the region have dogs that assist with livestock herding and security. This was the first evidence of R. felis infection of humans in Kenya, and the first in eastern sub-Saharan Africa [37].

Unanswered Questions

Much of the ecology of the agent of flea-borne spotted fever remains unresolved, including R. felisC. felis interactions. The role of other arthropods in the life cycle of R. felis and the epidemiology of human infections also deserves more investigations. R. felis infection cases have been reported globally. This may be linked to the worldwide distribution of the as yet only known flea vectors. However, the occurrence is relatively rare when compared with the high frequency of R. felis infections related to flea infestation.

On the other hand, two studies conducted with robust methods in western and eastern sub-Saharan Africa reported in 2010 a significant incidence of R. felis infection within patients with fever of unknown origin [36,37]. It has been speculated in both studies that these patients might have been exposed to fleas. However, recent studies raised the question of potential arthropod vectors, different to those that have been already associated with R. felis. Arthropod cell lines capable of supporting R. felis growth include those of Toad tadpole (Xenopus laevis), ticks (Ixodes scapularis) and mosquitoes (Aedes albopictus and Anopheles gambiae). Recently, R. felis has been detected and then isolated in psocid Liposcelis bostrychophila (Psocoptera: Liposcelidae), which are common and cosmopolitan household pest insects often intimately associated with humans and other vertebrates [38,39].


R. felis is an emergent rickettsial pathogen with a worldwide distribution. Cases remain poorly characterized and are apparently under-appreciated, possibly because of the lack of specific signs and symptoms, and the low availability of appropriate laboratory tests. Ideally, patients in sub-Saharan Africa with unexplained fever should be tested for R. felis infection using molecular methods. To date, this technique is, however, poorly available locally, with the exception of specific projects. The current understanding of R. felis biology leaves several issues that need to be addressed. More studies are needed, including the study of other arthropod vectors, but it can be speculated that R. felis infection might be an important neglected agent of fever in sub-Saharan Africa.

Transparency Declaration

The author declares no conflicts of interest.