Molecular Diagnostics of Rickettsia africae Infection in Travelers Returning from South Africa to The Netherlands

Abstract

Background:

African tick-bite fever (ATBF) is frequently diagnosed in The Netherlands in travelers returning from South Africa. It is caused by Rickettsia africae and diagnosis is based on travel history and clinical presentation and usually confirmed by detecting serum antibodies against rickettsiae of the spotted fever group. However, these typically occur late in the course of the disease, and a mild clinical course or early antibiotic treatment can diminish antibody production.

Methods and Results:

Four travelers presented with (sub)febrile temperatures and eschar(s), several days after returning from South Africa. R. africae DNA was amplified and sequenced from skin biopsies of the eschars of all patients. Initial immunofluorescence assays yielded no immunoglobulin M (IgM)/IgG antibodies directed against spotted fever group rickettsiae; however, serology in the convalescent phase—several weeks after the patients had fully recovered—was positive.

Conclusions:

ATBF should be considered in travelers returning from South Africa to The Netherlands with febrile illness and (multiple) skin lesions. The diagnosis can be confirmed by (paired) serology; however, polymerase chain reaction and sequencing on skin biopsies could be a (faster and more accurate) confirmatory test. Advantages of molecular methods over serology are species identification and high sensitivity early in the course of the disease.

Introduction

A frican tick-bite fever (ATBF) has a high prevalence in Africa and is a frequently diagnosed disease amongst visitors to this region (Raoult et al. 2001). It is the only tick-transmitted rickettsiosis in which several inoculation eschars are observed in a high proportion of cases (Raoult et al. 2001, Jensenius et al. 2003). ATBF is caused by Rickettsia africae, a tick-borne Gram-negative bacterium, belonging to the spotted fever group of the genus Rickettsia. Spotted fever group rickettsiae are intracellular microorganisms and may cause zoonotic infection (Raoult 2005). ATBF was described in the 1930s by Pijper (Troup and Pijper 1931). Kelly (1992) isolated the agent from an eschar. The principal vectors for ATBF are Amblyomma hebraeum, a tropical tick species that transmits R. africae in southern Africa, and Amblyomma variegatum, which is prevalent mainly in the sub-Saharan area (Raoult et al. 2001). Sporadic cases of R. africae have been described in Guadeloupe and the West Indies, where it is probably transmitted by A. variegatum (Parola et al. 1998).

R. africae is the second most frequent cause of systemic febrile illness among travelers returning from sub-Saharan Africa (Freedman et al. 2006). Rickettsia conorii is also prevalent in Africa and has a comparable, but more severe, clinical presentation compared with ATBF. The incidence, however, is lower and infection usually presents with a single eschar (Raoult et al. 1986). Further, the differential diagnosis of multiple eschars comprises serious, but rare, diseases in returning travelers, such as other rickettsial infections or cutaneous anthrax, warranting swift and accurate diagnostic tests. More commonly, insect/spider/arthropod bites or cutaneous infections with staphylococci or streptococci can mimic eschars.

Although ATBF is mainly a clinical diagnosis, it is usually confirmed by detecting antibodies in serum against rickettsiae of the spotted fever group. However, these typically occur late in the course of the disease and early antibiotic treatment or a mild course of the disease can diminish antibody production. In addition, in many countries, including The Netherlands, differentiation between R. conorii and R. africae infection by serology is not possible. The confirmation of the diagnosis of ATBF relies on detection of cross-reactive antibodies against other spotted fever group rickettsiae (e.g., R. conorii and R. rickettsii). In the outpatient clinic of the department of Tropical Medicine of the Academic Medical Center (AMC), from 2008 until 2010, 19 patients presented with fever and one eschar, and 6 patients presented with fever and multiple eschars. Thirteen of these patients had detectable antibodies against spotted fever group rickettsiae (unpublished data). Twenty-one patients had visited South Africa, and four had visited other countries in sub-Saharan Africa. Based on geographical and epidemiological considerations, these patients were probably infected with R. africae. This number is an underestimation, because (telephonic) consultations by colleagues or general practitioners were not recorded. Data collected via Geosentinel, of which the AMC is a surveillance center, showed that 197 patients returned from South Africa to 41 Geosentinel sites with a spotted fever group rickettsiosis from 1998 to 2006 (Jensenius et al. 2009).

In contrast to serology, polymerase chain reaction (PCR) is a rapid, sensitive and specific test to diagnose R. africae infection (Fournier and Raoult 2004; Fenollar and Raoult 2007). In several case reports or series, PCRs of eschar biopsies in patients with clinically confirmed ATBF were all positive (Bouvresse et al. 2008, Oteo et al. 2004, Tsai et al. 2009). In line with these observations, in a study by Raoult et al. (2001), a suicide PCR—in which primers are used only once to avoid contamination—on skin biopsies demonstrated 100% sensitivity; all of 15 patients with serologically confirmed R. africae infections were also positive with PCR. Fournier and Raoult (2004) described lower sensitivities for both regular and suicide PCRs, for example, in 25 of 34 clinically confirmed cases of ATBF (73.5%), DNA of R. africae was amplified using a suicide PCR. It should be noted that several patients were already taking antibiotics. Both regular and suicide PCR were more sensitive than culture.

We here describe four patients returning from South Africa with fever and skin lesions in which we confirmed the diagnosis of ATBF, that is, R. africae infection, by PCR and sequencing on DNA obtained from skin biopsies of the eschars, a technique, to the best of our knowledge, not routinely applied.

Patients

Patient 1

A 66-year-old man visited our travel clinic at 6 days after returning from a 15-day journey to South Africa. His medical history revealed benign prostate hypertrophy. His vaccination status was adequate. He had not been taking malaria prophylaxis and had not been in direct contact with freshwater bodies. He presented with headache, fever (38°C–39°C), and multiple skin lesions. Symptoms began at 1 week before presentation, which was 1 day after a visit to a private game park nearby Kruger National Park. He had walked through high grass surrounding a small body of water. He had not noticed insect bites or any ticks. On physical examination, six eschars on his legs, chest, and back were found (Fig. 1a). Further, a discrete spotted exanthema was observed on his trunk. His body temperature was 37.6°C aurally (after taking antipyretics). Laboratory investigation revealed a white blood cell count in the low normal range (leukocytes: 4.7×10⁹ cells/L), normal platelet count (thrombocytes: 175×10⁹ cells/L), a slightly elevated C-reactive protein level (15.2 mg/L), and normal liver and renal function. Thus, our patient was suspected of ATBF and treatment with doxycycline 100 mg twice daily orally for 1 week was initiated. To confirm our diagnosis, we took biopsies from the eschars for histological examination and PCR and performed serology (upon presentation and after convalescence). PCR was also performed on blood.

FIG. 1.

Eschars of all four patients. (A) One of the eschars on the trunk of patient 1 with an erythematous, yellow-centered lesion. (B) Eschar on the right thigh of patient 2 with an erythematous lesion with a necrotic center (“tache noir”). (C) Eschar on the left groin of patient 3. (D) Eschar on the back of patient 4.

The first patient has been described in detail. The other three patients had similar presentations (Table 1 and Fig. 1a–d).

Table 1.

Clinical Findings

	Patient 1	Patient 2	Patient 3	Patient 4
Age	66	66	65	56
Sex	Male	Male	Male	Male
Medical history	Benign prostate hypertrophy	Rickettsial infection (not further specified)	Aortic valve replacement	None
Travel details; risks	Game park nearby Kruger National Park (South Africa); walking through grass	Limpopo and Mpumalanga (South Africa); game hunting	Kruger National park (South Africa); removed a tick	Lion Farm northeast to Pretoria (South Africa); sitting on grass
Travel companions found to have rickettsiosis	No	Yes^a	No	No
Symptoms and signs	Headache, fever, skin lesions (Fig. 1a)	Fatigue, myalgia, skin lesions (Fig. 1b), regional lymphadenopathy	Fever, skin lesion (Fig. 1c), regional lymphadenopathy	Fever, fatigue, headache, skin lesions (Fig. 1d)
Days between return travel and onset of symptoms	1	5	3	6
No. of eschars	6	2	1	2
Aural temperature (°C)^b	37.6	38.2	37.1	37.4
Leukocytes (4–10.5×10⁹ cells/L)^b	4.7	5.4	5.0	6.0
Platelets (150–400×10⁹ cells/L)^b	175	172	157	228
C-reactive protein (<5 mg/L)^b	15.2	18.3	—	15.8
Therapy	Doxycycline 100 mg twice daily orally, for 7 days	Doxycycline 100 mg twice daily orally, for 7 days	Doxycycline 100 mg twice daily orally, flucloxaxillin 1000 mg three times daily orally, both for 7 days	Doxycycline 100 mg twice daily orally, for 7 days
Outcome	Complete recovery within 1 week	Complete recovery within 1 week	Complete recovery within 1 week	Complete recovery within 1 week

Rickettsiosis diagnosed and treated elsewhere.

On presentation.

Materials and Methods

Skin biopsies and histological examination

Two-millimeter punch biopsies were taken from the “tache noir” of the eschars after disinfection and local injection of 2% lidocaine. After fixation in buffered formalin and paraffin embedding, 4-μm-thick sections were stained with hematoxylin and eosin for histomorphological analysis and with Gram stain, Grocott's methenamine silver stain, periodic acid Schiff-diastase, and Ziehl-Nielsen's stain, respectively, for detection of microorganisms. In addition, transmission electron microscopy (EM) was performed on glutaraldehyde-fixed skin biopsies.

Serology

Blood was drawn upon presentation of patients (acute phase) and several weeks after presentation (convalescent phase). In both samples, immunoglobulin G (IgG) and IgM antibodies against R. conorii and R. rickettsii (spotted fever group) and Rickettsia typhi (typhus group) were tested as part of the routine diagnostic setting (Erasmus Medical Center, Rotterdam), using a commercially available immunofluorescence assay (IFA) (Focus Diagnostics, Inc., Cypress, CA). Routine diagnostic testing does not include testing for R. africae-specific antibodies. Tests were performed according to the manufacturer's instructions. All samples were screened with dilutions of 1:64 and 1:256. Positive samples were titrated to end-point fluorescence in twofold dilutions. The positive cutoff value for IgG and IgM was set at ≥1:64. As there was no specific R. africae IFA available, serologic confirmation of the diagnosis of ATBF relied on the detection of antibodies against R. conorii (or R. rickettsii), as there is cross-reactivity among spotted fever group rickettsiae, to which also R. africae belongs. Cross-adsorption assays were not performed.

PCR and sequencing

The primers used for PCR have been previously described (Christova et al. 2003, Tijsse-Klasen et al. 2010). The 16S rRNA gene of Rickettsia species (360 bp) was amplified using the HotStarTaq master mix (Qiagen, Hilden, Germany) under the following conditions: 15 min at 94°C; then cycles of 20 s at 94°C, 30 s at 72°C, and 30 s at 72°C, lowering the annealing temperature by 1°C in each cycle until reaching 62°C; and then 40 cycles at this annealing temperature, followed by a final elongation step for 7 min at 72°C. All samples were analyzed on agarose gels. Positive samples were subjected to PCR of an independent marker. The citrate synthase gene (glt), 850 bp in length, was amplified using primers CS409d and Rp1258n (Roux et al. 1997), using HotStarTaq master mix under the following conditions: 15 min at 95°C, then 40 cycles of 30 s at 94°C, 30 s at 54°C, and 55 s at 72°C, followed by a final elongation step for 7 min at 72°C. PCR products were sequenced by dideoxy-dye termination sequencing of both strands at the RIVM sequencing facilities using the ABI 3100 automated sequencer (Applied Biosystems, Foster City, CA) and compared with sequences in GenBank (www.ncbi.nlm.nih.gov/) using BLAST. The sequences were aligned and analyzed using BioNumerics 6.1 (Applied Maths, Kortrijk, Belgium). To minimize cross-contamination and false-positive results, negative controls were included in each batch tested by the PCR. As a positive control, Rickettsia helvetica DNA was used. Further, DNA extraction, PCR mix preparation, sample addition, and PCR analysis were performed in assigned separate laboratories.

Results

Histological examination of skin biopsies showed extensive superficial ulceration and angiocentric inflammatory infiltrates in the underlying dermis and subcutis, which were composed of mononuclear cells, granulocytes, and large amounts of nuclear fragments. Affected vessels showed fibrinoid necrosis of the vessel wall, endothelial swelling, and focal thrombus formation, representing a severe type of leukocytoclastic vasculitis (Fig. 2a). All histochemical staining tests for bacteria were negative. With the use of transmission EM, no viable bacteria could be detected. However, many macrophages and endothelial cells contained large amounts of myelin figures (condensed concentric lamellar remnants of lipoprotein layers), which can be interpreted as degenerate or necrotic cells from the lesion or as remnants of bacteria (Fig. 2b). No viable microorganisms were visualized, presumably because biopsies were taken after treatment had been initiated.

FIG. 2.

Histological examination. (A) Light microscopy detail of the dermis showing multiple severely damaged, partly necrotic vessels (arrowheads) surrounded by inflammatory infiltrate, extravasations of erythrocytes, and nuclear dust. Hematoxylin and eosin stain (magnification×100). The inset shows a fibrin thrombus inside a vessel lumen (×200). (B) Ultrastructural detail of two intradermally located macrophages showing large amounts of aggregated intracytoplasmic myelin figures (*), as shown by transmission electron microscopy. N=nucleus. Scale bar=2 μm.

Two to 3 weeks after initiation of treatment, several days after the patients had completely recovered, serologic tests on acute phase sera proved to be negative for spotted fever group rickettsiae. In convalescent sera, antibodies against spotted fever rickettsiae (i.e., R. conorii) were detected in all four (4/4) patients (Table 2).

Table 2.

Diagnostic Results

	Patient 1	Patient 2	Patient 3	Patient 4
Serology in acute phase^a
Rickettsia conorii IgM	<1:64	<1:64	<1:64	<1:64
R. conorii IgG	<1:64	<1:64	<1:64	<1:64
Rickettsia typhi IgM	<1:64	<1:64	<1:64	<1:64
R. typhi IgG	≥1:256	≥1:64	<1:64	<1:64
Serology in convalescent phase³
R. conorii IgM	<1:64	<1:64	<1:64	<1:64
R. conorii IgG	≥1:256	≥1:64	≥1:256	≥1:256
R. typhi IgM	<1:64	<1:64	<1:64	<1:64
R. typhi IgG	≥1:256	<1:64	<1:64	≥1:64
Rickettsia rickettsii IgM	≥1:64	Not done	<1:64	<1:64
R. rickettsii IgG	≥1:256	Not done	≥1:256	≥1:256
PCR biopsy: 16S rRNA	Rickettsia africae positive	R. africae positive	R. africae positive	R. africae positive
PCR biopsy: gltA	R. africae positive	R. africae positive	R. africae negative	R. africae positive
Homology sequencing^b	100% (16S and gltA)	100% (16S and gltA)	100% (16S)	100% (16S and gltA)
PCR blood	Negative	Not done	Not done	Negative
Histology	Superficial ulceration, leukocytoclastic vasculitis	Superficial ulceration, leukocytoclastic vasculitis	Not done	Not done
Electron microscopy	No viable bacteria, myelin figures	Not done	Not done	Not done

Positive serologic results are indicated in boldface.

Sequencing of 16S rRNA DNA was compared with R. africae sequences present in GenBank (L36098 and CP001612). Sequencing of gltA was compared with the gltA gene of R. africae from GenBank (CP001612).

PCR, polymerase chain reaction.

In skin biopsies of all patients, only several days after initiation of treatment, the 16S rRNA gene of the Rickettsia species was detected by amplification of the 16S rRNA gene using a PCR with generic rickettsia-specific primers. After removal of the primer sites, sequences were obtained from all samples over a length of at least 339 bp and were 100% homologous to the R. africae sequences present in GenBank (L36098 and CP001612). In addition, the presence of R. africae was confirmed by amplification and sequencing of gltA in three of four patients. After removal of primer sites, high-quality sequences were between 688 and 727 bp in length and showed 100% similarity to the gltA gene of R. africae from GenBank (CP001612). Neither the 16S rRNA gene nor the gltA gene was detected in blood (cases 1 and 4). In Table 2, the diagnostic results of all four patients are presented.

Discussion

We describe four travelers returning from South Africa who were found to have ATBF with (multiple) eschars. In all patients, the presence of R. africae was readily demonstrated by PCR and sequencing, a technique that it is not widely used, especially not in The Netherlands, for confirmation of the diagnosis. These patients were treated with doxycycline and promptly recovered. One patient was also presumptively treated with flucloxacillin, as the eschar had an atypical presentation and Staphylococcus aureus infection was suspected.

People who travel the countryside in areas where R. africae is endemic and have contact with ticks, e.g., hunters such as our second patient (Table 1), are at risk of acquiring ATBF (Raoult et al. 2001, Jensenius et al. 2003). Amblyomma ticks mainly feed on cattle and wild ungulates (Norval 1979), but also on humans. As opposed to the Rhipicephalus species, the vector of R. conorii, Amblyomma ticks, aggressively seek their host and respond to stimuli such as CO₂ concentration, NH₃, and body temperature (Norval et al. 1989). This may explain the multiple eschars and the appearance of ATBF in groups of returning travelers, as demonstrated by our second case, in which two of the patients' travel companions, all were hunters, also had symptoms consistent with ATBF (Table 1).

ATBF is usually confirmed by detecting antibodies in serum against rickettsiae of the spotted fever group with IFA using glass slides. Other serologic tests include latex agglutination, enzyme immunoassay (ELISA), western blot, or complement fixation. Antibodies against R. africae typically occur late in the course of the disease (Fournier et al. 2002). Also, early antibiotic treatment or a mild course of the disease can diminish antibody production. Indeed, previous studies have described ATBF patients of which a fairly large proportion was PCR positive, but no antibodies against rickettsiae could be detected (Raoult et al. 2001, Stephany et al. 2009, Althaus et al. 2010). In all our cases (4/4), initial serologic assays, performed early after onset of symptoms, were negative. In convalescent sera, antibodies against spotted fever group rickettsiae (i.e., R. conorii) were detected in all (4/4) patients. In contrast, a preliminary analysis of serology performed in the Erasmus Medical Center on sera of approximately 150 travelers returning from South Africa and clinically suspected of having contracted a rickettsiosis (and thus most likely ATBF) showed that only approximately one third could be confirmed by serology (data not shown).

In The Netherlands, serologic tests specific for R. africae are not available and cross-adsorption tests, previously described as highly specific, are not routinely used (Raoult et al. 2001). The combination of a clinical presentation with multiple eschars and positive R. conorii serology is considered sufficient for diagnosis. Thus, R. conorii infections with multiple eschars, R. africae infections with a single eschar (as our third patient), or R. africae infections with multiple eschars but an absent antibody response may be missed or remain unconfirmed. The therapeutical consequences of exact identification of the rickettsial species might be limited. However, there have been reports of R. conorii associated with severe complications such as meningo-encephalitis, septic shock, and multiorgan failure and mortality can be as high as 20% (DeMeester et al. 2010). Generally, R. africae has been described as a benign disease with a quick recovery following tetracycline therapy (Raoult et al. 2001), although Roch and collaborators (2008) have described a severe clinical course and slow recovery of ATBF in elderly patients. Additionally, from an epidemiological point of view it might be useful to identify the exact rickettsial species by PCR.

A major advantage of PCR and sequencing over serology is that it can be applied to detect microorganisms in skin biopsies in the acute phase, allowing for a quick diagnosis, differentiation between all Rickettsia species, and then additional testing to confirm or exclude differential diagnoses or additional therapy unnecessary. The PCR and sequencing techniques described here allowed for a diagnosis within several days, whereas novel molecular diagnostics, such as multiplex real time PCR, could result in even faster diagnosis or confirmation.

PCR assays on blood samples may be considered in patients without skin lesions returning from an area where ATBF is endemic. However, sensitivity and specificity are much lower compared with PCR on eschar biopsies (Raoult et al. 2001, Jensenius et al. 2004). Finally, culture of skin biopsies can also be used to identify rickettsia but sensitivity is low (Raoult et al. 2001) and should only be performed in specialized laboratories routinely working with intracellular bacteria (Norval et al. 1989).

In conclusion, we here demonstrate molecular evidence of R. africae infection causing ATBF with (multiple) eschars in four patients returning from South Africa to The Netherlands. PCR on skin biopsies and subsequent sequencing is only minimally invasive and should be considered as an important alternative or additional diagnostic method for ATBF, leading to quick diagnosis, initiation of treatment, and identification of the rickettsial species.

Footnotes

Acknowledgments

The authors express thanks to M.A. van den Bergh Weerman of the Department of Pathology in the Academic Medical Center, Amsterdam, for assistance with transmission EM and to Dr. H. Zaaijer, virologist of the Academic Medical Center, Amsterdam, for providing serological data obtained in the AMC. Also, the authors thank Dr. A.M. Nijhof of the Institute for Parasitology and Tropical Veterinary Medicine, Freie Universität Berlin, Germany, for his valuable comments on this manuscript.

Disclosure Statement

The authors declare that they have no conflict of interest.

References

Althaus

, Greub

, Raoult

, Genton

. African tick-bite fever: a new entity in the differential diagnosis of multiple eschars in travelers. Description of five cases imported from South Africa to Switzerland. Int J Infect Dis, 2010; 14,suppl 3:e274–e276.

Bouvresse

, Giudice

, Franck

, Buffet

et al.

Two cases of cellulitis in the course of African tick bite fever: a fortuitous association?

Dermatology, 2008; 217:140–142.

Christova

, Van De Pol

, Yazar

, Velo

, Schouls

. Identification of Borrelia burgdorferi sensu lato, Anaplasma and Ehrlichia species, and spotted fever group Rickettsiae in ticks from Southeastern Europe. Eur J Clin Microbiol Infect Dis, 2003; 22:535–542.

Demeester

, Claus

, Hildebrand

, Vlieghe

et al. Diversity of life-threatening complications due to Mediterranean spotted fever in returning travelers. J Travel Med, 2010; 17:100–104.

Fenollar

, Raoult

. Molecular diagnosis of bloodstream infections caused by non-cultivable bacteria. Int J Antimicrob Agents, 2007; 30,suppl:s7–s15.

Freedman

, Weld

, Kozarsky

, Fisk

et al. Spectrum of disease and relation to place of exposure among ill returned travelers. N Engl J Med, 2006; 354:119–130.

Fournier

, Jensensius

, Laferl

, Vene

et al. Kinetics of antibody responses in Rickettsia africae and Rickettsia conorii infections. CVI, 2002; 9:324–328.

Fournier

, Raoult

. Suicide PCR on skin biopsies for diagnosis of Rickettsiosis. J Clin Microbiol, 2004; 42:3428–3434.

Jensenius

, Davis

, von Sonnenburg

, Schwartz

et al. Multicenter geosentinel analysis of rickettsial diseases in international travelers, 1996–2008. Emerg Infect Dis, 2009; 15:1791–1798.

10.

Jensenius

, Fournier

, Raoult

. Tick-borne rickettsioses in international travelers. Int J Infect Dis, 2004; 8:139–146.

11.

Jensenius

, Fournier

, Vene

, Hoel

et al. African tick bite fever in travelers to rural sub-equatorial Africa. Clin Infect Dis, 2003; 36:1411–1417.

12.

Kelly

, Matthewman

, Beati

, Raoult

et al. African tick bite fever- a new spotted fever group rickettsiosis under an old name. Lancet, 1992; 340:982–983.

13.

Norval

. Tick-borne infestations and tick-borne diseases in Zimbabwe Rhodesia. J S Afr Vet Assoc, 1979; 50:289–292.

14.

Norval

, Andrew

, Yuner

. Pheromone mediation of host selection in bont ticks (Amblyomma hebraeum Koch) Science, 1989; 243:364–365.

15.

Oteo

, Portillo

, Blanco

, Ibarra

et al. Rickettsia africae infection. Three cases confirmed by PCR. Med Cliln (Barc), 2004; 122:786–788.

16.

Parola

, Jourdan

, Raoult

. Tick-borne infection caused by Rickettsia africae in the West Indies. N Engl J Med, 1998; 338:1391.

17.

Raoult

. Rickettsioses and Ehrlichoses. Mandell

, Bennett

, Dolin

et al. Principles and Practice of Infectious Diseases, 5th. Philadelphia: Elsevier, 2005; 2284–2295.

18.

Raoult

, Fournier

, Fenollar

, Jensenius

et al. Rickettsia africae, a tick borne pathogen in travellers to sub-saharan Africa. N Engl J Med, 2001; 334:1504–1509.

19.

Raoult

, Weiller

, Chagnon

, Chaudet

et al. Mediterranean spotted fever: clinical, laboratory and epidemiological features of 199 cases. Am J Trop Med Hyg, 1986; 35:845–850.

20.

Roch

, Epaulard

, Pelloux

, Pavese

et al. African tick bite fever in elderly patients: 8 cases in French tourists returning from South Africa. Clin Infect Dis, 2008; 47:e28–e35.

21.

Roux

, Rydkina

, Eremeeva

, Raoult

. Citrate synthase gene comparison, a new tool for phylogenetic analysis, and its application for the rickettsiae. Int J Syst Bacteriol, 1997; 47:252–261.

22.

Stephany

, Buffet

, Rolain

, Raoult

et al.

Rickettsia africae infection in man after travel to Ethiopia [letter]

Emerg Infect Dis, 2009; 15:11.

23.

Tijsse-Klasen

, Fonville

, Reimerink

, Spitzen-van der Sluijs

et al. Role of sand lizards in the ecology of Lyme and other tick-borne diseases in the Netherlands. Parasit Vectors, 2010; 3:14.

24.

Troup

, Pijper

. Tick-bite fever in Southern Africa. Lancet, 1931; 1183–1186.

25.

Tsai

, Lu

, Huang

, Fournier

et al. African tick bite fever in a Taiwanese traveller returning from South Africa: molecular and serologic studies. Am J Trop Med Hyg, 2009; 81:735–739.