Emerging rickettsial infections in Sri Lanka: the pattern in the hilly Central Province

Transcription

1 Tropical Medicine and International Health volume 8 no 9 pp september 2003 Emerging rickettsial infections in Sri Lanka: the pattern in the hilly Central Province S. A. M. Kularatne 1, J. S. Edirisingha 2, I. B. Gawarammana 1, H. Urakami 3, M. Chenchittikul 4 and I. Kaiho 5 1 Department of Medicine, Faculty of Medicine, University of Peradeniya, Peradeniya, Sri Lanka 2 Department of Parasitology, Faculty of Medicine, University of Peradeniya, Peradeniya, Sri Lanka 3 Department of Microbiology, Faculty of Pharmacy, Niigata University of Pharmacy and Applied Life Science, Niigata, Japan 4 Rickettsial Section, National Institute of Health, Nonthaburi, Thailand 5 Division of Virology, Chiba Prefectural Institute of Public Health, Chiba, Japan Summary objectives To identify different rickettsial infections using a specific immunofluorescent technique in patients clinically diagnosed as typhus fever in the Central Province of Sri Lanka, and to define the clinical picture, assess the severity of infection and to determine the pattern of geographical distribution of the infections of the hospital-based patients. methods A specific indirect immunofluorescent antibody technique was used on the sera of two groups of patients in laboratories in Japan and Thailand. results We serodiagnosed infections with Orientia tsutsugamushi, Rickettsia typhi and spotted fever group in 56 of 118 clinically investigated patients. There were eight infections with O. tsutsugamushi, two with R. typhi and 10 spotted fever group patients with IgM antibodies suggestive of acute infection. Nineteen patients had antibodies against these three rickettsial species, suggestive of past exposure, co-infection or cross-reactivity of antigens. Discrete, erythematous maculopapular rash was common to all three types of infection except for five patients who had no rash. Five patients positive for spotted fever antibodies developed fern-leaf type skin necrosis with severe illness. Duration of the febrile period ranged from 4 to 23 days with defervescence occurring after specific antibiotic treatment. conclusions The study has shown the presence of different types of rickettsial infections in the Central Province of Sri Lanka. The characterization of the clinical picture and the severity of infection provide useful information for the proper management of the patients in the future. keywords rickettsial infections, Orientia tsutsugamushi, spotted fever group, murine typhus, Sri Lanka Introduction Rickettsial infections or rickettsioses (also called typhus) are a major cause of febrile illness throughout the world, but especially in the Asia-Pacific region. They are caused by obligate intracellular coccobacilli within the family Rickettsiaceae and infect humans chiefly through arthropod vectors (Raoult & Roux 1997; Cowan 2000). Rickettsioses include (i) diseases caused by bacteria of the genus Rickettsia, including the spotted fever group (SFG) and the typhus group, and (ii) scrub typhus caused by Orientia tsutsugamushi (OT) (formerly Rickettsia tsutsugamushi). Clinically, a characteristic feature of rickettsial infection is the maculopapular rash (Raoult & Roux 1997; Cowan 2000). However, the extent and the severity of the rash are highly variable. Pathologically rickettsial proliferation occurs in the endothelium of small blood vessels and multiplication of the bacteria in these cells results in vasculitis. Diagnosis of rickettsioses is often made clinically as the specific serological tests are only available in a few laboratories of the world. The therapeutic response to specific antibiotics with a rapid defervescence is strongly suggestive of the correct diagnosis. Reliability of the routinely available Weil Felix serology test for diagnosis of rickettsioses is debatable. It is based on the detection of antibodies to various Proteus antigens that cross-react with rickettsiae (P. vulgaris OX2 with spotted fever rickettsiae, P. vulgaris OX19 with typhus group rickettsiae, and P. mirabilis OXK with O. tsutsugamushi). This test is neither sensitive nor specific (Chenchittikul et al. 1995; ª 2003 Blackwell Publishing Ltd 803

2 Jiatana et al. 1997). Other diagnostic tests include (i) serology, especially immunofluorescence as a reference method for both diagnosis of acute cases and seroepidemiology, (ii) molecular biology: identification method based on PCR, the technique for early diagnosis before seroconversion and for screening arthropods, and (iii) isolation in cell culture or in animal models, which is essential for identification of new rickettsial pathogens (La Scola & Raoult 1997). Information on the status of rickettsioses in Sri Lanka is scarce. Van Peenen et al. (1976) demonstrated low prevalence (prevalence rate <6%) of antibodies against O. tsutsugamushi, the agent of scrub typhus in 1500 venous blood samples obtained from patients, medical students, nurses and blood donors in Colombo, Sri Lanka. The low seroprevalence in urban Colombo is possible as scrub typhus is essentially a rural disease. However, patients presenting with fever and erythematous maculopapular rash have been a common occurrence in some regions of Sri Lanka (S.A.M. Kularatne, personal communication) over the last decade. The clinical impression is that these are rickettsial in nature and many resort to the Weil Felix test for confirmation (S.A.M. Kularatne, personal communication). No large studies have been conducted in Sri Lanka using more specific diagnostic techniques for the diagnosis of rickettsial infections. Hospital-based data show that suspected rickettsial infections are endemic on the western slope of the hilly Central Province of Sri Lanka. In this area, people live on tea estates, in villages, small towns and suburbs, cultivating mainly tea, rice and vegetables. One of the main hospitals that serve the region is the Teaching Hospital Peradeniya (THP). A considerable number of patients are treated for typhus fever in this hospital each year. Patients attending THP were enrolled in this study. Our main objective was to identify different rickettsial infections using a more specific indirect immunofluorescent test in hospital-based patients of Sri Lanka and to describe their geographical distribution. Then the study was extended to define the clinical picture and to assess severity of acute seropositive cases. Indirect immunofluorescent antibody assay (IFA), the reference serology method for the diagnosis and seroepdemiologic studies in rickettsioses, was used in this study. Collaboration with two reference centers (Japan and Thailand) to carry out IFA was considered in order to minimize the laboratory errors and to increase the reproducibility of the results. However, cut-off values of IFA titres were different in the two centres due to the protocols followed by individual laboratories based on their experiences regarding the appropriate cut-off values. Materials and methods Study population All patients with suspected rickettsial infections during the period January 2000 to December 2001 were included in the study. The patients were interviewed and examined on admission and observed periodically during the hospital stay. Further follow-up was made in the outpatient clinic after recovery. The clinical diagnosis was based on fulfilling any two of the following criteria: fever for more than 5 days, erythematous discrete maculopapular rash and rapid defervescence with specific antibiotic (tetracycline). A total of 118 patients fulfilled the above criteria. All details were recorded on individual formatted sheets. Basic demographic data such as age, gender, residence and information on socio-economic status, history of contact with animals and arthropod bites were recorded, and the extent and distribution of skin rash marked on printed figures of the human body. Each figure was divided into eight zones as follows: head and neck, anterior trunk, posterior trunk, arms and forearms, hands, thigh and gluteal regions, legs and feet. A score was given to each zone and total score was calculated for individual patient (Kularatne method, unpublished data). With each patient an effort was made to locate the eschar and to detect skin necrosis. Clinical signs and symptoms were assessed daily and the response to treatment was documented. Routine investigations done in order to facilitate management of patients were: white blood cell count-differential count (WBC-DC), Hb%, platelet count, erythrocyte sedimentation rate (ESR), alanine transaminase (ALT) and aspartate transaminase. The Weil Felix test, which was the only serology test available in Sri Lanka, was used for sera of 64 patients of the series. IFA assay Blood was sampled once between January and June 2000 in 20 patients (first batch) and January to July 2001 in 40 patients (second batch) for IFA. At the time of blood sampling all patients had stayed in hospital for 1 week. Thus serology was derived from single blood samples taken during the illness and not from acute or convalescent sera. Avenous blood sample of 5 ml was collected from each patients and allowed to clot at room temperature, centrifuged at 3000 rpm for 10 min, and the serum stored at )40 C prior to testing. The first batch of sera with added sodium azide was tested at the Department of Microbiology, Niigata College of Pharmacy, Japan, for IFA. The second batch of sera despatched in dry ice was tested at the Rickettsia Section of The National Institute of Health, Department of Medical Sciences, Nonthaburi, Thailand. In 804 ª 2003 Blackwell Publishing Ltd

3 both centers IFA was used for the determination of IgG and IgM antibody levels against three prototype strains of O. tsutsugamushi (serotypes Karp, Kato, Gilliam), Rickettsia typhi (RT) (Wilmington strain) and Thai tick typhus TT118. Rickettsia japonica YH strain was used for unclassified SFG rickettsiae. Additionally, Rickettsia conorii (Moroccan strain) antigen was tested for SFG at the Thai center. The IFA allows the detection of IgG and IgM antibodies or both. IFA identification of specific IgM antibodies to rickettsae species provides strong evidence of recent active infection (La Scola & Raoult 1997). Hence cut-off values were adopted as follows: for the first batch (Japan), (i) titres <20 were considered negative and (ii) titres ranging from >20 to >5120, in the presence of IgM, as indicative of acute infection. For the second batch (Thai) there were two categories. For the testing of O. tsutsugamushi and R. typhi antigens, (i) titres <50 were considered negative and (ii) titres >50 in the presence of IgM as showing acute infection. For the testing of SFG antigens, titres <100 were considered negative, and titres >100 in the presence of IgM as demonstrating acute infection. Furthermore, when sera were reacted with different antigens, the highest titre level above the cut-off values was considered to recognize the organism responsible for the infection. If several antigens gave high titres, they were considered undermined or cross-reacting infection. Data of 60 IFA-tested patients were studied in detail in order to elucidate the patterns of rickettsial infections in the Central Province of Sri Lanka. Results Serology on patients clinically diagnosed as having rickettsial infections Tables 1 and 2 show the results of the IFA in the two batches of sera (Japan and Thailand), respectively. In the first batch (Japan) of sera (n ¼ 20) there were 11 patients positive for O. tsutsugamushi antibodies with seven positive for IgM. Three patients (nos. 12, 13 and 14) were positive for SFG with high titres of both IgM and IgG against Japonica YH and TT118 Thai tick typhus Table 1 IFA results of the first batch of sera (Japan) Antibody titres against O. tsutsuegamushi Antibody titres against SFG and typhus group IgM IgG IgM IgG Case no. G P T G P T YH TT118 R. typhi YH TT118 R. typhi O. tsutsugamushi infection SFG infection Undetermined infections Negative results G, Gilliam; P, Karp; T, Kati; YH, Japonica YH; TT118, Thai tick typhus;, <20. ª 2003 Blackwell Publishing Ltd 805

4 Table 2 IFAT results of the second batch of sera (Thailand) O. tsutsugamushi R. typhi Wilmington R. conorii R. japonica TT-118 No. IgM IgG IgM IgG IgM IgG IgM IgG IgM IgG O. tsutsugamushi infection 1 1:50 * * * * * * R. typhi infection 2 1:200 * * * * * * 3 1:200 * * * * * * 4 1:800 1:400 * * * * * * 5 1:200 1:6400 * * * * * * SFG infection 6 1:50 * * * * * 1:200 7 * 1:200 * * * * 8 * 1:100 * * * * 9 * * * 1:200 1:100 1: * * * 1:100 * * 11 * 1:200 * 1:200 * 1: * 1:200 * * * 1: * * 1:100 1:100 1:100 1: * 1:1600 * 1:200 * * 15 1:200 1:100 1:200 1:200 1:100 * 16 1:50 * * 1:100 1:400 1:100 1: * * * * * 1: * * * * * 1: * 1:100 1:100 1:200 * 1: * * * * * 1: * * * * * 1: :100 1:200 * * * * 23 * * * * 1:100 1:200 Undetermined infections 24 1:50 * 1:100 * * * * 25 1:100 1:100 1:200 * 1:200 * 1: :1600 1:200 1:100 1:50 1:800 1:100 1:3200 1:200 1:200 * 27 1:50 * 1:100 * * * 1: :100 1:50 1:400 1:200 * * 1:200 * 29 1:100 1:200 * * * * * 1: :50 1:100 1:50 1:100 1:200 * * * * 31 1:100 1:100 1:200 1:200 1:400 1:100 * * 32 1:200 1:200 1:100 1:100 * * 1:100 * 33 1:100 1:100 1:200 1:200 * 1:100 * * 1:200 * 34 1:50 1:100 1:400 * * * 1: :100 1:400 * * * * * * 36 1:50 1:100 * * * 1:200 * 37 1:200 1:100 1:200 * 1:400 * 1: :100 * 1:200 * 1:400 * * Negative results 39 * * * * * * 40 * * * * * * G, Gilliam; P, Karp; T, Kati antigens were used for Orientia tsutsugamushi testing;, <1:50; *, <1:100. antigens. There were four patients (nos. 15, 16, 17 and 18) with high titres of antibodies, both IgM and IgG, against O. tsutsugamushi, SFG rickettsiae and R. typhi. Among the 20 sera of this batch there were two (nos. 19 and 20) with insignificant antibody titres against rickettsiae. In the second batch (Thailand) of sera (n ¼ 40) there were 18 sera positive for SFG rickettsiae with seven 806 ª 2003 Blackwell Publishing Ltd

5 Table 3 Types of rickettsial infections identified from first and second batches of sera and distribution of rickettsioses Rickettsiae Japan Thai Total n (%) Acute cases* n (%) Orientia tsutsugamushi (OT) (20) 8 (13) Spotted fever group (SFG) (35) 10 (16) Rickettsia typhi (RT) (7) 2 (3) OT and SFG together (5) OT and RT together (1.6) RT and SFG together (15) OT, RT and SFG together (10) Negative for all (7) * IgM positive patients. in two, and SFG rickettsiae in 11 cases positive for IgG only (negative for IgM). Clinical picture The demographic and clinical data were obtained by analysing clinical features of acute cases of O. tsutsugamushi, murine typhus and SFG rickettsiae (Table 4). There was no definitive pattern of fever but 40 (66%) patients had remittent spikes of temperature above 39 C. The patients who had arthritis belonged to SFG and the commonest joints involved were ankle and wrist with adjacent tendinitis. Three patients with O. tsutsugamushi infection had generalized lymphadenopathy. positive for IgM. Sera of four patients reacted only with R. typhi (Wilmington strain) suggestive of murine typhus, two having high titres of IgM. Only one serum sample was positive for O. tsutsugamushi IgM in this batch. Sera of 15 patients reacted with more than one type of rickettsial antigens. Among this last group there were six positive for O. tsutsugamushi antibodies. The results of the IFA after pooling the two batches (Japan and Thailand) are shown in Table 3. The study had detected antibodies against acute SFG rickettsiae in 10, acute O. tsutsugamushi in eight, acute murine typhus in two, and different combinations of SFG rickettsiae, O. tsutsugamushi and murine typhus rickettsiae in 19 cases. There were O. tsutsugamushi in four, murine typhus Skin rash The following data were obtained by analysing clinical features of acute cases only (Table 4). Seven with acute O. tsutsugamushi infections had discrete erythematous maculopapular skin rash. The rash was absent in one case and four patients had distinct involvement of palms and soles. In these patients, rash lasted for 5 9 days following treatment. None of these patients presented with signs of eschars or necrosis of the skin. All 10 patients with acute SFG had discrete erythematous, maculopapular rash which was dominant in peripheries including face, palm and sole. Three patients in this group developed a rash with patchy skin necrosis. The necrotic lesions were blackish and appeared fern leaf in Table 4 Demographic and clinical data of acute (IgM positive) cases Parameter No. of patients, duration and score of parameters OT (8 cases) RT (2 cases) SFG (10 cases) Age, mean (range) 28.3 (21 63) 54 (45 63) 32 (14 62) Gender (male, female) 6, 2 1, 1 6, 4 Fever duration, days (range) 9.6 (4 17) 18 (13 23) 10.4 (6 21) Rash score, mean (range) 5.5 (3 7) 5 (3 7) 4.7 (2 8) Rash negative cases, (n) Rash with skin necrosis, n (%) (30) Eschar (n) Headache, n (%) 7 (87) 2 (100) 8 (80) Myalgia, n (%) 6 (75) 2 (100) 7 (70) Vomiting, n (%) 4 (50) 2 (100) 5 (50) Cough, n (%) 2 (5) 0 2 (20) Diarrhoea (n) Conjunctival injection, n (%) 6 (75) 2 (100) 6 (60) Enlarged liver, n (%) 1 2 (100) 6 (60) Enlarged spleen, n (%) 1 2 (100) 2 (20) Arthritis, n (%) (60) Icterus, n (%) 0 2 (100) 3 (30) Lymhadenopathy, n (%) 3 (37) 0 1 (10) OT, Orientia tsutsugamushi; SFG, Spotted fever group; RT, Rickettsia typhi. ª 2003 Blackwell Publishing Ltd 807

6 shape with serrated margins, distributed mainly in the extensor aspect of the lower limbs. These patients were clinically ill and had associated severe myalgia and arthritis. With treatment the lesions took days to heal leaving black scabs. One patient had a healed ulcer caused by a tick bite. Two patients with acute murine typhus had erythematous, maculopapular skin rash but none of them had necrosis or eschars. Among the 19 patients with undermined infections (Tables 1 and 2) five were negative for rash while four patients developed fern leaf skin necrosis. These four patients were positive for spotted fever antibodies. Four patients who were negative for rickettsial antibodies (Tables 1 and 2) had typical skin rash (score 7 to 8) and fever for more than 5 days who responded dramatically to tetracycline. Routine laboratory investigations The results of the routine laboratory tests on patients whose sera were subjected IFA showed three patients with leucocytosis and neutrophilia irrespective of the type of rickettsia. However, three patients in the SFG with skin necrosis had leucopenia. Ten patients with skin necrosis had ESR above 50 mm/h and thrombocytopenia (count below ). Three patients with acute SFG who were clinically ill, but without skin necrosis had ESR above 50 mm/h, thrombocytopenia and elevated ALT. Seventeen (77%) of 22 patients who underwent liver biochemical profiles had elevated levels of liver enzyme ALT (range: U/l). The Weil Felix test was carried out on 34 (57%) patients and 25 of them fell into the identified OT, RT and SFG rickettsiae (Table 5). Table 5 Weil Felix test distribution of 25 cases (IFA confirmed) Depicted: positive cases/total cases of test done (%) OXK OX2 OX19 OT acute 2/4 (50) 0/2 (0) 2/6 (33) OT past 4/4 (100) 1/1 (100) 1/8 (12) RT acute 0/2 (0) 1/2 (50) 2/2 (100) RT past 1/1 (100) 1/1 (100) 1/1 (100) SFG acute 2/6 (33) 6/6 (100) 1/6 (16) SFG past 1/8 (12) 4/8 (50) 2/8 (25) Acute, IgM positive in IFA; Past, only IgG positive in IFA. Note: Proteus antigens OXK, OX2 and OX19 are known to cross react with O. tsutsugamushi, SFG rickettsiae, and TG rickettsiae, respectively. Treatment The antibiotics used in treatment were: tetracycline in 45, chloramphenicol in 12 and ciprofloxacin in three patients. Complete defervescence was observed in 1 3 days with antibiotic treatment. However, 12 patients required intravenous hydrocortisone in addition to antibiotics due to deterioration of the general condition with evidence of systemic vasculitis, severe skin rash with extensive necrosis; arthritis, tendinitis, high ESR, thrombocytopenia and deterioration of liver functions. Administration of hydrocortisone caused a dramatic response leading to improvement of general well-being and facilitating recovery. Geographical distribution, socio-economic status and seasonal variation All patients were inhabitants of a stretch of approximately 34 km between Gampola and Mawanella on the western slopes of the hilly Central Province of Sri Lanka. The average daytime temperature of the region fluctuates between 20 and 31 C (mean 29 C). Ten patients with O. tsutsugamushi infection and 14 patients with spotted fever were rural folks while the rest were urban and tea estate dwellers. Twenty patients (76%) from both groups were living in solidly constructed asbestos roofed houses. The majority of dwellings were surrounded by high grown wet zone vegetation. All patients who had dogs in their houses were of middle or lower socio-economical classes. Ten patients (83%) with O. tsutsugamushi infection presented during a period of 2 months from January to March 2000 and all were from locations close to Gampola. The occurrence of spotted fever and undermined infection were scattered throughout the year. However, 62% of cases presented from February to May 2001, suggestive of a seasonal preponderance. With regards to a history of insect or tick bite prior to the febrile illness, the majority was vague with their recollections. However, six patients had clear recollection of what they called ticks bite prior to the illness (four positive for O. tsutsugamushi and two for SFG). There were 10 patients who had regular contact with arthropods likely to be ticks. Of these, four were positive for O. tsutsugamushi infection while six had antibodies for the SFG. Outcome All 60 patients recovered without residual complications. They were followed up to 3 months as outpatients. In sixteen patients the stain of the skin rash persisted for over 28 days. 808 ª 2003 Blackwell Publishing Ltd

7 Discussion For the first time rickettsioses were documented in the Central Province of Sri Lanka using IFA, the reference serological method. The values for IFA titres determined at the two participating centres (Japan and Thailand) differed due to the different cut-off values used by the centres, based on their experience in expressing positive values. Nevertheless, the different cut-off values did not influence the interpretation of results or the conclusions made. Interpretation of IFA (Table 3) according to the significant titres indicates the existence of three rickettsial infections namely: OT (eight acute cases documented here) SFG (10 acute cases) and RT (two acute cases) in Sri Lanka. However, the prevalence and distribution of positive cases of scrub typhus in the two series (Japanese and Thai) differ widely. This could be due to the time difference of 1 year between the two series, despite both being done during the same season. Another possibility is that the 10 cases of OT presented from a single locality. Furthermore, our experience (Japanese) shows that the prevalence of scrub typhus varies very much in the same locality due to changes in the environment, natural and man-made, which will affect mite density. The other cases in the two series were widely scattered. This study detected 118 cases of rickettsioses clinically and of them 60 sera were sent for IFA which yielded 56 positive cases for rickettsioses, testifying to the accuracy of the clinical diagnosis. Also the detection of such a large number of rickettsioses over a 2-year period at a single centre alone points to the possibility of high prevalence of rickettsial infections in Sri Lanka. To confirm and to establish seroprevalence in Sri Lanka, community-based studies involving larger populations are needed. Ten cases of O. tsutsugamushi infection (eight acute and two IgG) appeared as a cluster at the outset of the study followed by low incidence. Significant number of patients had conjunctival injection and lymphadenopathy. Rash was a dominant feature except in one acute case and it healed without necrosis. The clinical illness was less severe and all patients recovered uneventfully. It is noteworthy that none of the patients in the scrub typhus group had detectable eschars. This disagrees with the observations of Cowan (2000) who states that eschars are detectable in 80% of the cases with O. tsutsugamushi infections. The presence of an eschar in scrub typhus is not invariable. It is absent in some infected indigenous people and is known to be rare in Malaysians (Christie 1980). The absence of eschars, pathognomonic of scrub typhus, in our series may be explained in two ways. On the one hand we can question whether these patients really had scrub typhus. The results obtained by an experienced laboratory leave little support for this assumption. On the other hand is the possibility of the presence of an acquired immunity in a population subjected to repeated subclinical infections. The possibility of the latter has to be unravelled only by further studies. The vectors of scrub typhus are known to be larval trombiculid mites. However, no specific information on the type vector of O. tsutsugamushi, transmission patterns or host factors are available for Sri Lanka. In a recent study in China, Wang et al. (2001) demonstrated that Rattus norvegicus and Leptotrombidium deliens as host and vector, respectively, for O. tsutsugamushi in Nan Peng Lie Island. It is also known that chiggers act as vectors and reservoirs of scrub typhus. Studies to determine the possibility of R. norvegicus acting as reservoir for O. tsutsugamushi in Sri Lanka would be helpful in this context. Furthermore, ecological factors for scrub typhus such as rural scrub jungles, jungle grasses and wild rodent populations are abundant in Sri Lanka. There were only two cases of acute murine typhus (R. typhi) cases detected in this study. The clinical picture is similar to scrub typhus except lymphadenopathy which was absent here. There were no eschars and the illness was not severe although the duration of fever was long. Rickettsia typhi is transmitted to man by Xenopsylla cheopis fleas living in Rattus rattus (the black rat) and R. norvegicus (the brown rat). Infected flea faeces either contaminate, human skin or enter, the respiratory tract to cause infection (Raoult & Roux 1997; Cowan 2000). However, reservoir, vector and transmission of R. typhi in Sri Lanka have not been studied. In contrast to scrub typhus and murine typhus, the incidence of SFG appeared to increase. There were 10 acute cases and 11 IgG positive cases detected in this study. Geographically, a few pockets (Kadugannawa, Mawanella) stand out as having a higher prevalence. It is likely that the prevailing environmental conditions, host and vector factors are better suited for the spread of spotted fever. This is another aspect which needs investigation in Sri Lanka. The most recognized vector of spotted fever is the hard ticks that parasitize every class of vertebrates in almost every region of the world (Cowan 1996). Ticks may also act as reservoirs of rickettsiae whereas the role of vertebrates as reservoirs of rickettsiae has yet to be determined (Raoult & Roux 1997). The R. conorii antigen strain Moroccan reacted with 21 sera (Table 2). Rickettsia conorii is known to be the aetiological agent of Mediterranean spotted fever (MSF) (La Scola & Raoult 1997). MSF has skin eruptions which are papular rather than macular, the disease is referred as boutonneuse fever (Raoult & Roux 1997). Furthermore, R. conorii has been ª 2003 Blackwell Publishing Ltd 809

8 reported from neighbouring India where it is transmitted by the dog tick Rhipicephalus sanguineus (Parola et al. 2001). The Moroccan strain of R. conorii which was used in this study differed from Indian tick typhus strain for reactivity with monoclonal antibodies suggesting antigenic diversity of R. conorii (Walker et al. 1992). There were 30 sera positive for TT118 which is the antigen of isolate named Thai tick typhus. Molecular similarities between R. honei and TT118 suggest that TT118 is a strain of a R. honei which is a pathogenic rickettsia in Flinders Island of Australia. However this strain has just been detected in Ixodes ticks in Thailand (Kollars et al. 2001). It is possible that R. honei may already exist in Sri Lanka. There were 20 sera positive for R. japonica (YH strain) which is a known pathogenic rickettsia in Japan (Raoult & Roux 1997). This is the first time that the antibodies to this strain has been detected in Sri Lanka. It is possible that unknown rickettsiae in the environment would have cross-reacted with the above strains. Nine genera of ticks of the family Ixodidae have been reported in Sri Lanka (Seneviratna 1965). The nine genera include the following: Rhipicephalus, Dermacentor, Amblyomma, Ixodes, Haemaphysalis and Hyalomma. Some of these ticks are known as vectors of rickettsiae in the world. However no studies in relation to human rickettsioses have been carried out in the country. Clinically, patients of the SFG had a severe illness compared with those with O. tsutsugamushi and R. typhi infections. Skin rash was exclusively seen in acute cases but three patients who had only IgG were negative for skin rash. Five patients with skin necrosis had antibodies against SFG antigens. It is our opinion that necrotic rash, elevated liver enzymes, thrombocytopenia, and high ESR are the warning signs of life threatening multi-organ dysfunction. We are aware of two patients with the above warning signs who died due to multi-organ dysfunction (not in the present study). Both these patients had positive Weil Felix tests but there were delays in diagnosis at the onset. Clinicians have to be aware that the patients suspected of spotted fever should be treated early with the appropriate antibiotics and with other supportive measures. Complications such as arthritis, arthralgia, myalgia and fascitis were common with spotted fever in our series. These findings are in agreement with the international literature (Raoult & Roux 1997). Our experience is that these complications improve with administration of hydrocortisone. The high ESR in these patients suggests an immune mediated process along with direct rickettsial damage. The presence of eschar (tache noire) is described as a constant feature in MSF (Cowan 1996). For some SFG rickettsioses other than MSF, eschar is not always seen (Raoult & Roux 1997). Only one patient in the SFG in the current study had an eschar. The presence of antibodies to more than one rickettsial group suggests the possibility of past exposure, co-infection, or cross-reactivity of rickettsial antigens. However, although cross adsorption studies and Western blot assays may be useful in specialized laboratories (La Scola & Raoult 1997), it is not possible to differentiate one from the other without using molecular methods or isolating the rickettsiae from the patients. As all three rickettsiae are sharing the same geographical environment, it is likely that co-infection or past exposure is a possibility for 19 patients in the current study to have more than one type of antibodies. The clinical picture of these patients was similar to others and four patients had necrotic skin rash similar to the SFG group. In the present study there were four patients who were negative for rickettsial antibodies but clinically compatible with rickettsial infection. A diagnosis of rickettsial infections in these four patients cannot be excluded without testing paired sera. Weil Felix test results of 25 IFA positive cases show a relationship as follows: OXK and O. tsutsugamushi, OX19 and R. typhi, OX2 and SFG. However, this correlation is not strong enough to make a recommendation for its use despite its availability in Sri Lanka. This test has poor sensitivity and specificity for the diagnosis of SFG, murine typhus, epidemic typhus and scrub typhus although it has a good correlation to IgM of IFA (La Scola & Raoult 1997). The results of the current study revealed the presence of O. tsutsugamushi, SFG and murine typhus in the island of Sri Lanka. The study also highlights the need for newer techniques for the identification of specific rickettsial agents. Doctors, health workers and visitors to the identified areas should be made aware of the emerging rickettsioses in these areas. Acknowledgements We wish to thank Dr T. Varagunam (former WHO consultant) for advice, Dr P. Abeykoon and Dr S. Kumari (WHO India) for helping to send serum samples to Thailand. We also thank Dr Pubudu Weerathunga and Dr Chandima Herath of the Faculty of Medicine, Peradeniya, Sri Lanka, for helping with the collection of data. References Chenchittikul M, Khlumklai S, Saisongkoth W & Bhumisawasdi J (1995) Comparison of the Weil Felix (Proteus mirabilis OXK) test and indirect fluorescent antibody test for serodiagnosis of 810 ª 2003 Blackwell Publishing Ltd

9 scrub typhus. Journal of Medical Technologists Association of Thailand 23, Christie AB (1980) In Infectious Diseases: Epidemiology and Clinical Practice, 3rd edn. Churchill Livingstone, Edinburgh, London, p Cowan GO (1996) Rickettsial infections. In: Manson s Tropical Disease, 20th edn. (ed. G Cook) W.B. Saunders, London, pp Cowan G (2000) Rickettsial disease: the typhus group of fevers areview. Postgraduate Medical Journal 76, Jiatana P, Khachornsakdi S, Helena P et al. (1997) Comparative evaluation of four serodiagnostic tests for scrub typhus in Thailand. Transactions of the Royal Society of Tropical Medicine and Hygiene 91, Kollars TM Jr, Tippayachai B & Bodhdatta D (2001) Shart report: Thai tick typhus, Rickettsia honei and a unique Rickettsia detected in Ixodes granulatus (Ixodidae: Acari) from Thailand. American Journal of Tropical Medicine and Hygiene 65, La Scola B & Raoult D (1997) Laboratory diagnosis of rickettsioses: current approaches to diagnosis of old and new rickettsial diseases. Journal of Clinical Microbiology 35, Parola P, Fenollar F, Badiaga S et al. (2001) First documentation of infection due to Rickettsia conorii (strain Indian tick typhus) in returned traveler. Emerging Infectious Diseases 7, Raoult D & Roux V (1997) Rickettsioses as paradigms of new or emerging infections diseases. Clinical Microbiological Reviews 10, Seneviratna P (1965) The Ixodoidea (ticks) of Ceylon. Ceylon Veterinary Journal xiii, Van Peenen PFD, See R, Soysa PE & Irving GS (1976) Seroepidemiological survey of hospital associated population in Colombo, Sri Lanka. Southeast Asian Journal of Tropical Medicine and Public Health 7, Walker DH, Liu QH, Yu XJ et al. (1992) Antigenic diversity of Rickettsia conorii. American Journal of Tropical Medicine and Hygiene 47, Wang S, Jiang P, Huang J et al. (2001) Demonstration of the natural foci of tsutsugamushi disease in Nan Peng Lie island in China. Southeast Asian Journal of Tropical Medicine and Public Health 32, Authors Dr S. A. M. Kularatne (corresponding author) and Dr I. B. Gawarammana, Department of Medicine, Faculty of Medicine, University of Peradeniya, Peradeniya, Sri Lanka. Tel.: ; Fax: , samkul@sltnet.lk, indigaw@sltnet.lk Dr J. S. Edirisingha, Department of Parasitology, Faculty of Medicine, University of Peradeniya, Peradeniya, Sri Lanka. Tel.: ; Fax: ; jsediri@yahoo.com Dr Hiroshi Urakami, Department of Microbiology, Faculty of Pharmacy, Niigata University of Pharmacy and Applied Life Science, Kamishin-ei-cho, Niigata , Japan. Tel.: ; Fax: ; urakami@niigata-pharm.ac.jp Mongkol Chenchittikul, Rickettsial Section, National Institute of Health, Department of Medical Sciences, Ministry of Public Health, 88/7, Tivanond Road, Muang, Nonthaburi, Thailand. Tel.: ; Fax: ; jmongkol@dmsc.moph.go.th Dr Ikuo Kaiho, Division of Virology, Chiba Prefectural Institute of Public Health , Nitona-cho, Chuo-tu, Chiba , Japan. Tel.: ; Fax: ; i.kih@ma.pref.chiba.jp ª 2003 Blackwell Publishing Ltd 811