Spotted Fever Group Rickettsioses in Israel, 2010–2019

In a multicenter, nationwide, retrospective study of patients hospitalized with spotted fever group rickettsiosis in Israel during 2010–2019, we identified 42 cases, of which 36 were autochthonous. The most prevalent species was the Rickettsia conorii Israeli tick typhus strain (n = 33, 79%); infection with this species necessitated intensive care for 52% of patients and was associated with a 30% fatality rate. A history of tick bite was rare, found for only 5% of patients; eschar was found in 12%; and leukocytosis was more common than leukopenia. Most (72%) patients resided along the Mediterranean shoreline. For 3 patients, a new Rickettsia variant was identified and had been acquired in eastern, mountainous parts of Israel. One patient had prolonged fever before admission and clinical signs resembling tickborne lymphadenopathy. Our findings suggest that a broad range of Rickettsia species cause spotted fever group rickettsiosis in Israel.

In a multicenter, nationwide, retrospective study of patients hospitalized with spotted fever group rickettsiosis in Israel during 2010-2019, we identifi ed 42 cases, of which 36 were autochthonous. The most prevalent species was the Rickettsia conorii Israeli tick typhus strain (n = 33, 79%); infection with this species necessitated intensive care for 52% of patients and was associated with a 30% fatality rate. A history of tick bite was rare, found for only 5% of patients; eschar was found in 12%; and leukocytosis was more common than leukopenia. Most (72%) patients resided along the Mediterranean shoreline. For 3 patients, a new Rickettsia variant was identifi ed and had been acquired in eastern, mountainous parts of Israel. One patient had prolonged fever before admission and clinical signs resembling tickborne lymphadenopathy. Our fi ndings suggest that a broad range of Rickettsia species cause spotted fever group rickettsiosis in Israel.
The etiology of SFGR in Israel is thought to be R. conorii ITTS on the basis of limited molecular identification of this strain from clinical cases and from ticks (9,15,25). However, the yearly variations in disease severity and clinical manifestations are intriguing and may suggest involvement of >1 species of spotted fever group (SFG) Rickettsia. Using a large database from the national reference center, we studied the specific species of Rickettsia that cause SFGRs in Israel and characterized their unique epidemiology and clinical features. Institutional review board approval was granted at the principal investigator site (0002-19-LND) and for each participating institute.

Study Design
We conducted a multicenter, retrospective study of hospitalized patients with an SFGR during 2010-2019. The study included SFGR diagnosed by molecular methods at the Israel Institute for Biological Research (IIBR; Ness Ziona, Israel), which serves as the national reference center for Rickettsia. Blood and tissue samples from hospitalized patients with a suspected SFGR are occasionally submitted to the IIBR for molecular diagnosis, at the discretion of the treating physician. For cases of successful molecular identification, the referring medical center was requested to provide patient demographic and clinical data from the medical charts at each participating site. Deidentified data were integrated into a central database.

Serology and Molecular Diagnoses
The IIBR tested serum samples for antibodies against R. conorii and R. typhi by an in-house immunofluorescence assay (cutoff for IgG of 1:100), as previously described (26). Skin biopsy samples, whole blood, cerebrospinal fluid, and other tissues were tested by PCR (17,27). Stored SFG-positive Rickettsia DNA samples and sequenced unique regions from 4 conserved Rickettsia genes were batch tested (primers listed in Table 1). R. africae was identified by real-time PCR targeting an internal transcribed spacer (30). These analyses enabled identification of ISF, R. africae, R. conorii Malish strain, and a new Rickettsia variant. Each set of reactions included a positive control and nontemplate as a negative control. The same primers were used for sequencing as for amplification. For species-level identification, we compared sequence results by using BLAST (http:// www.ncbi.nlm.nih.gov/BLAST).

Mapping Locations of SFGR Acquisition
We recorded the site of presumed rickettsiosis acquisition for autochthonous cases. When the site was unknown, we used the patient's address.

Statistical Analyses
We used descriptive statistics to summarize patient characteristics. We calculated differences between categorical and continuous variables by using Pearson χ 2 , Student t, and Mann-Whitney tests, as appropriate. We used 1-way analysis of variance to analyze differences among groups. For statistical analyses, we used SPSS Statistics 25 (https://www.ibm.com). We considered 2-sided p<0.05 to be significant.

SFGR Cases Diagnosed by IIBR in the Past Decade
In the 10-year study period, 1,985 cases of rickettsioses from the community and hospitals countrywide were diagnosed at IIBR by serologic testing; 811 were SFGR and 1,174 were murine typhus. Another 89 cases were positive by PCR, 66 for SFGR and 23 for murine typhus (Figure 1, panels A Of these, results were negative for 16 (61%) and the median time between disease onset to last serology test was 6.2 days (range 0-11 days); for 10 (39%), the result was either positive (5 patients) or borderline (5 patients) and median time from disease onset to last serologic test 15.5 days (range 2-29 days).

Geographic and Seasonal Distribution of Cases
The 36 autochthonous cases were reported from 12 hospitals representing most areas of Israel: 6 from central Israel, 4 from northern Israel, and 2 from southern Israel. R. conorii ITTS was reported from all but 1 hospital. All cases of R. africae infection were reported from 1 hospital. The most abundant concentration of cases (26/36, 72%, all R. conorii ITTS) was along the Mediterranean shoreline, with the highest aggregation of cases in the Sharon and Haifa districts ( Figure 2, panel A). Ten cases were acquired inland, of which 6 were caused by R. conorii ITTS and 3 the new Rickettsia variant. Those 3 cases were presumably acquired in more mountainous areas of Israel. The R. conorii Malish strain was acquired in the desert region near Beer-Sheba.

Clinical Features
All patients were hospitalized except for 1 with ATBF, who had mild disease ( Table 3). The mean duration of stay was 11.9 days, and median (IQR) was 5 (3-10) days. Fever affected 41/42 (98%) patients. The mean number of days with fever until hospitalization was 5.2, and the median (IQR) was 5 (3-6) days; 98% of patients were hospitalized by day 9 of fever onset. For patients infected with the new Rickettsia variant, the mean interval was significantly longer (13.6 ± 13.3 days, range 5-29 days; p = 0.002 when compared with the other groups); 1 patient in this group had fever of unknown origin for 29 days before hospitalization.
The time interval between fever onset and appearance of rash differed among the groups: for ISF patients, the mean (range) was 3.2 (0-8) days, and for patients infected with the new Rickettsia variant, the mean and median were 15 days. Of the 2 patients with R. conorii Malish strain infection, 1 had rash that reportedly appeared 1 day before fever onset.
Systemic symptoms (e.g., myalgia and headache) were common for all patients with non-ATBF Rickettsia infection, and meningoencephalitis was evident in 8/33 (24%) of patients with ISF. For ATBF patients, rates of systemic symptoms were lower. Similarly, about one third of patients with non-ATBF rickettsiosis but none with ATBF experienced severe disease with shock, multiorgan failure, need for mechanical ventilation, disseminated intravascular coagulation, acute respiratory distress syndrome, and need for intensive care.

Laboratory DataÈ
xcept for ATBF, laboratory findings did not differ between the groups (Table 4). During the first 3 days of hospitalization, acute kidney injury was common (>50%); other common findings included hepatic injury accompanied with mild to moderate jaundice, mild rhabdomyolysis, mild international normalized ratio prolongation, thrombocytopenia, and lymphocytopenia. Leukocytosis was more common than leukopenia, and C-reactive protein levels were >100 mg/L (reference <5 mg/L). ATBF cases were distinctly different and patients showed much milder systemic reactions: no hepatocellular injury,   (Figure 1, panel A). Because we lack clinical information for many of these cases, such as severity of illness and hospitalization, we can only partially infer the role of this strain in causing SFGR among hospitalized and ambulatory patients. Although R. conorii ITTS was suspected as the causative agent of SFGR in Israel, this suspicion has been supported only by very limited data: rare case reports of fatal human cases (9,15) and a few studies of ticks (25,32). Testing for spotted fever by serology and recently by PCR increased during 2017-2019 compared with the preceding 3 years. This rise probably represents increased clinical suspicion of SFGR and a true increase in disease activity. Despite missing molecular data for ambulatory patients with mild cases of SFGR, we believe that during the past 3 years, R. conorii ITTS has led a silent outbreak of SFGR in Israel. Similarly, an unprecedented 46% increase in SFGR was reported during 2016-2017 by the US Centers for Disease Control and Prevention from states where SFGR is known to be endemic (33). US authorities were faced with the dilemma of whether this increased incidence should be ascribed to a true increase, increased testing, or inappropriate use of a single serologic test instead of paired tests, reflecting past infection in a disease-endemic area.
We report the epidemiologic features of SFGR patients, although comparison between R. conorii ITTS with the other groups was limited because of small numbers. In Israel, SFGR affects mainly young adults and has a mild predilection for men. SFGR has been endemic to the Sharon and Haifa Districts since the 1990s (34); in our study, these 2 districts accounted for 44% (16/36) of all autochthonous cases. These results are similar to those reported in 2014 by Rose et al., who also investigated the association between geographic data and SFGR-positive ticks. Areas with SFG Rickettsia-infected ticks were associated with browntype soil, higher land surface temperatures, and higher precipitation (31). We observed a wide geographic distribution of human ISF cases with aggregation in northern Israel.
Although R. conorii ITTS seems to be the main Rickettsia causing clinical disease in Israel, why this strain is rarely found in ticks collected from Israel (25,35) and other countries (7,36,37) remains unclear. Studies from Israel have found Rickettsia massiliae to be more prevalent (≈10 fold) than R. conorii ITTS among questing ticks and ticks feeding on animals (31,35,38). This discrepancy could be explained by underreporting of R. massiliae infection in humans with mild or subclinical disease. Most patients in our study were hospitalized with severe disease and may represent a reporting bias of R. conorii ITTS, which causes more severe disease. Patients with milder illness, potentially caused by other rickettsiae, may not be hospitalized, and illness may resolve undiagnosed, without the need for molecular studies.
Rose et al. (31) collected Rhipicephalus sanguineus and Rh. turanicus ticks from geographic locations similar to the presumed areas of the clinical autochthonous cases in our study (Figure 2, panel B). However, R. conorii ITTS was found only rarely (1.8%) and strictly in Rh. sanguineus ticks. Hence, the role of Rh. turanicus ticks as possible vectors of R. conorii ITTS and R. massiliae as a cause of rickettsiosis in humans should be further explored. Although clearly reported as a cause of SFGR, R. massiliae is still rarely isolated from human patients (39). Deleterious effects of R. conorii on Rh. sanguineus tick fitness, resulting in infected ticks not surviving the winter, may explain its low prevalence among ticks in nature (40,41).
The new Rickettsia variant was found in the eastern and more mountainous parts of the country: the Golan Heights, the Galilee region, and the West Bank of the Palestinian Authority. This distribution may suggest a geographic niche for either this new Rickettsia or its vector and should be further explored in studies of tick collections from these mountainous areas. The single case of R. conorii Malish strain infection acquired locally in this study was in a 50-year-old man from the desert area, who had a necrotic eschar in the thigh and severe systemic disease.
Most cases were reported during the summer and peaked in August. This finding is consistent with previous reports (12) and may be attributed to increased activity of the vectors and to the aggressiveness and host indiscrimination of Rh. sanguineus ticks when exposed to higher temperatures (42).
Patients rarely remembered seeing or being bitten by ticks (only 2 remembered); however, exposure to animals was common (25/42, 59% of cases), mainly to dogs (20/25, 80%), the principal hosts of Rh. sanguineus ticks (the main reservoir of R. conorii). This finding implies that exposure to domestic pets is more relevant than exposure to ticks.
The clinical and laboratory features for patients in our case series were typical of SFGR, although eschar, which is considered rare in patients with ISF, was seen in 12% of patients, 1 of whom had 3 lesions. ISF caused purpura fulminans in 9% and meningoencephalitis in 24%. About half of ISF patients experienced multiorgan involvement that included kidney and liver injury, jaundice, rhabdomyolysis/myositis, and coagulopathy. Severe disease requiring intensive care was strikingly common (52%), and 30% of ISF patients died in hospital. The high mortality rate, previously reported for ISF infection (10), contrasted with lower rates from historical reports from Israel during the 1990s. This discrepancy may result from reporting bias with increased awareness in recent years, as well as improved laboratory capabilities. An additional possibility is an outbreak of a more virulent strain, such as R. conorii ITTS. Risk factors for death included only alcohol abuse, as previously described (12). Admission-to-treatment (doxycycline) interval was not significant.
The small number of cases in this investigation makes drawing conclusions or comparisons difficult; however, the new Rickettsia variant may lead to prolonged fever before care seeking and may resemble tickborne lymphadenopathy usually related to Rickettsia slovaca or Rickettsia raoultii (11). Patients with ATBF had a distinct clinical syndrome of a milder clinical disease; for 25%, systemic symptoms were limited to fever, myalgia, and headache with no systemic rash.
In conclusion, we report a nationwide case series of hospitalized patients with molecularly diagnosed SFGR over a decade in Israel, of which R. conorii ITTS was the principal cause of severe disease, multiorgan failure, and high mortality rates. We also describe a new Rickettsia variant, which may be associated with unique epidemiologic and clinical features. This study suggests that a broader range of species causes SFGR in Israel and that this possibility should be explored in larger, prospective studies, especially in light of the potential candidates found in ticks.