Humans Infected with Relapsing Fever Spirochete Borrelia miyamotoi, Russia

Disease may occur throughout the world because of the widespread prevalence of this pathogen in ixodid ticks.

Borrelia miyamotoi is distantly related to B. burgdorferi and transmitted by the same hard-body tick species. We report 46 cases of B. miyamotoi infection in humans and compare the frequency and clinical manifestations of this infection with those caused by B. garinii and B. burgdorferi infection. All 46 patients lived in Russia and had infl uenzalike illness with fever as high as 39.5°C; relapsing febrile illness occurred in 5 (11%) and erythema migrans in 4 (9%). In Russia, the rate of B. miyamotoi infection in Ixodes persulcatus ticks was 1%-16%, similar to rates in I. ricinus ticks in western Europe and I. scapularis ticks in the United States. B. miyamotoi infection may cause relapsing fever and Lyme disease-like symptoms throughout the Holarctic region of the world because of the widespread prevalence of this pathogen in its ixodid tick vectors. B orrelia miyamotoi, discovered in Japan in 1995, belongs to the relapsing fever group of Borrelia (1). Relapsing fever borreliae infections are characterized by infl uenzalike illness and >1 relapse episode of bacteremia and fever. B. miyamotoi is more distantly related to B. burgdorferi, a group of spirochetes that includes B. burgdorferi s.l. strains (B. afzelii; B. garinii; and B. burgdorferi s.s., the causative agent of Lyme disease) (2,3). In Eurasia and North America, B. miyamotoi is found in a small percentage of all species of ixodid tick vectors of B. burgdorferi, including Ixodes persulcatus (1,3,4), I. ricinus (5)(6)(7), I. scapularis (2,3,8,9), and I. pacifi cus (10). It is transmitted transovarially and transstadially by ticks and coexists with B. burgdorferi (2,3). Recently, we discovered B. miyamotoi in I. persulcatus and I. ricinus ticks in the European and Asian regions of Russia. In these areas, human ixodid tickborne infections, including those caused by B. afzelii, B. garinii, and viral tick-borne encephalitis virus (TBEV; genus Flavivirus) are endemic and transmitted by the same tick species.
Despite the presence of B. miyamotoi in vector ticks, to our knowledge, human disease caused by this spirochete has not been defi nitively established. We previously noted presumptive B. miyamotoi infection in residents of central Russia with infl uenza-like illness but were uncertain whether their clinical manifestations were caused by co-infecting B. burgdorferi s.l. species (11)(12)(13). To confi rm those fi ndings and develop initial estimates of the prevalence and severity of B. miyamotoi infection, we conducted a comparative cohort study. We used improved antibody assays and PCRs to compare the relative frequency and clinical manifestations of B. miyamotoi infection with those of B. garinii infection in Russia and B. burgdorferi infection in the United States.

Study Design
We enrolled patients admitted to  Figure 1). Ticks were collected by drag cloth, visually identifi ed to species level, and analyzed by PCR to identify specifi c Borrelia species.

Case Defi nitions
Diagnosis of B. miyamotoi infection required the report of a tick bite, presence of clinical manifestations consistent with borreliosis, and laboratory evidence of B. miyamotoi infection. Clinical manifestations included fever, headache, chills, fatigue, vomiting, and myalgia. Confi rmation of active infection consisted of amplifi cation of B. miyamotoi DNA/RNA in blood by species-specifi c PCR and detection of anti-borreliae immunoglobulin (Ig) M in acute-and/or convalescent-phase serum samples.
In Russia, diagnosis of B. garinii infection required report of a tick bite, physician diagnosis of erythema migrans (EM; an expanding, ring-like erythematous rash >5 cm in diameter), or an infl uenza-like illness. Confi rmation of infection consisted of amplifi cation of B. garinii DNA/RNA in blood by specifi c PCR, followed by direct sequencing of 5S-23S ribosomal RNA (rRNA) intergenic spacers, and detection of anti-borreliae IgM in acute-and/ or convalescent-phase serum samples.
In the United States, diagnosis of B. burgdorferi infection required a physician's diagnosis of EM or an infl uenza-like illness. For all cases, confi rmation of infection consisted of a >4-fold increase in anti-B. burgdorferi antibody in acute-and convalescent-phase serum samples. The diagnosis of TBEV infection was based on a viral-like illness, including headache (with or without meningitis or encephalitis), amplifi cation of TBEV RNA in blood by species-specifi c PCR, and/or detection of anti-TBEV IgM in an acute-phase serum sample.

PCR
The PCR we used enabled detection of DNA and RNA sequences. DNA/RNA was extracted from 2 mL of whole venous blood with EDTA or from tick suspensions by using an AmpliSens Riboprep Kit (Central Institute of Epidemiology, Moscow, Russia) according to the manufacturer's instructions. Of the blood samples used for PCR, 81% were obtained at the time of hospital admission and 96% within 2 days of admission. To assay the inhibitory effect of blood and tick extracts on the PCR, all samples were spiked with a universal RNA recombinant control having a known number of RNA copies per milliliter. Reverse transcription of RNA to cDNA was performed by using an Amplisens Reverta-L Kit (Central Institute of Epidemiology). The cDNA samples were assayed for B. miyamotoi and other tick-borne pathogens by using realtime quantitative PCR (qPCR) assays in a Rotor Gene 6000 cycler (Corbett Life Science, Concorde, New South Wales, Australia).
The cDNA samples were divided into 2 aliquots, and different types of real-time qPCR were performed on each. The fi rst used in-house primers and a probe that targeted the 16S rRNA gene of B. miyamotoi. The inclusion of the reverse transcription procedure improved the detection sensitivity because the 16S rRNA that also became detectable is present in higher copy numbers than the 16S rRNA gene. The detection limit of at least 5 × 10 3 copies/ mL was determined by using positive recombinant DNA of the B. miyamotoi 16S rRNA gene fragment with a known number of copies. The B. miyamotoi-specifi c forward and reverse primers at 360 nmol/L were, respectively, Brm1

5′-CGCTGTAAACGATGCACACTTGGTGTTAATC-3′
and Brm2 5′-CGGCAGTCTCGTCTGAGTCCCCAT CT-3′. The corresponding dye-labeled probe (fi nal concentration 100 nmol/L) was R6G-CCTGGGGA GTATGTTCGCAAGAATGAAACTC-BQH1. The PCR conditions were 95°C for 15 min; followed by 10 cycles at 95°C for 20 s, 67°C for 50 s, and 72°C for 20 s; then by 40 cycles at 95°C for 20 s, 60°C for 50 s, and 72°C for 20 s. The fl uorescence signal was recorded at the 60°C step for the last 40 cycles. Each run included negative controls and positive recombinant control DNA of the B. miyamotoi 16S rRNA gene fragment as a standard. PCR-based detection of B. burgdorferi s.l., Anaplasma phagocytophilum, Ehrlichia chaffeensis, Ehrlichia muris, and TBEV was performed on the second cDNA aliquot by using a commercial multiplex PCR (AmpliSens TBEV, B. burgdorferi s.l., A. phagocytophillum, E. chaffeensis/E. muris-FL; Central Institute of Epidemiology) (16), according to the manufacturer's instructions. This assay was designed to detect, but not discriminate between, B. afzelii, B. burgdorferi s.s., and B. garinii. The same assays were used to detect specifi c DNA/RNA in ticks and humans.
The specifi city of B. miyamotoi and B. burgdorferi s.l. assays was confi rmed by direct sequencing of fl agellin gene fragments and/or 16S rRNA gene fragments and/or 5S-23S rRNA intergenic spacer amplifi ed from blood samples of the same patients or from the same ticks (GenBank accession nos. GU797331-GU797350, JF951378-JF951392). Of the 97 borreliae sequenced, results of DNA amplifi cation using species-specifi c PCR were entirely consistent with the sequencing results. Absence of false-positive PCR results means that our PCRs were highly specifi c.

ELISA
Serum samples collected at the time of admission and 1-2 weeks later were tested for anti-borrelial IgM and IgG. Serologic evidence of exposure to borreliae was detected by ELISA EUROIMMUN EI 2132-9601 M and EI 2132-9601-2 G (EUROIMMUN AG, Lübeck, Germany). The ELISA consisted of a mixture of whole antigens from B. afzelii, B. burgdorferi, and B. garinii and thus could detect but not discriminate specifi c antibody against any of these species. Seroconversion in patients infected with the relapsing fever borrelia B. persica also has been detected by EUROIMMUN assay (18). Serum from most B. miyamotoi-positive patients reacted to the antigen(s) in this assay. Anti-TBE IgM was detected by the semiquantitative EUROIMMUN ELISA EI 2661-9601 M.

Statistical Analyses
Comparisons were performed by using the Mann-Whitney U test (independent numeric interval variables), χ 2 test (categorical variables), and corresponding exact tests, if necessary; p<0.01 was statistically signifi cant. Data were analyzed by using SPSS version 11.0.1 (SPSS Inc., Chicago, IL, USA).

Study Population
Among 302 patients evaluated for presumptive tickborne infection, B. miyamotoi infection was found for 51 (17%) ( Table 1). Of these 51 patients, anti-borrelial IgM was found at the time of admission for 6, anti-borrelial IgM seroconversion was demonstrated 2 weeks after admission for 40, no such antibody was found for 2, and laboratory evidence of TBEV co-infection was found for 3. Only the 46 patients who had amplifi able B. miyamotoi DNA and anti-borrelial IgM and who were not coinfected with TBEV were included in further analyses. For 18 of these patients, anti-borrelial IgG was absent at admission but detected 2 weeks after admission; subsequent IgG testing was not performed because all patients had IgM by 2 weeks. Attempts to detect B. miyamotoi on standard Giemsastained blood smears and in culture in 6 patients yielded negative results. These tests were performed 4-6 days after onset of illness but before initiation of antimicrobial drug therapy.
Of Although mean peripheral leukocyte and platelet counts were lower for patients with B. miyamotoi than B. garinii infection, they were within the reference range. Proteinuria and transient elevation of serum alanine aminotransferase and aspartate aminotransferase concentrations were found for 3× more B. miyamotoi patients than B. garinii patients (51% and 68% vs. 15% and 20%, respectively, p<0.01), but no nephritis or hepatitis was clinically apparent. We found similar clinical and laboratory results when we omitted from analysis the 4 B. miyamotoi patients with EM who might have been co-infected with B. burgdorferi s.l.

Therapy and Clinical Outcome
Antimicrobial drug therapy for the B. miyamotoi patients was started ≈1 week after admission when IgMbased serologic tests results confi rmed the diagnosis (median 7 days, IQR 6-10 days). Therapy consisted of ceftriaxone, 2 g intravenously every 24 hours for 2 weeks (42 patients) or doxycycline 100 mg orally every 12 hours for 2 weeks (2 patients). Two patients received no antimicrobial drug while hospitalized; 1 later received doxycycline at home, and the other was readmitted to the hospital for relapse and received ceftriaxone. Patients with B. garinii infection received doxycycline (71%) or ceftriaxone (29%) immediately after admission because diagnosis of borreliosis, based on presence of EM, was made at the time of admission. B. burgdorferi patients all received doxycycline, 100 mg orally every 12 hours, or

Relapsing Infection
Of the 46 B. miyamotoi patients, 5 (11%, 95% confi dence interval 2%-20%) experienced relapse of febrile illness; 1 patient experienced 2 relapses before hospital admission, and 4 experienced 1 relapse after hospitalization but before the start of antimicrobial drug therapy. Thus, antimicrobial drugs might have prevented relapse in those who received this therapy. The mean time between relapses was 9 days (range 2 days to 2 weeks). The maximum fever and duration of illness did not differ signifi cantly for the fi rst and second episodes of illness ( Figure 2). No clinical or laboratory fi ndings indicated other infections (including blood-borne, skin, neurologic, respiratory, cardiac, gastrointestinal, and urologic) or medical conditions that could account for these febrile episodes.

Genetic Characteristics of B. miyamotoi
The nucleotide sequences of 16S rRNA and fl agellin gene fragments of all B. miyamotoi isolates from humans and I. persulcatus ticks were almost indistinguishable from the corresponding sequences of the prototype Japan HT-31 strain (1) (Figure 3). B. miyamotoi from I. ricinus ticks collected in the European part of Russia were more closely related to European B. miyamotoi strains (5,6).

Discussion
We provide confi rmatory evidence of B. miyamotoi infection in humans. Most patients experienced clinical manifestations similar to those caused by B. burgdorferi s.l. and relapsing fever Borrelia infections, a fi nding consistent with the genetic characteristics of this novel spirochete. Febrile relapses occurred in only 1 of 10 B. miyamotoi patients, 2 days to 1 month after the initial illness; however, early treatment may have prevented subsequent relapse for other patients. Although the febrile episodes at home might have been caused by other illnesses, the onset of fever within 2 weeks after a tick bite was consistent with the incubation period of infection with borreliae. Furthermore, no clinical or laboratory fi ndings indicated other infections or medical conditions that could account for the febrile episodes. EM was found in ≈1 of 10 B. miyamotoi patients, but these patients might have had unrecognized B. burgdorferi s.l. co-infection. A single course of ceftriaxone or doxycycline seemed to clear B. miyamotoi infection. Although effective therapy is available, appropriate diagnosis and therapy are  B. miyamotoi infection may cause substantial health problems in the regions of Russia where it has been found, given its relatively high incidence and associated severity of disease. On the basis of the number of patients with B. miyamotoi infection in Yekaterinburg Hospital in 2009 and the populations of Yekaterinburg Province (4,395,000), we estimate that the minimal incidence of B. miyamotoi infections is 1 case per 100,000 population. According to offi cial federal notifi cation during the past 10 years, ≈8,000 cases of human borreliosis occur in Russia annually (12).

B. miyamotoi infection seems to constitute at least 1/4 of all clinical tick-borne borreliosis cases in Yekaterinburg.
If other Borrelia spp.-endemic areas have a similar rate of B. miyamotoi infection as Yekaterinburg (and our tick data suggest that this assumption is reasonable), >1,000 B. miyamotoi cases might occur in Russia each year. More studies are necessary to determine if this projection is accurate.
Acute B. miyamotoi infection was more severe than early stage B. burgdorferi infection. The time from symptom onset to hospital admission was shorter, and the number of clinical manifestations was greater for patients with B. miyamotoi infection than with B. garinii infection. Relapsing febrile episodes were only reported for B. miyamotoi patients. Such multiple disease episodes not only have an adverse effect on a patient's health but also may result in costly medical bills, many days or weeks of lost wages, and medical misdiagnosis (19)(20)(21)(22). Co-infection of B. miyamotoi with other ixodid ticktransmitted agents may increase disease severity (15,23). Additional problems that might occur with B. miyamotoi infection are ocular, neurologic, respiratory, cardiac, and pregnancy complications associated with relapsing fever (19)(20)(21)(22).
Our study had several limitations. Attempts to detect B. miyamotoi on blood smear or in culture were not successful, although we confi rmed B. miyamotoi infection with a combination of qPCR, genetic sequencing, clinical, and seroconversion evidence. The comparison of clinical manifestations of Borrelia spp. infection of patients from Russia and the United States was complicated by enrollment at different times and from different locations, although we assessed the same 11 clinical manifestations at each location. The possibility that the clinical description of our B. miyamotoi cases was compromised by unrecognized coinfection with B. burgdorferi s.l. is unlikely. The expected number of cases of co-infection depends on the prevalence of the pathogens in ticks in the region (3,11,24), and this number is even fewer than the 4 B. miyamotoi patients with EM we found. Inclusion or exclusion of these 4 cases had no effect on our comparative analysis with patients who did not have B. miyamotoi infection. We limited our description of B. garinii cases to those that were confi rmed by detection of amplifi able B. garinii DNA/RNA, although such cases may be more severe than those in which such DNA/RNA cannot be detected (25,26). Patients with B. burgdorferi s.l. PCR-negative results experienced fewer symptoms and milder fever than did patients with B. burgdorferi s.l. PCRpositive results. Our analysis of patients with B. miyamotoi and B. garinii infection was limited to those who were hospitalized, although hospital admission policy in these regions of Russia is liberal because of concern about TBE and problems associated with B. burgdorferi infection. The geographic dispersion and extent of B. miyamotoi disease in humans are unclear, but the infection probably occurs outside of Russia, given the comparative infection rates of vector ticks in Russia and at several locations in Europe and the United States (2)(3)(4)(5)(6)(7)(8). In the northeastern United States, ≈15% of all spirochetes carried by I. scapularis ticks are B. miyamotoi (2). Cases may remain undiagnosed because of the nonspecifi c nature of the illness, which might be confused with viral infections or such tick-borne infections as Lyme disease, babesiosis, anaplasmosis, or ehrlichiosis, and because of the lack of laboratory tests for confi rmatory diagnosis (19)(20)(21)(22).
B. miyamotoi infection may have negative health consequences, including relapsing disease that may last for months and may not respond to inappropriate antimicrobial drug therapy. The discovery of a Borrelia sp. that is pathogenic in humans and transmitted by an array of ixodid ticks greatly expands the potential geographic distribution of this disease (1)(2)(3)(4)(5)(6)(7)(8)(9)(10)(11). Further investigation of possible B. miyamotoi infection in humans is warranted wherever I. pacifi cus, I. persulcatus, I. ricinus, and I. scapularis ticks are found.