A signed consent form was obtained from the mother of the patient. The patient had reported recurrent joint pain, most prevalent in his left knee, since he was 6 years old. In December 2012, when he was 10 years old, his left knee became swollen, and he was treated locally with nonsteroidal antiinflammatory drugs.

In February 2014, the patient had effusions of both knees and the left ankle. The first joint aspiration of the left knee was performed in February 2014, but the patient was discharged without a diagnosis. Soon after that, he became febrile (temperature 39°C) and was referred to another hospital because of persistent effusions of all 3 joints. Clinical investigation revealed swelling of both knees, which were not warm or red, and a swollen, hot, red left ankle.

A routine blood test showed a normal leukocyte count, elevated C-reactive protein levels (79.4 mg/L, positive threshold 5 mg/L), and elevated erythrocyte sedimentation rates (85 mm/h, reference <7 mm/h, and 109 mm/2 h, reference <12 mm/2 h). Tests for rheumatoid factor and other autoantibodies (antinuclear antibodies, double-stranded DNA, proteinase 3, myeloperoxidase antibodies) showed negative results. The patient underwent needle aspiration of all 3 joints under general anesthetic to obtain synovial fluid: 18 mL from the right knee, 60 mL from the left knee, and 6 mL from the left ankle. The patient was given a preliminary diagnosis of juvenile idiopathic arthritis and treated with nonsteroidal antiinflammatory drugs. A few days later, another aspiration of the left knee was performed, followed by an intraarticular injection of steroids.

Infection with pathogens associated with reactive arthritis was ruled out by negative serologic test results for Chlamydia spp. (IgG and IgA enzyme immunoassays), Mycoplasma pneumoniae (IgG and IgA enzyme immunoassays), Salmonella spp. (agglutination assay), and Yersinia spp. (agglutination assay). Antibodies against B. burgdorferi s.l. were detected by Borrelia ELISA (Medac, Hamburg, Germany); IgG levels were highly elevated (IgG ELISA >200 AU/mL, cutoff 10.8 AU/mL; IgM ELISA results were negative). Test results obtained by using the Anti-Borrelia Euroline Westernblot (Euroimmun, Lübeck, Germany) were positive for IgG with strong band intensities for VlsE, p83, p39, p30, p21, p19, and p17 and weak band intensity for p25 (OspC). Cytologic test results for synovial fluid showed an inflammatory infiltrate with lymphocytes and segmented neutrophils. Culture for bacterial pathogens was negative.

Synovial fluid samples from all 3 joints were tested by using PCR. DNA was extracted by using the PeqGOLD Tissue DNA Mini Kit (Peqlab, Erlangen, Germany). Two TaqMan-based real-time PCR assays targeting the 16S rDNA gene (6) and the flagellin gene (7) were performed; primer and probe sequences are listed in Table 1. The DNA of an in-house B. burgdorferi s.l. strain was used as a positive control, and PCR-grade water was used as a negative control. To check for PCR inhibition, we used samples spiked with borrelial DNA in an extra well. In 2 of the 3 joint aspirates (left knee and left ankle), borrelial DNA was detected by both assays. For genotype identification, samples were subjected to a previously described nested PCR targeting the 5S-23S intergenic spacer region (8).

Amplicons were purified by using the QIAquick PCR Purification Kit (QIAGEN, Hilden, Germany) and sent to MWG Eurofins (Munich, Germany) for bidirectional sequencing by using primers B5S-23S_Fn and B5S-23–S_Rn (8). Sequencing revealed the same Borrelia strain in both joints. When compared with known sequences by using BLAST (http://www.ncbi.nlm.nih.gov/blast), the sequences showed 100% identity with that of the PBi strain.

Because the PBi strain and all PBi-like strains are now known as a distinct genospecies within the B. burgdorferi s.l. complex, namely B. bavariensis (9), and because B. bavariensis strains are associated with OspA serotype 4, another PCR for identifying the OspA serotype was carried out. To determine the OspA type of the Borrelia spp. found in the joint aspirates, we performed another previously described nested PCR targeting the ospA gene (10), then sequenced and identified a virtual restriction fragment polymorphism by using CLC Main Workbench version 7.0 (CLC bio, Aarhus, Denmark). Because the obtained restriction fragment length polymorphism pattern matched OspA serotype 4, the presence of a B. bavariensis strain in both Borrelia-positive joints could be confirmed.

After the patient was treated with amoxicillin (500 mg 3×/d for 28 days) (4), all joint effusions resolved. At his 9-month follow-up visit, the patient did not report any symptoms. No erythema migrans was observed, and the patient’s mother reported only 1 tick bite for the patient when he was 2 years old.

In Europe, Lyme arthritis can be caused by several genospecies of B. burgdorferi s.l. Published studies have most commonly identified B. burgdorferi sensu stricto in joints from Lyme arthritis patients in Europe (Table 2). B. bavariensis, the pathogen that caused the illness in the patient we describe, was formerly classified as B. garinii genospecies characterized by OspA serotype 4 (9). On the basis of multilocus sequence analysis of chromosomal housekeeping genes, this group was found to be genetically distinct from other B. garinii strains. Furthermore, the 2 genotypes differ in their hosts: B. bavariensis is a rodent-associated strain, whereas other B. garinii serotypes can be found in birds.

Most Lyme arthritis patients respond well to a single course of treatment with antimicrobial drugs, although in a small percentage of cases persistent synovitis can develop months or even years after treatment. For those patients whose synovial fluid PCR result is negative, intraarticular application of corticosteroids can be beneficial (15).

This case illustrates that Lyme arthritis must be taken into account in patients in Europe who have persisting joint effusions. Treatment with antimicrobial drugs is highly effective. We did not find any other report of cases in which the pathogen was detected in multiple joints by using a direct identification method. This case is further evidence for the systemic characteristics of Lyme borreliosis.