Diagnostic challenge of Q fever osteoarticular infection

Abstract

A 74-year-old man from Logan, Queensland, presented with lower back pain and 10 kg unintentional weight loss and was diagnosed with culture-negative L3/4 discitis, left epidural abscess and bilateral large volume psoas abscesses seen on T1 fat saturated images (Fig. 1), which did not clinically, radiologically or biochemically respond to a month of empiric piperacillin–tazobactam. He had brief contact with kid goats, but no other specific risk factors. He had significant muscle wasting in his lower limbs and decreased reflexes, but no other abnormalities on examination. C-reactive protein was raised at 45 mg/L. The 16S ribosomal RNA (rRNA) gene nucleic acid amplification was negative on two abscess aspirates a month apart. Serology for melioidosis, brucellosis, bartonella and Q fever was ordered as part of a culture-negative screen. The Q fever serology completed at Queensland Pathology was positive with phase 1 IgG of >1280 and phase 2 IgG of 1280 on immunofluorescence assay (IFA). In retrospect, Q fever polymerase chain reaction (PCR) on both abscess aspirates was positive. As chronic Q fever often manifests as infective endocarditis or mycotic aneurysms, echocardiography and computed tomography angiography were performed, but were negative. Positron-emission tomography was not performed. The patient was treated with doxycycline 100 mg twice daily and hydroxychloroquine 200 mg three times daily, which is continuing at the time of publication, a year from initial diagnosis. His symptoms and inflammatory markers have substantially improved. Q fever Phase 1 IgG (IFA) has fallen on therapy from 819 200 to 51 200 over the course of the therapy to date, suggesting a successful response. Chronic Q fever is a zoonosis caused by Coxiella burnetii that was first identified in Queensland. Transmission to humans can occur through aerosolisation of contaminated dust, contact with infected animals (typically livestock) or tissue, or unpasteurised milk ingestion. Chronic Q fever is characterised by elevated Phase I antibody responses (IgG ± IgA): an I IgG IFA titre of ≥1/800 is used as a cut-off for diagnosis. Osteoarticular manifestations are rare; to date, there are only 14 case reports of isolated osteoarticular infection with abscess formation. Risk factors can be difficult to elicit due to the innocuous nature of the exposure, as was the case in our patient. These infections are therefore more challenging to diagnose. Immunohistochemistry is less useful due to granuloma formation, which results in a decreased bacteria number often below the limit of detection in this test. Broad range 16S rRNA gene amplification and sequencing also has a low sensitivity due to fewer target gene copy numbers and reduced amplification cycles (used to reduce contamination risk). It may also be negative with prior empiric antibiotics. Conversely, real-time PCR targeting bacteria-specific multicopy gene sequences such as IS1111 (used in the present case study) have been found to be highly sensitive. This case report demonstrates the difficulty in diagnosing Q fever osteoarticular infection and illustrates the pitfalls of relying on 16S rRNA gene sequencing in cases of possible Q fever osteoarticular infection. Given its largely favourable outcome with long-term antibiotic treatment and surgical intervention, C. burnetii should be considered as a differential in patients who remain culture-negative despite appropriate investigation, including those with negative 16S rRNA gene sequencing.