An evaluation of the usefulness of Staphylococcus aureus serodiagnosis in clinical practice

The measurement of serum antibodies to Staphylococcus aureus has been used in the diagnosis of individuals with suspected deep-seated infection, particularly when culture is difficult or unproductive. The Health Protection Agency (HPA), UK, currently offers testing to detect antibodies to alpha haemolysin (staphylolysin) and the nuclease enzyme in parallel. However, a lack of high specificity and sensitivity of the anti-staphylolysin test has been demonstrated repeatedly in occult S. aureus infections [1–3], whilst there is little information on the usefulness of the anti-nuclease test. It is understood that serological tests, even in combination, are an imperfect marker of staphylococcal infection. It is unclear whether staphylococcal serological testing is useful in clinical practice. We undertook a retrospective evaluation of the use of staphylococcal serodiagnosis in the Hull and East Yorkshire NHS Trust. Our objectives were to determine the extent of testing, to identify the clinical context in which tests were requested, to ascertain whether the results contributed towards patient management and to inform the production of local guidelines. We retrospectively identified the results of serum antistaphylolysin and anti-nuclease titres collected from individuals presenting to a 1,500-bed teaching hospital over a 3-year period. Information concerning the demographics of the individual, their relevant hospital and/or outpatient attendances, past medical history, clinical diagnoses, microbiology, antimicrobial and medical management, and outcome was obtained from their clinical records and the trust pathology database. Test results were interpreted in accordance with HPA guidance [4]. A positive test was defined as an anti-staphylolysin titre of ≥8 units/ml and/or an anti-nuclease titre of >32 units/ml; an equivocal test (suggestive of staphylococcal infection), an antistaphylolysin titre of 4–6 units/ml and/or anti-nuclease titre of 32 units/ml; and a negative test (no evidence of staphylococcal infection or repeat testing required), an anti-staphylolysin titre of ≤2 units/ml and anti-nuclease titre of ≤16 units/ml. Serological testing was performed on 121 serum samples collected from 113 individuals between 1st September 2004 and 31st August 2007. Of these, 73 (65%) were male and their median age was 55 years (range 2–90). Testing was performed in both inpatient (81, 72%) and outpatient (32, 28%) hospital settings. Testing was requested on individuals managed by a variety of specialties, including infectious diseases (78, 70%), orthopaedics (11, 10%), cardiology (3, 2.7%), neurosurgery (3, 2.7%) and the acute medical admitting teams (16, 14.5%). Ninety (80%) individuals were managed as suspected or proven deep-seated infection, including cases of discitis, (20, 18%), prosthetic joint (15, 13%) and native (10, 9%) septic arthritis, osteomyelitis (18, 16%), soft tissue infection (15, 13%), infective endocarditis (10, 9%) and with bacteraemia of uncertain source (2, 1.7%). In the remainder (23, 20%), there was no evidence of infection. Test results were available a median of 14 days from the time of the serum sample collection. Repeat serology was performed on only six individuals. J. Elston (*) :M. Ling :B. Jeffs :K. Adams :H. Thaker : P. Moss :G. Barlow Department of Infection and Tropical Medicine, Main Administration block, Castle Hill Hospital, Cottingham, East Yorkshire HU16 5JQ, UK e-mail: jameselston2003@yahoo.co.uk