ABSTRACT
BACKGROUND: Variations in surgical technique of temporal artery biopsy (TAB) performed for diagnosis of giant cell arteritis (GCA) may contribute to high false-negative rates. This was a retrospective analysis of a large database that explored potential associations between specimen length and diagnostic sensitivity of TAB. METHODS: Histopathological reports and medical records of patients who underwent TAB in six hospitals between 2004 and 2009 were reviewed. RESULTS: A total of 966 biopsies were analysed. The median postfixation specimen length was 1 (range 0·1-8·5) cm and 207 biopsies (21·4 per cent) were positive for GCA. Significant variation in prebiopsy erythrocyte sedimentation rate (ESR), arterial specimen length and positive results was noted amongst hospitals. Multivariable analysis revealed that patient age, ESR value and specimen length were independent predictors of GCA. Positive biopsies had significantly longer median specimen length compared with negative biopsies: 1·2 (range 0·3-8·5) versus 1·0 (0·2-8·0) cm respectively (P = 0·001). Receiver operating characteristic (ROC) analysis identified postfixation specimen length of at least 0·7 cm as the cut-off length with highest positive predictive value for a positive biopsy (area under ROC curve 0·574). Biopsies with specimen length of 0·7 cm or more had a significantly higher rate of positive results than smaller specimens (24·8 versus 12·9 per cent respectively; odds ratio 2·17, P = 0·001). CONCLUSION: Specimen length and ESR were independent prognostic factors of a positive TAB result. A uniform referral practice and standard specimen length of approximately 1 cm could help eliminate discrepancies in the results of TAB.
