Optimal length and usefulness of temporal artery biopsies in the diagnosis of giant cell arteritis: a 10-year retrospective review of medical records.

Abstract

In giant cell arteritis, temporal artery biopsies often show vasculitis with giant cell formation, but optimal biopsy length for diagnosis is debated. We reviewed temporal artery biopsies from a 10-year period in the province of Alberta, Canada, to identify an ideal biopsy length in the diagnostic process for giant cell arteritis. We retrospectively reviewed electronic medical records of patients who had undergone a temporal artery biopsy procedure in Alberta between Jan 1, 2008, and Jan 1, 2018, as reported in the Data Integration and Management Repository of Alberta Health Services. We extracted data on baseline demographic characteristics (sex and age), inflammatory markers (erythrocyte sedimentation rate [ESR] and C-reactive protein [CRP]), temporal artery biopsy characteristics (side of biopsy and postfixation length), and final pathological diagnoses. All positive biopsies were reviewed by a single pathologist to ensure uniformity of pathological interpretation, with subsequent discordant results removed from analysis. Predictors of positive pathological diagnosis of giant cell arteritis were modeled by logistic regression, and the Akaike information criterion was used to compare logistic regression models with varying biopsy length cutoffs (0·5, 1·0, 1·5, 2·0, and 2·5 cm) to determine a change point for diagnostic sensitivity in postfixation length. We extracted data on 1203 temporal artery biopsies; after removal of 13 discordant biopsies, 1190 biopsies from 1163 patients were reviewed. The mean age of patients was 72·0 years (SD 10·3) and 799 (68·7%) patients were women. 222 (18·7%) temporal artery biopsies were positive for giant cell arteritis. In univariable analysis, increases in age (71·3 years [SD 10·6] in negative biopsies vs 75·3 years [8·3] in positive biopsies; odds ratio [OR] 1·04 [95% CI 1·02-1·06]; p<0·0001)), ESR (36 mm/h [IQR 18-62] in negative biopsies vs 57 [31-79] in positive biopsies; 1·01 [1·01-1·02]; p<0·0001), CRP (12·1 mg/L [IQR 3·3-35·1] in negative biopsies vs 41·8 [14·6-82·4] in positive biopsies; 1·01 [1·01-1·01]; p<0·0001), and biopsy length (1·2 cm [IQR 0·9-1·7] in negative biopsies vs 1·6 [1·1-2·0] in positive biopsies; 1·28 [1·09-1·51]; p=0·0025) were associated with a positive pathological diagnosis. In multivariable analysis adjusted for age, ESR, and CRP, age (adjusted OR 1·04 [95% CI 1·02-1·05]; p=0·0001), CRP (1·01 [1·00-1·01]; p=0·0006), and biopsy length (1·22 [1·00-1·49]; p=0·047) remained statistically significant predictors. The Akaike information criterion determined a change point of 1·5 cm for diagnostic sensitivity. Accounting for postfixation shrinkage, our findings suggest a 1·5-2·0 cm prefixation length as the optimal biopsy length to diagnose patients with giant cell arteritis, with greater lengths unlikely to provide significant additional diagnostic yield to justify risks associated with surgery. None.