The predictive value of high sensitivity C-reactive protein for suspected giant cell arteritis
Z HOT TOPIC Z
Author Disclosure Block: A.K. Banwell: None. I. Irrcher: None. W. Hopman: None. M. ten Hove: None. J. Gonder: None.
Abstract Body:
Purpose: Visual loss is a significant concern in Giant Cell Arteritis (GCA), a vasculitis common in older adults. GCA is an ophthalmic emergency. In Kingston, Ontario, high sensitivity C-Reactive Protein (hsCRP), erythrocyte sedimentation rate (ESR), and platelet count are used to risk stratify patients with suspected GCA. The relationship between standard CRP and GCA is well established. HsCRP is determined via a quantitative immunoturbidimetric method to detect CRP in serum and plasma with variable assay ranges and at lower limits than standard CRP. The use of hsCRP and CRP in other conditions is established. The purpose of this study is to ascertain hsCRP’s predictive reliability and validity for GCA, as it is yet unquantified.
Study Design: This retrospective review of charts from September 2011 to September 2016 formed a database of all patients that underwent temporal artery biopsy (TAB) at Kingston General Hospital or Hotel Dieu Hospital. Statistical analysis of hsCRP levels was performed.
Ethics approval was granted by the local Research Ethics Board.
Methods: Age at biopsy, gender, hsCRP, ESR, platelet count, and other confounders such as past medical history and corticosteroid use were abstracted from patient charts. The primary outcome was to establish the predictive value of various hsCRP cutoffs in order to determine the most significant cutoff value for positive TAB.
The sensitivity and specificity of hsCRP was evaluated by 2x2 table. Odds ratios were computed for hsCRP, ESR, and platelet counts. Cutoff values were selected based on current laboratory cutoffs, literature surrounding standard CRP, and a receiver operator curve.
Results: Thirty-seven (19%) of 197 patients (127 male, 70 female) with a mean age at biopsy of 71 (range 36-92; SD 9.4) years had a positive TAB. Four cutoff values were established: >3.0mg/L, >8.0mg/L, >15 mg/L, and >35 mg/L. The most statistically significant odds of a positive TAB were found at hsCRP values >35 mg/L (OR 5.86, 95%CI 2.675-12.826; p<0.001). Twenty-six of 72 patients (36%) with hsCRP values >35 mg/L had a positive TAB. At hsCRP values >35mg/L, negative predictive value was 91% and positive predictive value was 36%. Sensitivity and specificity were 70% and 71%, respectively.
Conclusions: HsCRP is useful for risk stratifying patients with suspected GCA, with higher suspicion for values over 35 mg/L, and lower suspicion for values under 3.0 mg/L. The sensitivity and specificity of these values impact clinical decisions around temporal artery biopsies and treatment of GCA in high risk patients.
