Concordance between CRP and SAA in familial Mediterranean fever during attack-free period: A study of 218 patients

Abstract

Introduction: Monitoring SAA level in attack-free FMF patients is recommended in order to adjust colchicine dose, and minimize the risk of AA amyloidosis. In countries where this test is not available, C-reactive protein (CRP), another acute phase reactant, is used instead. However, CRP is low and SAA is increased in some patients and vice versa.

Objectives: To determine the threshold of CRP corresponding to SAA<10mg/L in patients with FMF and to assess their concordance at the patient level.

Patients and methods: Consecutive FMF patients in attack-free period and no other cause of intermittent inflammation including infections were recruited during their regular visits in the French reference center for FMF. Demographic and genetic data were recorded; CRP and SAA were tested simultaneously. The threshold value of CRP corresponding to 10mg/L for SAA was determined and the concordance between the two markers was assessed with Cohen's kappa index.

Results: 399 samples were obtained from 218 patients, mean age of 27years (33% under 18years old), 55% of female, from Sephardic Jewish origin in 71%. MEFV mutation was M694V homozygous or compound heterozygous in 52%, and simple heterozygous in 18%. Six patients had AA amyloidosis. The appropriate CRP threshold was found to be 5mg/L in children and 8.75mg/L in adults. Global agreement with SAA<10mg/L was 84% [95% confidence interval: 82 to 86%], leading to a kappa index at 0.62 [95% confidence interval: 0.57 to 0.68].

Conclusion: CRP<5mg/L in FMF children or 8.75mg/L in FMF adults during attack-free periods might be a convenient substitute to guide therapeutic decisions when SAA is unavailable.

Keywords: C reactive protein; Familial Mediterranean fever; Serum amyloid A.