Background Familial Mediterranean fever (FMF) is characterized by recurrent, self limited episodes of fever and serosal inflammation. The diagnosis still based on clinical criteria. We presented the decreased specificity and positive predictive value (PPV) of serum amyloid A (SAA) in predicting FMF in PRES 2014 and concluded that SAA can be used as an independent laboratory parameter to support FMF diagnosis.
Objectives The objective of this study is to compare SAA levels with mutation results and disease severity scores of the patients to find out if there is a relationship.
Methods We reviewed the medical files of 120 patients with FMF followed up in our center in which SAA was measured and yet untreated. The diagnosis of FMF was established according to Livneh criteria. All of them also met Yalçınkaya criteria. Patients were classified according to the presence or absence of attack while SAA measurement was performed. The levels of ESR, CRP and fibrinogen were measured simultaneously with SAA. For each parameter the level above the normal range accepted as increased. Assessment of the disease severity was performed using the scoring systems of Mor et al. and Pras et al. Mutational analysis was available in 109 patients. The mutations were classified as absence of mutations, pathogenic mutations (M694V, M694I, M680I, V726A), uncertain mutations (E148Q, P369S) and single-nucleotide polymorphisms (R202Q). Differences among the mutations groups and SAA levels were evaluated by Kruskal-Wallis variance analysis. Spearman's rank correlations were used to evaluate the correlation between disease severity scoring systems and SAA levels.
Results Thirty five patients were evaluated during FMF attack and 85 were evaluated during the attack free period. The median (inter quartiles range) levels of evaluated parameters during the attack and attack free period were: SAA 170 (475,9) mg/L, CRP 6.14 (6,25) mg/dl, fibrinogen 4.92 (2,38) g/L, ESR 34 (30) mm/hour; SAA 26.4 (81.02) mg/L, CRP 0.3 (0.95) mg/dl, fibrinogen 3.34 (1.11) g/L, ESR 10 (13) mm/hour respectively. The sensitivity, specificity, PPV and negative predictive value (NPV) of SAA and CRP were: 94%, 25%, 34%, 91%; 97%, 57%, 48% and 98% respectively. Spearman's correlation analysis showed that the correlation coefficient between SAA and CRP was 0.489 (p<0.05). Any correlations between the disease severity scoring systems and SAA levels were not. Also the two scoring systems (Pras and Mor) were not statistically consistent with each other. Additionally there was no significant difference determined for the SAA levels among the three groups of mutations. According to the groups of genetic mutations and presence or absence of attack, SAA levels did not show any significant difference (p=0.191, p=0.626).
Conclusions As a conclusion SAA and CRP is highly sensitive diagnostic tests for FMF attack, where as relatively lower specificity of SAA during attack free period indicates it as an independent laboratory predictor of subclinic inflammation in FMF. However SAA is neither a predictor of disease severity nor shows a relation with pathogenic mutations.
Disclosure of Interest None declared