Ratio of serum amyloid A to C-reactive protein is constant in the same patients but differs greatly between patients with inflammatory diseases.

C-reactive protein (CRP) is commonly monitored to track the activity of inflammation and has become the gold standard in the management of all inflammatory diseases. Indeed, serum amyloid A (SAA) have seemed to correlate moderately with CRP, but the discrepancy of CRP and SAA levels has often been reported, especially in rheumatoid arthritis. Then, we examined CRP reflects a real magnitude of inflammation in patients with rheumatic and infectious inflammatory diseases. A total of 414 patients with infectious and non-infectious inflammatory diseases were enrolled. At initial visit, each patient underwent a clinical assessment and had also laboratory tests such as SAA and CRP. In each patient, we carried out a longitudinal analysis of CRP and SAA levels. We determined the inter-individual correlation between SAA and CRP and also clarified intra-individual changes of SAA/CRP ratio. SAA and CRP levels changed approximately linearly over time within individuals irrespective of rheumatic and infectious inflammatory diseases. However, SAA/CRP ratios differed dramatically between patients (from 0.117 to 50.8, median 5.71). In patients with high SAA/CRP ratio (>8.44), SAA is a better predictor of inflammation than CRP. In contrast, CRP is a better predictor in patients with low ratio (<3.52). Our results suggest that the SAA/CRP ratio differed greatly between individuals but was constant in intra-individuals. Low CRP levels could be accompanied by SAA levels predicting any degree of inflammation, implying that CRP is not reflecting a real magnitude of inflammation. To evaluate the real magnitude of inflammation, to access the SAA/CRP ratio in advance is essential.