Serum amyloid A versus C-reactive protein in acute pancreatitis:: Clinical value of an alternative acute-phase reactant

Objectives: The acute-phase reactant C-reactive protein (CRP) is currently the serum variable of choice for an early, accurate, and cost-effective severity assessment of acute pancreatitis in the daily clinical routine. Serum amyloid A (SAA) proteins comprise a family of apolipoproteins that constitute another major acute-phase reactant and thus could be a potential alternative to CRP assessment. In the present study we investigated the clinical usefulness of SAA determinations in acute pancreatitis using an automated immunoassay technique. Design: Cohort study, comparing patients with complicated and mild acute pancreatitis; control groups included individuals with further abdominal disorders and healthy volunteers. Setting: A collaborative study between the department of general surgery and the routine laboratory of the department of clinical chemistry/pathobiochemistry. Patients: We enrolled 66 patients with acute pancreatitis in the present study. Control groups consisted of healthy subjects (n = 30), patients with chronic pancreatitis (n = 20), patients with pancreatic carcinoma (n = 20), and patients with acute appendicitis (n = 20). Interventions: Blood samples were collected during 14 consecutive days in patients with acute pancreatitis. A single blood specimen was taken in all control groups after the diagnosis was established. Measurements and Main Results: SAA concentrations were 3 mg/L (median; range, 3-93) in healthy subjects. although SAA and CRP both reached their maximum within 4 days after onset of symptoms in patients with acute pancreatitis, SAA concentrations rose faster above normal ranges and reached 676 mg/L (median; range, 12-1880), higher than CRP, which reached 313 mg/L (median; range, 29-613). As observed for CRP, SAA was significantly higher in patients who developed complications such as necrosis, infection of necrosis, or multiple organ dysfunction syndrome or in patients who died. SAA achieved best results in discriminating between necrotizing pancreatitis and interstitial edematous pancreatitis. However, CRP provided an earlier differentiation between both entities and a significantly better overall accuracy, as shown by receiver operating characteristics analysis. SAA concentrations in patients with chronic pancreatitis were 6 mg/L (median; range, 3-756). In patients with pancreatic carcinoma, SAA concentrations were 7 mg/L (median; range, 3-492), and in patients with acute appendicitis, they were 50 mg/L (median; range, 3-2140). Conclusion: SAA is a nonspecific and rapidly produced variable in inflammatory abdominal disorders with a wider dynamic range than GRP. The current assay technique renders SAA an applicable and readily available variable under clinical routine conditions. In cases of acute pancreatitis, however, CRP is still superior to SAA for early and accurate stratification of patients with a complicated course.