Amylase and Lipase Co-ordering in the Workup of Pancreatic Disorders Is Unnecessary and Costly: An Institutional Lab Utilization Study
Abstract Background Acute pancreatitis is the most common gastrointestinal cause of hospitalization in the United States. Amylase and lipase are commonly ordered together in the workup of suspected pancreatitis at our institution (a large academic medical center), despite meta-analyses and evidence-based guidelines that recommend the use of lipase alone. Lipase has more diagnostic accuracy, rises earlier, and stays elevated longer than amylase. Co-ordering amylase has not been shown to increase sensitivity or specificity. This study aims to explore how well amylase and lipase correlate in our patient population, and which marker more accurately detects acute pancreatitis when the results are discrepant. Methods The laboratory information system was queried for co-orders of amylase and lipase during a 2-month period (November 1, 2016, to December 31, 2016). The overall agreement and Cohen’s kappa were calculated. Discrepant results were separated from concordant results. Clinical data were extracted from the electronic medical record, including the rationale for ordering and relevant diagnosis code. Results There were 962 instances of amylase and lipase co-orders in the 2-month period. Of these, 138 had high amylase and lipase, and 694 had normal amylase and lipase concentrations. The overall agreement was 86.5% and Cohen’s kappa was 0.596 (moderate agreement) (P < .0001). There were 88 cases in which the lipase was elevated and the amylase was normal. Of these, 12 were acute pancreatitis and 76 were not. There were 42 cases in which the amylase was elevated and the lipase was normal; only one case out of 42 was acute pancreatitis. Further investigation into this case revealed that the patient was approximately 7 months status-post pancreas transplant. During this admission, he was diagnosed with “graft pancreatitis” and, while amylase remained elevated above the reference range for longer than lipase after treatment, both enzymes were increased at diagnosis (lipase 831 U/L [73-393 U/L] and amylase 194 U/L [30-110 U/L]). These data show that, had only lipase been ordered, no cases of acute pancreatitis would have been missed. Conclusion Amylase has been shown to be an inferior marker of acute pancreatitis compared to lipase, and many guidelines recommend the use of lipase alone. However, these tests are commonly co-ordered at our institution; in a given 2-month period, there were 962 co-order events. Of these, there were no cases of pancreatitis that would have been missed by ordering lipase alone. Our laboratory charges $50 per amylase assay; therefore $48,100 was spent on unnecessary amylase orders during the 2 months of this study. Lab test overutilization has been identified as an important problem; if addressed and curtailed, labs and hospitals can increase efficiency and cost savings.