The treatment of people with pancreatic necrosis differs from that of people with oedematous pancreatitis. It is important to know the diagnostic accuracy of serum C-reactive protein (CRP), serum procalcitonin, and serum lactate dehydrogenase (LDH) as a triage test for the detection of pancreatic necrosis in people with acute pancreatitis, so that an informed decision can be made as to whether the person with pancreatic necrosis needs further investigations such as computed tomography (CT) scan or magnetic resonance imaging (MRI) scan and treatment for pancreatic necrosis started. There is currently no standard clinical practice, although CRP, particularly an increasing trend of CRP, is often used as a triage test to determine whether the person requires further imaging. There is also currently no systematic review of the diagnostic test accuracy of CRP, procalcitonin, and LDH for the diagnosis of pancreatic necrosis in people with acute pancreatitis.
To compare the diagnostic accuracy of CRP, procalcitonin, or LDH (index test), either alone or in combination, in the diagnosis of necrotising pancreatitis in people with acute pancreatitis and without organ failure.
We searched MEDLINE, Embase, Science Citation Index Expanded, National Institute for Health Research (NIHR HTA and DARE), and other databases until March 2017. We searched the references of the included studies to identify additional studies. We did not restrict studies based on language or publication status, or whether data were collected prospectively or retrospectively. We also performed a 'related search' and 'citing reference' search in MEDLINE and Embase.
We included all studies that evaluated the diagnostic test accuracy of CRP, procalcitonin, and LDH for the diagnosis of pancreatic necrosis in people with acute pancreatitis using the following reference standards, either alone or in combination: radiological features of pancreatic necrosis (contrast-enhanced CT or MRI), surgeon's judgement of pancreatic necrosis during surgery, or histological confirmation of pancreatic necrosis. Had we found case-control studies, we planned to exclude them because they are prone to bias; however, we did not locate any. Two review authors independently identified the relevant studies from the retrieved references.
Two review authors independently extracted data, including methodological quality assessment, from the included studies. As the included studies reported CRP, procalcitonin, and LDH on different days of admission and measured at different cut-off levels, it was not possible to perform a meta-analysis using the bivariate model as planned. We have reported the sensitivity, specificity, post-test probability of a positive and negative index test along with 95% confidence interval (CI) on each of the different days of admission and measured at different cut-off levels.
A total of three studies including 242 participants met the inclusion criteria for this review. One study reported the diagnostic performance of CRP for two threshold levels (> 200 mg/L and > 279 mg/L) without stating the day on which the CRP was measured. One study reported the diagnostic performance of procalcitonin on day 1 (1 day after admission) using a threshold level of 0.5 ng/mL. One study reported the diagnostic performance of CRP on day 3 (3 days after admission) using a threshold level of 140 mg/L and LDH on day 5 (5 days after admission) using a threshold level of 290 U/L. The sensitivities and specificities varied: the point estimate of the sensitivities ranged from 0.72 to 0.88, while the point estimate of the specificities ranged from 0.75 to 1.00 for the different index tests on different days of hospital admission. However, the confidence intervals were wide: confidence intervals of sensitivities ranged from 0.51 to 0.97, while those of specificities ranged from 0.18 to 1.00 for the different tests on different days of hospital admission. Overall, none of the tests assessed in this review were sufficiently accurate to suggest that they could be useful in clinical practice.
The paucity of data and methodological deficiencies in the studies meant that it was not possible to arrive at any conclusions regarding the diagnostic test accuracy of the index test because of the uncertainty of the results. Further well-designed diagnostic test accuracy studies with prespecified index test thresholds of CRP, procalcitonin, LDH; appropriate follow-up (for at least two weeks to ensure that the person does not have pancreatic necrosis, as early scans may not indicate pancreatic necrosis); and clearly defined reference standards (of surgical or radiological confirmation of pancreatic necrosis) are important to reliably determine the diagnostic accuracy of CRP, procalcitonin, and LDH.