ERCP in cholangitis.

No Thumbnail Available
Authors
Menon, Shyam
Issue Date
2014-02-01
Journal
Type
Letter/Correspondence
Keywords
Delayed diagnosis
Patient readmission
Cholangitis
Endoscopic retrograde cholangiopancreatography (ERCP)
Adverse effects
Journal Title
Gastrointestinal Endoscopy
Volume
79
Issue
2
Begin page
359
End page
Abstract
Navaneethan et al1 present elegant data examining the variables associated with 30-day readmission in patients undergoing ERCP for cholangitis. However, the methodology used for their statistical analysis deserves critique and has a crucial impact on their findings. The authors introduce the concept of “door to ERCP” time, which seems intuitively to be a powerful measure of clinical success and attempt to define it as being 48 hours, which on face value is logical. However, there are no data on how this time frame was captured from the univariate analysis. Did the authors analyze other time periods (eg, 24 hours, 72 hours)? Moreover, univariate analysis of the “door to ERCP” variable yields a significance of P = .09, which should not ordinarily not merit further exploration in a multivariate model if the level of inclusive significance was set at P < .05. However, the authors use P < .1 as their minimum significance for inclusion into their multivariate model and do not explain this strategy sufficiently. The results of the multivariate model seem to suggest that “door to ERCP” time is a significant outcome variable, and it is not clear additionally whether a conditional regression model was used in this analysis. I believe that this statistical oversight presents a headline that is appealing to therapeutic endoscopists practicing ERCP but that may not necessarily be accurate. Ultimately, as the authors note, appropriate resuscitation facilitates the speed at which biliary decompression can be scheduled in this critically ill cohort of patients
Citation
Menon S. ERCP in cholangitis. Gastrointest Endosc. 2014 Feb;79(2):359. doi: 10.1016/j.gie.2013.09.005. PMID: 24433633.
Rights