EPISOD revisited …..and validated, thank you Zach

My gastroenterologist readers, if any, (thank you) may recall that I led the team that developed a research study called EPISOD (Evaluating Predictors and Interventions in Sphincter of Oddi Dysfunction). The results were published in 2014 (Cotton PB, Durkalski V, Romagnuolo J, et al. A multicenter, randomized trial of endoscopic sphincterotomy for suspected sphincter of Oddi dysfunction in patients with pain after cholecystectomy – the EPISOD trial. JAMA 2014;311:2101-09).

We hoped that publication of the arresting results would affect future practice. Whether it did or not has been addressed by a new publication fronted by Zachary Smith, just this month.  All will be revealed, but I need to take a few steps back for those fortunate enough to be unfamiliar with the little sphincter and its tantrums. Please hold your excitement while I set the stage and paint the scenery.

As a medical student in 1887, Ruggero Oddi described muscle fibres surrounding the termination of the biliary and pancreatic ducts where they drain into the duodenum though the ampulla of Vater.  

Oddi got the credit for the sphincter but, like many eponyms, it was described much earlier, actually by Francis Glisson, an English physician.

The sphincter opens and closes to control the flow of bile and pancreatic juices into the intestine when needed to assist digestion.

“Sphincter of Oddi dysfunction (SOD)” is the label given to the concept that failure of the sphincter to relax when the juices are flowing can result in increased pressure in the bile duct and/or pancreatic duct with resulting pain (especially in patients who have had the gall bladder removed, which otherwise can accommodate some of the back pressure).

The logical treatment for this is to disable the sphincter, which is now easy to do during the ERCP procedure. Specialists can pass a flexible endoscope through the mouth, find the ampulla, enter the bile duct and then use cautery to cut a “sphincterotomy”. Over the last 30 years, endoscopic sphincterotomy has become a popular treatment for patients with suspected SOD, and I did thousands.

Not all patients benefitted, and the procedure was/is not without hazard. In particular, it causes immediate pancreatitis, in at least 10% of patients, some rather serious.  Clearly that meant and means that it is essential to offer the treatment only to those patients most likely to benefit. What do we know about that?

It is logical to assume that the increased pressure in the bile duct would result in two measurable phenomena. The bile duct would increase in size, and liver blood tests would rise. These ideas led Walt Hogan, Joe Geenen and colleagues in Milwaukee more than 30 years ago to postulate 3 types or levels of SOD. Type 1 patients had both enlarged bile ducts and elevated liver tests, Type 2 had one or other but not both, and Type 3 had neither (simply pains that sounded “biliary”). Many studies suggested that the outcomes of sphincterotomy mirrored those categories, with more patients reporting benefit in Types I and 2. The benefits in Type 3 (“pain only”) patients were less clear (and the risks more worrisome), resulting in many being referred to “expert centers” like ours at MUSC. Another reason for referral was that we were able to measure the pressure in the sphincter at ERCP, there being some evidence that “sphincter manometry” was useful in predicting the outcomes of sphincterotomy. These patients began to dominate the practice of this particular “expert”, and I felt increasingly the need to get a better handle on which patients should be offered sphincterotomy, and, just as important, which should not.

So, we assembled a team at MUSC and collaborators in 8 other interested centers in USA, and, in 2005, submitted a grant application to the National Institutes of Health (NIH) to study these “Type 3” patients. After a lot of hassle and re-writing it was eventually funded, and enrollment began in 2007. The protocol was simple in concept, a classic sham/placebo-controlled randomized study. After of course approval by institutional research review boards and appropriate informed consent, all post-cholecystectomy patients with severe attacks of “biliary-type” pains but no dilation of the bile duct or significant elevations of liver tests, underwent the ERCP procedure with sphincter pressure measurements.  They were then randomly allocated to undergo sphincterotomy or no treatment (at the same procedure). The patients, their subsequent care givers and the research staff following them to assess the outcomes were “blinded”, that is they did not know which patients had been treated (to avoid bias in the assessment).

To our surprise, at one year, the two groups had similar outcomes. Indeed, at five years, the untreated patients were doing better! 73% of them were satisfied as opposed to only 37% of those who had active treatment. We could not identify any patient characteristic (age, sex, pain type, intensity etc) that predicted the outcomes, and the results of manometry were equally unhelpful.

These results indicated that ERCP sphincterotomy should no longer be offered to “Type 3” patients, indeed the prestigious Rome Foundation (International arbiter of functional GI disorders) subsequentially abolished the 1,2,3 classification.

Several questions jumped out as we attempted to digest the results.

1. Did publication of the results affect subsequent practice?

That’s where Zach comes in (at last). Zachary Smith and colleagues just published the results of a study of that question. (Smith ZL, Elmunzer BJ. Chak A. The Next EPISOD: Trends in Utilization of Endoscopic Sphincterotomy for Sphincter of Oddi Dysfunction from 2010-2019. Clinical Gstroenterology and Hepatology. https://doi/10.1016/j.cgh.2020.11.008)

The figure (supplied by Zach, thank you) shows that the proportion of patients labelled as SOD and undergoing sphincterotomy fell progressively after 2013 (dotted line) when the EPISOD results were first presented at national conferences.

So, it certainly looks like EPISOD made an impact, at least reducing the number of patients being offered a procedure that was both unhelpful and dangerous. That is encouraging, although the authors could not tell who might have been in the type 3 category. It is an interesting wrinkle that Zach works in Milwaukee where enthusiasm for SOD and sphincterotomy was first generated.

• Why did so many sham-treated patients do rather well?

Trying to answer that question led me belatedly into the extensive literature about the placebo response, and resulted in a publication (Why did the sham treated EPISOD study subjects do so well? Important lessons for research and practice? (Cotton PB. Doi: 10.1016/j.gie.2018.11.006) and a ramble in my blog almost exactly 2 years ago “Thinking about placebos and healing. Thankyou Dad” (www.drpetercotton.com/blog).

The answer is simple. The patients were happy because of the quality and intensity of the CARE, not the CUT. The outcomes of some procedures are much influenced by who does it, in what context, and with how much conviction. The placebo response is particularly strong when treatments are applied for pain. Lots of implications for practice there.

• Do these findings raise questions about the value of sphincterotomy for “SOD patients” with some evidence of biliary obstruction, ie old Types 1 and 2?

Belief in the value of sphincterotomy in that context has been based largely on two early and rather small randomized trials from Milwaukee and Australia (by a Milwaukee trainee) which emphasized the value of manometry. That maneuver has largely been discarded and sphincterotomy is nowadays offered based largely on the severity of symptoms.  No one would claim 100% success, so we still need to refine the indications. Which combination of patient characteristics best predicts success? This question is currently being addressed by the RESPOnD (Results of Ercp for SPhincter of Oddi Disorders) study based at MUSC, led by my colleague Greg Cote. Patients in 12 US centers undergoing sphincterotomy for suspected SOD (of any type) are being followed for a year. It will be a while before we know the results. We will either have much better data on which to advise patients, or direction for more studies, perhaps even a sham-controlled trial. Could there be yet another EPISOD in this story?

All part of life’s rich tapestry.