Sphincter of Oddi manometry: still a valid gold standard?

In this month’s issue of Endoscopy there are two articles, from the same university center, that address the limitations and technical problems of sphincter of Oddi manometry (SOM) as a tool to diagnose sphincter of Oddi dysfunction (SOD) [1] [2]. Since the initial description of manometric evaluation of the motility of the human sphincter of Oddi [3] there has been a lack of universal acceptance of this modality as a clinical tool for evaluation of patients with unexplained pancreaticobiliary pain with or without accompanying biochemical or imaging abnormalities. Even the group whose articles are discussed here refer to SOM on the one hand as a ”gold standard,” the implication being that SOM findings are incontrovertible, and on the other hand simply as a ”reference standard,“ suggesting a somewhat less definitive diagnostic test [1] [2]. Whether SOM can be considered a definitive test or merely an adjunctive investigation whose value is limited is under debate.

In the study by Blaut et al. [1], the influence of variable-stiffness guide wires on the measurement of basal sphincter of Oddi pressures is evaluated in a prospective partially controlled study. Three different wires were used arbitrarily at the discretion of experienced endoscopists, with measurement of sphincter of Oddi basal pressure in a nonrandomized sequence, with assessment of pressure always first with the wire in place and then repeated after wire withdrawal. The stiffness of the three wires was bench-tested using a proprietary Taber Stiffness Tester Model. The wires used were a 0.018-inch Roadrunner with a calibrated Taber stiffness unit (TSU) value of 0.740, a 0.020-inch Glidewire stiff shaft with a TSU of 0.153, and a 0.018-inch Standard Teflon with a TSU of 0.077. There was a convincing and significant difference between the recorded mean basal sphincter pressures, 52 vs 34.4 mmHg, with and without a wire in place. The pressure differences did not correspond in a linear fashion to the wire stiffness, with a delta of 27.5 mmHg in basal pressure comparisons using the stiff Roadrunner, but a delta of 6.4 mmHg with the medium stiff Glidewire, and of 10.1 mmHg with the least stiff Standard Teflon wire. The concern with regard to wire use relates to the potential misdiagnosis of patients, given that the range of normal values was established with a perfusion catheter that was not wire-guided. Indeed, 39 % of patients in the Roadrunner group crossed over from showing normal to abnormal values when the wire was in place, with a 23.5 % rate for the Glidewire and nil with Standard Teflon. The authors correctly emphasize that a false-positive diagnosis would have been made in 24.4 % of patients in this study, and a false-negative diagnosis in 4.4 %.

The study was clearly initiated after an astute observation by these experienced endoscopists of manometric changes following the introduction of guide wire-assisted cannulation. This muddies the waters considerably, given that undoubtedly a large number of manometric procedures in the last several years have been performed with guide wire assistance. The manner in which a simple variation in technique can have such a profound impact on the results is a reflection of the tenuous nature of manometric recordings.

However that was only one variable which came to the attention of the investigators: there could be several other uncontrolled variables in so-called routine manometric studies. For instance the curvature on the tip of the perfusion catheter may impact the pressure measurement through the perfusion ports if one or more of the sideholes is pushed against the wall of the sphincter with increased force. The greater the arc of the perfused ports on the catheter tip, the greater the risk of falsely raising the basal pressure based on the mechanical force exerted on the catheter tip. The lie of the catheter is a function of the location of the papilla, anatomic variations such as diverticula, depth of scope insertion, and the like. Experienced manometrists will be careful to have a freely gliding catheter without mechanical stress, to ensure an accurate measure repeatable on multiple pull-throughs. On the other hand the objectivity of the measurement loses credibility when we realize that the pressures recorded can be easily altered with challenging anatomy, varied depth of scope insertion, or different angles of sphincter orientation, not to mention the potential for the individual endoscopist to inadvertently raise the pressure when the proper angle of entry is not achieved.

This study of variable-stiffness wires emphasizes how seemingly minimal variation in technique challenges the concept of a manometric reference standard. Although this study addressed biliary measurements, I can only think that a similar study of pancreatic sphincter pressure would also add to our concerns.

The second study, by Khashab et al., looks at the frequency of SOD found in patients with a second manometry when the previously performed SOM study was normal. The question posed by the authors is whether a short-term pressure recording will reflect the ”24-hour pathophysiology” of the sphincter. Of course this question remains, as the examination does not evaluate the sphincter function over a 24-hour period and the implication that it might is somewhat misleading. What this study does is to look at the reproducibility of SOM results in a subset of patients, with normal findings in prior pressure studies of the biliary and pancreatic sphincters, who are referred back to the investigators for persistent symptoms. These are self-selected patients in the sense that they were re-studied after presenting themselves to their primary or referral physicians. One of the difficulties in understanding the implications of the findings of this study is that there is no parallel systematic review of the remainder of the cohort who did not return for follow-up studies and whose outcomes are not available as a comparison.

Before going into the details regarding these 30 patients among 5352 studied over a 13-year period, I am somewhat intrigued by the mathematics as applied to these problematic patients. The authors were able to make a diagnosis of SOD in the pancreatic and/or biliary sphincter in 65.8 % of the 5352 patients, with an additional 14.9 % having an elevated basal pressure in one lead only, giving a total of 80.7 %. Among the remainder, those with normal findings at first SOM, the authors had the opportunity to study 30, among whom 60 % had abnormal findings at the second SOM; if this was extrapolated to the entire cohort, it would mean that almost 93 % of the initial 5352 would have abnormal findings. Not to push the point but if a third SOM were to be performed and the results were similar, the total number with an abnormal manometry would reach 97 % and continue to rise with each successive manometry until virtually 100 % would be found to be abnormal if the proportions remained constant. Of course this is absurd and the assumptions applied to this mathematical model are not valid, but it highlights the difficulty in extrapolating the results of a tiny cohort of self-selected patients reported here to the entire group of 1037 patients with previously normal findings. It is curious that such a small minority returned, given the assumption that a single examination is only a glimpse into the 24-hour timeline of sphincter function and that it is unlikely that the other 1007 would have had resolution of their symptoms. Be that as it may, what was found with the repeat study in the cohort reported here?

Let me delve into the details. Among the 30 patients with previously normal pressure studies of both sphincters, 18 were found to have abnormal basal sphincter pressures, 17 of whom had abnormal pancreatic sphincter pressure with or without an accompanying elevation of the biliary pressure. Why should there be such a high proportion of patients with pancreatic sphincter hypertension, especially as 22 of the 30 were originally classified as having biliary SOD prior to the index SOM. Delving deeper, it is notable that among the 17 patients with a hypertensive pancreatic sphincter, 12 had a history of a prior pancreatic stent or post-ERCP pancreatitis and only five had neither. This raises the specter of the possibility that inflammatory changes at the sphincter level following the initial study were responsible for the subsequent findings, in other words, the abnormal pancreatic duct pressure at the second manometry was ultimately iatrogenic.

I have a healthy degree of skepticism when it comes to the validity of a diagnosis of SOD of the pancreatic sphincter when the normal values have been established largely in patients in whom cannulation was relatively easy. Guelrud, in his landmark publication of manometry in 50 normal volunteers, among whom 20 had simultaneous pancreatic pressure measurements, excluded individuals in whom ”cannulation was difficult or when deep insertion of the catheter was not possible“ [4]. I suspect that patients with small ducts or small sphincter lumens would fall into the category of being difficult to cannulate. Therefore sticking a 1.7-mm catheter into a native sphincter lumen less than this size would of necessity yield a high pressure zone on a purely structural basis and unrelated to muscle contraction or spasm. This may also be a reason for the inordinately high percent of hypertensive minor papillary pressure measurements using the standard manometric catheter. The bottom line is that the norm for pancreatic pressure is questionable especially when there is no adjustment for duct size or ease of cannulation as defined in the validating study.

Lastly, the impact of chronic pain and narcotic use should be factored into the argument as to why the sphincter pressure had risen. There are two studies in which SOD has been attributed to chronic opiate ingestion [5] [6] and, given the intractability of pain in this subset of patients, narcotic use may be an unrecognized contributor to the SOD diagnosed on the repeat study.

The authors reference several studies that showed a strong concordance in repeated measures of biliary manometry. The discordant results in the study of Khashab et al. are for the most part attributable to changes in pancreatic pressures. The only other study which showed a high proportion of abnormal SOM on a repeat exam was one in which all patients with abnormal findings on the repeat exam had a hypertensive pancreatic sphincter [7]. When pancreatic sphincter hypertension is implicated in unexplained unassigned abdominal pain, I believe caution should be exercised before embarking on sphincterotomy of the pancreatic sphincter. This intervention mandates pancreatic stenting, with the attendant risks of stent-induced inflammation and pancreatitis may ensue regardless, despite the use of a prophylactic stent. In fact, in the study of Khashab et al. there was a 16.7 % incidence of post-ERCP pancreatitis in spite of placement of a protective pancreatic stent in all five patients with this complication. Re-stenosis or de novo iatrogenic stenosis is a major concern following pancreatic sphincterotomy, especially with an otherwise healthy pancreas.

Although this is not a central issue, I dispute the statement in the introduction to the study of Khashab and colleagues that there is a histopathologic correlation of SOD with sphincter abnormalities identified in tissue obtained by wedge resection from surgical sphincteroplasty specimens. The paper referenced is an older surgical series from UCLA in which sphincteroplasty or surgical sphincterotomy were performed for pancreaticobiliary pathology [8]. This was a report on 109 patients treated over 26 years and in the publication there is a single sentence in the results that states ”Sphincter tissue was examined pathologically in 25 patients and was abnormal in 21 (84 %)”. However the 109 patients included 53 with choledocholithiasis, 28 with dyskinesia or stenosis, and 28 with recurrent and chronic pancreatitis. There is no discussion as to which of the entire group of patients had the histologic changes and certainly no statement regarding the histology in the 26 % of patients diagnosed as having dyskinesia or stenosis. It is suggested that this surgical study showed that the pathology may worsen over time but no evidence was provided in the paper to substantiate such a claim. As SOD remains a controversial diagnosis, great care must be taken not to attribute features which are only speculative.

Although manometry may be an imperfect tool, I do believe that these authors have worked assiduously and dedicated entire careers for the better understanding of the function and dysfunction of the sphincter of Oddi. Their experience is enormous and I have confidence as to their expertise and devotion in trying to help their patients with unexplained abdominal pain. The high rate of post-ERCP pancreatitis is undoubtedly an underpinning of the diagnosis of SOD as a real entity. The skepticism is not so much about the existence of dysfunction but rather the diagnosis and definition of the entity. The lack of widespread adoption of pancreaticobiliary manometry is a reflection of the lingering doubts as to its utility as a clinical tool after three decades of evaluation.

This issue has been addressed in a decision analysis model in suspected type II SOD patients that compared empirical sphincterotomy with initial manometry followed by sphincterotomy only in those with hypertensive sphincters. Empirical sphincterotomy provided cost savings with slightly better clinical outcomes compared with manometry-based therapy, but the model was based on hypothetical premises more reflective of the practice of the authors than that of other centers [9]. A practice-based algorithm of empirical treatment of type II SOD without manometry has been published in abstract form [10]. Among 262 patients, follow-up was available in 215 at a mean of 30 months post sphincterotomy; there was resolution of symptoms in 59 % and improvement in 27 %, results which are comparable to or better than those of manometry-based therapy.

I do not believe we have a ”gold” standard in pancreaticobiliary manometry. We have a tool whose utility is yet to be proven to the majority of endoscopists. The studies reported in this issue of Endoscopy continue to put into perspective the vagaries of the technology of SOM and the complexity of the clinical conundrum of dysfunction of the sphincter of Oddi.

Competing interests: None

References

• 1 Blaut U, Alazmi W, Sherman S. et al . The influence of variable-stiffness guide wires on basal biliary sphincter of Oddi pressure measured at endoscopic retrograde cholangiopancreatography.  Endoscopy. 2010;  42 375-380
• 2 Khashab M A, Watkins J L, McHenry L. et al . Frequency of sphincter of Oddi dysfunction in patients with previously normal sphincter of Oddi manometry studies.  Endoscopy. 2010;  42 369-374
• 3 Geenen J E, Hogan W J, Dodds W J. et al . Intraluminal pressure recording from the human sphincter of Oddi.  Gastroenterology. 1980;  78 317-324
• 4 Guelrud M, Mendoza S, Rossiter G, Villegas M I. Sphincter of Oddi manometry in healthy volunteers.  Dig Dis Sci. 1990;  35 38-46
• 5 Sharma S S. Sphincter of Oddi dysfunction in patients addicted to opium: an unrecognized entity.  Gastrointest Endosc. 2002;  55 427-430
• 6 Mousavi S, Toussy J, Zahmatkesh M. Opium addiction as a new risk factor of sphincter of Oddi dysfunction.  Med Sci Monit. 2007;  13 528-531
• 7 Varadarajulu S, Hawes R H, Cotton P B. Determination of sphincter of Oddi dysfunction in patients with prior normal manometry.  Gastrointest Endosc. 2003;  58 341-344
• 8 Anderson T M, Pitt H A, Longmire W P. Experience with sphincteroplasty and sphincterotomy in pancreatobiliary surgery.  Ann Surg. 1985;  201 399-406
• 9 Arguedas M R, Linder J D, Wilcox C M. Suspected sphincter of Oddi dysfunction type II: empirical biliary sphincterotomy or manometry-guided therapy?.  Endoscopy. 2004;  36 174-178
• 10 Lawrence C, Howell D A, Conklin D E. et al . ERCP sphincterotomy without initial manometry for type lI sphincter of Oddi dysfunction patients: a safe and effective strategy.  Gastrointest Endosc. 2004;  59 AB355