How Often Should an Endoscopy Be Normal?

An interesting retrospective study (S Wang et al. JPGN 2018; 66: 876-81) looks at a total of 218 endoscopies in 164 children and examines findings in relation to gastrointestinal symptoms.  The results focus mainly on upper endoscopy as there were only 20 who had isolated colonoscopy.

Key findings.

  • 164 of 612 (26.8%) of all new patients had an endoscopy
  • Among upper endoscopy, 72% were histologically normal, and 56% were macroscopically normal.  The authors suspect those with abnormal macroscopic appearance and normal histology were mainly errors in interpretation.
  • In those with isolated colonoscopy, 25% were histologically and macroscopically abnormal; thus, complete concordance among the 20 cases.
  • In those with combined procedures, the likelihood of abnormalities was higher at 53% for both histology and macroscopically being abnormal.  This higher rate was driven mainly by the increased suspicion and diagnosis of inflammatory bowel disease in this subset of 74 patients.
  • Overall, 49% of all first diagnostic endoscopies were completely normal, though 65% were histologically normal.
  • For upper endoscopy, in those who had performance for isolated abdominal pain, the histologic yield was 11%.  It was also 11% for reflux.  The symptom with highest yield was vomiting, yet even for this, the findings were normal in about 80%.
  • For colonoscopy, rectal bleeding had the highest yield, but 72% were normal in this small cohort, indicating that a trial of conservative management may be appropriate.

An important point in the discussion.  The authors take an exception to the idea that normal findings are reassuring.  “When procedures are performed despite the expectation of normality this has not been shown to lead to better clinical outcomes of patients…abdominal pain was unaffected by whether or not the patient had undergone endoscopy>”

My take: I agree with the authors that the histologic findings are more likely significant than macroscopic findings in the majority of patients.  However, it is worth noting that mild histologic findings are of dubious importance in many cases.  Ultimately, identifying strategies to maximize diagnostic yield is needed to provide more cost-effective care and minimize the risks of unnecessary procedures.


Improving the Value of Pediatric Colonoscopy

Two recent studies examine the diagnostic utility of pediatric gastrointestinal endoscopy:

  • PS Kawada et al. JPGN 2017; 64: 898-902
  • M Thomson, S Sharma. JPGN 2017; 64: 903-06

Before looking at these studies more closely, I would say that I was struck by contrasting remarks in their discussions. The first study: “a negative colonoscopy has not been shown to improve outcomes in those with functional pain” and references: Bonilla S et a. Clin Pediatr (Phila) 2011; 50: 396-401.  The second study states that “a negative endoscopic finding, with effective reassurance, can prevent unnecessary medicalization of many children in whom other nonorganic causes may present with GI symptoms.” The latter study does not provide any data to support their claim.

In terms of the specifics, the first study is a retrospective examination of 999 colonoscopies.  The indications for colonoscopy were suspected IBD; in this circumstance, 143 of 449 (32%) were normal.  For isolated rectal bleeding, 141 of 197 (72%) were normal.  For recurrent abdominal pain, all 46 were normal.  The cecal or beyond completion rate was only 52%, potentially lowering diagnostic yield.  The perforation rate during the 10 year timeframe (2001-2010) was 0.2%. The authors conclude that the yield of colonoscopy for recurrent abdominal pain (without other features) is very low and that many children with isolated rectal bleeding “should have a trial of conservative management before undergoing endoscopy.”

The second study retrospectively examined 153 endoscopic cases from a database of 2471 children (2012-2014).  The median age was 9.58 years. The authors found a diagnostic yield of 18.9% for upper endoscopy alone, 32.6% for ileocolonoscopy alone, and 39.2% for combined upper endoscopy/ileocolonoscopy. The terminal ileum intubation rate was 98%.

My take: Both of these studies look at pediatric endoscopy and reach opposite conclusions. The first study suggests that many colonoscopies could be avoided and the latter suggests that whether normal or not, endoscopy contributes to improved management. What is your conclusion?

Related blog posts:

Jean Hugues 1890, Edipe a Colone, Marbre taille d’apres le platre expose au Salon des Artistes fracaise. Musee d’Orsay


Is Propofol Safe in Pediatric Patients with Food Allergy and Eosinophilic Esophagitis?

According to a recent study (P Mehta et al. JPGN 2017; 64: 546-49), propofol was safe in pediatric patients with eosinophilic esophagitis (EoE) and food allergy.

This finding was based on a retrospective study of 1365 upper endoscopies (2013-2014).  Though, propofol was used less frequently, “there was no difference in complication rates relative to propofol use.”

Specifically, egg or soy allergy patients had 38 procedures; 114 children had EoE (without known egg or soy allergy) and 27 and EoE and egg or soy allergy.

This study is important because propofol is used frequently in patients with egg and soy allergies despite a contraindication warning on the package insert. Nevertheless, this study does not provide a definitive answer due to the very low rates of allergic reactions to propofol (~1:10,000 to 1:20,000).  In addition, the diagnosis of food allergy in this study relied on review of the medical record.

My take: This study is limited in scope but did not identify any significant safety concerns with propofol in patients who had EoE and/or egg/soy allergies.

Palace of Versailles

Why Are So Many “Low Value” Endoscopies Performed?

After reading a few commentaries regarding value in medicine (which I will summarize tomorrow), it made me think a little more about value in pediatric gastroenterology.

I recently observed that a pediatric gastroenterologist in another group had a pattern of scheduling a lot of procedures.  In pediatric gastroenterology, we are not doing endoscopies to screen for malignancy.  The majority of children evaluated in our offices do not have organic disease.  In addition, there are a number of variables that can be used to select patients who are most likely to benefit from evaluation. In fact, much of our value comes from this selection process, because non-physicians can be taught to be endoscopic technicians.

My reaction to this volume of cases was that I thought either this practitioner was seeing a ton of patients, had been away and had accumulated a number of cases, or that this was low value care.  Though, another possibility is that the physician may be influenced by the “illusion of control” or “therapeutic illusion.” (NEJM full text: The Science of Choosing Wisely –Overcoming the Therapeutic Illusion).  According to a recent editorial, “When physicians believe that their actions or tools are more effective than they actually are, the results can be unnecessary and costly care.”

“The therapeutic illusion is reinforced by a tendency to look selectively for evidence of impact — one manifestation of the “confirmation bias” that leads us to seek only evidence that supports what we already believe to be true.”

Whatever the circumstances with regard to endoscopy volume, my intent is not to single out an individual or specific group.  My impression is that there are a lot more pediatric endoscopies being done these days and many are not needed.  While I recognize that clinicians recommend endoscopy with a great deal of variation, my suspicion is that those who use endoscopy less frequently are likely to see similar outcomes.  So, why are there so many low value endoscopies performed?

  1. The entire system is incentivized to do more procedures.  Physicians and hospitals are compensated more for doing these procedures.
  2. Families and sometimes referring physicians think these procedures are necessary.  In fact, there are studies that generally indicate higher levels of patient satisfaction when more diagnostic tests are done even if they are unnecessary.
  3. Physicians have a great deal of knowledge asymmetry in healthcare compared with families and it is expected that they will use their knowledge to help families pursue appropriate care.  While all physicians may have some lapses, some physicians skirt this part of their job.  One physician described this type of pediatric GI practice to me: “Scope first, think second.”

This blog has highlighted numerous aspects of health care economics.  Pharmaceutical companies and hospitals have been criticized for gaming the system.  The blog has discussed efforts to improve value like the “Choosing Wisely” campaign.  Though, it is interesting to note that even with this campaign, most physician groups rarely identified areas that would affect their financial bottom-line.  Among pediatric gastroenterologists, a frequent concern that I hear regards the overuse of CT scans by emergency room physicians.

When I take my car for repairs, I don’t want them doing an expensive overhaul unless it is really needed.  If a car needs a muffler change, but the repairman recommended a few thousand dollars of repairs, that would be outrageous.  Yet, in many cases with children, who are more precious than cars, the main difference with excessive endoscopic procedures, is that health insurance covers the majority of the costs.

I wonder too whether the frequency of endoscopy procedures actually discourages some families from having endoscopic procedures when they are clearly needed (eg. suspected celiac disease, suspected inflammatory bowel disease).

My take: Financial resources are limited.  When physicians do not help utilize resources well, this results in poor care, whether families realize this or not.  Ultimately, this will result in increased regulatory burdens for all physicians to more carefully justify what they are doing and/or result in efforts to eliminate financial incentives for unnecessary care.  However, as noted previously (Do deductibles work to improve smart spending on health care?), financial incentives often affect both low value and high value care.

Any readers care to comment?

Related blog posts:

ViK Muniz Art -done completely from chocolate syrup

ViK Muniz Art -done completely from chocolate syrup -see the picture below for comparison.

The Vik Muniz piece is modeled after this photograph of Jackson Pollack

The Vik Muniz piece above is modeled after this photograph of Jackson Pollack

Lower Endoscopic Thresholds for Thrombocytopenia

According to an advances in endoscopy report (Ross, WA. Gastroenterology Hepatology 2015; 11: 115-17), lower platelet thresholds are indicated for many endoscopic procedures.  The author works at MD Anderson Cancer Center in Houston.

Key points:

  • We feel that the traditional threshold of 50,000 platelets/microliter that many doctors adhere to or aim for should be put aside, and a lower platelet threshold of perhaps 25,000 or 30,000 platelets/microliter should be employed for endoscopic procedures, including biopsies.
  • “We found that therapeutic maneuvers could be performed to control bleeding.”
  • “This change would require fewer platelet transfusions to prepare a patient for endoscopy.”
  • Based on their published experience (Krishna SG, et al. “Saftey of endoscopic interventions in patients with thrombocytopenia.” Gastrointest Endosc 2014; 80: 425-34), the author notes that “polypectomy could probably be performed with a platelet count under 50,000/microliter, likely in the 30,000-40,000/microliter range, particularly if the polyp was small (<10 mm).”  They caution that cold snare technique may be safer in this setting but is not suitable for larger polyps.
  • Other preventative measures include stopping aspirin use, limiting the number and size of biopsies, and using non thermal means to help stop bleeding, such as clips or injections.
  • “Performing an endoscopic procedure in a patient with an extremely low platelet count, such as 5000/microliter, is associated with a high risk of bleeding.”

“If the procedure is just a simple biopsy, a platelet count of 25,000/microliter to 30,000/microliter should suffice.”

Take-home message: While the data that the author references is derived from adults, it is likely that in pediatrics that endoscopy, if needed, can be performed in patients with platelet counts less than 50,000/microliter.