Briefly noted: MA Lowry et al. JPGN 2018; 67: 198-203. This study showed that active eosinophilic esophagitis (EoE) was associated with much lower impedance values that inactive EoE, NERD, and controls. At 2, 5 and 10 cm above the squamo-columnar junction, median values of impedance with active EoE were 1069, 1368, and 1707 respectively. In comparison, inactive EoE had median values were 3663, 3657, and 4494, respectively. My take: Since impedance was also performed during endoscopy with sedation, this does not represent a significant advance in current management.
Another study (Funderburk et al. JPGN 2016; 62: 556-61) has shown that gastroesophageal reflux disease is infrequent in infants with a “strong clinical suspicion for reflux.” This is a good to know since we also know that pharmacologic therapy for gastroesophageal reflux has not been proven to be effective in infancy either.
This retrospective study with 58 infants, including 40 preterm infants, evaluated for GERD with MII-pH studies. Characteristics of cohort: median gestational age 31 weeks, median birth wt 1683 gm, and median age at study: 70 days. 10 patients were receiving acid suppression therapy.
Indications for testing:
- Irritability 55%
- Bradycardia 34%
- Desaturation 31%
- Cough 21%
- Gagging 12%
- Difficulty feeding 12%
- Arching 10%
- Apnea 5%
- Only 6 infants (~10%) had abnormal MII-pH studies (defined as >95th percentile for reflux episodes/hours or >95th percentile for acid exposure time)
- None of the symptom indices correlated with symptoms. SI, SSI, or SAP
- The majority of reflux episodes did not correlate with clinical “reflux” behaviors
- Small bore (5 Fr) NG tubes were not associated with increased reflux.
In the related commentary by Rachel Rosen (pgs 517-18), she noted that “there is little to no evidence to show that the 3 indices predict any meaningful clinical outcome…including response to fundoplication, or medications.” “The current literature fails to support the use of symptom indices to prove causality when resolution of symptoms with medical or surgical therapies is used as the criterion standard.”
My take: The vast majority of infants with “reflux behaviors” do not have reflux. Even if they do, current pharmacologic therapies have not been shown to work. So, there is little value in reflux testing in most infants. Finally, given the failure of symptom indices, does the addition of the impedance data to the pH data add any value?
Related blog posts:
Every now and then a dentist sends a kid to our GI practice due to eroded teeth because of concerns about reflux damaging the enamel. While it is recognized that reflux may damage teeth, the exact frequency is unclear. Other questions:
- Which asymptomatic kids with poor dentition require GI evaluation?
- What is the best way to evaluate these children?
- If reflux is identified, how long should they remain on treatment? Forever?
- How effective is reflux treatment in reducing tooth damage?
While none of these questions have been definitely answered, Rosen et al (JPGN 2016; 62: 309-13) show that acid reflux rather than nonacid reflux is predictive of tooth erosion. In this study, the authors used a prospective cohort of 27 children (age ≥3 years)–ALL of them were ON acid suppression (for >1 year) at the time of pH-MII testing. Key findings:
- Prevalence of tooth erosion was 10 or 27 (37%)
- There was correlation with acid reflux episodes (& time in reflux) and tooth erosion, r=0.44, P=0.02
- There was correlation with reflux index as well, r=0.54, P=0.004, In the tooth erosion group, the mean reflux index was 7.3% compared with 1.6% in no dental erosion group.
- There was no correlation with nonacid reflux with tooth erosion
The authors’ discussion highlights many prior relevant studies and indicates that a pH-metry study alone (rather than pH-MII) “may be adequate.” They note some of the limitations of this study which included a small number of patients and potential referral bias, as these children had suspected GERD. In the methods section, the authors state that their standard practice, at the time of the study, was to maintain patients on prior acid suppression medication. It would be useful to acknowledge that many experts, at this time, recommend doing pH-MII studies as well as standard pH studies off all acid suppression due to improved sensitivity/accuracy.
My take: This study shows that in the 10 children with tooth erosion who had suspected GERD, there was correlation with acid reflux but not with nonacid reflux.
Related blog post: Notes from PPI Webinar GutsandGrowth
Unrelated but interesting: Are medical errors really the 3rd leading cause of death in U.S.? Here’s NPR’s summary of a recent BMJ article which makes that claim: Only Heart Disease and Cancer Exceed Medical Errors As Cause of U.S. Death
With some studies, the abstract may suggest a more compelling result than is truly evident. That’s how I feel about a recent report (Nobile S, et al. J Pediatr 2015; 167: 279-85).
Here’s the conclusion (verbatim) from the abstract: “The increased number of (and sensitivity for) pH-only events among infants with BPD may be explained by several factors, including lower milk intake, impaired esophageal motility, and a peculiar autonomic nervous system response pattern.”
To me, it sounds like this prospective study of pH-multichannel intraluminal impedance (pH-MII) of 46 infants born ≤32 weeks gestation (12 with bronchopulmonary dysplasia (BPD) and 34 without BPD) must have identified something important linking gastroesophageal reflux disease (GERD) and BPD. But, the real findings, in my view, are that this is a negative study. Period.
Here are the results reported in the abstract:
- “Infants with BPD…had increased numbers of pH-only events (median number 21 v 9) and a higher symptom symptom sensitivity index for pH-only events (9% vs. 4.9%)”
- They also state: “the number and characteristics of acid, weakly acid, nonacid and gas gastroesophageal reflux events, acid exposure, esophageal clearance, and recorded symptoms did not significantly differ between the 2 groups.”
Here’s a little more data –not in the abstract:
- The P value for the difference in pH-only events was .360
- The authors could just have easily pointed out (in the abstract) that infants without BPD had increased acid exposure: 40.5 min compared with 27.0 min (P = .599)
What should have been in the abstract conclusion? Perhaps, the first line of their discussion: “Infants with BPD did not have significantly higher GER features compared with infants without BPD as measured by esophageal pH-MII monitoring, except for higher occurrence of pH-only events and higher SSI for pH-only events.”
The authors try to explain the differences in the BPD patients by highlighting some of the potential mechanisms of reflux and/or autonomic dysfunction. I think the limitations of this study deserve careful scrutiny. This was a small study with only 12 BPD infants. There was a significant selection bias -only ‘symptomatic’ infants were included. Some of the factors affecting BPD directly could have an indirect effect on reflux (eg. caffeine).
The authors make one other point: “we believe pharmacologic treatment for GER should be initiated only after the demonstration of pathologic pH-MII monitoring to avoid unnecessary drug therapy, adverse events, and costs.”
Related blog posts:
Several years ago, an “obituary” was written for the pH probe (Putnam PE,J Pediatr. 2010; 157(6):878-80) due to the presumed superiority of pH-impedance (pH-MII) studies in detecting gastroesophageal reflux disease (GERD) As noted in previous blogs (see below), there have remained a number of concerns with the assumption that pH-MII is an improvement over pH studies without impedance. Several recent studies elaborate on those concerns:
- Cheng F-K F, et al. Clin Gastroenterol Hepatol 2015; 13: 867-73
- Patel A, et al. Clin Gastroenterol Hepatol 2015; 13: 884-91
- Vaezi MF. Clin Gastroenterol Hepatol 2015; 13: 892-94 (editorial)
In the first study, the authors identified 221 patients and retrospectively reviewed GERD testing from 2006-2011. Prior to testing, 97% had received prescribed PPIs before testing; however, PPIs were discontinued for at least 1 week prior to evaluation which included upper endoscopy, esophageal manometry, and pH-MII.
- 21 (10%) had erosive esophagitis
- 61 (27%) had nonerosive reflux disease with increased pH
- 18 (8%) had nonerosive reflux disease with abnormal impedance
- 30 (14%) had hypersensitive esophagus
- 18 (8%) had functional heartburn
- 30 (14%) had other functional disorders
- 43 (19%) were undetermined
Thus, this retrospective study showed that the majority (roughly 2/3rds) of patients with GERD symptoms on PPI therapy did not have GERD based on objective testing. The authors chose to test off PPI therapy “because we postulated that the pretest probability of GERD diagnosis was low, primarily given their lack of response to PPI.”
In the second study, 187 subjects (≥18 years) underwent pH-MII testing in a prospective study from 2005-2010. 49.7% were tested off proton pump inhibitor therapy. Abnormal acid exposure time consistently predicted symptomatic outcome. The authors note that performing pH-MII off PPI therapy best predicts response to antireflux therapy
In the third reference, the editorial which commented on the second, there are several useful points:
- “There is little doubt that pH-impedance testing provides a more sensitive means of comprehensively identifying reflux events in a given patient. However, to date, studies have failed to demonstrate that it provides any significant additional clinical benefit.”
- “Caution must be exercised when incorporating the added objective data from non-acidic or weakly acidic reflux events into treatment decision-making…Studies including Patel et al have not shown that knowledge regarding continued non-acid or weakly acid reflux events alter patient outcomes.”
- “Wireless pH testing is generally better tolerated and provides longer measurement duration”
- The use of symptom indices are too subjective. “Recent data question the use of these indices especially in those with refractory symptoms and minimal reflux by pH or impedance testing.” SI and SAP could be altered by chance occurrences….”A colleague expert in esophageal diseases …once said: “I know the tests are no good but I don’t know what else to use.'”
- “Let us simplify our approach on the basis of available data and not use measures that we know are suboptimal at best.”
After looking at these studies and the previous pH probe obituary, I’m reminded of a story. Several religious leaders were asked what they wanted someone to say at their funeral. A few stated that they wanted their congregants/flock to comment on their values, like piety and charity. However, one said, “I hope they say, ‘Look he’s moving!'”
Bottomline: There is no reliable evidence that pH-MII testing improves outcomes over conventional pH probe testing. In fact, the use of pH-MII, by lowering the specificity for GERD, could have a detrimental effect. With either test, holding acid suppression for 1 week (with PPIs) is likely to be helpful in interpreting the results.
Related blog posts:
- Even the Experts Agree: pH-MII is a “Flawed Test” | gutsandgrowth
- Why didn’t patient with documented reflux get better with PPI?
- Gastroesophageal Reflux: I know it when I see it | gutsandgrowth
Disclaimer: These blog posts are for educational purposes only. Specific dosing of medications/diets (along with potential adverse effects) should be confirmed by prescribing physician/nutritionist. This content is not a substitute for medical advice, diagnosis or treatment provided by a qualified healthcare provider. Always seek the advice of your physician or other qualified health provider with any questions you may have regarding a condition.
A recent study (JPGN 2014; 58: 22-26) reports on the combination of a new technique of intraesophageal pressure recording (IEPR) along with multichannel intraluminal impedance with pH (pH-MII). While this prospective study is small with only 20 children who had a history of chronic intractable cough, some of its observations are important, especially for those who have embraced pH-MII.
In determining whether the pH-MII studies were abnormal the authors relied on symptom index (SI) defined as the number of symptoms associated with reflux/total number of symptoms. SI is considered positive if >50%. In addition, the authors calculated the symptoms sensitivity index (SSI) which is defined as the total number of reflux episodes associated with symptoms/total number of reflux episodes; it is considered positive if it is >10%. The authors note SAP and SI have a comparable positive predictive value and “our experience suggests that SAP calculation using software is unreliable.”
- IEPR changed the diagnosis in 15-20% of patients depending of scoring index used. That is, IEPR assisted the detection of reflux-associated cough.
- IEPR detected 106% more coughs than patient report alone. Thus, this study, if accurate, indicates that “symptom reporting during pH or pH-MII testing is significantly flawed and, if possible, should not be used alone for clinical decision making.”
- “We did not find a significant association between cough production and the height of the refluxate.”
- The authors argue that since nonacid reflux can be associated with cough and is not always detected with pH-MII, that this could “explain why studies that have tried to use pH criteria to predict clinical outcome after acid suppression therapy have been negative.” The two studies cited at that point by the authors were landmark studies (referenced below) showing that proton pump inhibitors are not effective in children or adults in improving asthma. I think the authors’ comment misses the importance of these studies entirely. There are no proven effective GERD (acid or nonacid) therapies that alter the course of asthma.
Take-home message from authors: “Studies are now needed to determine whether this increased detection improves therapeutic outcomes, but clearly, relying on symptom reporting by patients is flawed and clinical decision making based on patient report alone should be done with caution.”
- JAMA 2012; 307: 373-81
- NEJM 2009; 360: 1487-99
Related blog posts:
Determining whether reflux is causing an infant to have trouble sleeping is quite difficult as a practical matter. A recent retrospective study adds information to this topic but opens up a can of worms (JPGN 2013; 56: 431-35).
The main problem is that clinical reflux occurs in the majority of infants and that sleeping problems are ubiquitous as well. This leaves the door open to testing a lot of infants. At the same time, effective therapeutic options are limited. So, identifying that reflux is causing trouble sleeping, when feasible, may be akin to getting the license plate of the truck that ran you over.
In this retrospective study (2008-2010) of 24 infants with a median age of 5 months, 18 were receiving acid-suppressing medications prior to evaluation. Determination that reflux was causally associated with awakenings and arousals was determined with symptom association probabilities (SAP) based on 2-minute measurement intervals with multichannel intraluminal impedance/pH monitoring/simultaneous polysomnography; SAP was considered significant if ≥95%.
Findings: Seven patients had a positive SAP for arousals due to GER (5 exclusively related to non-acid GER). Nine patients had a positive SAP for awakenings due to GER (4 exclusively related to non-acid GER).
There were several limitations of the study. Besides the small size, the main limitation of this retrospective study was a selection bias. Other limitations included a large number of patients with comorbid conditions and the coincident usage of acid-suppressing medications. 19 of 24 patients had one or more significant comorbid conditions: laryngomalacia in 10, prematurity in 5, genetic syndromes in 3, esophageal atresia in 1, and SLE in 1.
Take-home point: This study is in agreement with the general consensus that GER (acid and non-acid) may trigger sleep interruptions in infants. Nevertheless, given the lack of impact on management, only rarely will infants with poor sleep benefit from these investigations.
Related blog entries: