Combined Esophageal Multichannel Intraluminal Impedance and pH Monitoring After Repair of Esophageal Atresia

Abstract

Objective: 

Gastroesophageal reflux disease and impaired esophageal motility is a frequent problem after repair of esophageal atresia (EA). Combined multichannel intraluminal impedance (MII) and pH measurement identifies acidic and weakly acidic esophageal bolus movements. Our aim was to evaluate gastroesophageal reflux and its association with gastrointestinal or extraintestinal symptoms using combined MII-pH measurement.

Patients and Methods: 

In all, 24 patients (4 months–23 years; 17 male, 7 female) with repaired EA were included in this study. A single catheter with 6 impedance channels and 1 pH channel was used to perform the 24-hour MII-pH study. Symptoms were recorded during the study as events and with a standardized questionnaire. In a subgroup of patients, an impedance swallowing test evaluating liquid and viscous bolus movements was performed.

Results: 

We detected a total of 911 episodes of retrograde bolus movement, 379 acidic and 532 weakly acidic bolus movements. In all, 201 symptom events were recorded. Of these events, 42% were associated with retrograde bolus movements. More symptom events were associated with weakly acidic reflux (26%) than with acid reflux episodes (16%). There was a poor correlation between symptoms and MII-pH study findings. In comparison with patients without esophageal surgery, EA patients showed significantly fewer complete swallows of liquid (42% vs 98%) and viscous material (18% vs 87%).

Conclusions: 

In patients with corrected EA, half of the reflux events could be detected only by MII. We also demonstrated that weakly acidic reflux can be responsible for the patients' symptoms. However, patients may have few or no symptoms despite poor esophageal function and extensive gastroesophageal reflux disease.

Esophageal atresia (EA), with an incidence of 1 in 2500 births, is one of the most common gastrointestinal malformations. Reconstructive surgery is successful in almost all patients, who now must cope with several long-term complications that were unknown in these patients in the middle of the 20th century when the first operations were reported ⁽¹⁾. Gastroesophageal reflux disease (GERD) affects about half of the patients with repaired EA and may be held responsible for many of the symptoms that arise after surgical correction ^(2,3). GERD often manifests itself early after the primary repair and frequently requires long-term medical therapy or antireflux surgery. In adults with repaired EA, reflux esophagitis was present in 58% and Barrett esophagus in 11% ⁽⁴⁾.

Monitoring of pH is still the most widely used diagnostic tool for gastroesophageal reflux (GER) in children. However, this method detects only the presence of acidic material in the esophagus. The use of combined pH and intraluminal multichannel impedance (MII) allows the detection of both acidic (pH <4) and weakly acidic (pH ≥4) GER episodes, the height of the refluxate, and the total acid and bolus clearance time in a patient ⁽⁵⁾. Combining MII with pH monitoring on a single catheter has proved to be a valuable tool for diagnosing GERD and associated symptoms in infants and children and increases the sensitivity and specificity of the detection of GER ⁽⁶⁾.

The technique of multichannel intraluminal impedance was first described by Silny ⁽⁷⁾ more than a decade ago. The principles of impedance testing are based on measuring differences in resistance to alternating current of the intraluminal contents, using multiple impedance measuring sites. The technique allows detection and quantification of bolus movement. Furthermore, the combination of multichannel intraluminal impedance and manometry has been shown to be a valuable tool in the investigation of esophageal motility disorders ^(8,9).

We conducted a follow-up study based on combined multichannel intraluminal impedance and pH monitoring to investigate the incidence of GERD and associated symptoms in a cohort of patients who had formerly undergone surgical repair of EA. In a subgroup of these patients, we conducted additional impedance swallowing testing for liquid and viscous bolus transit.

PATIENTS AND METHODS

In 3 centers of pediatric gastroenterology associated with the German Pediatric Impedance Group, 24 patients with a median age of 3.5 years (4 months–23 years, 17 male, 7 female) with repaired EA were studied by a combination of multichannel intraluminal impedance (MII) and pH measurement. According to the classification by Vogt ⁽¹⁰⁾, 21 patients (87.5%) were classified as type IIIb, 1 (4%) as type II, 1 (4%) as type IIIc, and 1 (4%) as type IV. The patients were recruited at routine follow-up appointments. In patients who were currently being treated for GERD with proton pump inhibitors or other antacids, the medications were discontinued 1 week before the investigation. Written informed consent was obtained for all of the patients.

Before the measurements were made, all of the patients or their parents received a questionnaire to evaluate symptoms related to GER or esophageal function ⁽¹¹⁾. In the questionnaire the patients were asked to score the intensity (0 = not present, 1 = mild, 2 = moderate, 3 = severe) and the frequency (1 = 2–3 times/month, 2 = 1 time/week, 3 = 2–3 times/week, 4 = daily) of the following symptoms: heartburn, belching, cough, odynophagia, dysphagia for liquids and solids. For each symptom, a specific symptom score (intensity × frequency) and a total symptom score (sum of all specific symptom scores) was calculated.

Performance of combined MII and pH measurement for 24 hours was performed with age-appropriate probes (Sandhill Scientific) according to the manufacturer's instructions. The pH sensor was placed 2 to 5 cm above the lower esophageal sphincter, depending on the age of the patient, and correct placement was confirmed by fluoroscopy. Recording of the MII was performed with a Sleuth mobile impedance recorder (Sandhill Scientific). Patients or their parents were instructed in the use of the symptom marker keys during the monitoring. The following symptoms were recorded: heartburn or agitation (in infants and toddlers), cough, and regurgitation. Symptoms within a 2-minute range before or after the bolus movement were regarded as associated.

Analysis of the MII-pH tracings was performed by computer with software from Sandhill Scientific (BioView Analysis version 5.0.9). The tracings were revised visually and manually for bolus and symptom events and for bolus–symptom correlation by 2 investigators ⁽⁵⁾. The reflux index (RI) (percentage of recording time with pH <4) and bolus index (percentage of recording time with esophageal exposure to a refluxate) were included in the final analysis reports.

The symptom index (SI) and symptom sensitivity index (SSI) were calculated for pooled data as SI = (reflux-related symptom events/total symptom events) × 100% and SSI = (reflux-related symptom events/total reflux episodes) × 100%. According to the literature, SI values below 50% were considered abnormal, and SSI values above 10% were considered abnormal ⁽¹²⁾. The SI and SSI were calculated on the basis of the pH probe results alone (RI) and then recalculated on the basis of the combined MII-pH results for nonacid and total reflux events.

All of the reflux events were evaluated manually for their proximal extent. Retrograde bolus movements that reached at least channel 2 (the second most proximal channel) in the upper esophagus were considered high refluxes.

An impedance swallowing test was performed in 5 EA patients and in a cohort of 6 patients with GER symptoms without any previous surgery (EA patients ages 9–23 years, median 13 years; non-EA patients ages 8–12 years, median 10 years). Administration of swallows and diagnosis of abnormal bolus transit for either liquid or viscous material was based on criteria established by a combination of MII manometry with 10 liquid and 10 viscous swallows in a study with healthy adult volunteers by Tutuian et al ⁽⁹⁾. In the supine position patients were given 10 swallows of 5 mL normal saline and 10 swallows of 5 mL viscous material with standardized impedance value (Sandhill Scientific). Swallows were considered as complete bolus transit if the bolus entry occurred at the most proximal site and bolus exit points were recorded subsequently in all 3 distal impedance-measuring sites (channels 4, 5, and 6). Incomplete bolus transit was defined if the bolus exit could not be identified at the 3 distal channels. Also, the difference in the amplitude (impedance baseline in ohms) in channels 5 and 6 was analyzed during the swallowing studies.

Statistical analysis was performed as follows. Correlation analysis was done with the Pearson correlation coefficient, and comparison of RI in different patient groups (low vs high symptom scores) was done with the Mann-Whitney U test. To analyze the percentage of complete versus incomplete bolus movements for the functional testing, we used the independent t test. For all of the statistical tests, significance was set at α = 0.05.

RESULTS

Evaluation of the patient questionnaires showed heartburn in 4 patients (17%), belching in 9 patients (38%), cough in 14 patients (58%), dysphagia for liquids in 3 patients (13%), dysphagia for solids in 14 patients (58%), and odynophagia in 3 patients (13%). See Table 1 for details. The median of the total symptom score of all symptoms combined was 7.5 (range 0–34).

TABLE 1: Symptoms of 24 patients with repaired EA according to standardized questionnaire

Combined MII and pH Measurement

In the group of 24 EA patients, 911 episodes of retrograde bolus movement, 379 acidic (42%) and 532 weakly acidic (58%), were detected. By the pH probe alone, 1073 episodes of pH below 4 were detected independently by MII. The median pH RI was 2.5% (range 0%–42.3%). Eight patients (33%) had a pathological pH measurement, defined as RI above 5%. In contrast to the RI, the bolus index measured by impedance showed a median of 1.7% (range 0.4%–12.2%). Normal values for children have not been established, as far as we are aware. For adults, a bolus index of 1.4% is considered normal ⁽¹³⁾. When this reference was applied to our group, a total of 16 patients (67%) had an abnormal bolus index.

Correlation of Symptom Scores and RI or Bolus Index

There was no significant correlation between the total symptom score and the RI or bolus index. To evaluate the clinical value of the total symptom score, patients were divided into 2 groups, 1 with only minor symptoms (total symptom score <14) and the other with major symptoms (total symptom score ≥14). Applying the Mann-Whitney U test, we found that patients with a total symptom score of 14 or higher showed a statistically significant higher RI (18.15% vs 1.95%, P < 0.05) and bolus index (4.5% vs 1.5, P < 0.01) than did patients with only minor symptoms (Fig. 1A,B).

FIG. 1: A, Results for bolus index with MII-pH for patients with a total symptom score <14 (left) and ≥14 (right). Horizontal line depicts the median. B, Results for the RI with pH alone for patients with a total symptom score <14 (left) and ≥14 (right). Horizontal line depicts the median.

Symptom Event Correlation

In all, 201 symptom events were recorded. In 6 patients, no symptoms were recorded during the investigation. Of all 201 symptom events, 85 (42%) were associated with retrograde bolus movement. The number of associated weakly acidic refluxes was greater than the associated acid refluxes: weakly acidic 55 (26%) versus acidic 35 (16%). The use of combined MII-pH compared with the pH probe alone increased the SI in the pooled data from 14% to 36%. The SI for non–acid reflux events alone was 22%. No difference in the SSI using pH/MII (9.88%) compared to pH probe alone (9.2%) was identified. SSI for non–acid reflux events was 10.36%. By definition ⁽¹²⁾, only the SSI for non–acid reflux events can be considered positive.

Looking into the single symptom events, regurgitation showed the highest association (92%) to retrograde bolus movement (55% of them being acid related), agitation showed the second highest association (65% of symptom events being related to reflux episodes). All other symptoms showed an association of less than 50% with reflux events detected by impedance measurement (Table 2). Of the 911 reflux events detected during combined MII and pH measurement, 90 (9.9%) were associated with a symptom event perceived by the patients. There was no difference between acidic and weakly acidic reflux episodes.

TABLE 2: Total symptom events in 24 patients and their association to reflux events in general and divided into acidic and weakly acidic reflux events as detected by combined MII and pH measurement

Height of Bolus Movement

All of the reflux events were evaluated manually for their proximal extent. Retrograde bolus movement that reached at least channel 2 in the upper esophagus (36.8% of refluxes of all 24 patients) were considered high refluxes. Of these high refluxes, 47% were weakly acidic and 53% were acidic.

Of 4 patients who showed more than 50% high-reflux episodes, 1 had no symptoms, and 3 had chronic pulmonary problems with frequent postprandial coughing and dyspnea. Statistically there was no correlation between the amount of high refluxes and symptom scores or RI.

Impedance Swallowing Test

In a subgroup of EA patients, we conducted an impedance swallowing test as described in Materials and Methods. The 5 patients tested did not stand out from all of our patients with regard to total symptom score (median 7, range 0–10), RI (median 9, range 2.2–17.7), or bolus index (median 1.5, range 0.6–1.8). The control group of non-EA patients showed results of conventional MII and pH data that were not different from those of the EA group with respect to symptom score (median 4.5, range 0–36), RI (median 6.65, range 2–28.9), or bolus index (median 1.65, range 1.1–3). However, in the swallowing test, significant differences became obvious. Less than half of all swallows (liquid and viscous material) undergone by the EA patients showed an MII pattern according to the definition of a complete bolus transit. In our non-EA reference group, almost all of the individuals showed a normal complete bolus transit (Fig. 2).

FIG. 2: Percentage of complete bolus transit for liquid and viscous of swallows in non-EA patients (black bars) and EA patients (gray bars). For both materials (liquid and viscous), a significant difference between the groups was found.

With regard to the morphology of the single swallows, the EA patient tracings, in contrast to the control patient tracings, showed mostly uncoordinated and often hardly recognizable patterns of bolus entry and exit at the different impedance channels.

At the most proximal channels (channel 1), bolus entry and exit showed a normal pattern. In the detailed analysis of the tracings, there is a significant difference in the amplitude (impedance baseline in ohms) in channels 5 and 6 between the EA patients and the control group. The baseline of the EA patients was approximately 25% of the values in the control group. In the proximal channel there was no such difference (Fig. 3 and Table 3).

FIG. 3: Sample tracing of an incomplete bolus transit after 10 mL of liquid material in a patient with EA (A) and a non-EA patient (B). The EA patient tracing depicts an uncoordinated and hardly recognizable pattern of bolus movement, most prominent in the distal channels with low amplitudes.

TABLE 3: Baseline of impedance (median and range) before bolus transit of a liquid or viscous swallow for non-EA and EA patients, determined from channels 1, 5, and 6

DISCUSSION

In patients with corrected EA, the focus is increasingly on possible long-term complications and quality of life. The most frequent complication after EA repair is GERD disease ⁽¹⁾. GER itself can entail other complications, such as intestinal metaplasia, dysphagia ⁽⁴⁾, or pulmonary symptoms. To our knowledge, the present study is the first to search for GER in EA patients with the new method of combined MII and pH measurement. Inasmuch as the patients in our study were enrolled during routine follow-up care, a bias toward patients with more severe symptoms must be taken into account because these patients are more likely to remain in routine follow-up care and to consent to further investigations. However, when the results of the patient questionnaire are considered, the patients enrolled showed a wide range of symptoms, and 3 patients consented although they claimed to have no symptoms at the moment. In our patient cohort, we found a pathological RI in 33% with conventional pH monitoring, whereas a bolus index above the normal values derived from adults was found in 67% of the patients.

So far there are no normative data from children on impedance parameters such as the bolus index. Values for bolus exposure and clearance were compared for the moment with available adult data. In our study we therefore focused on symptom correlation with bolus exposure.

More than half (58%) of the reflux events detected in our patient group were weakly acidic and would not have been detected by pH probe alone. That means that the sensitivity of pH monitoring alone was only 42%. Only 35% of the episodes detected by pH monitoring were confirmed by the MII. Our SI results underline these findings for MII-pH compared with the pH probe alone. Although SI and SSI have been demonstrated to be valuable clinical parameters in other patient groups, the results in our patients were negative except for SSI for nonacidic reflux events. We believe that this is partially due to the lack of sensitivity and ability to communicate symptoms in patients with repaired EA, as discussed below. These results are consistent with studies showing better sensitivity and positive predictive value for the combined method in comparison with conventional pH monitoring ^(5,14,15).

It has also been recognized that EA patients may be insensitive to symptoms of GER or dysphagia and that a correlation between symptoms and criteria for GERD such as histological esophagitis is difficult to create ^(16,17). In our patient cohort, there was no correlation between an overall symptom score derived from a questionnaire, the RI, and the bolus index. Even an association between specific symptoms and reflux events could not be demonstrated. It must be considered that the patient questionnaire we used has not been formally established in pediatric patients. However, its general character may still make it a valuable tool to assess esophageal symptoms. Allocating patients into 2 groups with major (overall symptom score ≥14) or minor symptoms, we found a statistically significant correlation with higher RI and bolus index in the first group. Thus, the information from the questionnaire may help in judging the severity of GERD in EA patients.

Looking at individual patient recordings, we found that EA patients with GER rarely describe having typical GERD symptoms such as heartburn. It seems that patients with repaired EA are, to some extent, accustomed to (chronic) GER or dysphagia.

Weakly acidic reflux showed a higher tendency to be related to symptoms than did acid reflux. In a study of postprandial reflux ⁽¹⁸⁾, weakly acidic reflux was also more responsible for symptoms such as heartburn than was acid reflux. Our results confirm the importance of weakly acidic reflux as being pathogenic and the benefit of a combined method to detect both acidic and weakly acidic reflux.

Evaluating the reflux height, we found that 37% of all reflux events in our patient cohort ascended to the 2 most proximal channels. Nonetheless, in our patient group there was no relevant correlation of high-reflux events with, for instance, respiratory symptoms, as shown in infants by Wenzl ⁽⁵⁾. Motility problems of the esophagus are also a frequent problem in EA patients, with a prevalence of 75% to 100% ⁽²⁾. There is no improvement with age, and many EA patients experience bolus obstruction or need to flush down solids with water. The causes of disturbed esophageal motility are still not clear because there are indications of abnormal extrinsic innervation at birth and of nerval disruption due to surgical intervention ^(19–21). Our results of the impedance swallowing test with impaired bolus transit in the distal esophagus of all of the EA patients are in line with studies of pediatric EA patients with manometry or manometry and videofluoroscopy ^(22,23) showing uncoordinated swallowing function in the distal esophagus. With the swallowing test we could for the first time visualize the morphology of the uncoordinated or incomplete bolus transit through the distal esophagus of EA patients without using radiography. On the basis of the results in healthy adults, complete bolus transit of more than 80% for liquid swallows and more than 70% for viscous swallows can be expected ⁽⁸⁾. All 5 EA patients tested had abnormal results according to these data. The major reason why only a few patients underwent bolus challenges was that it proved difficult to let children below the age of 6 years (15 of 24 patients) perform controlled swallows of saline or viscous material.

The predominantly lower impedance amplitude in the distal channels of our patients may be explained by a general esophageal hypomotility in the distal esophagus with a distended esophageal wall and liquid material (saliva or reflux) resting in the distal esophagus because of the impaired motility. Dysphagia due to impaired bolus transit is difficult to relieve. Performing an impedance swallowing test in EA patients may help to assign the dysphagia to the underlying diseases. Also, the findings may help the patient understand the symptoms and possibly aid in the evaluation of new (prokinetic) therapeutic strategies in the future.

In the group of EA patients studied, half of the GER events could be detected only by MII. The role of weakly acidic reflux in EA patients and its implication on therapeutic options remains to be determined. Even patients with no symptoms or few symptoms may have severe GER. The high rate of Barrett esophagus (11%) and strictures in (42%) adult patients with repaired EA ⁽⁴⁾ further emphasizes the lifelong need for esophageal surveillance strategies.