Pilic D, Fröhlich Th, Nöh F, et al. Detection of Gastroesophageal Reflux in Children Using Combined Multichannel Intraluminal Impedance and pH Measurement: Data from the German Pediatric Impedance Group. J Pediatr 2011;158:650-4.
Detection of Gastroesophageal Reflux in Children Using Combined Multichannel Intraluminal Impedance and pH Measurement: Data from the German Pediatric Impedance Group
Denisa Pilic, MDa, Thorsten Fröhlich, MDb, Frank Nöh, MDc, Angeliki Pappas, MDd, Anjona Schmidt-Choudhury, MDa, Henrik Köhler, MDb, Heino Skopnik, MDc, and Tobias G. Wenzl, MDd
a Department of Pediatric Gastroenterology, Children’s University Hospital Bochum at St Josef Hospital, Ruhr University, Bochum, Germany
b Children’s Hospital, University Medical Center, Friedrich-Alexander University Erlangen-Nuremberg, Germany
c Children’s Hospital, Klinikum Worms, Germany
d Klinik für Kinder- und Jugendmedizin, Universitätsklinikum RWTH Aachen, Germany
Objective. To validate the use of combined multichannel intraluminal impedance (MII)–pH (MII-pH) monitoring for detecting gastroesophageal reflux (GER) in children in daily clinical practice.
Study design. The patients were divided into 3 symptom groups based on the main indication for the procedure. MII-pH monitoring was performed in 700 children presenting with symptoms suggestive of GER, including 329 children with pulmonary symptoms, 325 with gastrointestinal symptoms, and 46 with neurologic symptoms. The MII-pH results were compared with pH monitoring alone, and retrograde bolus movements, symptom index, and symptom association probability were compared.
Results. Overall, 270 measurements were abnormal: 101 (37%) showed abnormal MII-pH study, 49 (18%) showed only pathological pH measurements and 120 measurements (45%) had an abnormal MII recording only. Extraintestinal symptoms of GER were seen more often in younger children and were more often related to a normal pH study but an abnormal MII study. Infants had a significantly higher number of retrograde bolus movements than older children. Symptom index and symptom association probability showed moderate agreement (Cohen kappa, 0.54).
Conclusions. From this large systematically standardized data collection of MII-pH measurements in children, we conclude that 45% of the patients with abnormal GER would not have been recognized by 24-hour pH measurement alone. Our findings confirm that MII-pH is superior to pH monitoring alone in detecting GER.
In many children’s hospitals, pH monitoring remains the primary diagnostic tool for detecting gastroesophageal reflux (GER). This method detects changes in acid content in the esophageal lumen; however, it does not actually document reflux as a retrograde bolus movement (RBM). For this reason, a new method, combined multichannel intraluminal impedance (MII)–pH monitoring (MII-pH) is being increasingly used. In this technique, developed in the early 1990s by Silny,1 multiple impedance measurement of segments along a catheter within the esophagus detects bolus movements through changes in electric resistance/impedance between two segments. RBMs can be distinguished from antegrade bolus movements. An additional pH-sensitive antimony electrode measures pH changes during bolus passage, allowing differentiation of acid reflux and nonacid reflux.2
In 2005, the German Pediatric Impedance Group was formed by 4 of the first German children’s hospitals to use MII-pH monitoring in reflux diagnostics: Children’s University Hospital Aachen, Children’s University Hospital Bochum, Children’s University Hospital Erlangen, and Children’s Hospital Worms. This group’s initial aims included to pool measurement data, to validate the importance of this new method in clinical practice, and to create a growing database for future studies.
MethodsStudies were preformed on patients referred to one of the 4 hospitals with suspected GER. A total of 700 patients were recruited between January 2004 and June 2008.
The patients were divided into 3 symptomgroups based on the main indication for the diagnostic procedure: patients with gastrointestinal (GI) symptoms (eg, heartburn, abdominal pain, sickness, vomiting, belching, dysphagia, globus sensation, failure to thrive, food refusal), patients with pulmonary symptoms (eg, chronic cough, chronic bronchitis, tachydyspnea, aspiration pneumonia, apnea, apparent life-threatening event, stridor, chronic laryngitis, dysphonia, otitis), and patients with neurologic symptoms (eg, dystonia, Sandifer syndrome, opisthotonus, tic disorder). The patients had to be between 0 and 16 years of age, and parents had to provide written informed consent for the procedure and inclusion of the study data. The study design was approved by the local Ethics Committee of the University of Aachen Medical Faculty.
Data were recorded by an ambulatory MII-pH system (Sleuth; Sandhill Scientific, Denver, Colorado). One of 3 age-appropriate catheters was used depending on the patient’s height, each of which was mounted with 7 impedance electrodes and one pH-sensitive antimony electrode. The catheter was placed transnasally into the esophagus so that the pH electrode was located at the third vertebral body above the diaphragmatic angle, following European Society for Pediatric Gastroenterology, Hepatology, and Nutrition (ESPGHAN) guidelines.3 The location was confirmed by either fluoroscopy or chest x-ray. The minimum study duration was 18 hours. During the study, each patient or his or her parents were to document experienced symptoms, meal periods, and body position (upright or recumbent).
The recorded data were analyzed using the most current version of BioView analysis software (Sandhill Scientific). After the software-aided analysis, each measurement was validated visually for RBM in the impedance channels. An RBM was defined as a consecutive retrograde drop in impedance by >50% from baseline in at least the distal two channels.
An association between symptoms and RBM was considered positive if a symptom occurred within 30 seconds after an RBM for pulmonary or neurologic symptoms or within 120 seconds after an RBM for GI symptoms. All pH drops exceeding 5 seconds, excluding meal periods, were registered independently from the impedance reflux events. The pH study was defined as abnormal if the reflux index (RI) (% time with pH⩽4, excluding meal periods) was ⩾5% in patients aged >1 year and ⩾10% in those aged <1 year.
The MII study was defined as abnormal if the measurement fulfilled the following criteria: symptom index (SI; number of symptoms associated with reflux/number of all symptoms ×100)⩾50% or a high number of reflux episodes (arbitrarily defined as >70 episodes in 24 hours in patients aged ⩾1 year and >100 episodes in those aged <1 year).
For each symptom group, data on the number of measurements with normal results, abnormal pH with or without abnormal MII results, and abnormal MII results only were compared. In a subgroup of measurements, a retrospective analysis was performed, and information on criteria leading to pathological classification, numbers of recorded symptoms and RBMs, and median and maximal bolus clearance times were collected. Mean values for numbers of RBM were compared for different age groups and between patients with normal RI and those with pathological RI using one-way analysis of variance and the t test.
In a subgroup of patients, the symptom association probability (SAP) was calculated retrospectively using the most current BioView software (Sandhill Scientific). An SAP⩾95% was considered positive. Agreement between the symptom index (SI) and SAP was calculated using the Cohen kappa test. SPSS version 17.0 (SPSS Inc., Chicago, Illinois) was used for all statistical analyses.
ResultsBetween January 2004 and June 2008, MII-pH measurements were performed in 700 patients in the 4 German Pediatric Impedance Group centers, ranging in age from 3 weeks to 16 years (median age, 4 years). These 700 patients included 291 females and 409 males. A total of 329 patients presented with pulmonary symptoms (of whom 39 had additional GI symptoms), 325 patients complained of mainly GI symptoms, and 46 patients reported neurologic symptoms.
Overall, 430 of the 700 measurements performed demonstrated no abnormal reflux indices or MII results. A total of 270 measurements were considered abnormal; 101 (37%) showed abnormal MII and pH values, 49 (18%) showed only pathological pH values, and 120 (45%) had only abnormal MII values (Table I and Figure; available at www .jpeds. com). The median age of the patients in this group was 2 years, 9 months.
Abnormal values were found in 114 of the 325 patients who presented with GI symptoms (35%), in 133 of the 329 patients with pulmonary symptoms (40%), and in 23 of the 46 patients with neurologic symptoms (50%). Median patient age and abnormal values in each subgroup are shown in Table I. Seventeen of the 133 patients with pulmonary symptoms and abnormal values also had additional GI symptoms. In this subgroup, 5 (29%) had a pathological MII-pH study, 4 (24%) had only a pathological pH study, and 8 (47%) had only a pathological MII study.
A retrospective subgroup analysis of the available data was performed in 511 patients. In this subgroup, 204 patients had normal MII-pH measurements, 84 had a pathological MII-pH study, 175 had only a pathological MII study, 48 had only a pathological pH study. Table II shows the criteria that led to a pathological MII classification in the first two of these groups. The median value of symptoms in all patients was 5 (range, 0-363; mean, 13 ± 25). In the patients with a positive SI, the median value of symptoms was 6 (range, 1-174; mean, 11 ± 16). Fifty-two patients with a positive SI had fewer than 4 symptoms during the measurement. In 19 of these 52 measurements, an additional parameter was abnormal (13 with pathological pH measurement, 4 with a high number of reflux episodes, and 2 with both). In 313 of the 511 patients in the subgroup analysis, SAP (defined as positive at ⩾95%) was calculated and compared with the SI (Table III). Moderate agreement between the two indices was seen (Cohen kappa, 0.54).
The 511 patients had a median of 33 RBMs (range, 0-213; mean, 42 ± 34). Five different age groups were created, and the median and mean numbers of RBM were compared among these groups. These numbers did not include ‘‘gas’’ episodes (retrograde air movement) and ‘‘pH-only’’ episodes (pH drop without RBM on MII) (Table IV; available at www. jpeds.com). The mean number of RBMs was significantly higher in the first age group (0-12 months) compared with the others (P <.001, one-way analysis of variance). No significant difference was seen in the other age groups. Using a smaller age window for the older age groups did not change these results.
The mean number of reflux episodes was compared in patients with a normal RI and those with a pathological pH study. Patients with normal RI had a mean of 39 ± 31 episodes, compared with 58 ± 43 episodes in those with a pathological pH study (P < .001, t test).
DiscussionOur data from a large, systematically standardized collection of MII-pH measurements in children indicate that 45% of the patients identified with abnormal GER would not have been recognized by 24-hour pH measurement alone. In our data collection we did not determine whether the abnormal MII findings were due to nonacidic (pH >7), weakly acidic (pH 4-7), or acidic (pH <4) reflux episodes. Comparing our 3 symptom groups, the mean ages suggest that younger patients tend to have more extraintestinal symptoms, whereas adolescent children more often have classical GER symptoms. Exceptions are found in each group.
In our data, it seems that especially extraintestinal symptoms (ie, pulmonary symptoms due to GER) are more often associated with a normal pH measurement. In the group with pulmonary symptoms, 58% of the abnormal findings were detected by MII only. The subgroup of 17 patients with pulmonary symptoms as well as GI symptoms more often had a pathologic pH study compared with those patients with pulmonary symptoms only (53% vs 42%).
These findings are in contrast to those of a previous MII-pH study in 25 patients with chronic pulmonary symptoms.4 In that study, the patients mainly had acidic RBMs (19:1 ratio of acidic to nonacidic). This might have been due to the relatively small cohort with a selected group of patients. In another study of 24 children with difficult-to-treat asthma, the proportions of acid and nonacid reflux were equal.5 A study of children with pulmonary problems being treated with proton pump inhibitors showed an association between respiratory symptoms and nonacidic reflux.6 Compared with these previous studies, our study has a much larger population and thus can be considered more representative.
The problem with this new MII method remains in data interpretation. Normal values for the number of RBMs in children detected by MII are not yet available. Based on our experience in analyzing our data, we considered an MII measurement abnormal if it showed a high number of RBMs or a positive symptom association as defined earlier.
Our subgroup analysis demonstrated that most patients with only pathological MII (74%) had mainly a pathological SI (Table II). However, 30% of those patients reported fewer than 4 symptoms during the measurement. Thus, this number may be too small to allow for reliable SI calculation. Patients with pathological MII and pH studies were more likely to have a high number of RBMs than a positive SI (50% vs 37%) (Table II).
Although our patient cohort does not represent a normal population, we compared the number of RBM episodes in the various age groups to look for an age-dependent difference. Infants are known to have a higher incidence of reflux episodes physiologically.7 As expected, we found a significantly higher number of RBMs in infants than in older children. This is in contrast to Francavilla et al,8 who found a higher mean number of weakly acidic RBMs in infants but no difference in the mean of all reflux episodes (total, 175; infants, 61). Loots et al9 also reported more RBMs in infants than in older children. We did not find a significant difference among the other age groups even when a smaller age window was used. These findings support our choice of a higher cutoff for RBM episodes in infants to avoid overestimation of abnormal measurements in this patient group.
Patients with an abnormal RI had a significantly higher number of RBMs than those with a normal RI, suggesting that a high number of RBMs may be an indicator of an abnormal MII measurement.
The arbitrary definition of symptom association with RBMs raises several issues.10 First, in most previous studies, the parameters and cutoffs for symptom analysis (SI, symptom sensitivity index, or SAP) were defined for pH analysis, but not for MII. Second, optimal time windows between symptom onset and RBM occurrence have not yet been validated. Third, symptom analysis is adaptable only to discontinuous symptoms and cannot be performed in a patient who was asymptomatic during the measurement. It demands precise recording of symptoms by the patient or by the parents of younger children. In our experience, not all patients and parents manage to do this. Sifrim et al11 described similar findings in adults, with only 39% of coughing episodes detected by manometry recorded by the patients. Thilmany et al4 noted that the patient protocols were not reliable in their study. An objective automatic symptom registration (eg, for cough or apnea) would be desirable, but is not available at the moment. All of the aforementioned problems with interpretation of symptom association apply for the current standard reflux test, pH probe analysis, as well.12
The software package provided with the MII-pH system uses a symptom association window of 5 minutes. The only validated time window is a 2-minute window between heartburn and acidic reflux/pH drop of <4 on pH measurement.12 A wider time window would increase the risk of a chance association being the reason why smaller time windows were chosen in this study. The optimum parameter for symptom association analysis seems to be the SAP based on the Fisher exact test. At the time of data collection, SAP calculation was not included in the available software. In 313 patients, the SAP was calculated retrospectively. In 257 measurements (82%), SI and SAP were equal leading to a moderate agreement between the two indices (SI and SAP) (Cohen kappa, 0.54). This is in contrast to the study of Lüthold et al,13 which found only a weak agreement between SI and SAP in 23 infants.
Compared with conventional pH measurement, MII-pH measurement provides more information about RBM episodes. In children with extraintestinal symptoms, an association of symptoms with nonacidic reflux is possible. Several authors have reported an increase in SI when using combined MII-pH analysis compared with pH analysis only.5,6,8 Other studies have shown that nonacidic and weakly acidic reflux episodes represent more than one-third of all reflux episodes,11,14,15 and even outnumber acid reflux episodes in infants.16,17
Conclusively evaluating the clinical relevance of these findings is difficult. There is some evidence, however; Tutuian et al18 described 6 patients on proton pump inhibitor therapy with persistent cough and a positive SI on MII who underwent fundoplication and were asymptomatic after surgery. Even if the clinical outcome does not change because of lack of sufficient therapy for every situation, it remains clinically important to know whether the reflux might be a reason for the presenting symptoms, to spare patients from further investigation and ineffective therapy.
Based on our data, we can presume that nonacidic or weakly acidic reflux plays an important role in GER. Our data confirm the superiority of MII-pH monitoring over pH monitoring alone in detecting a GER–symptom association. Consistent with the new ESPGHAN/North American Society for Pediatric Gastroenterology, Hepatology, and Nutrition guidelines,7 MII-pH should be the standard test for detecting GER in children of all age groups.
Назад в раздел
Популярно о болезнях ЖКТ читайте в разделе "Пациентам"
Информация на сайте www.gastroscan.ru предназначена для образовательных и научных целей. Условия использования.