Until recently, 24-hour pH monitoring had long been considered the gold standard for the diagnosis of gastroesophageal reflux disease (GERD) (1). This technique has significant limitations as concerns detection of nonacid reflux events (NAREs), likely involved in postprandial or atypical extraesophageal manifestations of GERD (2). In infants, in particular, nearly 90% of the REs have been reported to be NA because of the frequent neutralization of the gastric content following milk or formula feeding (3). Multichannel intraluminal impedance (MII) has recently been introduced as a pH-independent method to investigate the bolus transport in hollow organs. This technique allows detection of RE and definition of its chemical (acid, weakly acid [WA], and weakly alkaline) and physical composition (liquid, mixed, or gas) (4). MII also defines the proximal extent of the refluxate and the bolus presence time in the esophagus (bolus clearance time [BCT]). In adults, combined pH-MII monitoring has a sensitivity of at least 90% for the detection of all REs (4,5); few data are available in children (6,7). The temporal association between gastroesophageal reflux (GER) and symptoms has been commonly assessed by the symptom index (SI), the symptom sensitivity index (SSI), and the symptom association probability (SAP). SI and SSI have been validated by means of MII-pH monitoring in adults but no data exist for SAP in infants and children, separately (8–10).
Aims of the present study were to compare 24-hour MII-pH monitoring versus pH probe, as first-line assessment of GER in a large pediatric population and to define the relation between GER and types of symptoms (typical and atypical); moreover, for the first time, an attempt has been made to establish whether the results of the tests are influenced by the age of the children.
From June 2005 to April 2007, all consecutive patients referred to the Pediatric Unit of the University of Bari for suspected GERD were prospectively enrolled. Patients complaining of atypical symptoms (chronic cough, recurrent respiratory symptoms, asthma, hoarseness) were evaluated by a pediatric pulmonologist and/or ENT specialist; laryngoscopy, sweat, immunological, and allergy tests were performed in all of these individuals, and patients with underlying diseases were excluded. Apparent life-threatening events and previous treatment with medication for GERD (proton pump inhibitors, H2 blockers, motility medications) were considered exclusion criteria. The local ethics committee approved the study. Written consent was obtained from all of the parents. Patients were divided in 2 age groups: younger (infants) and older than 12 months (children). All of them underwent a 24-hour combined MII-pH recording (z-lab, Sandhill Scientific Inc, Highlands Ranch, CO). Two height-appropriate impedance catheters were used: infant (height <75 cm) and pediatric (height >75 cm). Each flexible polyethylene catheter had 7 impedance electrodes, representing 6 bipolar impedance channels, and a pH-sensitive electrode situated in the middle of the most distal impedance channel. A catheter was introduced transnasally through the esophagus, and pH sensor placement was confirmed by radiograph. Children received a regular diet, with a minimum of 3 hours between each meal; acid drinks were excluded. Patients were asked not to use antacid and/or antisecretory therapy 72 hours before and during the test. Parents filled in a diary recording times of meals, body position, and any symptom suggesting GER occurrence during the recording period.
Analysis of pH Data
The pH probe reports were generated by GERD analysis software (BioVIEW analysis, Sandhill Scientific). An acid pH-RE was defined by threshold pH of less than or equal to 4 for more than or equal to 5 seconds (all pH-RE) irrespective of the latency time between individual episodes. A reflux index >10.7% (infants) and 6% (children) was considered pathological (1).
Analysis of MII-pH Combined Data
Impedance data were automatically evaluated using the BioVIEW analysis software and each tracing was manually reviewed by 1 of the authors (A.M.M.).
A liquid RE was defined by MII when a fall in impedance more than or equal to 50% from baseline occurred in at least 2 consecutive channels in an aboral direction. A gas-only RE was defined as a simultaneous increase in impedance >3000 Ω in any 2 consecutive impedance sites, with 1 site having an absolute value >7000 Ω; a mixed RE met both the liquid and the gas criteria (11). On the basis of the lowest pH value REs were defined as acid (RE decreasing pH across 4, or occurring when esophageal pH was already acidic [all acid-MII-RE]), WA (RE resulting in pH between 4 and 7), or weakly alkaline (RE during which pH nadir did not drop below 7) (4). The total of WA and weakly alkaline RE was also defined as NARE. BCT was defined as the time elapsing from a drop in impedance to 50% of its baseline value, to its recovery (to 50%) in the most distal impedance channel. Refluxes were defined as proximal or full-column (reaching channels 1 and/or 2), intermediate (reaching channel 3 and/or 4) or distal (reaching channel 5). For each RE, its relation to feeding (postprandial time less than or equal to 2 hours) was reported. A pH-only episode was defined as a pH <4 for >5 seconds without a retrograde bolus detected by MII monitoring (pH-only RE). The ability of MII to detect an acid RE irrespective of its duration allowed the identification in the manual analysis of a subset of RE that would otherwise have passed undetected by pH-probe analysis software (acid-MII only RE). All of the data were analyzed for all of the patients and for the 2 age groups.
Diagnostic Accuracy in Detecting RE and Symptom Association
Because a gold standard is still lacking in evaluating the diagnostic accuracy of MII-pH and pH monitoring in detecting RE, we assumed that the sum of all acid and NA MII-RE plus pH-only RE represents the totality of all detectable RE (6), whereas the sum of all acid MII-RE plus pH-only RE represents the totality of all acid detectable RE. We defined the diagnostic accuracy of MII-pH in detecting RE as (all acid and NA MII-RE/all detectable RE), whereas the diagnostic accuracy of MII-pH and pH monitoring in detecting acid RE was defined as (all acid MII-RE/all detectable acid RE) and (all acid pH-RE/all detectable acid RE), respectively. Data have been compared between the 2 age groups. pH-only RE associated with swallows of acid contents were excluded (true false-positive RE).
A symptom was considered associated with an RE if it occurred within a 2-minute time window of its onset (12). The SI ([number of reflux-associated symptoms/total number of symptoms during 24 hours] × 100) and the SSI ([number of reflux-associated symptoms/total number of RE in 24 hours] × 100) were calculated for each patient on the basis of the pH probe alone and on the combined MII-pH results. SI and SSI were calculated for each individual symptom. Symptom index more than or equal to 50% (13) and SSI more than or equal to 10% were defined as pathological (14).
The SAP was calculated by dividing the 24-hour pH data into 2-minute segments. Each segment was studied whether reflux occurred within it and whether a symptom occurred during the segment. A 2 × 2 table was prepared in which the number of 2-minute segments with and without reflux and with and without symptoms were tabulated. The probability of the distribution of symptoms and RE in 2-minute windows was then calculated to determine whether it could have occurred by chance (χ2 test) (SAP = 1–p). The SAP was positive if >95% (15).
Normally distributed grouped data were expressed as the mean (±SD) and compared using the paired/unpaired t tests. Nonparametric grouped data are expressed as median (interquartile range [IQR], and 95th percentile) and compared with Mann-Whitney test (paired) or Wilcoxon test (unpaired). Spearman test was used for correlations. Proportionate data were compared with Fisher exact test or the χ2 test. A logistic regression model was used to explore the occurrence of a particular symptom with age of patient, chemical and physical composition, height of refluxate, and time elapsing since the last meal as independent variables given the detection of an RE. Significance was established at P < 0.05. The SPSS 13.0 program (SPSS Inc, Chicago, IL) was used.
Characteristics of the Study Sample
Of the total 291 patients referred for suspected GER in the study period, 109 were not included in the investigations for the presence of almost 1 exclusion criterion, 61 patients complaining of atypical symptoms with diagnosis of underlying diseases on the basis of investigations, 18 previous treatment with medication for GERD and 30 apparent life-threatening events; the remaining 182 patients were enrolled. Albeit, 7 patients were excluded because of artifacts/technical problem in MII-pH tracings, the total number available for final evaluation was, therefore, 175 (112 mol/L, median age 1.5 years [range 0.2 month–15.9 years]) (Fig. 1). Overall, 61 (35%) were younger than 12 months; 99 (57%) presented atypical symptoms, 32 (18%) typical, and 44 (25%) patients both symptoms (Table 1). The referred atypical symptoms were chronic cough in 51 patients, asthma in 44, recurrent respiratory symptoms in 42, and hoarseness in 6.
Analysis of pH Data
In the total recording period of 3858.8 hours (median 21.4 hours, range 20.3–26.3 hours), a total of 7510 REs were detected by pH monitoring (all pH-RE); of these, 2355 (31%) were not associated with a retrograde bolus movement (pH-only RE). The mean of all REs and of pH-only REs did not differ between the 2 age groups (Table 2). Of the total pH-only REs, 1201 (51%) were associated with pH fluctuation and 942 (40%) with acid swallows on the basis of MII data; in the remaining 212 cases (9%), REs were associated with a drop in impedance more than or equal to 50% from baseline occurring only at channel 6 and not reaching channel 5 required for impedance RE definition. A pathological RI was detected in 51 patients (29%): 10 (16%) infants and 41 (36%) children (P < 0.01).
Analysis of MII-pH Combined Data
In the 175 patients, 13 631 RE were recorded by MII-pH monitoring; 7361 were acid (54%) and 6270 REs were NA (6134 WA [45%] and 136 weakly alkaline [1%]). On the basis of impedance analysis, 6952 (51%) of the total REs were liquid, 3408 (25%) gas, and 3271 (24%) were mixed. Out of the total REs, 9951 (73%) were proximal, 2862 (21%) intermediate, and the remaining 818 (6%) were distal. REs were prevalently detected in the postprandial period (8860; 65%); of these, 4784 (54%) were acid, and 4076 were NA (WA 3987 [45%], weakly alkaline 89 [1%]).
Because MII identifies RE irrespective of duration, we revealed, by manual reading, a subset of acid RE <5 seconds (n = 1390; 19% of the total acid MII-RE) detected but then excluded by the pHmetry software (acid MII-only RE).
Infants Versus Children
The mean of all REs per recording did not differ according to age, but the mean of WARE was significantly higher in infants (P < 0.001) (Table 2). A negative correlation was observed between WARE and age (ρ = −0.41, P < 0.001) (see the online-only figure at http://links.lww.com/A1481). The physical composition of refluxate was similar in the 2 age groups, being prevalently liquid. The proximal extension of refluxate was prevalent in infants (P < 0.01); a negative correlation was found between percentage of full-column REs and age (ρ = −0.24, P < 0.02). Even if REs were mainly postprandial, in the 2 hours after meals the prevalence of WARE was significantly greater in infants (P < 0.001). Bolus clearance time was significantly higher in children (P < 0.03) and a positive correlation was found between BCT and age (ρ = 0.33, P < 0.003).
Diagnostic Accuracy in Detecting RE and Symptom Association
MII-pH and traditional pH monitoring did not differ in detecting acid RE (Table 3).
Infants Versus Children
The diagnostic accuracy of combined MII-pH in revealing all detectable RE was significantly higher in infants as compared with children (92% vs 82%, P < 0.01), similar to the detection of acid RE (83% vs 76%, P < 0.04).
Diagnostic Accuracy in Detecting Symptom Association
During the investigation, 149 patients reported a total of 3004 symptoms; 26 reported no symptoms. Therefore, 2607 symptoms were analyzed: 2197 respiratory (cough) and 410 typical (vomiting [n = 188], regurgitation [n = 137], heartburn [n = 85]). The mean number of symptoms per patient did not differ between the 2 age groups.
The addition of MII to conventional pH monitoring resulted in a higher yield of respiratory symptom association analysis, as revealed by a higher rate of RE association (11% vs 5% P < 0.001), mean SI (63% vs 25% P < 0.001), number of patients with positive SI (77 [69%] vs 21 [19%], P < 0.001), and mean SSI (15% vs 12%, P < 0.03). With regard to typical symptoms, comparing the yield of association between MII-pH and pH probe was found as higher rate of RE association (5% vs 2%; P < 0.001), mean SI (75% vs 44%, P < 0.001), and number of patients with a positive SI (57 vs 31, P < 0.001) (Table 4, Fig. 2). In calculating SAP, both for typical and atypical symptoms, the number of patients with a positive index was found to be significantly superior on the basis of MII-pH as compared with pH monitoring data (21 [20%] vs 9 [8%], P < 0.02). Analyzing single symptoms, as compared with traditional pH probe, MII-pH detected a higher SAP only for regurgitation (26% vs 63%, P < 0.01) and vomiting (17% vs 57%, P < 0.003).
Infants Versus Children
The higher diagnostic yield of symptom association analysis of combined MII-pH monitoring as compared with traditional pH probe was confirmed in the 2 age groups for atypical symptoms, whereas for typical ones it was maintained only in infants (Table 4).
The logistic regression analysis showed that composition (liquid RE) (P < 0.001) and proximal extension of refluxate (P < 0.001) were correlated with an increased association between RE and atypical symptoms.
Data emerging from the present investigation clearly suggest that the addition of MII to conventional pH monitoring significantly increases the diagnostic accuracy of the test as first-line assessment in infants or children with suspected GERD. Indeed, this new method has a higher accuracy in revealing REs and allows their association with symptoms, increasing at least 2-fold the SI as compared with traditional pH monitoring. This observation is of particular importance in infants and in subjects with atypical symptoms.
This study represents the first attempt, on a large scale, to evaluate MII-pH monitoring for the assessment of children, at presentation, naïve for antisecretory therapy; until now, children investigated by combined MII-pH recording represented, prevalently, a selected population having failed conventional medical therapy (8–10). Our data show that the use of MII-pH, as first choice, may improve the management of patients, allowing detection of a higher rate of association between REs and symptoms, decrease the number of children undergoing second-line investigations, and define final diagnosis and even therapeutic approach.
Esophageal pH monitoring has several limitations, mainly related to inability to detect NA refluxate, underestimating the amount of REs. This limit is particularly the case in the postprandial period and in infants, in whom this technique is blind for half the recording time (16). Our data clearly show the existence of differences in chemical composition rather than in frequency of REs according to age. Indeed, in our series the mean REs per recording does not vary between infants and children and is similar to that reported in preterm neonates (17,18). However, the prevalence of NARE varies considerably according to age, being about 40% of the total RE in children and 60% in infants. The majority of REs occur in the postprandial period and these are probably NA (9,10); these are significantly higher in infants than in children, likely due to the frequent and prolonged gastric buffering after milk intake in infants.
In our experience, REs are prevalently liquid and reach the most proximal channels especially in infants. These findings probably depend on the shorter esophagus and on the positive association between the reflux level and the gastric filling state, which, as recently reported, is higher in infants (19).
MII monitoring has recently been described as the only method offering high sensitivity for the detection of all types of REs (4). Until now, different criteria to assess the diagnostic accuracy of this technique have been used, the major issue being the inclusion of pH-only REs, the clinical significance of which is still debated. The reported prevalence of pH-only REs varies between 2% and 70% (5,7,20,21). These REs are largely secondary to acid swallows or pH fluctuation, not associated with a bolus movement in the esophagus and likely because of body movements, respiration, or electrode drift. It is, however, possible that some of these represent REs restricted to the distal esophagus. The existence of these short segment refluxes, identified only by manual reading, is confirmed by our finding that approximately 10% of the total pH-only REs were associated with a drop in impedance more than or equal to 50% from baseline occurring only at the most distal channel. In our experience, the diagnostic yield of combined MII-pH in detecting acid REs is not superior to the traditional pH probe; indeed, both techniques reveal a number of acid REs not seen by the other techniques (pH-only REs for pH probe and acid MII-only REs for MII-pH).
In clinical practice, one of the most used applications of pH monitoring is to prove the association between a symptom and an RE by the calculation of SI, SSI, and SAP. In the present report, we confirm that MII-pH significantly increases the detection of an association between symptoms and REs in a pediatric population. Other authors have reported that MII-pH increases 6-fold the number of children with a positive SI (8), as also found in the present study, although to a lesser extent; this may depend on a different selection of patients because we have studied naïve as compared with antisecretory unresponsive patients, thus more prone to having NARE. The advantages of MII-pH monitoring in the study of symptom association is maintained in the 2 age groups only for respiratory symptoms, whereas MII-pH is superior for typical symptoms only in infants.
The logistic regression analysis shows that having a liquid or a full-column RE significantly increases the likelihood of experiencing cough, supporting the hypothesis that a direct stimulation of airway structures is involved in the genesis of RE-associated cough.
Several aspects of this technique may still limit its clinical application. Indeed, no normal impedance data are available in children, and the technique is expensive and time-consuming, needing well-trained personnel. A therapeutic trial treating acid/nonacid RE is mandatory to validate the association of RE symptoms.
Our study shows that combined MII-pH monitoring represents a powerful first-line test for the assessment of GERD. The addition of impedance to conventional pH monitoring significantly increase the diagnostic yield in detecting REs, prevalently in infants, and in revealing an association between refluxes and symptoms, prevalently respiratory ones and in the infants group.
The authors thank Marian Shields for help with preparation of manuscript and Ms N Iavernaro for expert technical collaboration.
1. Rudolph CD, Mazur LJ, Liptak GS, et al
. North American Society for Pediatric Gastroenterology and Nutrition. Guidelines for evaluation and treatment of gastroesophageal reflux
in infants and children: recommendations of the North American Society for Pediatric Gastroenterology and Nutrition. J Pediatr Gastroenterol Nutr 2001; 32(suppl 2):S1–S31.
2. Rudolph CD. Supraesophageal complications of gastroesophageal reflux
in children: challenges in diagnosis and treatment. Am J Med 2003; 115(suppl 3A):150S–156S.
3. Wenzl TG, Silny J, Schenke S, et al
. Gastroesophageal reflux
and respiratory phenomena in infants: status of the intraluminal impedance technique. J Pediatr Gastroenterol Nutr 1999; 28:423–428.
4. Sifrim D, Castell D, Dent J, et al
. Gastro-oesophageal reflux monitoring: review and consensus report on detection and definitions of acid, NA, and gas reflux. Gut 2004; 53:1024–1031.
5. Shay S, Richter J. Direct comparison of impedance, manometry, and pH-probe in detecting reflux before and after meal. Dig Dis Sci 2005; 50:1584–1590.
6. Rosen R, Lord C, Nurko S. The sensitivity of Multichannel Intraluminal Impedance
and the pH-probe in the evaluation of gastroesophageal reflux
in children. Clinical Gastroenterol Hepatol 2006; 4:167–172.
7. Wenzl TG, Moroder C, Trachterna M, et al
. Esophageal pH monitoring and impedance measurement: a comparison of two diagnostic tests for gastroesophageal reflux
. J Pediatr Gastroenterol Nutr 2002; 34:519–523.
8. Rosen R, Nurko S. The importance of multichannel intraluminal impedance
in the evaluation of children with persistent respiratory symptoms. Am J Gastroenterol 2004; 99:2452–2458.
9. Condino AA, Sondheimer J, Pan Z, et al
. Evaluation of infantile acid and non acid gastroesophageal reflux
using combined pH monitoring and impedance measurement. J Pediatr Gastroenterol Nutr 2006; 42:16–21.
10. Condino AA, Sondheimer J, Pan Z, et al
. Evaluation of gastroesophageal reflux
in pediatric patients with asthma using impedance-pH monitoring. J Pediatr 2006; 149:216–219.
11. Shay S, Tutuian R, Sifrim D, et al
. Twenty-four hour ambulatory simultaneous impedance and pH monitoring: a multicenter report of normal values from 60 healthy volunteers. Am J Gastroenterol 2004; 99:1037–1043.
12. Lam HG, Breumelhof R, Roelofs JM, et al
. What is the optimal time window in symptom analysis of 24-hour esophageal pressure and pH data? Dig Dis Sci 1994; 39:402–409.
13. Wiener GJ, Richter JE, Copper JB, et al
. The symptom index
: a clinically important parameter of ambulatory 24-hour esophageal pH monitoring. Am J Gastroenterol 1988; 83:358–361.
14. Breumelhof R, Smout AJ. The symptom sensitivity index
: a valuable additional parameter in 24-hour esophageal pH recording. Am J Gastroenterol 1991; 86:160–164.
15. Weusten BL, Roelofs JM, Akkermans LM, et al
. The symptom-association probability: an improved method for symptom analysis of 24-hour esophageal pH data. Gastroenterology 1994; 107:1741–1745.
16. Washington N, Spensley PJ, Smith CA, et al
. Dual pH probe monitoring versus single pH probe monitoring in infants on milk feeds: the impact on diagnosis. Arch Dis Child 1999; 81:309–312.
17. Lopez-Alonso M, Moya MJ, Cabo JA, et al
. Twenty-four-hour esophageal impedance-pH monitoring in healthy preterm neonates: rate and characteristics of acid, weakly acidic, and weakly alkaline gastroesophageal reflux
. Pediatrics 2006; 118:e299–e308.
18. Magistà AM, Indrio F, Baldassarre M, et al
. Multichannel intraluminal impedance
to detect relationship between gastroesophageal reflux
and apnoea of prematurity. Dig Liv Dis 2007; 39:216–221.
19. Cresi F, de Sanctis L, Savino F, et al
. Relationship between gastro-oesophageal reflux and gastric activity in newborns assessed by combined intraluminal impedance, pH metry and epigastric impedance. Neurogastroenterol Motil 2006; 18:361–368.
20. Hila A, Agrawal A, Castell OD. Combined multichannel intraluminal impedance
and pH esophageal testing compared to pH alone for diagnosing both acid and weakly acid gastroesophageal reflux
. Clin Gastroenterol Hepatol 2007; 5:172–177.
21. Woodley FW, Mousa H. Acid gastroesophageal reports in infants: a comparison of esophageal pH monitoring and multichannel intraluminal impedance
measurements. Dig Dis Sci 2006; 51:1910–1916.