Dig Dis Sci (2014) 59:623–630 DOI 10.1007/s10620-013-2950-0
ORIGINAL ARTICLE
Children with Cystic Fibrosis Have Prolonged Chemical Clearance of Acid Reflux Compared to Symptomatic Children Without Cystic Fibrosis Frederick W. Woodley • Rodrigo S. Machado • Don Hayes Jr. Carlo Di Lorenzo • Ajay Kaul • Beth Skaggs • Karen McCoy • Alpa Patel • Hayat Mousa
•
Received: 9 September 2013 / Accepted: 11 November 2013 / Published online: 28 November 2013 Ó Springer Science+Business Media New York 2013
Abstract Objectives Few studies compare gastroesophageal reflux (GER) parameters of cystic fibrosis (CF) children and symptomatic non-CF children. We aimed to compare the impedance-pH (IMP-pH) parameters for these two groups and to test the hypothesis that prolonged acid exposure in CF patients is due to delayed chemical clearance (CC). Methods IMP-pH tracings from 16 CF children (median 8.2 years) and 16 symptomatic non-CF children (median 8.3 years) were analyzed. Software was used to generate IMP-pH reports and parameter data were extracted. IMPpH was used to calculate the mean CC for each patient.
F. W. Woodley (&) C. Di Lorenzo B. Skaggs H. Mousa Division of Gastroenterology, Hepatology, and Nutrition, Nationwide Children’s Hospital, Columbus, OH, USA e-mail: [email protected] F. W. Woodley D. Hayes Jr. C. Di Lorenzo K. McCoy A. Patel H. Mousa Department of Pediatrics, The Ohio State University College of Medicine, Columbus, OH 43205, USA
Results pH studies showed no difference in acid GER (AGER) frequency (p = 0.587); however, mean AGER duration, duration of longest AGER, AGER index, and DeMeester scores were all significantly higher for CF patients. IMP showed no difference in GER frequency [neither acidic (p = 0.918) nor non-acidic (p = 0.277)], but total bolus clearance was more efficient in CF patients (p = 0.049). A larger percentage of total GER reached the proximal esophagus in non-CF children (p = 0.039). Analyses of two-phase AGER episodes showed that these events were more acidic (p = 0.003) and the CC phase was significantly prolonged in the CF cohort (p = 0.001). Conclusions Compared to symptomatic non-CF children, CF children do not have more frequent reflux. Actually, they have better bolus clearance efficiency following reflux and may even have better control over the number of GER episodes that reach the proximal esophagus. CC of AGER, however, is significantly prolonged in the CF cohort, likely due to hyperacidity of refluxed gastric contents. Keywords Cystic fibrosis Acid reflux Chemical clearance Volume clearance Proximal extent
R. S. Machado Universidade Federal de Sa˜o Paulo, Sa˜o Paulo, Brazil D. Hayes Jr. K. McCoy A. Patel Pulmonary Medicine, Nationwide Children’s Hospital, Columbus, OH, USA A. Kaul Division of Gastroenterology, Hepatology, and Nutrition, Cincinnati Children’s Hospital Medical Center, Cincinnati, OH, USA A. Kaul Department of Pediatrics, University of Cincinnati College of Medicine, Cincinnati, OH, USA
Introduction Gastroesophageal reflux (GER) is common among patients with cystic fibrosis (CF). The prevalence of acid GER (AGER) in CF patients, in particular, has been estimated at 35–81 % [1–6]. Few studies compare CF children to nonCF children who have similar symptoms of GER. Despite reports of increased AGER indices in CF patients [1], there have been no studies determining whether this increased acid exposure is due to delays in volume clearance (VC)
123
624
Dig Dis Sci (2014) 59:623–630
pH 4
VC
CC
Fig. 1 IMP-pH tracing illustrating a two-phase acid reflux episode. VC volume clearance, CC chemical clearance
alone, chemical clearance (CC) alone, or to both. VC is the first phase of AGER clearance during which the bulk of refluxed gastric contents is cleared from the esophagus (back into the stomach) by esophageal peristalsis, and CC is the second phase of AGER clearance during which acidified mucosa is neutralized by bicarbonate-rich saliva [7] and bicarbonate ions released directly into the esophageal lumen from submucosal glands [8–10]. Increased acid exposure, as quantified by the AGER index, could result from frequent AGER, prolonged VC, or prolonged CC. We previously showed that combined multichannel intraluminal impedance and esophageal pH monitoring (IMP-pH) can be used to distinguish the VC phase of twophase AGER episodes from the CC phase [11, 12]. Synchronization of impedance and pH measurements permits a clear distinction between the dynamics of intraluminal flow (impedance) and neutralization of the acidified esophageal mucosa (pH probe) (Fig. 1). This investigation compared CF children and symptomatic non-CF children on the basis of the IMP-pH parameters that are commonly used to assess patients suspected of having GER. We also tested the hypothesis that prolonged acid exposure in children with CF is due to delayed CC.
123
Patients and Methods Patients The IMP-pH databases at Nationwide Children’s Hospital (NCH) and Cincinnati Children’s Hospital Medical Center (CCHMC) were searched for CF patients who were not taking anti-reflux medications and who had not received a fundoplication prior to testing. To be included in the study, patients had to have been off anti-reflux medications such as proton pump inhibitors, histamine-receptor antagonists, and prokinetics for at least 7, 5, and 2 days, respectively, prior to each study. A non-CF age-matched patient was selected from the NCH database each time a qualified CF patient was identified. The selection process (January 2002 to March 2009) involved choosing the first subject from the list of names/ages of previously studied patients (in the order of newer to older tracings) who satisfied the inclusion criteria and who was closest in age to the paired CF patient. This study was approved by the NCH and CCHMC Institutional Review Boards. Catheter and Placement IMP-pH catheters and their placement have been described previously [13]. Age-appropriate catheters included the
Dig Dis Sci (2014) 59:623–630
pediatric (age 2–10 years) and the adolescent ([10 years) catheters (Sandhill Scientific, Highlands Ranch, Colorado). Analyses of IMP-pH Tracings
625
AGER indices B50 % of the upper end of values considered normal by pH probe (i.e. B3 %)[13], did not have positive temporal associations of GER with symptoms, were not taking any anti-reflux medications, and had not received a fundoplication prior to testing.
pH Probe Proximal Extent of GER Acid gastroesophageal reflux (AGER) episodes were characterized by a drop in distal esophageal pH to \4. The IMP-pH reports were generated using the software provided by Sandhill Scientific, with pH probe variables that included the number of AGER, the mean AGER duration, the longest AGER duration, the DeMeester score, and the AGER index. Only events lasting 5 s or longer were registered and analyzed. Acid exposure was considered increased when the AGER index was [6 % [13]. Impedance Impedance-detectable events are defined by a drop of intraluminal impedance to below 50 % of baseline values beginning in the distal-most impedance channel (channel 6) and propagating proximally into at least two impedance channels. Only those impedance events lasting 5 s or longer were registered. GER events are defined as AGER episodes when the pH falls to\pH 4 at any point during the impedance-detectable episode. Non-acid GER events are impedance-detectable events during which the pH is B4; these episodes included both weakly-acid GER and weakly-alkaline GER [14]. For both GER detection tools (impedance and pH probe), GER episodes occurring during feeding were excluded from analysis. IMP-pH tracings were analyzed by a single author (FW).
For each patient, the number (and percentage of total) of each impedance-detectable GER event (AGER and NAGER) that reached the proximal esophagus was determined. GER that reached one or both of the two proximalmost impedance channels (channels 1 and 2) were tagged as proximal reflux episodes. Temporal Association of Symptoms with GER For each patient, an IMP-pH report was generated and printed. Using a 2-min window, the symptom association probability (SAP) was calculated for all measured symptoms. A SAP was considered positive when it equaled [95 %. We compared CF and non-CF symptomatic children based on the presence of a positive SAP, either by pH or by impedance. CF and non-CF groups were compared by chi-squared analysis. Statistical Analysis The Wilcoxon test was used to assess the difference for each parameter between the two groups (CF children and non-CF age-matched children). Frequency of pathological GER [17] in cases and controls was compared with the McNemar test. Data are presented as median values (25–75th percentile) unless otherwise indicated. Statistical significance was established at p \ 0.05.
Assessing Chemical Clearance Two-phase AGER episodes (Fig. 1) [15], detected using combined IMP-pH, serve as the ideal model system [11] for assessing the efficiency of acid neutralization. During two-phase AGER episodes, CC begins promptly at the point at which VC (also known as bolus contact time, BCT) is completed. For each patient, the IMP-pH tracing was manually screened for two-phase AGER episodes. For each episode, CC duration was calculated as previously described [11]. A mean was then calculated for each patient and a median (25–75th quartiles) was calculated for each group (CF children and non-CF age-matched children). CC duration for CF and non-CF patients in this study was compared to 50 ‘‘normal’’ children [M/F 36/14, median 7.0 years (range 2.8–11.0 years)] who were previously described in abstract form [16]. These control children had
Results Clinical Results We enrolled 16 CF patients, aged 3.0–17.7 years. Three of the 16 patients were tested at CCHMC. Sixteen age-matched non-CF patients were chosen from a registry of children who were tested at NCH. All patients were off anti-reflux medications and had not received a fundoplication prior to IMP-pH testing. Each child was referred for symptoms suggestive of GER as an etiology. Across all participants, median study length was 23.8 h (range 22.5–26.5 h) for CF patients and 23.9 h (range 21.3–27.6 h) for non-CF patients. Patient demographics are shown in Table 1.
123
626
Dig Dis Sci (2014) 59:623–630
Table 1 Patient demographics
Table 3 Characteristics of GER by IMP-pH
Demographic
CF (N = 16)
Non-CF (N = 16)
Impedance-pH
CF
Non-CF
p value
Age, median (range)
8.2 years (3.1–17.7 years)
8.3 years (3.0–17.2 years)
#AGER
42.0 (36.0–60.7)
47.5 (25.7–69.5)
0.918
#NAGER
4.0 (2.0–11.2)
6.0 (2.2–14.0)
0.277
12.0 (9.2–21.0)
19.2 (13.3–37.3)
0.088
Sex, M/F
6/10
9/7
#GER
Race, CA/AA
16/0
14/2
%AGER
1.3 (0.7–1.7)
1.7 (0.8–2.4)
0.272
Genotype-DF508(2)/ DF508(l)/other
10/4/2
N/A
%NAGER
0.1 (0.02–0.2)
0.2 (0.1–0.5)
0.157
%BCT
1.5 (0.9–1.9)
2.1 (0.9–2.8)
0.187
Median BCT (s)
12.0 (11.0–13.7)
15.5 (11.7–18.0)
0.049*
Symptoms presented Cough
100 %
75 %
Vomiting
0%
25 %
Regurgitation
12 %
6%
AGER acid gastroesophageal reflux, NAGER non-acid gastroesophageal reflux, GER gastroesophageal reflux, BCT bolus contact time (also known as volume clearance), CF cystic fibrosis
Heartburn
25 %
12 %
Values are expressed as medians (25–75th interquartile)
Gasping
0%
6%
* Significance at p \ 0.05
Night awakenings
6%
6%
Retching
0%
6%
Choking
6%
12 %
Throat clearing
0%
12 %
Abdominal pain
6%
12 %
Decline in PFT
12 %
0%
Wheezing
12 %
6%
CA Caucasian, AA African American, CF cystic fibrosis, PFT pulmonary function tests
Table 2 Characteristics of GER by pH probe
symptoms that could be positively associated with AGER (SAP[95 %), three patients had symptoms associated with both AGER and NAGER (23 %), and one patient (\1 %) did not have any symptoms that could be associated with GER. For the non-CF children, four of the ten children (40 %) with pathological GER had symptoms that could be positively associated with AGER, one patient (10 %) had symptoms associated with both AGER and NAGER, and four patients (40 %) did not have any symptoms that could be associated with GER.
pH probe
CF
Non-CF
p value
Number of AGER episodes
67.5 (50.7–83.5)
54.5 (38.2–120.5)
0.587
Impedance
Mean AGER duration (s)
129.0 (77.2–197.5)
86.5 (37.5–108.5)
0.017*
Longest AGER duration (min)
29.8 (19.0–39.7)
9.9 (4.0–14.6)
0.004*
DeMeester score
38.9 (24.3–60.7)
23.7 (11.2–37.9)
0.006*
AGER index (%) Pathological AGER indicesa
11.3 (6.9–18.2) 13/16 (81 %)
5.6 (2.6–9.7) 10/16 (62 %)
0.003* 0.125
From each IMP-pH report, we extracted the number of episodes, total time, and percent times for AGER, NAGER, and total GER (Table 3). As indicated in Table 3, only the median BCT was different between groups; the median BCT was significantly higher in non-CF patients (p = 0.049). The number of each event type that reached the proximal esophagus was also extracted directly from the IMP-pH report. The percentages of total AGER, NAGER and GER reaching the proximal esophagus were calculated (Table 4). As indicated in Table 4, the only significant difference between groups was the percent total proximal GER, which was higher in non-CF patients.
Values are expressed as medians (interquartile range), unless otherwise indicated AGER acid gastroesophageal reflux, CF cystic fibrosis * Significance at p \ 0.05 AGER indices [5 %
a
pH Probe With the exception of AGER frequency (p = 0.587), all pH probe parameters were significantly greater for the CF patients (p \ 0.05, Table 2). Thirteen of 16 CF children (81 %) had pathological GER (i.e. AGER Index [5 %) [17] compared to ten of 16 non-CF children (62 %). Nine of 13 CF children (69 %) with pathological GER had
123
IMP-pH to Detect Two-Phase AGER Episodes Each IMP-pH tracing was manually inspected for twophase AGER episodes. While there was no difference in the frequency of two-phase AGER (p = 0.979), and the more efficient VC for the CF children (25.7 vs 30.3 s) failed to achieve statistical significance (p = 0.057), CC duration was almost twofold less efficient in the CF children (122.8 vs 65.9 s) (p = 0.001) (Table 5). CC duration data distribution was positively skewed. Following
Dig Dis Sci (2014) 59:623–630
627
Discussion
Table 4 Proximal extent of GER Impedance-pH
CF
Non-CF
p value
#Proximal AGER
22.5 (11.5–35.0)
35.0 (14.0–57.0)
0.205
%Total proximal AGER
57.9 (34.2–72.1)
80.2 (59.5–86.5)
0.056
#Proximal NAGER %Total proximal NAGER
1.0 (0–4.5) 29.1 (0–58.6)
3.0 (1.2–8.5) 62.1 (19.0–91.7)
0.151 0.061
#Proximal GER
27.0 (14.0–38.2)
38.5 (21.5–73.5)
0.105
%Total proximal GER
55.7 (33.7–72.4)
78.8 (61.2–87.4)
0.039*
AGER acid gastroesophageal reflux, NAGER non-acid gastroesophageal reflux, GER gastroesophageal reflux, CF cystic fibrosis Values are expressed as medians (25–75th interquartile) * Significance at p \ 0.05
Table 5 Impedance-pH parameters for two-phase AGER episodes Impedance-pH
CF
Non-CF
p value
VC (s)
25.7 (18.5–28.7)
30.3 (25.9–34.2)
0.057
CC (s)
122.8 (96.5–197.7)
65.9 (30.0–101.4)
0.001*
Total duration (s)
159.0 (120.2–234.9)
95.0 (65.3–134.8)
0.002*
pH nadir
1.1 (0.8–1.6)
1.8 (1.4–2.1)
0.003*
Number of episodes
23.2 (16.8–35.7)
24.7 (14.6–36.2)
0.979
AGER acid gastroesophageal reflux, VC volume clearance, CC chemical clearance, CF cystic fibrosis Values are expressed as medians (25–75 % interquartile) * Significance at p \ 0.05
logarithmic transformation, CC durations of CF and nonCF children were compared to a previously reported set of 50 children with no evidence of GER-associated symptoms [16]. CF children presented a significantly longer CC duration compared to normal controls [Log10 CC duration—mean difference 0.553 (95 % CI 0.410–0.697), p \ 0.010]. The median crude CC duration was 122.8 s (IQR 96.5–197.7 s) versus 37.8 s in controls (IQR 29.2–52.6 s). Symptomatic non-CF children also presented a significantly longer CC duration compared to normal controls [Log10 CC duration—mean difference 0.209 (95 % CI -0.0370 to 0.381), p = 0.036], with the crude median CC duration for non-CF children and controls, respectively, 65.9 s (IQR 30.0–101.4 s) and 37.8 s (IQR 29.2–52.6 s). The mean nadir pH of two-phase AGER episodes was significantly more acidic in the CF cohort than in the non-CF cohort [median pH 1.1 (IQR 0.8–1.6) vs median pH 1.8 (IQR 1.4–2.1), p = .003] (Table 5).
This study found that, when compared to other patients referred for symptoms suggestive of GER, CF patients have increased AGER that is not due to more frequent episodes of AGER or to delayed bolus clearance but rather to the prolonged CC of AGER. It was reported several years ago (using pH monitoring) that children with CF have increased AGER when compared to healthy children [1–6]; however, the pH probe monitoring used in these studies was unable to distinguish bolus clearance from CC. Our data show that compared to non-CF patients who presented with symptoms similar to those of our CF cohort, CC lasts twice as long in CF children. VC efficiency, on the other hand, was as good as (or better) in CF children than in the non-CF age-matched children. Recently, in a study of 24 CF children (median 4 years [range 0.3–13 years]), Blondeau et al. [6] reported a median AGER index of 3.5 %, a median volume exposure of 1.3 %, a median total number of GER episodes (per patient) of 41, a median proximal extent of 25 %, and an AGER:NAGER ratio of 70:30. In the present study of 16 CF children (median 8.2 years [range 3.1–17.7 years]), we report a median AGER index of 11.3 %, a median volume exposure of 1.5 %, a median total number of GER episodes (per patient) of 47, a median proximal extent of 56 %, and an AGER:NAGER ratio of 95:5. Differences in AGER index and AGER:NAGER ratio can be attributed to differences in age; Blondeau et al. included infants in their study and formula-fed infants experience more frequent NAGER episodes (and fewer AGER) than older children. Differences in genotype may also be important; the genotype ratio [DF508(homozygote):DF508 (heterozygote):other] was 10:4:2 for the present cohort and 5:8:11 for the Blondeau et al. cohort. In the present study, more than 62 % of the CF cohort were homozygous for the DF508 genotype compared to approximately 21 % in the Blondeau cohort. The difference in proximal extent between the two cohorts may be explained by the fact that we defined proximal events as those GER reaching either one of the two proximal impedance channels (1 or 2), while Blondeau and colleagues defined proximal events as those GER reaching only channel 1. VC, also known as bolus clearance or bolus contact time (BCT), is the process by which refluxed material is extruded back into the stomach by swallow-induced primary peristalsis and/or mechano-receptor-induced secondary peristalsis. CC is the process by which the acidified esophageal mucosa is neutralized following passage of the bolus of an acid reflux episode. Neutralization is subsequently accomplished when bicarbonate-rich saliva is transported to the distal esophagus by primary and secondary peristalsis and also by intrinsic esophageal
123
628
bicarbonate that is secreted directly into the esophageal lumen by submucosal glands [18–22]. In addition to bicarbonate, saliva also contains proteins and phosphates that also have buffering potential. One possible explanation for the prolonged CC in CF children is a decreased buffering capacity of the saliva. Reduced bicarbonate secretion into CF saliva, for example, would reduce the efficiency with which CF patients could neutralize the esophageal mucosa following an AGER episode. In situ hybridization studies show that cystic fibrosis transmembrane conductance regulator (CFTR) is expressed in salivary gland ductal cells [23]; it is within these ducts that, with the functional wild-type CFTR gene, sodium is reabsorbed and potassium and bicarbonate ion are secreted. More than five decades ago, in a study aimed at comparing bicarbonate levels in saliva between CF and non-CF controls, no significant differences in bicarbonate levels were detected. Using a titration method [24] to quantify bicarbonate levels in parotid salivary secretions, Chauncey et al. [25] measured 15.4 ± 6.0 mEq/L [range 6.9–32.8] of bicarbonate in CF patients and 13.4 ± 3.6 mEq/L [range 7.7–33.5] (p = 0.40) in the nonCF patients. However, whole saliva is composed of secretions from three major glands (parotid, submandibular, and sublingual) as well as from several minor glands [26]. While the parotid gland is by far the largest anatomically, parotid contribution to total saliva secretion is only about 21 %, compared to approximately 70 % from the submandibular gland [27]. Therefore, equivalent levels of bicarbonate within the parotid portion of whole saliva between CF and non-CF patients may not significantly influence the overall bicarbonate content in whole saliva. Davies et al. [28] later compared the regulation of electrolyte secretion in submandibular saliva in CF and control subjects. Using a colorimetric system (Technicon SMAC II system), they found decreased bicarbonate levels in CF resting saliva and when the submandibular gland was stimulated using vitamin C, bicarbonate levels increased in control subjects but not in CF subjects. Recently, Aps et al. [29] compared the electrolyte levels in whole saliva among CF homozygotes, CF heterozygotes, and non-CF subjects. Bicarbonate levels were determined enzymatically using a commercially available test kit (BioMerieux, Marcey-lesEtoiles, France). Unfortunately, these authors excluded calcium and bicarbonate levels from their published tables, since there was no difference in calcium or bicarbonate levels between the three groups. Most recently, in a study of 80 CF patients [median age 12.0 years (range 4.0–34.0 years)] and 84 healthy volunteers [median age 17.0 years (range 5.0–28.0 years)], Goncalves et al. [30] found no significant difference in bicarbonate levels between groups (p = 0.459). Considering the influence of diet and different types of salivary stimulation on saliva
123
Dig Dis Sci (2014) 59:623–630
composition [27], combined with possible circadian rhythm or diurnal influences [31, 32], further studies involving carefully controlled experimental designs are needed to assess potential differences in the buffering capacity of whole saliva produced by patients with and without CF. Saliva transport efficiency may be compromised in CF patients; reduced salivary flow rates and high viscosity have been reported [32, 33]. Further, esophageal hypomotility in CF could also reduce efficient transport of the viscous saliva through the esophageal lumen [34]. Since VC efficiency in our CF cohort was equal to or better than VC in the non-CF patients, hypomotility is not the likely explanation for our CF population. Alternatively, differences in the volume of saliva may have an effect; Goncalves et al. [30] found that saliva volume is significantly reduced in CF patients. Saliva volume was not analyzed in this study. Reduced secretion of bicarbonate into the esophageal lumen by submucosal glands may also be responsible for reduced CC efficiency. Using immunofluorescence, RTPCR, and in situ hybridization in a porcine model, Abdulnour-Nakhoul et al. [22] demonstrated expression of the CFTR in submucosal gland duct cells and localized CFTR and SLC26A6 on the luminal membrane. In other tissues (including the pancreas), it has been reported that the SLC26 family of anion exchangers interacts with CFTR to regulate secretion of bicarbonate [35, 36]. Expression of CFTR in the esophageal submucosal glands of humans has not yet been demonstrated. Reduced CC efficiency in CF patients could also be related to differences in gastric acidity, as larger numbers of hydrogen ions refluxed onto the esophageal mucosa would be expected to require longer periods of time to be neutralized. Gastric hypersecretion is common among CF patients [37, 38]. Consistent with these reports, in the present study, two-phase AGER events were significantly lower in CF patients. We present a novel method for assessing CC efficiency in patients evaluated by combined IMP-pH. Significant delays in CC could be a finding in numerous established conditions that include CF, scleroderma [39], esophageal hypomotility [40], defects in saliva production and secretion as in Sjogren’s syndrome [41], and progressing neurological disorders [42, 43]. Of the four AGER types detectable by IMP-pH (two-phase, single-phase, pH-only events, and re-reflux AGER episodes)[12], only the twophase episode is suitable for CC assessment because the start and stop time points of CC are discretely demarcated; CC begins when VC concludes and ends when the pH reaches 4. Assessing CC of spontaneous two-phase AGER episodes detected during IMP-pH testing is less invasive than other classic methods [15, 40] involving the use of
Dig Dis Sci (2014) 59:623–630
radioactive hydrochloric acid instilled into the esophageal lumen. In the present study, we simply identified and analyzed all two-phase AGER episodes and then calculated a mean duration of CC for each patient. Further studies involving larger sample sizes are needed to determine the minimum number of two-phase AGER episodes that would be needed to derive a valid metric to compare to reference values for CC in infants, children, or adults. The most surprising finding of this study was the fact that fewer GER reached the proximal esophagus in children with CF. While it is tempting to speculate that CF children might have an innate reflexive response to prevent aspiration, it is equally possible that gastric contents in CF patients become thicker due to constantly swallowed mucous, making it less likely to go against gravity to the proximal esophagus. Also, it is possible that the differences in proximal extent are unique to the present study and cannot be extrapolated to other patient populations. Other compensatory mechanisms in CF patients have also been suggested [44–46]. This study has several limitations. First, the size of the cohort is not ideal. We were limited by the number of CF patients (B18 years) who were off anti-reflux medications and who had not received a fundoplication prior to their IMP-pH study. Second, our subjects were not studied in a sleep lab where sleep status, symptoms, and precise body positioning could have been more precisely monitored. Third, we realize that our selection of non-CF control patients could have influenced our results. Finally, analyses of IMP-pH tracings were not blinded to patient genotype (CF vs non-CF). In conclusion, while the average CF child has an AGER index that is significantly higher than other children presenting with similar symptoms, he or she does not generally experience more frequent AGER events than other GERD patients. However, the CF child has significantly delayed CC that is likely due to hyperacidity of refluxed gastric contents. Currently, there are no anti-reflux drugs that specifically target CC. In order to treat CF patients who have GERD, aggressive anti-secretory drugs and possibly anti-reflux surgery seem to be a reasonable clinical approach. Further research is needed to study the impact of different clinical management strategies for treating GERD in patients with CF. Conflict of interest
None.
References 1. Scott RB, O’Loughlin EV, Gall DG. Gastroesophageal reflux in patients with cystic fibrosis. J Pediatr. 1985;106:223–227.
629 2. Vinocur CD, Marmon L, Schidlow DV, et al. Gastroesophageal reflux in the infant with cystic fibrosis. Am J Surg. 1985;149: 182–186. 3. Ledson MJ, Tran J, Wallshaw MJ. Prevalence and mechanisms of gastroesophageal reflux in adult cystic fibrosis patients. J R Soc Med. 1998;91:7–9. 4. Brodzicki J, Trawinska-Bartnicka M, Korzon M. Frequency, consequences and pharmacological treatment of gastroesophageal reflux in children with cystic fibrosis. Med Sci Monit. 2002;8: CR529–CR537. 5. Cucchiara S, Santamaria F, Andreotti MR, et al. Mechanisms of gastro-oesophageal reflux in cystic fibrosis. Arch Dis Child. 1991;66:617–622. 6. Blondeau K, Pauwels A, DuPont LJ, et al. Characteristics of gastroesophageal reflux and potential risk of gastric content aspiration in children with cystic fibrosis. J Pediatr Gastroenterol Nutr. 2010;50:161–166. 7. Helm JF, Dodds WJ, Riedel DR, et al. Determinants of esophageal acid clearance in normal subjects. Gastroenterology. 1983;85:607–612. 8. Boyd DD, Carney CN, Powell DW. Neuro-hormonal control of esophageal epithelial electrolyte transport. Am J Physiol. 1980;239:G5–G11. 9. Hamilton BH, Orlando RC. In vivo alkaline secretion by mammalian esophagus. Gastroenterology. 1989;97:640–648. 10. Myers RL, Orlando RC. In vivo bicarbonate secretion by human esophagus. Gastroenterology. 1992;103:1174–1178. 11. Woodley FW, Fernandez S, Mousa H. Diurnal variation in the chemical clearance of acid gastroesophageal reflux in infants. Clin Gastroenterol Hepatol. 2007;5:37–43. 12. Woodley FW, Hayes J, Mousa H. Acid gastroesophageal reflux in symptomatic infants is primarily a function of classic 2-phase and pH-only acid reflux event types. J Pediatr Gastroenterol Nutr. 2009;48:550–558. 13. Rudolph CD, Mazur LJ, Liptak GS, et al. Guidelines for evaluation and treatment of gastroesophageal reflux in infants and children: recommendations of the North American Society for Pediatric Gastroenterology and Nutrition. J Pediatric Gastroenterol Nutr. 2001;32:S1–S31. 14. Sifrim D, Castell D, Dent J, et al. Gastro-oesophageal reflux monitoring: review and consensus report on detection and definitions of acid, non-acid, and gas reflux. Gut. 2004;53: 1024–1031. 15. Helm JF, Dodds WJ, Pelc LR, et al. Effect of esophageal emptying and saliva on clearance of acid from the esophagus. N Engl J Med. 1984;310:284–288. 16. Woodley FW, Machado RS, Benner J, et al. Normal values for chemical clearance of acid gastroesophageal reflux in infants and children. Gastroenterol. 2013;144:399–400. 17. Kim BJ, Choi SC, Kim JJ, et al. Pathological bolus exposure plays a significant role in eliciting non-cardiac chest pain. J Gastroenterol Hepatol. 2010;25:1855–1860. 18. Meyers RL, Orlando RC. In vivo bicarbonate secretion by human esophagus. Gastroenterology. 1992;103:1174–1178. 19. Boyd DD, Carney CN, Powell DW. Neurohumoral control of esophageal epithelial electrolyte transport. Am J Physiol. 1980;239:G5–G11. 20. Hamilton BH, Orlando RC. In vivo alkaline secretion by mammalian esophagus. Gastroenterology. 1989;97:640–648. 21. Abdulnour-Nakhoul S, Nakhoul NL, Orlando RC. Lumen-tosurface pH gradients in opossum and rabbit esophagi: role of submucosal glands. Am J Physiol Gastrointest Liver Physiol. 2000;278:G113–G120. 22. Abdulnour-Nakhoul S, Nakhoul HN, Medhat K, et al. Ion transport mechanisms linked to bicarbonate secretion in the
123
630
23.
24.
25.
26. 27.
28.
29.
30.
31.
32. 33.
34.
Dig Dis Sci (2014) 59:623–630 esophageal submucosal glands. Am J Physiol Regul Integr Comp Physiol. 2011;301:R83–R96. Trezise AEO, Buchwald M. In vivo cell-specific expression of cystic fibrosis transmembrane conductance regulator. Nature. 1991;353:434–437. Van Slyke DD. Studies of acidosis –XVII. Determination of the bicarbonate concentration of the blood and plasma. J Biol Chem. 1922;52:495–499. Chauncey HH, Levine DM, Kass G, et al. Composition of human saliva; parotid gland secretory gland rate and electrolyte concentration in children with cystic fibrosis. Arch Oral Biol. 1962;7:707–713. Kaufman E, Lamster IB. The diagnostic applications of saliva—a review. Crit Rev Oral Biol Med. 2002;13:197–212. Bradley PJ, Guntinas-Lichius O. Salivary gland disorders and diseases: diagnosis and management. New York: NY. Georg Thieme Verlag; 2001. Davies H, Bagg J, Goodchild M, et al. Defective regulation of electrolyte and protein secretion in submandibular saliva of cystic fibrosis. Acta Paediatr Scand. 1991;80:1094–1095. Aps JK, Delanghe J, Martens LC. Salivary electrolyte concentrations are associated with cystic fibrosis transmembrane regulator genotypes. Clin Chem Lab Med. 2002;40:345–350. Goncalves AC, de Lima Marson FA, de Holanda Mendona RM et al. Saliva as a potential tool for cystic fibrosis diagnosis. Diagn Pathol. 2013;8:46. Hardt M, Witkowska HE, Webb S, et al. Assessing the effects of diurnal variation on the composition of human parotid saliva: quantitative analysis of native peptides using iTRAQ reagents. Anal Chem. 2005;77:4947–4954. Blomfield J, Allars HM, Rush AR, et al. Parotid serous hypersecretion in cystic fibrosis. Aust Paediatr J. 1974;10:75–81. Blomfield J, Van Lennep EW, Shorey CD, et al. Ultrastructure of the in vitro formation of hydroxyapatite in submandibular saliva of children with cystic fibrosis. Arch Oral Biol. 1974;19: 1153–1160. Lidums I, Holloway RH. Motility abnormalities in the columnarlined esophagus. Gastroenterol Clin. 1997;26:519–531.
123
35. Ko SB, Zeng W, Dorwart MR, et al. Gating for CFTR by STATS domain of SLC26 transporters. Nat Cell Biol. 2004;6:343–350. 36. Stewart AK, Yamamoto A, Nakakuki M, et al. Functional coupling of apical Cl-/HCO3- exchange with CFTR in stimulated secretion by guinea pig interlobular pancreatic duct. Am J Physiol Gastrointest Liver Physiol. 2009;296:G1307. 37. Rozmanic V, Tjesic-Drinkovic D, Banac Srdan, et al. Gastroesophageal reflux and gastric hyperacidity in cystic fibrosis patients. Pedijatrija Danas 2010;6(1):45–52. 38. Cox KL, Isenberg JN, Ament MEI. Gastric acid hypersecretion in cystic fibrosis. J Pediatr Gastroenterol Nutr. 1981;1:559–565. 39. Sawaf NW, Orzel JA, Weiland FL. Gastroesophageal reflux demonstrated by hepatobiliary imaging in scleroderma. J Nucl Med. 1987;28:387–389. 40. Simren M, Silny J, Holloway R, et al. Relevance of ineffective esophageal motility during oesophageal acid clearance. Gut. 2003;52:784–790. 41. Pauley KM, Lee BH, Gauna AE, et al. Mechanisms of salivary gland secretory dysfunction in Sjorgren’s syndrome. In: Harrison A ed. Insights and perspectives in rheumatology. 2012. ISBN:978-953-3007-846-5. 42. de Veer AJ, Bos JT, Niezen-de Boer RC, et al. Symptoms of gastroesophageal reflux disease in severely mentally retarded people: a systematic review. BMC Gastroenterol. 2008;8:23–35. 43. Ruigo´mez A, Wallander MA, Lundborg P, et al. Gastroesophageal reflux disease in children and adolescents in primary care. Scand J Gastroenterol. 2010;45:139–146. 44. Kellogg DL, Hodges GJ, Orozco CR, et al. Cholinergic mechanisms of cutaneous active vasodilation during heat stress in cystic fibrosis. J Appl Physiol. 2007;103:963–968. 45. Zheng W, Kuhlicke J, Jackel K, et al. Hypoxia inducible factor-1 (HIF-1)-mediated repression of cystic fibrosis transmembrane conductance regulator (CFTR) in the intestinal epithelium. FASEB J. 2009;23:204–213. 46. Ribeiro CMP, Paradiso AM, Carew MA, et al. Cystic fibrosis airway epithelial Ca2? I signaling. J Biol Chem. 2005;280: 10202–10209.
ORIGINAL ARTICLE
Children with Cystic Fibrosis Have Prolonged Chemical Clearance of Acid Reflux Compared to Symptomatic Children Without Cystic Fibrosis Frederick W. Woodley • Rodrigo S. Machado • Don Hayes Jr. Carlo Di Lorenzo • Ajay Kaul • Beth Skaggs • Karen McCoy • Alpa Patel • Hayat Mousa
•
Received: 9 September 2013 / Accepted: 11 November 2013 / Published online: 28 November 2013 Ó Springer Science+Business Media New York 2013
Abstract Objectives Few studies compare gastroesophageal reflux (GER) parameters of cystic fibrosis (CF) children and symptomatic non-CF children. We aimed to compare the impedance-pH (IMP-pH) parameters for these two groups and to test the hypothesis that prolonged acid exposure in CF patients is due to delayed chemical clearance (CC). Methods IMP-pH tracings from 16 CF children (median 8.2 years) and 16 symptomatic non-CF children (median 8.3 years) were analyzed. Software was used to generate IMP-pH reports and parameter data were extracted. IMPpH was used to calculate the mean CC for each patient.
F. W. Woodley (&) C. Di Lorenzo B. Skaggs H. Mousa Division of Gastroenterology, Hepatology, and Nutrition, Nationwide Children’s Hospital, Columbus, OH, USA e-mail: [email protected] F. W. Woodley D. Hayes Jr. C. Di Lorenzo K. McCoy A. Patel H. Mousa Department of Pediatrics, The Ohio State University College of Medicine, Columbus, OH 43205, USA
Results pH studies showed no difference in acid GER (AGER) frequency (p = 0.587); however, mean AGER duration, duration of longest AGER, AGER index, and DeMeester scores were all significantly higher for CF patients. IMP showed no difference in GER frequency [neither acidic (p = 0.918) nor non-acidic (p = 0.277)], but total bolus clearance was more efficient in CF patients (p = 0.049). A larger percentage of total GER reached the proximal esophagus in non-CF children (p = 0.039). Analyses of two-phase AGER episodes showed that these events were more acidic (p = 0.003) and the CC phase was significantly prolonged in the CF cohort (p = 0.001). Conclusions Compared to symptomatic non-CF children, CF children do not have more frequent reflux. Actually, they have better bolus clearance efficiency following reflux and may even have better control over the number of GER episodes that reach the proximal esophagus. CC of AGER, however, is significantly prolonged in the CF cohort, likely due to hyperacidity of refluxed gastric contents. Keywords Cystic fibrosis Acid reflux Chemical clearance Volume clearance Proximal extent
R. S. Machado Universidade Federal de Sa˜o Paulo, Sa˜o Paulo, Brazil D. Hayes Jr. K. McCoy A. Patel Pulmonary Medicine, Nationwide Children’s Hospital, Columbus, OH, USA A. Kaul Division of Gastroenterology, Hepatology, and Nutrition, Cincinnati Children’s Hospital Medical Center, Cincinnati, OH, USA A. Kaul Department of Pediatrics, University of Cincinnati College of Medicine, Cincinnati, OH, USA
Introduction Gastroesophageal reflux (GER) is common among patients with cystic fibrosis (CF). The prevalence of acid GER (AGER) in CF patients, in particular, has been estimated at 35–81 % [1–6]. Few studies compare CF children to nonCF children who have similar symptoms of GER. Despite reports of increased AGER indices in CF patients [1], there have been no studies determining whether this increased acid exposure is due to delays in volume clearance (VC)
123
624
Dig Dis Sci (2014) 59:623–630
pH 4
VC
CC
Fig. 1 IMP-pH tracing illustrating a two-phase acid reflux episode. VC volume clearance, CC chemical clearance
alone, chemical clearance (CC) alone, or to both. VC is the first phase of AGER clearance during which the bulk of refluxed gastric contents is cleared from the esophagus (back into the stomach) by esophageal peristalsis, and CC is the second phase of AGER clearance during which acidified mucosa is neutralized by bicarbonate-rich saliva [7] and bicarbonate ions released directly into the esophageal lumen from submucosal glands [8–10]. Increased acid exposure, as quantified by the AGER index, could result from frequent AGER, prolonged VC, or prolonged CC. We previously showed that combined multichannel intraluminal impedance and esophageal pH monitoring (IMP-pH) can be used to distinguish the VC phase of twophase AGER episodes from the CC phase [11, 12]. Synchronization of impedance and pH measurements permits a clear distinction between the dynamics of intraluminal flow (impedance) and neutralization of the acidified esophageal mucosa (pH probe) (Fig. 1). This investigation compared CF children and symptomatic non-CF children on the basis of the IMP-pH parameters that are commonly used to assess patients suspected of having GER. We also tested the hypothesis that prolonged acid exposure in children with CF is due to delayed CC.
123
Patients and Methods Patients The IMP-pH databases at Nationwide Children’s Hospital (NCH) and Cincinnati Children’s Hospital Medical Center (CCHMC) were searched for CF patients who were not taking anti-reflux medications and who had not received a fundoplication prior to testing. To be included in the study, patients had to have been off anti-reflux medications such as proton pump inhibitors, histamine-receptor antagonists, and prokinetics for at least 7, 5, and 2 days, respectively, prior to each study. A non-CF age-matched patient was selected from the NCH database each time a qualified CF patient was identified. The selection process (January 2002 to March 2009) involved choosing the first subject from the list of names/ages of previously studied patients (in the order of newer to older tracings) who satisfied the inclusion criteria and who was closest in age to the paired CF patient. This study was approved by the NCH and CCHMC Institutional Review Boards. Catheter and Placement IMP-pH catheters and their placement have been described previously [13]. Age-appropriate catheters included the
Dig Dis Sci (2014) 59:623–630
pediatric (age 2–10 years) and the adolescent ([10 years) catheters (Sandhill Scientific, Highlands Ranch, Colorado). Analyses of IMP-pH Tracings
625
AGER indices B50 % of the upper end of values considered normal by pH probe (i.e. B3 %)[13], did not have positive temporal associations of GER with symptoms, were not taking any anti-reflux medications, and had not received a fundoplication prior to testing.
pH Probe Proximal Extent of GER Acid gastroesophageal reflux (AGER) episodes were characterized by a drop in distal esophageal pH to \4. The IMP-pH reports were generated using the software provided by Sandhill Scientific, with pH probe variables that included the number of AGER, the mean AGER duration, the longest AGER duration, the DeMeester score, and the AGER index. Only events lasting 5 s or longer were registered and analyzed. Acid exposure was considered increased when the AGER index was [6 % [13]. Impedance Impedance-detectable events are defined by a drop of intraluminal impedance to below 50 % of baseline values beginning in the distal-most impedance channel (channel 6) and propagating proximally into at least two impedance channels. Only those impedance events lasting 5 s or longer were registered. GER events are defined as AGER episodes when the pH falls to\pH 4 at any point during the impedance-detectable episode. Non-acid GER events are impedance-detectable events during which the pH is B4; these episodes included both weakly-acid GER and weakly-alkaline GER [14]. For both GER detection tools (impedance and pH probe), GER episodes occurring during feeding were excluded from analysis. IMP-pH tracings were analyzed by a single author (FW).
For each patient, the number (and percentage of total) of each impedance-detectable GER event (AGER and NAGER) that reached the proximal esophagus was determined. GER that reached one or both of the two proximalmost impedance channels (channels 1 and 2) were tagged as proximal reflux episodes. Temporal Association of Symptoms with GER For each patient, an IMP-pH report was generated and printed. Using a 2-min window, the symptom association probability (SAP) was calculated for all measured symptoms. A SAP was considered positive when it equaled [95 %. We compared CF and non-CF symptomatic children based on the presence of a positive SAP, either by pH or by impedance. CF and non-CF groups were compared by chi-squared analysis. Statistical Analysis The Wilcoxon test was used to assess the difference for each parameter between the two groups (CF children and non-CF age-matched children). Frequency of pathological GER [17] in cases and controls was compared with the McNemar test. Data are presented as median values (25–75th percentile) unless otherwise indicated. Statistical significance was established at p \ 0.05.
Assessing Chemical Clearance Two-phase AGER episodes (Fig. 1) [15], detected using combined IMP-pH, serve as the ideal model system [11] for assessing the efficiency of acid neutralization. During two-phase AGER episodes, CC begins promptly at the point at which VC (also known as bolus contact time, BCT) is completed. For each patient, the IMP-pH tracing was manually screened for two-phase AGER episodes. For each episode, CC duration was calculated as previously described [11]. A mean was then calculated for each patient and a median (25–75th quartiles) was calculated for each group (CF children and non-CF age-matched children). CC duration for CF and non-CF patients in this study was compared to 50 ‘‘normal’’ children [M/F 36/14, median 7.0 years (range 2.8–11.0 years)] who were previously described in abstract form [16]. These control children had
Results Clinical Results We enrolled 16 CF patients, aged 3.0–17.7 years. Three of the 16 patients were tested at CCHMC. Sixteen age-matched non-CF patients were chosen from a registry of children who were tested at NCH. All patients were off anti-reflux medications and had not received a fundoplication prior to IMP-pH testing. Each child was referred for symptoms suggestive of GER as an etiology. Across all participants, median study length was 23.8 h (range 22.5–26.5 h) for CF patients and 23.9 h (range 21.3–27.6 h) for non-CF patients. Patient demographics are shown in Table 1.
123
626
Dig Dis Sci (2014) 59:623–630
Table 1 Patient demographics
Table 3 Characteristics of GER by IMP-pH
Demographic
CF (N = 16)
Non-CF (N = 16)
Impedance-pH
CF
Non-CF
p value
Age, median (range)
8.2 years (3.1–17.7 years)
8.3 years (3.0–17.2 years)
#AGER
42.0 (36.0–60.7)
47.5 (25.7–69.5)
0.918
#NAGER
4.0 (2.0–11.2)
6.0 (2.2–14.0)
0.277
12.0 (9.2–21.0)
19.2 (13.3–37.3)
0.088
Sex, M/F
6/10
9/7
#GER
Race, CA/AA
16/0
14/2
%AGER
1.3 (0.7–1.7)
1.7 (0.8–2.4)
0.272
Genotype-DF508(2)/ DF508(l)/other
10/4/2
N/A
%NAGER
0.1 (0.02–0.2)
0.2 (0.1–0.5)
0.157
%BCT
1.5 (0.9–1.9)
2.1 (0.9–2.8)
0.187
Median BCT (s)
12.0 (11.0–13.7)
15.5 (11.7–18.0)
0.049*
Symptoms presented Cough
100 %
75 %
Vomiting
0%
25 %
Regurgitation
12 %
6%
AGER acid gastroesophageal reflux, NAGER non-acid gastroesophageal reflux, GER gastroesophageal reflux, BCT bolus contact time (also known as volume clearance), CF cystic fibrosis
Heartburn
25 %
12 %
Values are expressed as medians (25–75th interquartile)
Gasping
0%
6%
* Significance at p \ 0.05
Night awakenings
6%
6%
Retching
0%
6%
Choking
6%
12 %
Throat clearing
0%
12 %
Abdominal pain
6%
12 %
Decline in PFT
12 %
0%
Wheezing
12 %
6%
CA Caucasian, AA African American, CF cystic fibrosis, PFT pulmonary function tests
Table 2 Characteristics of GER by pH probe
symptoms that could be positively associated with AGER (SAP[95 %), three patients had symptoms associated with both AGER and NAGER (23 %), and one patient (\1 %) did not have any symptoms that could be associated with GER. For the non-CF children, four of the ten children (40 %) with pathological GER had symptoms that could be positively associated with AGER, one patient (10 %) had symptoms associated with both AGER and NAGER, and four patients (40 %) did not have any symptoms that could be associated with GER.
pH probe
CF
Non-CF
p value
Number of AGER episodes
67.5 (50.7–83.5)
54.5 (38.2–120.5)
0.587
Impedance
Mean AGER duration (s)
129.0 (77.2–197.5)
86.5 (37.5–108.5)
0.017*
Longest AGER duration (min)
29.8 (19.0–39.7)
9.9 (4.0–14.6)
0.004*
DeMeester score
38.9 (24.3–60.7)
23.7 (11.2–37.9)
0.006*
AGER index (%) Pathological AGER indicesa
11.3 (6.9–18.2) 13/16 (81 %)
5.6 (2.6–9.7) 10/16 (62 %)
0.003* 0.125
From each IMP-pH report, we extracted the number of episodes, total time, and percent times for AGER, NAGER, and total GER (Table 3). As indicated in Table 3, only the median BCT was different between groups; the median BCT was significantly higher in non-CF patients (p = 0.049). The number of each event type that reached the proximal esophagus was also extracted directly from the IMP-pH report. The percentages of total AGER, NAGER and GER reaching the proximal esophagus were calculated (Table 4). As indicated in Table 4, the only significant difference between groups was the percent total proximal GER, which was higher in non-CF patients.
Values are expressed as medians (interquartile range), unless otherwise indicated AGER acid gastroesophageal reflux, CF cystic fibrosis * Significance at p \ 0.05 AGER indices [5 %
a
pH Probe With the exception of AGER frequency (p = 0.587), all pH probe parameters were significantly greater for the CF patients (p \ 0.05, Table 2). Thirteen of 16 CF children (81 %) had pathological GER (i.e. AGER Index [5 %) [17] compared to ten of 16 non-CF children (62 %). Nine of 13 CF children (69 %) with pathological GER had
123
IMP-pH to Detect Two-Phase AGER Episodes Each IMP-pH tracing was manually inspected for twophase AGER episodes. While there was no difference in the frequency of two-phase AGER (p = 0.979), and the more efficient VC for the CF children (25.7 vs 30.3 s) failed to achieve statistical significance (p = 0.057), CC duration was almost twofold less efficient in the CF children (122.8 vs 65.9 s) (p = 0.001) (Table 5). CC duration data distribution was positively skewed. Following
Dig Dis Sci (2014) 59:623–630
627
Discussion
Table 4 Proximal extent of GER Impedance-pH
CF
Non-CF
p value
#Proximal AGER
22.5 (11.5–35.0)
35.0 (14.0–57.0)
0.205
%Total proximal AGER
57.9 (34.2–72.1)
80.2 (59.5–86.5)
0.056
#Proximal NAGER %Total proximal NAGER
1.0 (0–4.5) 29.1 (0–58.6)
3.0 (1.2–8.5) 62.1 (19.0–91.7)
0.151 0.061
#Proximal GER
27.0 (14.0–38.2)
38.5 (21.5–73.5)
0.105
%Total proximal GER
55.7 (33.7–72.4)
78.8 (61.2–87.4)
0.039*
AGER acid gastroesophageal reflux, NAGER non-acid gastroesophageal reflux, GER gastroesophageal reflux, CF cystic fibrosis Values are expressed as medians (25–75th interquartile) * Significance at p \ 0.05
Table 5 Impedance-pH parameters for two-phase AGER episodes Impedance-pH
CF
Non-CF
p value
VC (s)
25.7 (18.5–28.7)
30.3 (25.9–34.2)
0.057
CC (s)
122.8 (96.5–197.7)
65.9 (30.0–101.4)
0.001*
Total duration (s)
159.0 (120.2–234.9)
95.0 (65.3–134.8)
0.002*
pH nadir
1.1 (0.8–1.6)
1.8 (1.4–2.1)
0.003*
Number of episodes
23.2 (16.8–35.7)
24.7 (14.6–36.2)
0.979
AGER acid gastroesophageal reflux, VC volume clearance, CC chemical clearance, CF cystic fibrosis Values are expressed as medians (25–75 % interquartile) * Significance at p \ 0.05
logarithmic transformation, CC durations of CF and nonCF children were compared to a previously reported set of 50 children with no evidence of GER-associated symptoms [16]. CF children presented a significantly longer CC duration compared to normal controls [Log10 CC duration—mean difference 0.553 (95 % CI 0.410–0.697), p \ 0.010]. The median crude CC duration was 122.8 s (IQR 96.5–197.7 s) versus 37.8 s in controls (IQR 29.2–52.6 s). Symptomatic non-CF children also presented a significantly longer CC duration compared to normal controls [Log10 CC duration—mean difference 0.209 (95 % CI -0.0370 to 0.381), p = 0.036], with the crude median CC duration for non-CF children and controls, respectively, 65.9 s (IQR 30.0–101.4 s) and 37.8 s (IQR 29.2–52.6 s). The mean nadir pH of two-phase AGER episodes was significantly more acidic in the CF cohort than in the non-CF cohort [median pH 1.1 (IQR 0.8–1.6) vs median pH 1.8 (IQR 1.4–2.1), p = .003] (Table 5).
This study found that, when compared to other patients referred for symptoms suggestive of GER, CF patients have increased AGER that is not due to more frequent episodes of AGER or to delayed bolus clearance but rather to the prolonged CC of AGER. It was reported several years ago (using pH monitoring) that children with CF have increased AGER when compared to healthy children [1–6]; however, the pH probe monitoring used in these studies was unable to distinguish bolus clearance from CC. Our data show that compared to non-CF patients who presented with symptoms similar to those of our CF cohort, CC lasts twice as long in CF children. VC efficiency, on the other hand, was as good as (or better) in CF children than in the non-CF age-matched children. Recently, in a study of 24 CF children (median 4 years [range 0.3–13 years]), Blondeau et al. [6] reported a median AGER index of 3.5 %, a median volume exposure of 1.3 %, a median total number of GER episodes (per patient) of 41, a median proximal extent of 25 %, and an AGER:NAGER ratio of 70:30. In the present study of 16 CF children (median 8.2 years [range 3.1–17.7 years]), we report a median AGER index of 11.3 %, a median volume exposure of 1.5 %, a median total number of GER episodes (per patient) of 47, a median proximal extent of 56 %, and an AGER:NAGER ratio of 95:5. Differences in AGER index and AGER:NAGER ratio can be attributed to differences in age; Blondeau et al. included infants in their study and formula-fed infants experience more frequent NAGER episodes (and fewer AGER) than older children. Differences in genotype may also be important; the genotype ratio [DF508(homozygote):DF508 (heterozygote):other] was 10:4:2 for the present cohort and 5:8:11 for the Blondeau et al. cohort. In the present study, more than 62 % of the CF cohort were homozygous for the DF508 genotype compared to approximately 21 % in the Blondeau cohort. The difference in proximal extent between the two cohorts may be explained by the fact that we defined proximal events as those GER reaching either one of the two proximal impedance channels (1 or 2), while Blondeau and colleagues defined proximal events as those GER reaching only channel 1. VC, also known as bolus clearance or bolus contact time (BCT), is the process by which refluxed material is extruded back into the stomach by swallow-induced primary peristalsis and/or mechano-receptor-induced secondary peristalsis. CC is the process by which the acidified esophageal mucosa is neutralized following passage of the bolus of an acid reflux episode. Neutralization is subsequently accomplished when bicarbonate-rich saliva is transported to the distal esophagus by primary and secondary peristalsis and also by intrinsic esophageal
123
628
bicarbonate that is secreted directly into the esophageal lumen by submucosal glands [18–22]. In addition to bicarbonate, saliva also contains proteins and phosphates that also have buffering potential. One possible explanation for the prolonged CC in CF children is a decreased buffering capacity of the saliva. Reduced bicarbonate secretion into CF saliva, for example, would reduce the efficiency with which CF patients could neutralize the esophageal mucosa following an AGER episode. In situ hybridization studies show that cystic fibrosis transmembrane conductance regulator (CFTR) is expressed in salivary gland ductal cells [23]; it is within these ducts that, with the functional wild-type CFTR gene, sodium is reabsorbed and potassium and bicarbonate ion are secreted. More than five decades ago, in a study aimed at comparing bicarbonate levels in saliva between CF and non-CF controls, no significant differences in bicarbonate levels were detected. Using a titration method [24] to quantify bicarbonate levels in parotid salivary secretions, Chauncey et al. [25] measured 15.4 ± 6.0 mEq/L [range 6.9–32.8] of bicarbonate in CF patients and 13.4 ± 3.6 mEq/L [range 7.7–33.5] (p = 0.40) in the nonCF patients. However, whole saliva is composed of secretions from three major glands (parotid, submandibular, and sublingual) as well as from several minor glands [26]. While the parotid gland is by far the largest anatomically, parotid contribution to total saliva secretion is only about 21 %, compared to approximately 70 % from the submandibular gland [27]. Therefore, equivalent levels of bicarbonate within the parotid portion of whole saliva between CF and non-CF patients may not significantly influence the overall bicarbonate content in whole saliva. Davies et al. [28] later compared the regulation of electrolyte secretion in submandibular saliva in CF and control subjects. Using a colorimetric system (Technicon SMAC II system), they found decreased bicarbonate levels in CF resting saliva and when the submandibular gland was stimulated using vitamin C, bicarbonate levels increased in control subjects but not in CF subjects. Recently, Aps et al. [29] compared the electrolyte levels in whole saliva among CF homozygotes, CF heterozygotes, and non-CF subjects. Bicarbonate levels were determined enzymatically using a commercially available test kit (BioMerieux, Marcey-lesEtoiles, France). Unfortunately, these authors excluded calcium and bicarbonate levels from their published tables, since there was no difference in calcium or bicarbonate levels between the three groups. Most recently, in a study of 80 CF patients [median age 12.0 years (range 4.0–34.0 years)] and 84 healthy volunteers [median age 17.0 years (range 5.0–28.0 years)], Goncalves et al. [30] found no significant difference in bicarbonate levels between groups (p = 0.459). Considering the influence of diet and different types of salivary stimulation on saliva
123
Dig Dis Sci (2014) 59:623–630
composition [27], combined with possible circadian rhythm or diurnal influences [31, 32], further studies involving carefully controlled experimental designs are needed to assess potential differences in the buffering capacity of whole saliva produced by patients with and without CF. Saliva transport efficiency may be compromised in CF patients; reduced salivary flow rates and high viscosity have been reported [32, 33]. Further, esophageal hypomotility in CF could also reduce efficient transport of the viscous saliva through the esophageal lumen [34]. Since VC efficiency in our CF cohort was equal to or better than VC in the non-CF patients, hypomotility is not the likely explanation for our CF population. Alternatively, differences in the volume of saliva may have an effect; Goncalves et al. [30] found that saliva volume is significantly reduced in CF patients. Saliva volume was not analyzed in this study. Reduced secretion of bicarbonate into the esophageal lumen by submucosal glands may also be responsible for reduced CC efficiency. Using immunofluorescence, RTPCR, and in situ hybridization in a porcine model, Abdulnour-Nakhoul et al. [22] demonstrated expression of the CFTR in submucosal gland duct cells and localized CFTR and SLC26A6 on the luminal membrane. In other tissues (including the pancreas), it has been reported that the SLC26 family of anion exchangers interacts with CFTR to regulate secretion of bicarbonate [35, 36]. Expression of CFTR in the esophageal submucosal glands of humans has not yet been demonstrated. Reduced CC efficiency in CF patients could also be related to differences in gastric acidity, as larger numbers of hydrogen ions refluxed onto the esophageal mucosa would be expected to require longer periods of time to be neutralized. Gastric hypersecretion is common among CF patients [37, 38]. Consistent with these reports, in the present study, two-phase AGER events were significantly lower in CF patients. We present a novel method for assessing CC efficiency in patients evaluated by combined IMP-pH. Significant delays in CC could be a finding in numerous established conditions that include CF, scleroderma [39], esophageal hypomotility [40], defects in saliva production and secretion as in Sjogren’s syndrome [41], and progressing neurological disorders [42, 43]. Of the four AGER types detectable by IMP-pH (two-phase, single-phase, pH-only events, and re-reflux AGER episodes)[12], only the twophase episode is suitable for CC assessment because the start and stop time points of CC are discretely demarcated; CC begins when VC concludes and ends when the pH reaches 4. Assessing CC of spontaneous two-phase AGER episodes detected during IMP-pH testing is less invasive than other classic methods [15, 40] involving the use of
Dig Dis Sci (2014) 59:623–630
radioactive hydrochloric acid instilled into the esophageal lumen. In the present study, we simply identified and analyzed all two-phase AGER episodes and then calculated a mean duration of CC for each patient. Further studies involving larger sample sizes are needed to determine the minimum number of two-phase AGER episodes that would be needed to derive a valid metric to compare to reference values for CC in infants, children, or adults. The most surprising finding of this study was the fact that fewer GER reached the proximal esophagus in children with CF. While it is tempting to speculate that CF children might have an innate reflexive response to prevent aspiration, it is equally possible that gastric contents in CF patients become thicker due to constantly swallowed mucous, making it less likely to go against gravity to the proximal esophagus. Also, it is possible that the differences in proximal extent are unique to the present study and cannot be extrapolated to other patient populations. Other compensatory mechanisms in CF patients have also been suggested [44–46]. This study has several limitations. First, the size of the cohort is not ideal. We were limited by the number of CF patients (B18 years) who were off anti-reflux medications and who had not received a fundoplication prior to their IMP-pH study. Second, our subjects were not studied in a sleep lab where sleep status, symptoms, and precise body positioning could have been more precisely monitored. Third, we realize that our selection of non-CF control patients could have influenced our results. Finally, analyses of IMP-pH tracings were not blinded to patient genotype (CF vs non-CF). In conclusion, while the average CF child has an AGER index that is significantly higher than other children presenting with similar symptoms, he or she does not generally experience more frequent AGER events than other GERD patients. However, the CF child has significantly delayed CC that is likely due to hyperacidity of refluxed gastric contents. Currently, there are no anti-reflux drugs that specifically target CC. In order to treat CF patients who have GERD, aggressive anti-secretory drugs and possibly anti-reflux surgery seem to be a reasonable clinical approach. Further research is needed to study the impact of different clinical management strategies for treating GERD in patients with CF. Conflict of interest
None.
References 1. Scott RB, O’Loughlin EV, Gall DG. Gastroesophageal reflux in patients with cystic fibrosis. J Pediatr. 1985;106:223–227.
629 2. Vinocur CD, Marmon L, Schidlow DV, et al. Gastroesophageal reflux in the infant with cystic fibrosis. Am J Surg. 1985;149: 182–186. 3. Ledson MJ, Tran J, Wallshaw MJ. Prevalence and mechanisms of gastroesophageal reflux in adult cystic fibrosis patients. J R Soc Med. 1998;91:7–9. 4. Brodzicki J, Trawinska-Bartnicka M, Korzon M. Frequency, consequences and pharmacological treatment of gastroesophageal reflux in children with cystic fibrosis. Med Sci Monit. 2002;8: CR529–CR537. 5. Cucchiara S, Santamaria F, Andreotti MR, et al. Mechanisms of gastro-oesophageal reflux in cystic fibrosis. Arch Dis Child. 1991;66:617–622. 6. Blondeau K, Pauwels A, DuPont LJ, et al. Characteristics of gastroesophageal reflux and potential risk of gastric content aspiration in children with cystic fibrosis. J Pediatr Gastroenterol Nutr. 2010;50:161–166. 7. Helm JF, Dodds WJ, Riedel DR, et al. Determinants of esophageal acid clearance in normal subjects. Gastroenterology. 1983;85:607–612. 8. Boyd DD, Carney CN, Powell DW. Neuro-hormonal control of esophageal epithelial electrolyte transport. Am J Physiol. 1980;239:G5–G11. 9. Hamilton BH, Orlando RC. In vivo alkaline secretion by mammalian esophagus. Gastroenterology. 1989;97:640–648. 10. Myers RL, Orlando RC. In vivo bicarbonate secretion by human esophagus. Gastroenterology. 1992;103:1174–1178. 11. Woodley FW, Fernandez S, Mousa H. Diurnal variation in the chemical clearance of acid gastroesophageal reflux in infants. Clin Gastroenterol Hepatol. 2007;5:37–43. 12. Woodley FW, Hayes J, Mousa H. Acid gastroesophageal reflux in symptomatic infants is primarily a function of classic 2-phase and pH-only acid reflux event types. J Pediatr Gastroenterol Nutr. 2009;48:550–558. 13. Rudolph CD, Mazur LJ, Liptak GS, et al. Guidelines for evaluation and treatment of gastroesophageal reflux in infants and children: recommendations of the North American Society for Pediatric Gastroenterology and Nutrition. J Pediatric Gastroenterol Nutr. 2001;32:S1–S31. 14. Sifrim D, Castell D, Dent J, et al. Gastro-oesophageal reflux monitoring: review and consensus report on detection and definitions of acid, non-acid, and gas reflux. Gut. 2004;53: 1024–1031. 15. Helm JF, Dodds WJ, Pelc LR, et al. Effect of esophageal emptying and saliva on clearance of acid from the esophagus. N Engl J Med. 1984;310:284–288. 16. Woodley FW, Machado RS, Benner J, et al. Normal values for chemical clearance of acid gastroesophageal reflux in infants and children. Gastroenterol. 2013;144:399–400. 17. Kim BJ, Choi SC, Kim JJ, et al. Pathological bolus exposure plays a significant role in eliciting non-cardiac chest pain. J Gastroenterol Hepatol. 2010;25:1855–1860. 18. Meyers RL, Orlando RC. In vivo bicarbonate secretion by human esophagus. Gastroenterology. 1992;103:1174–1178. 19. Boyd DD, Carney CN, Powell DW. Neurohumoral control of esophageal epithelial electrolyte transport. Am J Physiol. 1980;239:G5–G11. 20. Hamilton BH, Orlando RC. In vivo alkaline secretion by mammalian esophagus. Gastroenterology. 1989;97:640–648. 21. Abdulnour-Nakhoul S, Nakhoul NL, Orlando RC. Lumen-tosurface pH gradients in opossum and rabbit esophagi: role of submucosal glands. Am J Physiol Gastrointest Liver Physiol. 2000;278:G113–G120. 22. Abdulnour-Nakhoul S, Nakhoul HN, Medhat K, et al. Ion transport mechanisms linked to bicarbonate secretion in the
123
630
23.
24.
25.
26. 27.
28.
29.
30.
31.
32. 33.
34.
Dig Dis Sci (2014) 59:623–630 esophageal submucosal glands. Am J Physiol Regul Integr Comp Physiol. 2011;301:R83–R96. Trezise AEO, Buchwald M. In vivo cell-specific expression of cystic fibrosis transmembrane conductance regulator. Nature. 1991;353:434–437. Van Slyke DD. Studies of acidosis –XVII. Determination of the bicarbonate concentration of the blood and plasma. J Biol Chem. 1922;52:495–499. Chauncey HH, Levine DM, Kass G, et al. Composition of human saliva; parotid gland secretory gland rate and electrolyte concentration in children with cystic fibrosis. Arch Oral Biol. 1962;7:707–713. Kaufman E, Lamster IB. The diagnostic applications of saliva—a review. Crit Rev Oral Biol Med. 2002;13:197–212. Bradley PJ, Guntinas-Lichius O. Salivary gland disorders and diseases: diagnosis and management. New York: NY. Georg Thieme Verlag; 2001. Davies H, Bagg J, Goodchild M, et al. Defective regulation of electrolyte and protein secretion in submandibular saliva of cystic fibrosis. Acta Paediatr Scand. 1991;80:1094–1095. Aps JK, Delanghe J, Martens LC. Salivary electrolyte concentrations are associated with cystic fibrosis transmembrane regulator genotypes. Clin Chem Lab Med. 2002;40:345–350. Goncalves AC, de Lima Marson FA, de Holanda Mendona RM et al. Saliva as a potential tool for cystic fibrosis diagnosis. Diagn Pathol. 2013;8:46. Hardt M, Witkowska HE, Webb S, et al. Assessing the effects of diurnal variation on the composition of human parotid saliva: quantitative analysis of native peptides using iTRAQ reagents. Anal Chem. 2005;77:4947–4954. Blomfield J, Allars HM, Rush AR, et al. Parotid serous hypersecretion in cystic fibrosis. Aust Paediatr J. 1974;10:75–81. Blomfield J, Van Lennep EW, Shorey CD, et al. Ultrastructure of the in vitro formation of hydroxyapatite in submandibular saliva of children with cystic fibrosis. Arch Oral Biol. 1974;19: 1153–1160. Lidums I, Holloway RH. Motility abnormalities in the columnarlined esophagus. Gastroenterol Clin. 1997;26:519–531.
123
35. Ko SB, Zeng W, Dorwart MR, et al. Gating for CFTR by STATS domain of SLC26 transporters. Nat Cell Biol. 2004;6:343–350. 36. Stewart AK, Yamamoto A, Nakakuki M, et al. Functional coupling of apical Cl-/HCO3- exchange with CFTR in stimulated secretion by guinea pig interlobular pancreatic duct. Am J Physiol Gastrointest Liver Physiol. 2009;296:G1307. 37. Rozmanic V, Tjesic-Drinkovic D, Banac Srdan, et al. Gastroesophageal reflux and gastric hyperacidity in cystic fibrosis patients. Pedijatrija Danas 2010;6(1):45–52. 38. Cox KL, Isenberg JN, Ament MEI. Gastric acid hypersecretion in cystic fibrosis. J Pediatr Gastroenterol Nutr. 1981;1:559–565. 39. Sawaf NW, Orzel JA, Weiland FL. Gastroesophageal reflux demonstrated by hepatobiliary imaging in scleroderma. J Nucl Med. 1987;28:387–389. 40. Simren M, Silny J, Holloway R, et al. Relevance of ineffective esophageal motility during oesophageal acid clearance. Gut. 2003;52:784–790. 41. Pauley KM, Lee BH, Gauna AE, et al. Mechanisms of salivary gland secretory dysfunction in Sjorgren’s syndrome. In: Harrison A ed. Insights and perspectives in rheumatology. 2012. ISBN:978-953-3007-846-5. 42. de Veer AJ, Bos JT, Niezen-de Boer RC, et al. Symptoms of gastroesophageal reflux disease in severely mentally retarded people: a systematic review. BMC Gastroenterol. 2008;8:23–35. 43. Ruigo´mez A, Wallander MA, Lundborg P, et al. Gastroesophageal reflux disease in children and adolescents in primary care. Scand J Gastroenterol. 2010;45:139–146. 44. Kellogg DL, Hodges GJ, Orozco CR, et al. Cholinergic mechanisms of cutaneous active vasodilation during heat stress in cystic fibrosis. J Appl Physiol. 2007;103:963–968. 45. Zheng W, Kuhlicke J, Jackel K, et al. Hypoxia inducible factor-1 (HIF-1)-mediated repression of cystic fibrosis transmembrane conductance regulator (CFTR) in the intestinal epithelium. FASEB J. 2009;23:204–213. 46. Ribeiro CMP, Paradiso AM, Carew MA, et al. Cystic fibrosis airway epithelial Ca2? I signaling. J Biol Chem. 2005;280: 10202–10209.
