Proximal Esophageal Impedance Contour Predicts Increased Reflux Burden in Patients With Laryngopharyngeal Symptoms

Abstract

INTRODUCTION:

The value of esophageal baseline impedance (BI) in assessing proximal reflux and laryngopharyngeal symptoms (LPSs) is unclear.

METHODS:

Two hundred eighteen patients with LPS underwent 24-hour combined hypopharyngeal-esophageal impedance-pH monitoring. Proximal/distal BI was obtained, and a slope-and-intercept model of proximal BI contour was constructed.

RESULTS:

Proximal BI correlated with proximal/pharyngeal reflux (r = −0.21, P < 0.01) and reflux symptom index (r = −0.14, P = 0.08). The proximal BI contour model incorporating both the BI change (slope) and BI just below upper esophageal sphincter (intercept) outperformed models using individual BI measures in predicting proximal (Akaike information criterion: 110 vs 251–253) or pharyngeal (akaike information criterion: 32 vs 141–148) reflux.

DISCUSSION:

Proximal esophageal impedance contour predicts proximal reflux in patients with LPS.

INTRODUCTION

Laryngopharyngeal reflux disease (LPRD) is a heterogeneous condition characterized by the retrograde flow of gastric contents into the upper airway. However, it remains challenging to diagnose as conventional reflux monitoring may not correlate with proximal/pharyngeal reflux and presenting laryngopharyngeal symptoms (LPSs) (1,2). Baseline impedance (BI) is a novel metric of esophageal mucosal integrity on the impedance-pH study, with previous research showing correlation between lower distal BI and greater esophageal reflux burden (3). However, there remain little data on proximal esophageal BI and its potential association with proximal reflux and LPS. Moreover, the mucosal impedance pattern along the esophagus has shown utility in assessing gastroesophageal reflux (4). We sought to characterize the predictive values of proximal BI and overall BI pattern, as measured by combined hypopharyngeal-esophageal multichannel intraluminal impedance-pH testing (HEMII-pH), for abnormal proximal reflux and symptoms in patients with suspected LPRD.

METHODS

Adults with suspected LPR presenting for HEMII-pH monitoring off acid suppression at a tertiary center from April 2015 to August 2019 were enrolled (see Supplementary Methods, Supplementary Digital Content, http://links.lww.com/CTG/B470 for inclusion/exclusion criteria). The HEMII-pH catheter was placed with the proximal pH sensor located within the upper esophageal sphincter (UES) using high-resolution manometry guidance, as described previously (1). BI on HEMII-pH was averaged across 3 quiescent 10-minute intervals (12 am/1 am/2 am) to determine the mean nocturnal BI (MNBI) (5). Proximal and distal MNBI was obtained from electrode pairs located 1–3/3–5 cm (channels 1–2) and 9–11/11–13 cm (channels 3–4) below the UES, respectively (Figure 1). Conventional measures of reflux burden on impedance-pH testing were recorded after exclusion of meal periods (see Supplemental Methods, Supplementary Digital Content, http://links.lww.com/CTG/B470). HEMII-pH tracings were manually analyzed by 2 expert readers (>5 years of experience) using a dedicated software package (BioView Analysis, version 5.6.3.0; Diversatek Healthcare, Milwaukee, WI). For LPS severity, participants prospectively completed the Reflux Symptom Index (RSI) at the time of testing (6).

Figure 1.

Schematic of the combined hypopharyngeal-esophageal multichannel intraluminal impedance and pH monitoring catheter. Proximal and distal esophageal mean nocturnal baseline impedance was obtained from impedance electrode pairs located at 1–3/3-5 cm (channels 1–2) and 9–11/11–13 cm (channels 3–4) below the upper esophageal sphincter, respectively.

The relationships between MNBI measured at each channel and reflux parameters, assessed as both continuous and dichotomized outcomes, were evaluated by Pearson's correlation and Student's t-test, respectively. Pearson's correlation was additionally used to assess the relationship between MNBI and RSI. Linear regression was performed to derive a novel parameter that estimates rate of change in MNBI across channels (ΔMNBI). We assessed the predictive strength for pharyngeal reflux of (i) models using ΔMNBI (slope) + impedance measured just below the UES (intercept) and (ii) models using discrete impedance measured at different individual channels. The Akaike information criterion was used to assess the comparative strength of these models.

RESULTS

Overall, we enrolled 208 participants, including 141 (68%) women, with a median age of 63 years (interquartile range [IQR]: 48–70), body mass index 26 kg/m² (IQR: 23–30), and RSI 17.5 (IQR: 12–25) (see Supplementary Table, Supplementary Digital Content, http://links.lww.com/CTG/B470). Lower distal MNBI (channel 4) significantly correlated with greater distal reflux, including acid exposure time (AET), distal acid reflux events, bolus exposure time, and total reflux events (Table 1). Proximal MNBI (channels 1 and2) was significantly associated with proximal AET and proximal acid reflux events. When reflux parameters were dichotomized, distal MNBI was associated with AET and bolus exposure time, whereas proximal MNBI correlated with proximal reflux events (Table 1). On evaluation of symptoms, lower proximal MNBI predicted higher scores on the RSI esophageal subscale (r = −0.16, P = 0.04) and total RSI (r = −0.14, P = 0.08). Proximal impedance contour models accounting for both ΔMNBI and MNBI just below UES generally outperformed (lower Akaike information criterion) models using individual distal or proximal MNBI metrics in predicting proximal or pharyngeal reflux (Table 2).

Table 1.

Association of MNBI at the proximal and distal esophagus with reflux parameters, expressed as continuous and dichotomized measures

Continuous measures	Proximal channels				Distal channels
	Channel 1		Channel 2		Channel 3		Channel 4
	r	P value	r	P value	r	P value	r	P value
Proximal parameters
Proximal AET	−0.18 a	0.01	−0.19 a	0.01	−0.14b	0.06	−0.12	0.10
Proximal acid events	−0.21 a	<0.01	−0.21 a	<0.01	−0.12	0.10	−0.13b	0.06
Proximal reflux events	−0.07	0.32	−0.09	0.24	−0.07	0.35	−0.14b	0.06
Distal parameters
AET	−0.15 a	0.04	−0.13b	0.08	−0.11	0.13	−0.19 a	0.01
Distal acid events	−0.12b	0.09	−0.17 a	0.02	−0.10	0.19	−0.20 a	0.01
BET	−0.11	0.12	−0.06	0.40	−0.15 a	0.04	−0.15 a	0.03
Total reflux events	−0.09	0.21	−0.09	0.23	−0.06b	0.39	−0.16 a	0.02

Dichotomized measures	Proximal channels			Distal channels
	Mean baseline impedance (Ω) among patients with
Reflux parameter	Abnormal reflux burden	Normal reflux burden	P	Abnormal reflux burden	Normal reflux burden	P
Proximal parameters
Proximal reflux events	2,514 a	2,918	0.04	2,133b	2,935	0.07
Distal parameters
AET	2,374b	2,640	0.10	1884 a	2,368	0.01
BET	2,469	2,681	0.14	2009 a	2,482	0.01

Bolded entries indicate those that are statistically significant (p < 0.05). AET, acid exposure time; BET, bolus exposure time.

^aP < 0.05.

^bP < 0.10.

Table 2.

Comparison of performance characteristics in predicting reflux burden between models with individual MNBI measures and the slope-and-intercept model of proximal impedance contour (ΔMNBI and MNBI at channel 1)

Reflux parameters	Ch 1	Ch 2	AIC for traditional MNBI modelsa						AIC for the novel modelb
			Ch 3	Ch 4	Ch 1 and 3	Ch 2 and 3	Ch 1 and 4	Ch 2 and 4	Slope and intercept
Pharyngeal reflux events	253	253	253	251	254	255	252	252	110c
Proximal reflux events	146	148	148	141	148	150	142	143	32c
Proximal AET	−7.3	−7.8	−4.6	−3.7	−5.8	−6.2	−6.0	−6.4	61
Proximal acid events	789	789	795	794	791	791	790	790	366c

AET, acid exposure time; AIC, Akaike information criterion; BET, bolus exposure time; MNBI, mean nocturnal baseline impedance.

^aAIC provided for models containing MNBI from individual channels.

^bAIC provided for the model containing estimates for change in MNBI across channels (ΔMNBI) and MNBI just below the upper esophageal sphincter (channel 1).

^cNovel model outperforms traditional MNBI models (ΔAIC > 10).

DISCUSSION

LPRD remains a diagnostic challenge as standard objective measures inconsistently correlate with pharyngeal reflux exposure and symptoms. Traditional reflux monitoring is also limited by day-to-day variability of reflux, which may be even more pronounced in the LPRD population (7). BI is an emerging metric that reflects esophageal mucosal integrity and has been shown to predict acid exposure and symptom improvement after antireflux therapy (3,8). The US Food and Drug Administration has approved an endoscopic device that measures real-time esophageal impedance for clinical use (9). Despite the growing literature on esophageal impedance, previous studies have focused on distal esophageal pathology (e.g., typical gastroesophageal reflux disease [GERD], eosinophilic esophagitis). We found that proximal MNBI significantly correlated with abnormal proximal reflux and more severe symptoms in patients presenting with LPS. Similarly, distal MNBI predicted distal reflux exposure. These relationships were observed for both acid and nonacid reflux. We also developed a novel model that accounted for both the pattern of BI from the proximal to distal channels (slope) and the BI at 1–3 cm from the UES (intercept). This model of proximal impedance contour showed promise in predicting abnormal proximal reflux, outperforming models using MNBI measured across individual channels alone.

Previous studies have correlated single measures of MNBI with reflux burden, biomarkers such as salivary pepsin, airway disease severity, and symptom response to proton pump inhibitor in patients with suspected extraesophageal reflux (10–15). Our findings suggest not only proximal BI as a potentially valuable tool in evaluating patients with LPS but also that the pattern of BI contour might more comprehensively reflect proximal reflux burden than isolated impedance measured at individual locations. This concept is similar to the slope-and-intercept model developed for the endoscopic mucosal impedance device, which was shown to predict probability of GERD vs eosinophilic esophagitis vs non-GERD (4). However, the slope or degree of MI change in the distal esophagus may, in theory, be blunted in patients with significant proximal migration of refluxate. The correlation between distal reflux burden alone and extraesophageal symptoms is often suboptimal (2). Our previous work also implicates the role of proximal esophageal pathology in LPRD (16). Therefore, we examined the concept of a proximal esophageal impedance contour model that may more closely reflect the dynamic physiological changes along the esophagus associated with proximal reflux exposure, even to the level of the hypopharynx. Pragmatically, our exploratory findings infer a potential extension of the endoscopic mucosal impedance device to evaluate extraesophageal reflux.

Our study has limitations. Given our cross-sectional data, we could not ascertain if proximal BI predicts longitudinal outcomes. Whether proximal impedance simply serves as a marker of proximal reflux exposure or whether impaired mucosal integrity itself leads to LPS is unclear. However, our exploratory findings demonstrated promise in a novel application of proximal impedance metrics to assess extraesophageal reflux.

In conclusion, we demonstrated that proximal BI measures, particularly with a slope-and-intercept contour model, correlate with proximal esophageal reflux burden in patients presenting with LPS. These exploratory findings implicate potential for a novel application of the endoscopic mucosal impedance device in extraesophageal reflux. Further studies are needed to prospectively evaluate the clinical utility of proximal impedance contour metrics for assessing LPS/LPRD and predicting response to therapy.

CONFLICTS OF INTEREST

Guarantor of the article: Walter W. Chan, MD, MPH, FACG.

Specific author contributions: D.R.S. and W.W.C. initiated study concepts and design. D.R.S., J.X.C., R.L., and W.W.C. contributed to acquisition of data. D.R.S. and W.W.C. performed analysis and interpretation of data. D.R.S. and W.W.C. drafted the manuscript. D.R.S., J.X.C., R.L., T.L.C., and W.W.C. contributed to critical revision of manuscript for important intellectual content. W.W.C. provided administrative support and overall study supervision.

Financial support: None to report.

Potential competing interests: None to report.

Supplementary Material

ct9-17-e00994-s001.pdf ^{(143.8KB, pdf)}

ABBREVIATIONS:

AET

acid exposure time

BI

baseline impedance

HEMII-pH

hypopharyngeal-esophageal multichannel intraluminal impedance-pH testing

IQR

interquartile range

LPRD

laryngopharyngeal reflux disease

LPS

laryngopharyngeal symptoms

MNBI

mean nocturnal baseline impedance

RSI

reflux symptom index

UES

upper esophageal sphincter