Compliance With Anti-Reflux Therapy in Patients With Otolaryngologic Manifestations of Gastroesophageal Reflux Disease^†

INTRODUCTION

Gastroesophageal reflux has been defined as the retrograde flow of gastric contents into the esophagus. This refluxate is responsible for the symptoms associated with gastroesophageal reflux disease (GERD). Many patients have the classic symptoms of heartburn and regurgitation. A smaller group of patients present with primarily otolaryngologic manifestations of GERD, which has been termed laryngopharyngeal reflux (LPR).¹ These patients may not have any of the classic symptoms of GERD and may in fact have normal esophagrams. Their symptoms include hoarseness, cough, frequent throat clearing, sensation of postnasal drip, and globus and their physical findings range from postcricoid edema and erythema to laryngeal findings such as vocal cord edema. Many patients with LPR are treated successfully with dietary restrictions, behavioral modifications, and drug therapy. The objective of this study is to evaluate patient compliance with all three aspects of antireflux therapy and to evaluate how patient compliance with these factors correlates with their symptomatic improvement.

MATERIALS AND METHODS

Twenty-nine consecutive patients who presented to the otolaryngology clinic at the Manhattan Veterans Affairs Medical Center with LPR were prospectively evaluated. These patients had a history of at least 3 months of one or more of the following symptoms; hoarseness, sensation of increased phlegm, chronic throat clearing, chronic cough, and globus pharyngeus. The patients underwent a complete head and neck examination including flexible laryngoscopy. The patients had no other apparent cause of their symptoms other than LPR. Barium esophagrams were performed. Each patient was given a standardized GERD diet and behavior modification form that was discussed extensively with the patient. Medical therapy included either an H₂ blocker (cimetidine, 400 mg BID) or omeprazole (20 mg every night). The most frequently prescribed drug was cimetidine (67%), which reflected the particular pharmacy requirements at the Veterans Affairs Medical Center. Patients were examined at 6- to 8-week intervals.

After 2 to 6 months of treatment, patients were given a questionnaire to assess improvement of symptoms as well as compliance with nine different parameters of antireflux therapy, including drug therapy and behavioral modification (Table I). The responses for each patient were assigned a numerical value and analyzed for trends. Symptomatic improvement as reported in the top question of Table I was quantified by assigning 0 to “none,” 1 to “a little better,” 2 to “a lot better,”and 3 to “gone.” Compliance with recommendations was quantified by assigning numbers to the responses to the rest of the questions in as follows: 0 to “never,” 1 to “sometimes,”and 2 to “always.” With this quantification schema, statistical analysis using Pearson's correlation coefficient was performed comparing the stated compliance with each question to the stated improvement on the same questionnaire.

Table TABLE I.. Patient Questionnaire.

Questions 2 to 5 pertained to positional habits that were combined and analyzed as a “position score.” Questions 6 to 9 pertained to food habits and were also combined and analyzed as a “food score.” Finally all the compliance parameters for each patient were added together to obtain a measure of overall compliance and analyzed as a “total score.” The degree of symptomatic improvement was correlated with compliance for each of the nine parameters, the position score, food score, and total score. Statistical analysis was conducted for each using Pearson's correlation coefficient.

RESULTS

Twenty-nine patients began this study but 4 patients did not have consistent follow-up and were excluded. One patient was found to have a Zenker's diverticulum on barium esophagram and was also excluded. Twenty-four patients were followed for 4 to 6 months. All the patients were elderly men, reflecting the population at the Veterans Administration Hospital. The average age of the study population was 64 years (range, 48-80 y). The most frequent symptoms reported were hoarseness (83%), phlegm (67%), throat clearing (58%), and sore throat (50%). Most patients reported three or more symptoms concomitantly.

Head and neck physical findings in 11 of the patients (46%) were considered to be unremarkable. Posterior commissure edema and erythema were the most common findings in the remaining patients. Twenty patients underwent barium esophagram examinations. Nine of these studies (45%) demonstrated reflux, 1 revealed a hiatal hernia without reflux, and 10 examinations (50%) were normal. The average length of treatment was 4.17 months (range, 2-6 mo). Nineteen of the 24 patients (80%) reported some improvement in symptoms and eight (33%) reported moderate improvement or resolution on treatment.

Each treatment modality was correlated with the degree of symptomatic improvement. Pearson correlation coefficient analysis (Table II) revealed that avoidance of eating and drinking 2 or more hours before bed time and elevation of the head of the bed had the most significant influence on improvement of LPR symptoms (P < .05; Table II) A Pearson's coefficient (r) greater than 0.413 was statistically significant at P < .05. The combination of avoidance of prone position before bed, avoidance of eating or drinking before bed, coupled with elevation of the head of bed (the position score) had a greater impact on improvement of symptoms than modification of food habits or compliance with medications. Figure 1 shows a graphic comparison of food, position, and medication as independent interventions. The steeper the slope of the best fit line (the slope equals the Pearson's correlation coefficient), the more highly correlated the symptomatic improvement was to the compliance with the specific intervention (position, food, or medication).

Table TABLE II.. Pearson's Correlation Coefficient Analysis (r) for Symptomatic Improvement Versus Patient Compliance As Indicated by Patient Responses to Questionnaire.

Only 12 of the patients (50%) reported taking their medications daily as prescribed. Compliance with the medication schedule contributed to the improvement of symptoms, but was not statistically significant. A meaningful comparison of cimetidine to omeprazole could not be accomplished with our study because too few patients received omeprazole.

DISCUSSION

Gastroesophageal reflux disease affects 25 to 75 million people in the United States and is associated with esophageal, laryngopharyngeal, and pulmonary symptoms.¹ Patients who present to otolaryngologists do not usually complain of the more typical symptoms of heartburn and regurgitation, but more commonly of hoarseness, chronic cough, excess phlegm production, throat clearing, and globus pharyngeus.^1-6 Laryngoscopic examination is variable and may include vocal fold edema and erythema and postcricoid edema and erythema. Sequelae of LPR may include vocal fold contact ulcers and granulomas, pachydermia larynges, subglottic stenosis, and even carcinoma.^{1, 3, 4, 7, 8}

The diagnostic modalities frequently used to test for LPR include radiographic examinations, ambulatory pH probe monitoring, flexible laryngoscopy, and upper endoscopy. Each examination has its advantages and disadvantages and none offer 100% accuracy.^{1, 5} Barium esophagram is probably the most frequently employed test, but definite reflux is demonstrated in only 15% to 25% of patients. In our study, barium esophagram demonstrated reflux in 45% of patients. The flexible laryngoscopic examination may be diagnostic of LPR, but confirmation requires esophagoscopy and biopsy.

The standard treatment for patients with LPR is directed at lifestyle or behavioral modifications, dietary modifications, and the use of H₂ blockers and proton pump inhibitors.¹ Recent studies have shown improvement or resolution of symptoms in 60% to 90% of patients who were treated with these interventions.^{1, 8-11} We saw the greatest symptomatic improvement in those who followed the behavioral modifications of avoiding food before bed time and sleeping with the head of the bed raised. Compliance with medical therapy and dietary restrictions did not have as great a correlation with symptom improvement in our study.

Laryngopharyngeal reflux treatment plans require a significant commitment. Undoubtedly, patients may find it difficult to follow these recommendations regularly, especially those relating to lifestyle and dietary modifications. Compliance with drug therapy may also be a problem: previous studies have shown poor compliance in 10% to 30% of the adult population and 50% to 60% of the elderly.^11-13 In our study of elderly men, we saw a rate of 50% noncompliance with medication. Studies evaluating compliance with antireflux drug therapy for esophagitis in the short term report a mean rate of 90%.¹⁴

There are several possible reasons for poor compliance. Many of our patients were already taking multiple medications. Many elderly have diagnoses of hypertension, diabetes, and coronary artery disease. With these diagnoses come medication regimens and dietary restrictions: low-salt, low-sugar, low-fat, and low-cholesterol restrictions. If one adds LPR dietary restrictions to this list, patients may have the perception that there is nothing left to eat. There is also a perception that these diseases are more serious or life-threatening than LPR, which possibly results in failure of compliance with many reflux recommendations. Additionally, patients are often not fully cognizant of the relation between their food habits and potential otolaryngologic manifestations of reflux to be able to note associations they could potentially correct.

CONCLUSION

In our study we found that compliance with a drug regimen was poor. Although dietary modifications undoubtedly play a role in LPR therapy, elevation of the head of the bed, avoidance of lying down after eating, and avoidance of liquids and food before bed appeared to have a greater impact. In this small study, it would appear that simplifying the treatment regimen to those elements most correlated with symptomatic improvement may increase patient compliance and therefore symptomatic improvement.