Validity and Reliability of the Filipino Reflux Symptom Index

ARTICLE IN PRESS Validity and Reliability of the Filipino Reflux Symptom Index *José Florencio F. Lapeña Jr., †Giancarla Marie C. Ambrocio, and ‡Ryner Jose D. Carrillo, *†‡Manila, Philippines Summary: Objectives. This study aimed to establish validity and reliability of the Filipino Reflux Symptom Index (FRSI) and to test it among patients with laryngopharyngeal reflux (LPR) before and after 6 months’ trial of rabeprazole. Study Design. A case-control study was carried out. Methods. There were 35 LPR patients and 30 controls who were twice-administered the FRSI and Filipino Voice Handicap Index (FVHI) for test-retest reliability, and videostroboscopy was performed to obtain baseline reflux finding scores (RFSs). Patients took rabeprazole 20 mg twice daily for 6 months. The FRSI and FVHI were readministered a third time, repeat videostroboscopy was performed, and repeat RFS was obtained. Reliability, validity, and internal consistency were computed. Results. A total of 58 participants, 29 patients and controls each, aged 22–65 years completed the study. FVHI 2:1 and FRSI 2:1 significantly correlated with no significant differences between FRSI 2:1. FRSI had good item-total correlations indicating psychometrically sound items. There were significant differences between patients and controls for FRSI scores and mean scores. FRSI 3 scores were significantly lower than FRSI 1 scores, suggesting symptoms improved after treatment. There were no significant differences between RFS 2 and 1. Significant differences between FRSI 3 and 1, but not between FVHI 3 and 1, suggest the FRSI was more sensitive to changes in reflux after 6 months’ intervention than the FVHI. Conclusions. The FRSI is a valid and reliable tool for assessing LPR symptoms and may be used for primary care screening among Filipinos. Initial response to a 2-week empirical proton pump inhibitor trial may support an impression of LPR; non-response warrants specialist referral for further investigation. Key Words: Laryngopharyngeal reflux–Reflux symptom index–Outcome measures–Rabeprazole–Filipino.

INTRODUCTION Laryngopharyngeal reflux (LPR) has a growing prevalence in up to 60% of gastroesophageal reflux disease (GERD) patients,1 which is more common in Asians than previously thought.2 Increasingly recognized by generalists, pulmonologists, and otolaryngologists,1,3 LPR affects 10% of patients consulting the latter.4,5 Together with diet and lifestyle modification, LPR is treated by proton pump inhibitors (PPIs) such as rabeprazole.6,7 However, clinical assessment of LPR may be difficult because laryngeal findings cannot always be reliably determined from clinician to clinician, and such variability may make precise laryngoscopic diagnosis highly subjective.1,8,9 LPR has been diagnosed by the Reflux Symptom Index (RSI)10,11 and reflux finding score (RFS).12 Because it is found in more than 50% of patients with hoarseness,4,13 it might also be diagnosed (albeit indirectly) by the Voice Handicap Index (VHI).14,15 The RSI is a self-administered nine-item instrument Accepted for publication September 13, 2016. Presented at: Philippine Society of Otolaryngology Head and Neck Surgery 59th Annual Convention and 1st Philippine International ORL-HNS Congress “#Head&Neck Be Connected: Otolaryngology in the Age of Information Technology.” Marriott Grand Ballroom, Pasay City, Philippines, December 1, 2015. From the *Professor of Otorhinolaryngology and Attending Otorhinolaryngologist, College of Medicine-Philippine General Hospital, University of the Philippines Manila, Manila, Philippines; †Medical Officer III (Resident Physician), Department of Otorhinolaryngology, Philippine General Hospital, University of the Philippines Manila, Manila, Philippines; and the ‡Associate Professor of Anatomy and Clinical Associate Professor of Otorhinolaryngology, College of Medicine-Philippine General Hospital, University of the Philippines Manila, Manila, Philippines. Address correspondence and reprint requests to José Florencio F. Lapeña, Jr., Department of Otorhinolaryngology, Ward 10, Philippine General Hospital, Taft Avenue, Ermita, Manila 1000 Philippines. E-mail: [email protected] Journal of Voice, Vol. ■■, No. ■■, pp. ■■-■■ 0892-1997 © 2016 The Voice Foundation. Published by Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.jvoice.2016.09.013

to detect and document LPR.10 Valid and highly reproducible,10 it has served as a standard for other studies.16 Resonating the need for culture-specific versions of such measures,17 it has been translated and validated in Hebrew,18 Italian,19 and Greek.20 Both the RSI and VHI have been translated into the Filipino RSI (FRSI)21 and Filipino VHI (FVHI),22 but the former has not been validated. Establishing FRSI validity and reliability may provide a simple screening tool for LPR diagnosis and clinical assessment of therapy that can be used among 100 million Filipinos worldwide, particularly in primary healthcare settings. This study aims to establish the validity and reliability of the FRSI among a sample of LPR patients and controls, and to test the validated FRSI before and after 6 months’ trial of rabeprazole. METHODS With Institutional Review Board approval and informed consent, 35 adult LPR patients and 30 age- and gender-matched controls were prospectively recruited from outpatients at the Philippine General Hospital, Manila, Philippines. Inclusion criteria of patients were as follows: aged 18–70 with LPR diagnosis by current or previous English RSI >13 or documented RFS >7 and ability to accomplish the FRSI unassisted. Patients of the principal investigator, those on PPI therapy within 6 months, or with comorbidities unrelated to LPR that could potentially compromise well-being by study participation were excluded. Antacids, H2 blockers, or prokinetics were ceased 2 weeks before videostroboscopy. Age- and gender-matched controls had no history or diagnosis of LPR or GERD. The study was registered on the Philippine Health Research Registry.

ARTICLE IN PRESS 2

Journal of Voice, Vol. ■■, No. ■■, 2016 Instruksyon: Sa loob ng nakaraang buwan, paano ka naapektuhan ng mga sumusunod na problema? Bilugan lamang ang iyong mga sagot batay sa sumusunod na iskala: 0 = Walang problema. 5 = Matinding problema. Pamamalat/pamamaos o problema sa iyong boses.

0

1

2

3

4

5

Pag-alis ng bara sa lalamunan (hal. pag-ehem).

0

1

2

3

4

5

0

1

2

3

4

5

Hirap sa paglunok/paglulon ng pagkain, inumin, tableta o kapsula.

0

1

2

3

4

5

Pag-ubo matapos kumain o matapos humiga.

0

1

2

3

4

5

Hirap sa paghinga o madalas na nasasamid/nabibilaukan

0

1

2

3

4

5

Pag-ubong nakaaabala o nakapeperwisyo.

0

1

2

3

4

5

Pakiramdam na parang may nakadikit o nakabara sa lalamunan.

0

1

2

3

4

5

Pangangasim/paghapdi ng sikmura, pananakit ng dibdib, hindi natunawan, o pagsikad ng asido mula sa tiyan.

0

1

2

3

4

5

Labis na plema sa lalamunan.

FIGURE 1. Filipino Reflux Symptom Index. The nine-item FRSI was independently thrice translated, and each translation was thrice back-translated by three otolaryngolgists, three non-otolaryngologist physicians, and three lay persons, with the final version approved by the National Commission on Filipino Language21 (Figure 1). The FRSI was pretested for internal consistency (FRSI 1), and FVHI was administered for comparison (FVHI 1). The FRSI and FVHI were repeated after 1–2 weeks for inter-rater test-retest reliability (FRSI 2, FVHI 2), and videostroboscopy was performed by a blinded examiner who obtained baseline reflux finding scores (RFS 1) using a Digital Videostroboscopy System 9295E and Light Source 9100B with 5.8-mm 70° rigid Laryngoscope 9106 (Kay Elemetrics, NJ, USA). Topical Lidocaine 10 mg/dose (10%) spray × 3 (AstraZeneca AB, Sweden) was administered as needed. Videostroboscopic images were captured using integrated system software, and the findings were scored by the same examiner and tabulated by an encoder. The 35 patients took rabeprazole 20 mg twice daily for 6 months, with standard behavior modification instructions. After 6 months, FRSI and FVHI were administered a third time (FRSI 3, FVHI 3), repeat videostroboscopy was performed, and repeat RFS was obtained (RFS 2). Data were encoded using Microsoft Excel 2010 Version 14 (Microsoft, Redmond, Washington). Only participants with complete data were included in the final sample for statistical analysis. Test-retest reliability and validity indices (Pearson r; Student t

test) and internal consistency (Cronbach alpha [α]) were computed using STATISTICA 12, 64-bit version (Dell StatSoft, Inc., Tulsa, OK). RESULTS Although 35 patients and 30 controls serially met inclusion criteria and consented to participate, 5 patients did not complete the study, and no controls were matched for them, whereas data for 1 patient were incomplete. Eliminating the matched control for this patient yielded 58 participants (29 patients and 29 controls; 7 males and 22 females each), aged 22–65 years old (mean 41.7, SD = 12.1). Nine misclassified patients (low FRSI 1 scores) and 1 misclassified control (high FRSI 1 score) were excluded post hoc from LPR vs control analysis but included in FRSI analysis before treatment. After matching remaining patients and controls, there were 20 matched pairs for each comparison of FRSI and FVHI scores. No adverse events or drug reactions were encountered during the study period. Reliability assessment Test-retest reliability (Pearson r) showed FRSI 2 scores significantly correlated with FRSI 1 scores; r(48) = .92, P < 0.001, demonstrating FRSI is a reliable test. The FVHI 2 moderately correlated with FVHI 1; r(48) = .66, P < 0.001, demonstrating the FVHI may also be reliable. There were no significant

ARTICLE IN PRESS José Florencio F. Lapeña, Jr., et al

3

Validity and Reliability of the FRSI

TABLE 1. Internal Consistency of the Filipino Reflux Symptom Index Item in RSI Hoarseness or a problem with your voice Clearing your throat Excess throat mucus Difficulty swallowing food, liquids, or pills Coughing after eating or lying down Breathing difficulties or choking episodes Troublesome or annoying cough Globus sensation or sensation of something sticking in your throat or a lump in your throat Heartburn, chest pain, indigestion, or stomach acid coming up

Item in Filipino RSI

Mean

Standard Deviation

Item-Total r Value

Ranking

Pamamalat o pamamaos o problema sa iyong boses Pag-alis ng bara sa lalamunan (hal. pag-ehem) Labis na plema sa lalamunan Hirap sa paglunok o paglulon ng pagkain, inumin, tableta, o kapsula Pag-ubo matapos kumain o matapos humiga Hirap sa paghinga o madalas na nasasamid o nabibilaukan Pag-ubong nakaaabala o nakapeperwisyo Pakiramdam na parang may nakadikit o nakabara sa lalamunan

0.92

1.23

0.76

7th

1.66

1.76

0.91

2nd

1.54 1.04

1.80 1.54

0.86 0.80

4th 5th

0.78

1.20

0.72

8th

0.96

1.38

0.77

6th

1.00

1.34

0.69

9th

1.98

2.04

0.94

1st

1.80

1.85

0.88

3rd

Pangangasim o paghapdi ng sikmura, pananakit ng dibdib, hindi natunawan, o pagsikad ng asido mula sa tiyan

differences between FRSI 2 and FRSI 1 on non-directional t test; t(49) = 1.91, P < 0.063. On average, patients scored 11.68 on FRSI 1 and 10.46 on FRSI 2, showing FRSI test-retest reliability. Cronbach α for internal consistency showed the nine-item FRSI is reliable (mean = 11.68, SD = 11.59, Cronbach α = .94, SEM = 2.91, n = 9). Item analysis showed good item-total correlation, indicating the FRSI has psychometrically sound items. The top three “best” FRSI items based on item-total r values were “globus sensation or sensation of something sticking in your throat or a lump in your throat,” r = .94; “clearing your throat,” r = .91; and “heartburn, chest pain, indigestion, or stomach acid coming up,” r = .88. Intermediate items were “excess throat mucous,” r = .86; “difficulty swallowing food, liquids, or pills,” r = .80; and “breathing difficulties or choking episodes,” r = .77. The bottom three items were “hoarseness or a problem with your voice,” r = .76; “coughing after eating or lying down,” r = .72; and “troublesome or annoying cough,” r = .69 (Table 1). Validity assessment There were significant differences between patients and controls for both FRSI and FVHI: (FRSI 1 LPR mean = 24.1, SD = 6.5; FRSI 1 control mean = 2.69, SD = 3.16; FVHI 1 LPR mean = 22.86, SD = 20.07; FVHI 1 control mean = 2.9, SD = 4.85). Validity indices (t test, using separate variance estimates) also showed significant differences in mean scores of patients and controls for both FRSI and FVHI (FRSI 1 LPR (n = 20) vs FRSI 1 control (n = 20), t(28.81) = 13.46, P < 0.001; FVHI 1 LPR (n = 20) vs FVHI 1 control (n = 20), t(22.3) = 3.98, P < 0.001). FRSI 3 t test scores were significantly lower than FRSI 1 scores, meaning symptoms improved after 6 months’ treatment with rabeprazole (FRSI 1 LPR mean = 24.50, SD = 6.44;

FRSI 3 LPR mean = 17.50, SD = 11.76). However, FVHI 3 scores were not significantly different from FVHI 1 scores (FVHI 1 LPR mean = 21.45, SD = 19.50; FVHI 3 LPR mean = 22.65, SD = 26.78). Validity indices using t test for dependent samples further showed significant differences between FRSI 1 LPR vs FRSI 3 LPR (t(19) = 3.22, P < 0.004), but no significant differences between FVHI 1 LPR vs FVHI 3 LPR (t(19) = −0.21, P < 0.837). That FRSI scores significantly differed after 6 months with no significant differences in FVHI scores suggest the FRSI may have been more sensitive to reflux changes than the FVHI. Additionally, significant correlations were obtained for FRSI 1 and FVHI 1 with rho(28) = .719; FRSI 1 with FRFS 2 with rho(28) = −.766, which meant that as FRSI 1 scores increased, FRFS 2 scores decreased; FRSI 2 with FVHI 1 (rho = .567) and with FVHI 2 (rho = .537); FRSI 3 with FVHI 3 (rho = .652) and with FRFS 2 (rho = −.428). Also FVHI 1 was significantly correlated with FRFS 2 with rho(28) = −.598, and the correlation between FRSI 3 and FRFS 2 was significant and in the negative direction (rho(28) = −0.766*); that is, as FRSI 3 scores increased, FRFS 2 scores decreased. This set of significant intercorrelations provides evidence supporting the construct validity of FRSI. DISCUSSION This study established the validity and reliability of the FRSI for assessing symptoms of LPR among our participants, consistent with previous studies.18–20 The FRSI is test-retest and psychometrically reliable based on internal consistency evidence, with good item-total correlational values indicating psychometrically sound items. Our Pearson test-retest reliability correlation coefficient (r) of .92 (P < 0.001) was the same as that obtained by Cohen et al18 and

ARTICLE IN PRESS 4 comparable to r > .90 of Schindler et al.19 Although Printza et al20 used different parameters (Spearman-Brown; Guttman splithalf reliability coefficients), they also obtained adequate testretest reliability. To be psychometrically reliable, Cronbach α should be ≥.90. Our α was .94 compared to the extremely high α = .99 of Schindler et al.19 Printza et al20 used an expanded RSI with a 10th item (“throat pain”) and obtained two groups of factors with relatively high but varying Cronbach α values—Factor 1: breathing difficulties or choking, coughing after eating or lying down, troublesome cough, difficulty in swallowing, and heartburn or indigestion (Cronbach α = .765); and Factor 2: throat clearing, globus, postnasal drip, and hoarseness or voice disorder (Cronbach α = .842). Belafsky et al10 originally obtained r = .81, which may be considered high even if it did not reach threshold value of .90. Our top “best” item was “globus, sensation of something sticking in the throat or a lump in the throat.” However, “globus pharyngeus” was only the fourth most common LPR symptom (tied with hoarseness at 95%) in a survey of American Bronchoesophagological Association members.23 “Foreign body sensation in the throat” was the third most common symptom (94.9%) found by De la Iglesia et al.24 “Throat-clearing” was our second best item, consistent with Koufman’s13 findings in 87% of LPR patients compared to 3% of GERD patients—but was the most common symptom (98%) surveyed by Book et al.23 Interestingly, our third best item was “heartburn, chest pain, indigestion or stomach acid coming up” as LPR is not usually associated with heartburn and regurgitation, differentiating it from GERD.4 Indeed, Koufman13 reported only 20% of LPR patients compared to 83% of GERD patients complained of heartburn, and the LPR statement of the American Academy of Otolaryngology Head and Neck Surgery (AAOHNS) Committee on Speech, Voice, and Swallowing Disorders emphasized the most significant difference between LPR and GERD is that most LPR patients do not have esophagitis or heartburn—the diagnostic sine qua non of GERD.5 More interestingly, our bottom three items were “hoarseness or a problem with your voice,” “coughing after eating or lying down,” and “troublesome or annoying cough” contra findings of Book et al23 that “persistent cough” and “hoarseness” (tied with globus pharyngeus) were second and fourth most common symptoms of LPR (97% and 95%, respectively), and De la Iglesia et al,24 where “loss of voice strength” and “persistent cough” were the most common symptoms (98.3% and 96.6%, respectively), with “change in voice tone” the fourth most common symptom at 94.9%. Perhaps Morice’s observation that LPR symptoms parallel those of GERD (in patients with chronic cough) validly questions “whether the reflux of stomach contents into the esophagus alone is sufficient to cause cough or whether reflux into the upper airway is required.”25 Although previous studies have shown LPR is found in 50% of patients with hoarseness,4 our findings suggest hoarseness may not conversely be found in as many patients with LPR. Our findings resonate those of Cathcart et al26 whose comparison of non–voice-related throat symptoms using the RSI vs Glasgow and Edinburgh Throat Scale revealed two similar domains of “coughing and blockage” and “globus or postnasal drip or throat-clearing.” They parallel two factor clusters (“breath-

Journal of Voice, Vol. ■■, No. ■■, 2016

ing difficulties or choking, coughing after eating or lying down, troublesome cough, difficulty in swallowing, and heartburn or indigestion” and “throat clearing, globus, postnasal drip, and hoarseness of voice”) of Printza et al20 except for “heartburn or indigestion” and “hoarseness or voice disorder.” Interestingly, Cathcart et al26 found the “heartburn or reflux” RSI item also mapped poorly (as did throat symptoms), concluding overreliance “on the presenting pattern of throat symptoms” and “a need to revisit the traditional clinical classification of throat symptoms.” Our high rating for “heartburn, chest pain, indigestion or stomach acid coming up” and low rating for “hoarseness or a problem with your voice,” “coughing after eating or lying down,” and “troublesome or annoying cough” suggest that findings of Cathcart et al26 and Printza et al20 warrant further consideration. In terms of validity, significant differences in FRSI scores between patients and controls, and between pre- and posttreatment scores demonstrate ability to detect LPR and reflect reflux changes. These echo findings of Belafsky et al10 that the RSI score of untreated LPR patients was significantly higher than controls (21.2 vs 11.6; P < 0.001). Whereas there were significant FVHI differences between patients and controls, pre- and posttreatment scores were not significantly different. This is not surprising, as the FVHI was not designed to measure LPR, and although it can discriminate patients from controls based on voice-related throat symptoms, it may not measure other symptom clusters. That voicerelated symptoms such as hoarseness were not prominent in our sample might further explain why FVHI may have been less sensitive to reflux changes. Our findings did not establish good FRSI and RFS correlation. Neither did a randomized placebo-controlled trial by Lam et al7 that showed no significant differences in RFS between treatment and placebo groups despite significant decrease in treatment group RSI scores vs placebo. Post hoc logistic regression analysis of each RFS item vs FRSI >13 labeled as LPR in our study yielded only two good items: posterior commissure hypertrophy (odds ratio 3.20, P = 0.01) and vocal fold edema (odds ratio 2.51, P = 0.04). Schindler et al19 obtained high combined correlations between RFS and RSI total scores but poor correlations between individual RSI items and total RFS scores, especially with the first three items: “hoarseness or a problem with your voice” (.11), “excess throat mucus or postnasal drip” (.20), “clearing your throat” (.24). Perhaps, as they suggest, “no single RSI symptom can be associated with LPR,” and correlation with RFS increases “as the presence of different symptoms increased.” As this study never aimed to validate the RFS, we did not perform inter- and intra-rater testing for internal validity. RFS was obtained without benefit of clinical history by a single examiner who may have been influenced by an outpatient clinic context where patients have videostroboscopy indications. Although the RFS has been posited to diagnose LPR,12 such diagnosis is not simple, nor universally agreed upon. A study among general otolaryngologists to determine whether RFS was influenced by reflux symptoms by Chang et al27 could not demonstrate its reliability and objectivity in diagnosing reflux. According to Simpson and Rosen28 laryngeal signs in normal subjects that are more likely to be LPR-related include

ARTICLE IN PRESS José Florencio F. Lapeña, Jr., et al

Validity and Reliability of the FRSI

erythema or edema of the true or false vocal folds, posterior commissure (all similar to our RFS findings) and posterior pharyngeal wall, posterior cricoid wall erythema, and frank reflux. A review of laryngopharyngeal mucosal signs in diagnosing LPR29 found that multiple signs may improve detection of reflux sufferers from asymptomatic controls, with Reinke edema (similar to our RFS findings) and pseudosulcus more prevalent in individuals with pH-proven LPR and those symptomatic of LPR. In addition, ventricular obliteration, vocal fold nodules, and granulomas were more prevalent in pH-proven LPR, whereas interarytenoid thickening (again, similar to our RFS findings) was more prevalent in those symptomatic of LPR. The study showed RFS sensitivity of 87.8% and specificity of 37.5% for detecting pH-proven LPR.29 Laryngeal visualization should scrutinize the posterior larynx (erythema, edema, vocal fold granuloma) and look for vocal fold edema (pseudosulcus) and blunting of supraglottic cartilage outlines.28 Like a positive Purified Protein Derivative (PPD) test,28 pachydermia may not change over time despite LPR therapy.30 Post hoc stepwise logistic regression analysis at accepted significant P value of 0.2 showed the most likely RFS findings to improve with FRSI symptoms in our study were arytenoid erythema or hyperemia (P = 0.081) and thick endolaryngeal mucus (P = 0.158). LPR diagnosis is “not simple” or “universally agreed upon” and involves “more art and less science.”28 It “relies on putting together many pieces of the puzzle” and needs “the mind to be thoughtful and open.”28 The high sensitivity and low specificity of the RSI accommodate scores >20 or <10, but scores between 10 and 20 pose problems in interpretation. Pace Simpson and Rosen,28 two scores of 5 and one score of 4 yield a score of 14 in Allergic Rhinitis (hoarseness or a problem with your voice, 5; clearing your throat, 4; excess throat mucous or postnasal drip, 5) or Paradoxical Vocal Fold Motion Disease (breathing difficulties or choking episodes, 5; troublesome or annoying cough, 5; sensations of something sticking in your throat or a lump in your throat, 4)!28 A good history can never be replaced by and should complement any tool such as the FRSI. Empirical PPI acid suppression is frequently used in initial diagnosis.1,4,8,9,31,32 The mainstay of treatment is at least 3 months’ medication, and a majority of LPR patients require at least twicedaily dosing.5,7,31,33 Although most report significant symptomatic relief within 2–4 months,34 twice-daily PPI treatment periods of 3,7,35 4,31 and 65 months (adopted in our study) have been recommended for laryngeal findings to resolve. Since completion of our study, increasing concerns with vitamin B12 deficiency, osteoporosis, infections, chronic kidney disease, dementia, and myocardial infarction behoove us to reconsider PPI prescribing patterns.36 Fortunately, Simpson and Rosen28 opine that an initial 2-week empirical PPI trial should suffice to confirm LPR diagnosis and non-response indicates the need for otolaryngologic referral to search for other pathologies, including laryngoscopy and an RFS. Kim et al6 demonstrated 2 weeks as the optimal duration for a rabeprazole 20-mg twice-daily empirical trial in patients with GERD-related non-cardiac chest pain. Although empirical PPI therapy is widely accepted as diagnostic and therapeutic for LPR, further research is needed to establish practice guidelines.37 Until

5

then, it may be prudent to err on the side of caution and refer patients for laryngoscopy if they do not respond to an initial 2-week empirical trial, as the “potentially serious consequences of under-treated reflux”38 must be weighed against “the importance of other entities that may be missed when they are misdiagnosed as reflux,”38 and PPIs are not nearly as benign as we once thought them to be. This study is not without its limitations, which include classifying participants based on test variables (FRSI and RSI) rather than an external gold standard. We initially assigned patients and controls based on LPR diagnosis by present or previous English RSI >13 or documented RFS >7 but failed to reclassify them after obtaining baseline FRSI 1 scores. This methodological flaw was overcome post hoc by excluding nine misclassified LPR patients (with low FRSI 1 scores) and one misclassified control (with high FRSI 1 score) from LPR vs control data analysis, although we included them in FRSI analysis before treatment. Our sample size and characteristics limit generalizability of our results, which do not apply to other populations. Although Filipino is the national language, there are 187 distinct languages in the Philippines, and not all speakers of these languages are equally fluent in Filipino (largely based on Tagalog). Further tests with larger samples may analyze correlation between one, the sum of two, and three (or more) FRSI items with total RFS scores, as well as correlation between one, the sum of two, and three (or more) RFS items with total FRSI scores, extending analyses of Schindler et al.19 Future RFS studies with multiple readers may establish inter- and intra-rater validity. Although such studies must be “excellently-designed, with rigorous inclusion criteria, involving close collaboration among laryngologists, gastroenterologists, research scientists, and reflux surgeons,”38 comparison with combined multichannel intraluminal impedance and pH monitoring is not feasible in low- and middle-income country settings such as ours, and true external validation of the FRSI and RFS, although ideal, are not possible. Nevertheless, we recommend further FRSI translations into our major language groups. CONCLUSIONS In summary, the FRSI is a valid and reliable tool for assessing LPR among Filipinos. It can discriminate between patients and controls, and is sensitive to changes in reflux following therapy. We recommend clinicians use the FRSI with a good history for primary care screening of LPR. Initial response to a 2-week empirical PPI trial may support an impression of LPR, but nonresponse warrants otolaryngologist referral for further investigation. Acknowledgements We acknowledge HI-Eisai Philippines Inc. (Study Grant Number: PRT-IIS-MO63-601) for funding this study and Ma. Angeles G. Lapeña, BS, MA, for the psychometric and statistical analysis. REFERENCES 1. Martinucci I, de Bortoli N, Savarino E, et al. Optimal treatment of laryngopharyngeal reflux disease. Ther Adv Chronic Dis. 2013;4:287–301.

ARTICLE IN PRESS 6 2. Goh KL. Changing epidemiology of gastroesophageal reflux disease in the Asian-Pacific region: an overview. J Gastroenterol Hepatol. 2004;19(suppl 3):S22–S25. 3. Richter JE. Extraesophageal presentations of gastroesophageal reflux disease: an overview. Am J Gastroenterol. 2000;95(suppl 8):S1–S3. 4. Ford CN. Evaluation and management of laryngopharyngeal reflux. JAMA. 2005;294:1534–1540. 5. Koufman JA, Aviv JE, Casiano RR, et al. Laryngopharyngeal reflux: position statement of the committee on speech, voice, and swallowing disorders of the American Academy of Otolaryngology-Head and Neck Surgery. Otolaryngol Head Neck Surg. 2002;127:32–35. 6. Kim JH, Sinn DH, Son HJ, et al. Comparison of one-week and two-week empirical trial with a high-dose rabeprazole in non-cardiac chest pain patients. J Gastroenterol Hepatol. 2009;24:1504–1509. 7. Lam PK, Ng ML, Cheung TK, et al. Rabeprazole is effective in treating laryngopharyngeal reflux in a randomized placebo-controlled trial. Clin Gastroenterol Hepatol. 2010;8:770–776. 8. Abou-Ismail A, Vaezi MF. Evaluation of patients with suspected laryngopharyngeal reflux: a practical approach. Curr Gastroenterol Rep. 2011;13:213–218. 9. Vaezi MF. Extraesophageal manifestations of gastroesophageal reflux disease. Clin Cornerstone. 2003;5:32–38. 10. Belafsky PC, Postma GN, Koufman JA. Validity and reliability of the reflux symptom index (RSI). J Voice. 2002;16:274–277. 11. Belafsky PC, Rees CJ. Laryngopharyngeal reflux: the value of otolaryngology examination. Curr Gastroeterol Rep. 2008;10:278–282. 12. Belafsky PC, Postma GN, Koufman JA. The validity and reliability of the reflux finding score (RFS). Laryngoscope. 2001;111:1313–1317. 13. Koufman JA. The otolaryngologic manifestations of gastroesophageal reflux disease (GERD): a clinical investigation of 225 patients using ambulatory 24-hour pH monitoring and an experimental investigation of the role of acid and pepsin in the development of laryngeal injury. Laryngoscope. 1991;101(suppl 53, 4 pt 2):1–78. 14. Jacobson BH, Johnson A, Grywalski C, et al. The Voice Handicap Index (VHI): development and validation. Am J Speech Lang Pathol. 1997;6: 66–70. 15. Roy N, Gray SD, Simon M, et al. An evaluation of the effects of two treatment approaches for teachers with voice disorders: a prospective randomized clinical trial. J Speech Lang Hear Res. 2001;44:286–296. 16. Andersson O, Ryden A, Ruth M, et al. Validation of the Swedish translation of LPR-HRQL. Med Sci Monit. 2010;16:CR480–CR487. 17. Wong BC, Kinoshita Y. Systematic review on epidemiology of gastroesophageal reflux disease in Asia. Clin Gastroenterol Hepatol. 2006;4:398–407. 18. Cohen JT, Gil Z, Fliss DM. The reflux symptom index—a clinical tool for the diagnosis of laryngopharyngeal reflux. Harefuah. 2005;144:826–829, 912. 19. Schindler A, Mozzanica F, Ginocchio D, et al. Reliability and clinical validity of the Italian Reflux Symptom Index. J Voice. 2010;24:354–358. [Epub 2009 Mar 20]. 20. Printza A, Kyrgidis A, Oikonomidou E, et al. Assessing laryngopharyngeal reflux symptoms with the Reflux Symptom Index: validation and prevalence

Journal of Voice, Vol. ■■, No. ■■, 2016

21. 22.

23. 24.

25. 26.

27. 28.

29.

30.

31.

32.

33. 34.

35.

36. 37.

38.

in the Greek population. Otolaryngol Head Neck Surg. 2011;145:974–980. [Epub 2011 Oct 10]. Lapeña JF. Filipino Reflux Symptom Index (Unpublished). 2012. University of the Philippines Manila. Lim ACE, Hernandez ML, Llanes EGDV. Measurement of the handicap of dysphonic patients using the Filipino Voice Handicap Index. Philipp J Otolaryngol Head Neck Surg. 2010;25:7–11. Book DT, Rhee JS, Toohill RJ, et al. Perspectives in laryngopharyngeal reflux: an international survey. Laryngoscope. 2002;112(8 pt 1):1399–1406. De la Iglesia FV, Gonzalez SF, Gomez Mde L. Laryngopharyngeal reflux: correlation between symptoms and signs by means of clinical assessment questionnaires and fibroendoscopy. Is this sufficient for diagnosis? Acta Otorinnolaringol Esp. 2007;58:421–425. Morice AH. Is reflux cough due to gastroesophageal reflux disease or laryngopharyngeal reflux? Lung. 2008;186(suppl 1):S103–S106. Cathcart RA, Steen N, Natesh BG, et al. Non-voice-related throat symptoms: comparative analysis of laryngopharyngeal reflux and globus pharyngeus scales. J Laryngol Otol. 2011;125:59–64. Chang B, MacNeil D, Morrison M, et al. The reliability of the reflux finding score among general otorhinolaryngologists. J Voice. 2015;29:527–577. Simpson CB, Rosen CA. Case-Based Approach to Voice Disorders: Pitfalls and Pearls. (PowerPoint Presentation). AAO-HNSF 2015 Annual Meeting and OTO Expo, 2015, Dallas, Texas, USA. Powell J, Cocks HC. Mucosal changes in laryngopharyngeal reflux—prevalence, sensitivity, specificity and assessment. Laryngoscope. 2013;123:985–991. Hill RK, Simpson CB, Velazquez R, et al. Pachydermia is not diagnostic of active laryngopharyngeal reflux disease. Laryngoscope. 2004;114:1557– 1561. Park W, Hicks DM, Khandwala F, et al. Laryngopharyngeal reflux: prospective cohort study evaluating optimal dose of proton-pump inhibitor therapy and pretherapy predictors of response. Laryngoscope. 2005;115:1230–1238. Altman KW, Prufer N, Vaezi MF. A review of clinical practice guidelines for reflux disease: toward creating a clinical protocol for the otolaryngologist. Laryngoscope. 2011;121:717–723. Remacle M, Lawson G. Diagnosis and management of the laryngopharyngeal reflux disease. Curr Opin Otolaryngol Head Neck Surg. 2006;14:143–149. Belafsky PC, Postma GN, Koufman JA. Laryngopharyngeal reflux symptoms improve before changes in physical findings. Laryngoscope. 2001;111:979– 981. Aviv JE, Liu H, Parides M, et al. Laryngopharyngeal sensory deficits in patients with laryngopharyngeal reflux and dysphagia. Ann Otol Rhinol Laryngol. 2000;109:1000–1006. Mössner J. The indications, applications and risks of proton pump inhibitors: a review after 25 years. Dtsch Arztebl Int. 2016;113:477–483. Campagnolo AM, Priston J, Thoen RH, et al. Laryngophryngeal reflux: diagnosis, treatment and latest research. Int Arch Otorhinolaryngol. 2014;18:184–191. Sataloff RT, Hawkshaw MJ, Gupta R. Laryngopharyngeal reflux and voice disorders: an overview on disease mechanisms, treatments, and research advances. Discov Med. 2010;10:213–224.