Gastroesophageal reflux disease (GERD) is a frequent yet often underrecognized cause of idiopathic hoarseness. Correlating voice-based acoustic changes with objective reflux measures may enable earlier diagnosis through non-invasive screening. This study aimed to develop and validate two diagnostic models for predicting GERD in patients presenting with idiopathic hoarseness, using comprehensive clinical, endoscopic, manometric, and pH-impedance data as the diagnostic standard

This was a prospective, single-centre, open-label, interventional study conducted between January 2021 and June 2023 across ENT, speech pathology, and gastroenterology departments. Patients aged 18–65 years with idiopathic hoarseness ≥6 weeks’ duration were included; those with structural, infectious, neurological, or functional laryngeal disorders were excluded. All participants underwent video laryngoscopy (to determine Reflux Finding Score [RFS] and Reflux Symptom Index [RSI]), acoustic voice analysis using computerized software (Multi-Speech 3700, MDVP, KayPentax), esophagogastroduodenoscopy (EGD), high-resolution esophageal manometry, and 24-hour impedance pH-metry. GERD was diagnosed based on Los Angeles grade ≥B esophagitis or abnormal acid exposure time (AET > 6%) or DeMeester score > 14.72.

Out of 242 screened subjects, 40 were enrolled. Mean age was 38.6 ± 14.8 years, with equal sex distribution. GERD was confirmed in 25 patients (62.5%); 15 were classified as non-GERD (including reflux hypersensitivity and normal findings). GERD patients had significantly higher RSI (14.2 ± 3.1 vs 11.4 ± 2.4; p=0.002), RFS (10.3 ± 3.2 vs 6.6 ± 3.4; p=0.04), absolute jitter (77.1 vs 50.4; p=0.03), jitter percentage (1.98 vs 1.04; p=0.04), and shimmer percentage (5.42 vs 3.62; p=0.02). Multivariate analysis identified erosive esophagitis (OR 4.32; p=0.04), RSI (OR 1.8; p=0.018), absolute jitter (OR 1.01; p=0.005), shimmer percentage (OR 0.93; p=0.04), and NHR (OR 19.11; p=0.03) as independent predictors. The G-HARP+ Score, integrating acoustic, clinical, and endoscopic predictors, achieved AUROC 0.87 with 88% sensitivity and 85% specificity. The G-HARP Score, based solely on acoustic markers (jitter, shimmer, NHR), achieved AUROC 0.82 with 85% sensitivity and 80% specificity, demonstrating robust non-invasive diagnostic performance.

This study demonstrates that voice-based acoustic markers, particularly jitter, shimmer, and NHR, significantly correlate with reflux severity in idiopathic hoarseness. The proposed G-HARP and G-HARP+ scores provide practical, quantitative tools for predicting GERD and can guide clinicians in identifying patients needing confirmatory reflux testing. These models offer a promising, non-invasive adjunct for early diagnosis and targeted management of reflux-associated voice disorders.