NOVEL CLINICAL RISK PREDICTION SCORE FOR LARYNGOPHARYNGEAL REFLUX

Background
Discerning whether laryngeal symptoms are related to reflux (laryngopharyngeal reflux; LPR) is a challenge. Consequently, patients see numerous specialists, undergo multiple diagnostic tests and empirically trial proton pump inhibitors (PPIs), often with inconsistent response. A clinically practical method to ascertain the likelihood of LPR among patients with laryngeal symptoms is critical. In this study we aimed to develop a clinical risk prediction score for LPR.

Methods
This prospective single center study included adult patients with laryngeal symptoms (throat clearing, sore throat, dysphonia, cough, globus, mucus in throat) referred for LPR evaluation over 3 years (07/19-08/22). Patients were categorized as LPR+ or LPR- based on presence or absence of conclusive GERD as follows: Los Angeles C/D esophagitis and/or long segment Barrett’s esophagus on endoscopy, acid exposure time (AET) ≥6% and/or ≥80 reflux events/24h on ambulatory reflux monitoring off PPI, or AET ≥2% and/or ≥40 reflux events on 24h impedance-pH on PPI.

Independent variables with p-value <0.15 for LPR+ vs LPR- informed development of permutations for multiple logistic regression models. An optimal predictive model was discriminated by area under the curve (AUC) of the receiver operating characteristics. Beta coefficients informed a weighted risk prediction model. The a priori goal was to identify a lower threshold with high confidence to rule out LPR (<20% false negative rate) and upper threshold with high confidence to rule in LPR (<20% false positive rate). Analyses were completed with R v4.2.0 (Vienna, Austria).

Results
Of 304 patients, 130 (43%) met criteria for LPR+ and 174 (57%) LPR- (Table 1). The optimal predictive model [AUC of 0.68 (95% CI 0.62, 0.74)] consisted of Cough symptom (OR 1.6 (1.0, 2.8)), Obesity/overweight (body mass index (BMI) ≥25kg/m2) (OR 1.4 (0.9, 2.3)), Globus symptom (OR 0.5 (0.3, 0.9)), Hiatal hernia ≥1cm (OR 1.9 (1.1, 3.2)), Regurgitation symptom (OR 1.9 (1.1, 3.2)), and male seX (OR 2.0 (1.2, 3.4)). The weighted prediction model, the COuGH RefluX score [AUC of 0.67 (0.61, 0.73)], ranged from -0.5 to 9.0 with a lower threshold of 3.0 (80% sensitive (95% CI 72%, 87%)) and upper threshold of 5.5 (81% specific (74%, 86%)) for LPR. Amongst the cohort 34% had a COuGH RefluX score ≤3.0, 37% had a score 3.5 to 5.0, and 29% had a score ≥5.5 (Figure 1).

Conclusions
The COuGH RefluX risk score consists of widely available clinical data (Cough, Obesity/overweight, Globus (negative predictor), presence of Hiatal hernia, Regurgitation, male seX) to enable stratification of patients with laryngeal symptoms referred for LPR evaluation as low likelihood of LPR (score ≤3.0), high likelihood of LPR (score ≥5.5), and inconclusive for LPR (score 3.5 to 5.0) for which further testing may be indicated. External validation of the COuGH RefluX score is underway.