A simple clinical score may help rule out deep laryngeal involvement in pemphigus vulgaris and reduce unnecessary laryngoscopy.
A new clinical scoring tool could help clinicians identify patients with pemphigus vulgaris at low risk of deep laryngeal involvement and avoid unnecessary laryngoscopy, researchers report.
The retrospective cohort study, published in in JAMA Dermatology, analysed 247 patients with confirmed pemphigus vulgaris treated at an Israeli medical centre between 2015 and 2022.
Flexible fibre-optic laryngoscopy confirmed deep laryngeal involvement in 49 patients, representing 19.8% of the cohort.
The researchers identified three independent clinical predictors of laryngeal disease: dysphagia, hoarseness and higher Pemphigus Disease Area Index (PDAI) scores.
Dysphagia was the strongest predictor, increasing the odds of laryngeal involvement more than sixfold, while hoarseness was associated with a 2.5-fold increase in risk.
Using these variables, researchers developed a simplified risk score that stratified patients into low-, intermediate- and high-risk groups.
Patients scoring 0 or 1 point(s) had no observed laryngeal involvement, compared with a 15.3% involvement rate in the intermediate-risk group and 66.7% in the high-risk group.
At a threshold score of 2 points or higher, the model achieved 78.2% sensitivity and a negative predictive value of 92.1%, suggesting it may be particularly useful for ruling out clinically significant laryngeal disease, the researchers wrote.
Overall model discrimination was good, with an area under the receiver operating characteristic curve of 0.78.
The researchers said the score may help address a persistent clinical dilemma in pemphigus vulgaris management.
While deep laryngeal lesions could progress to airway compromise and respiratory failure if missed, routine laryngoscopy remained invasive, uncomfortable for patients and resource intensive.
Related
Current international guidelines recommend laryngoscopy mainly for patients with respiratory symptoms, but evidence-based criteria for broader screening are lacking.
The researchers also found the model performed differently by gender. In male patients, the model achieved an AUROC of 0.87 with sensitivity exceeding 95%, while performance was lower in female patients, with an AUROC of 0.71.
They noted that standard PDAI scoring did not include a specific laryngeal category, meaning disease severity may be underestimated without direct visualisation of the airway.
The study was limited by its retrospective, single-centre design and relatively small number of laryngeal events, the researchers said.
“This study developed a clinical prediction model identifying patients with pemphigus vulgaris at low risk for deep laryngeal involvement, with good discrimination and high negative predictive value,” they concluded.
“Logistic regression performed comparably to machine learning. Although external validation is needed, this model may inform laryngoscopy decisions in pemphigus vulgaris when combined with clinical judgment.”
