Laryngeal ultrasound for the diagnosis of laryngomalacia in infants.

Abstract

Laryngomalacia is the most prevalent cause of congenital stridor. Flexible laryngobronchoscopy (FLB) is the gold standard for diagnosis. However, FLB requires venous access, deep sedation or general anesthesia, and is associated with patient's and parental discomfort and a considerable cost. Laryngeal ultrasound (LUS) has been shown to provide good evaluation of the normal anatomy and the dynamic motion of laryngeal structures. We investigated the yield of LUS in the diagnosis of laryngomalacia in infants with congenital stridor compared to FLB. A prospective blind study of consecutive infants referred for FLB due to congenital stridor with unknown cause. The presence of arytenoids adduction on LUS during inspiration was used to diagnose laryngomalacia. LUS was followed by FLB under anesthesia. The physician performing the FLB was blinded to the LUS results. Twenty-four patients participated in the study (54% males). Median (IQR) age and weight were 3.5 (1-7.7) months and 5.1 (4-8.4) kg, respectively. LUS was well tolerated in all infants. Compared to FLB, LUS correctly diagnosed laryngomalacia in 11/14 (78.6%) infants. In 10 infants FLB ruled out laryngomalacia of whom LUS concurred in 9 infants and one was falsely diagnosed with laryngomalacia. LUS had a sensitivity of 78.5% (CI 52-95%), specificity of 90% (CI 49-94%), NPV of 75%, and PPV of 92%. ROC analysis demonstrated AUC of 0.84 (P < 0.01, 95%CI: 0.67-1.0). We suggest that LUS has a role in screening congenital stridor in otherwise healthy infants when laryngomalacia is highly probable.