Laryngeal ultrasound detects a high incidence of vocal cord paresis after aortic arch repair in neonates and young children

Objectives: To determine the incidence of vocal cord paresis (VCP) after neonatal aortic arch repair/Norwood-type procedure, and the effectiveness of noninvasive laryngeal ultrasound in detecting VCP compared with gold standard invasive nasoendoscopy. Methods: Fifty-two patients who underwent an arch repair (39 of 52; 75%) or Norwood-type procedure (13 of 52; 25%) via sternotomy between April 1, 2015, and April 30, 2017 underwent laryngeal ultrasound (50 of 52; 96%) and/or flexible fiber optic nasoendoscopy (39 of 52; 75%) at 48 to 72 hours after endotracheal extubation. Primary arch diagnoses were coarctation in 56% (29 of 52), hypoplastic left heart syndrome in 17% (9 of 52), isolated hypoplastic arch in 17% (9 of 52), and interrupted aortic arch in 10% (5 of 52). The median patient age at surgery was 5.5 days (interquartile range, 4.0-12.5 days). Fifteen patients (15 of 52; 29%) required preoperative intubation. Results: Left VCP was present in 59% (23 of 39) of patients on nasoendoscopy and in 59% (27 of 46) of patients on laryngeal ultrasound, and 4 additional patients had inconclusive ultrasound results. There was agreement between the results of nasoendoscopy and conclusive ultrasound in all cases. The overall sensitivity, specificity, positive and negative predictive values, and Cohen's kappa coefficient of laryngeal ultrasound compared with nasoendoscopy for the detection of left VCP were 95%, 88%, 91%, 93%, and 0.83, respectively. On multivariable analysis, preoperative intubation and arch repair techniques other than the Norwood procedure were associated with left VCP (odds ratio, 12.7; P = .03; and 14.1; P = .03, respectively). Conclusions: There is a high incidence of VCP after arch repair via sternotomy. Laryngeal ultrasound seems to be an effective and noninvasive method for detecting VCP in neonates and young children.