Diaphragmatic electromyography during cryoballoon ablation: a novel concept in the prevention of phrenic nerve palsy

BACKGROUND Hemidiaphragmatic paralysis is the most frequent complication associated with cryoballoon ablation for atrial fibrillation. To date, no preventive strategy has proved effective. OBJECTIVE We sought to assess the feasibility of diaphragmatic electromyography during cryoballoon ablation, explore the relationship between altered signals and phrenic nerve palsy, and define characteristic changes that herald hemidiaphragmatic paralysis. METHODS Cryoballoon ablation was performed in the right superior pulmonary vein or superior vena cava in 16 mongrel dogs weighing 37.7 +/- 2.4 kg, at sites determined by phrenic nerve capture. During ablation, the phrenic nerve was paced at 60 bpm from the superior vena cava while recording diaphragmatic compound motor action potentials (CMAPs) by esophageal decapolar catheters. Diaphragmatic excursion was monitored by fluoroscopy and abdominal palpation. RESULTS Before ablation, the CMAP amplitude was 592 (interquartile range 504, 566) mu V, initial latency 21.5 +/- 4.2 ms, peak latency 64.7 +/- 21.1 ms, and duration 101.7 +/- 13.3 ms. Hemidiaphragmatic paralysis was obtained in all dogs 62 +/- 34 seconds into the cryoapplication. The CMAP amplitude decreased exponentially, with no patterned changes in latencies and duration. Discriminatory analyses by receiver-operating curve characteristics identified a 30% reduction in CMAP amplitude as the most predictive cutoff value for hemidiaphragmatic paralysis (c-statistic 0.965; P <.0001). This criterion presaged diaphragmatic paralysis, as detected by abdominal palpation, by 31 +/- 23 seconds. CONCLUSION Diaphragmatic electromyographic signals could be reliably recorded during cryoballoon ablation. An exponential decrease in CMAP amplitude precedes diaphragmatic paralysis, with a 30% reduction yielding the best discriminatory potential. A promising safety margin was detected, which merits prospective validation.