Objectives: The absence of N20 somatosensory evoked potential after cardiac arrest is related to poor outcome. However, discrimination between the low-amplitude and the absence of N20 is challenging. P25 and P30 are short-latency positive peaks with latencies between 25 and 30 ms following N20 (P25/30). P25/30 is evident even with an ambiguous N20 in patients with good outcome. Therefore, we evaluated the predictive value of P25/30 after cardiac arrest.
Design: A retrospective observational study.
Setting: University-affiliated hospital.
Subjects: Comatose survivors after out-of-hospital cardiac arrest treated by hypothermic targeted temperature management.
Intervention: None.
Measurements and main results: The specificity and the positive predictive value of P25/30 and N20 in predicting poor outcome were the same, showing a rate of 100%. The sensitivity of P25/30 in predicting poor outcome (90.12% [95% CI, 81.5-95.6%]) was higher than that of N20 (70.37% [95% CI, 59.2-80%]). Also, the negative predictive value of P25/30 in predicting poor outcome (81.4% [95% CI, 69.4-89.4%]) was higher than that of N20 (59.3% [95% CI, 51-67.1%]). The P25/30-based adjusted model showed a larger area under the curve (0.98 [95% CI, 0.95-1]) compared with the N20-based adjusted model (0.95 [95% CI, 0.91-0.98]) (p = 0.02).
Conclusions: The absence of P25/30 is related to poor outcome with a higher sensitivity, negative predictive value than the absence of N20.