SSEP amplitude accurately predicts both good and poor neurological outcome early after cardiac arrest; a post-hoc analysis of the ProNeCA multicentre study

Aim To assess if, in comatose resuscitated patients, the amplitude of the N20 wave (N20amp) of somatosensory evoked potentials (SSEP) can predict 6-months neurological outcome. Setting Multicentre study in 13 Italian intensive care units. Methods The N20amp in microvolts (μV) was measured at 12 h, 24 h, and 72 h from cardiac arrest, along with pupillary reflex (PLR) and a 30-min EEG classified according to the ACNS terminology. Sensitivity and false positive rate (FPR) of N20amp alone or in combination were calculated. Results 403 patients (age 69[58–68] years) were included. At 12 h, an N20amp >3 μV predicted good neurological outcome (Cerebral Performance Categories [CPC] 1–2) with 61[50−72]% sensitivity and 11[6–18]% FPR. Combining it with a benign (continuous or nearly continuous) EEG increased sensitivity to 91[82−96]%. For poor outcome (CPC 3–5), an N20Amp ≤0.38 μV, ≤0.73 μV and ≤1.01 μV at 12 h, 24 h, and 72 h, respectively, had 0% FPR with sensitivity ranging from 61[51−69]% and 82[76−88]%. Sensitivity was higher than that of a bilaterally absent N20 at all time points. At 12 h and 24 h, a highly malignant (suppression or burst-suppression) EEG and bilaterally absent PLR achieved 0% FPR only when combined with SSEP. A combination of all three predictors yielded a 0[0−4]% FPR, with maximum sensitivity of 44[36−53]%. Conclusion At 12 h from arrest, a high N20Amp predicts good outcome with high sensitivity, especially when combined with benign EEG. At 12 h and 24 h from arrest a low-voltage N20amp has a high sensitivity and is more specific than EEG or PLR for predicting poor outcome.