Dear Editor, We thank Dr. Aibiki [1] for his observation that the bispectral index (BIS) and suppression ratio (SR) may change over time in cardiac arrest survivors receiving neuromuscular blockade (NMB) during therapeutic hypothermia. We agree that the evolution of these electrophysiological parameters may affect the prognostic power of BIS. We believe at least five potential confounders must be considered when interpreting the prognostic utility of BIS (or any other EEGbased variable) in a cardiac arrest survivor treated with hypothermia. The first is interference from muscle activity such as shivering, myoclonus, and seizures. Patients must be fully paralyzed or the BIS1 is not reliable. The second is electrographic seizure activity—anecdotally, we have noted a sudden ‘‘spike’’ in the BIS score when a nonconvulsive seizure begins, and a return to near-baseline BIS values after the seizure ends. The third potential confounder is the variability induced by the time course of cerebral electrical reconstitution following hypoxic-ischemic injury. Because recovery of brain electrical activity is poorly characterized and may vary with severity of brain injury, great caution must be exercised in the interpretation of early BIS1 values. Our patients were evaluated at a mean of 280 min after resuscitation (return of spontaneous circulation, ROSC), but it is possible that patients evaluated very early after ROSC might show false low BIS1 values. Like Dr. Aibiki and colleagues, we have seen changing BIS values in some patients but not others, and agree that this complex issue warrants further study. Reconstitution of cerebral electrical activity has been studied in rats following asphyxial arrest. Electrical bursting is identified as early as 13 min after ROSC, and a longer delay to bursting correlates with worse neurocognitive outcome [2]. Increased bursting of electrical activity at 30–40 min after the ROSC correlates with subsequent development of a continuous EEG background at 90 min, and with good neurological outcome. Conversely, rats with decreased early bursting usually failed to develop a continuous EEG pattern, and had poor neurological outcomes [3]. These time frames and observations have only been reported as experimental rat data, with no known translation to human clinical data. A fourth potential confounder is temperature; again in rat models, hypothermia and fever increased early bursting activity [4, 5] compared to normothermia. The fifth potential confounder of early EEGbased prognostication after cardiac arrest is intensity of analgo-sedation, another area that requires further study. In humans that recover from cardiac arrest, a stereotypical pattern of intermittent cortical electrical activity followed by continuous activity with the sequential appearances of delta, theta, and alpha activities on EEG has been described, whereas the appearance of spikes and sharp waves, or a discontinuous or suppressed background is associated with poor recovery [6, 7]. The time course of this progression, and the influence of clinical factors that may critically influence cerebral blood flow or electrophysiology, such as blood pressure, cardiac index, brain temperature, anticoagulation, electrolyte levels, and partial pressures of oxygen and carbon dioxide in arterial blood, have not been established. BIS1 is a promising and simple technique for early estimation of the severity of brain injury after cardiac arrest, but has potential confounders not yet studied. Further research is required before it can be applied clinically for critical decision-making regarding type of therapeutic intervention and especially before any decisions to limit therapy are made after cardiac arrest. Patients must be given the opportunity to benefit from improved resuscitation, neuroprotective, and critical care therapies, and we must avoid the temptation to declare the neurological prognosis hopeless unless we know the data leading to that conclusion is worthy of such a life-and-death decision.
Read full abstract