Author + information
- Keith A. Comess, MD and
- Kirk W. Beach, PhD, MD⁎ ()
- ↵⁎Reprint requests and correspondence:
Dr. Kirk W. Beach, Department of Surgery and Bioengineering, Box 356021, University of Washington, 1959 Pacific Street, Seattle, Washington 98195-6021
- biological markers
- brain injury
- cardiac arrest
- cardiopulmonary resuscitation
- clinical decision making
- NSE protein
- S100 protein
Doing and saying nothing are great powers but they should not be abused.
—Charles Maurice de Talleyrand-Périgord, First Prince de Bénévent (1)
Management of patients presenting with out-of-hospital cardiac arrest is a perennially vexing problem because no categorical predictors of neurological recovery exist. Predicting the extent of brain damage permits more sensible management efforts. Adding to the complexity of clinical decision making by both family and medical staff is a stew of ethical, economic, and emotional factors admixed with legal issues. Almost 80% of patients with return of spontaneous circulation after cardiopulmonary resuscitation remain unconscious for variable lengths of time (2). Of those, approximately 20% will enter a vegetative state (3), a condition most patients sensibly prefer to avoid. This management conundrum is amplified by anecdotal stories of dramatic and unexpected recovery after prolonged unresponsiveness wherein patients supposedly arise Lazarus-like from “coma to consciousness” and by reporting a subtle and confusing spectrum of consciousness (4). Lay media frequently omit or minimize potentially ruinous social and financial aspects of management: cost estimates for the care of severely brain-damaged survivors run in billions of dollars annually including extended hospital stays and rehabilitation (5). Not surprisingly, as applied by the friends and family of individual patients, societal concerns such as resource utilization and economic “best practice” are often irrelevant or frankly offensive topics of discussion. Since the advent of cardiopulmonary resuscitation and its widespread out-of-hospital adoption, many more patients survive to hospital admission, prompting the question “Now what?”
Clearly, prospects for meaningful recovery are enhanced after a witnessed event in a young patient given prompt and appropriate resuscitation who demonstrates signs of wakefulness on or within a short time of hospital admission. Despite the attractiveness of the history, when available, it is often unreliable, sometimes contradictory, frequently confounding and sometimes misleading. Given all this, it would be extremely useful to have some sort of objective standard for predicting functional recovery that would divide patients in a binary fashion (“will recover completely” vs. “dismal prognosis”) or at least assign them along a functional recovery continuum extending from a temporary “scrambling” of consciousness to severe brain damage. It would be even better if a determination of full recovery could be made within the first 24 h of hospital admission. Right now, there is no such measure.
In this issue of the Journal, Einav et al. (6) evaluate serum-100B (S100B) and neuron-specific enolase (NSE) sampled at intervals beginning on day 0 (representing initial evaluation in the emergency department), extending through the third day of hospitalization in combination with clinical information to predict neurological recovery from out-of-hospital cardiac arrest. The authors contend that functional recovery in patients experiencing out-of-hospital cardiac arrest within “ethically acceptable safety margins” can be predicted thusly.
Physical examination has been used to predict outcome in comatose survivors of cardiac arrest: it is inexpensive, easy, and traditionally valued. Booth et al. (5) performed a MEDLINE search extending from 1966 to 2003 to determine examination precision and accuracy for predicting outcomes. Five findings, all of which are part of a routine evaluation, were found to strongly predict death or poor neurological outcome at 24 h: absence of corneal reflexes; pupillary responses; withdrawal response to pain; no motor response on admission and none at 72 h. The most useful signs occur at 24 h after cardiac arrest, but they caution that prognosis should not be rendered by clinical examination alone; something else is necessary, especially when induced hypothermia and sedation are considered.
S100B is secreted by an astrocyte subtype and by NG2-expressing cells. It is detectable in the periphery by simple and relatively inexpensive laboratory testing during the acute phase of central nervous system injury. Its clinical utility in the evaluation of hypoxic and other central nervous system damage results from its ability to cross the blood-brain barrier and appear in the serum where it can be detected and quantified. S100B levels reportedly increase before changes in intracerebral pressure, neuroimaging, and physical examination are evident. The major advantage of S100B is that abnormal levels provide a sensitive measure for detecting central nervous injury before gross changes develop.
NSE, a protein contained in neurons and neuroendocrine cells, is released in anoxic brain damage. The level of NSE presumably correlates with the extent of injury and thus with prognosis. Plasma concentrations of NSE have been studied in out-of-hospital cardiac arrest (7–15). In the aggregate, data suggest a direct correlation between serum NSE levels and outcome categories. For example, Reisinger et al. (15) reported a cohort study of 177 patients with NSE measured on admission and 1 to 3 days afterward. Levels generally peaked 2 to 3 days post-admission. Poor outcome (death or persistent coma) occurred in all 37 patients with a peak NSE level of ≥0.80 ng/ml (specificity 100%); NSE levels were lower in 22 additional cases (sensitivity 63%).
What does the study by Einav et al. (6) add to all this? Serial NSE and S100B measurements and the reasonably large sample size are strengths of the study as is the study endpoint (good vs. bad outcomes); in fact, this endpoint is clinically critical. Of 195 eligible patients, only 43 (22%) survived to hospital discharge, 26 (13%) of whom had a good outcome as defined by the Cerebral Performance Scale (CPC) (16). The levels of both S100B and NSE were significantly lower in those with a CPC score of 1 to 2 than those with poor outcomes (CPC score 3 to 5). Their model found that “the level of NSE became non-significant and the model retained only three significant variables: age, VT/VF and the level of S100B at admission.” But S100B measurements obtained on day 0 added only ∼5% predictive value to clinical factors with an overall accuracy of ∼93%. Evidently NSE measurements can be jettisoned when S100B, age, and ventricular tachycardia/ventricular fibrillation (VT/VF) are accounted for, although S100B contributes a small additional improvement (odds: 2.14; p = 0.03).
Einav et al. (6) omit various “tried and true” physical findings with demonstrated prognostic significance (5). So it is not clear what influence, if any, incorporation of corneal and pupillary reflexes and others might have on their model and what effect the relatively high mortality rate in their population might have produced on their results. Neuroimaging and other testing were not reported, so their influence on the model is also unknown. The authors state that “[a]lthough biomarker data independently contributed an ostensibly modest 5.2% to the AUC, they substantially reduced the probability of misclassification error compared with that based solely on clinical criteria.” One might therefore presume that age >66 years and VT/VF are sufficient for rendering a clinical decision.
Everyone involved in post-resuscitation management wants to know 2 things: when to cease supportive efforts and what the prospects are for a good recovery given continued care. The paper does not directly address exactly how biomarkers would affect the decision. Clearly age >66 years (odds: 10 univariate and 6 multivariate) and VT/VF (odds: 17 univariate and 11 multivariate) are the salient predictive factors, and they are available at admission. On arrival or during the first day, S100B is influential. The crux issue is and remains society's (and physicians') priorities regarding both when and how the decision is made to stop resuscitation. It is only reasonable to conclude, as the authors do, that “[p]arallel to searching for the infallible brain biomarker, policy makers should determine the level of risk of misclassification acceptable to their society within this clinical setting.”
Both authors have reported that they have no relationships relevant to the contents of this paper to disclose.
↵⁎ Editorials published in the Journal of the American College of Cardiology reflect the views of the authors and do not necessarily represent the views of JACC or the American College of Cardiology.
- American College of Cardiology Foundation
- Milton J.
- Bates D.
- Gorman J.A.
- Einav S.,
- Kaufman N.,
- Algur N.,
- Kark J.D.
- Martens P.,
- Raabe A.,
- Johnsson P.
- Zingler V.C.,
- Krumm B.,
- Bertsch T.,
- Fassbender K.,
- Pohlmann-Eden B.
- Tiainen M.,
- Roine R.O.,
- Pettilä V.,
- Takkunen O.
- Reisinger J.,
- Höllinger K.,
- Lang W.,
- et al.
- Safar P.