Outcome of cardiac arrests attended by emergency medical services staff at community outpatient dialysis centers.

Cardiac arrest is the leading cause of death among dialysis patients in the United States. We measured the outcome of cardiac arrests attended by Emergency Medical Services (EMS) staff at hemodialysis facilities in a 14-year population-based retrospective study to identify cardiac arrest cases at a dialysis unit. Associated factors were determined using unconditional logistic regression. Of the 102 cardiac arrests identified around the time of dialysis, 10 occurred before, 72 during, and 20 after hemodialysis. The initial measured abnormality was ventricular fibrillation or tachycardia in 72 cases. Of those who survived transportation to a hospital, survival to discharge was 24 with 15% survival at 1 year. Compared to arrests that occurred prior to dialysis, the odds of ventricular fibrillation were 5-fold greater in patients on dialysis but 14-fold greater in those arresting after dialysis. One-third of cases occurred after the introduction of automated external defibrillators, and in half of the cases these devices were attached prior to EMS arrival. Once these devices were attached, most were used for defibrillation. We conclude that ventricular arrhythmias are the predominant features among arrested in-center dialysis patients with most occurrences during dialysis. The role of these devices in dialysis units will need a larger study to evaluate their efficacy.

Apolipoprotein E genotypes and outcome from out of hospital cardiac arrest.

Genetic factors may influence outcome from cardiac arrest. In Seattle, WA, paramedics collected blood specimens from patients who had suffered cardiac arrest outside of a medical institution (out of hospital cardiac arrest). We examined associations between apolipoprotein E (APOE) genotype and outcome in 134 who died "in the field", 131 who died in the hospital, 198 patients who were discharged from hospital alive, and 64 control subjects. APOE genotype was not significantly related to outcome, including being alive at and being independent by 3 months after the arrest. Specifically, having one or two alleles of APOE epsilon4 or having APOE epsilon3/epsilon3 was not related to outcome, even after controlling for age, sex, race, and initial rhythm. We failed to confirm previous studies and found no significant associations between APOE genotype and outcome from out of hospital cardiac arrest.

Randomized clinical trial of magnesium, diazepam, or both after out-of-hospital cardiac arrest.

OBJECTIVE: To evaluate the feasibility, safety, and efficacy of interventions aimed at improving neurologic outcome after cardiac arrest. METHODS: The authors conducted a double-blind, placebo-controlled, randomized clinical trial with factorial design to see if magnesium, diazepam, or both, when given immediately following resuscitation from out-of-hospital cardiac arrest, would increase the proportion of patients awakening, defined as following commands or having comprehensible speech. If the patient regained a systolic blood pressure of at least 90 mm Hg and had not awakened, paramedics injected IV two syringes stored in a sealed kit. The first always contained either 2 g magnesium sulfate (M) or placebo (P); the second contained either 10 mg diazepam (D) or P. Awakening at any time by 3 months was determined by record review, and independence at 3 months was determined by telephone calls. Over 30 months, 300 patients were randomized in balanced blocks of 4, 75 each to MD, MP, PD, or PP. The study was conducted under waiver of consent. RESULTS: Despite the design, the four treatment groups differed on baseline variables collected before randomization. Percent awake by 3 months for each group were: MD, 29.3%; MP, 46.7%; PD, 30.7%; PP, 37.3%. Percent independent at 3 months were: MD, 17.3%; MP, 34.7%; PD, 17.3%; PP, 25.3%. Significant interactions were lacking. After adjusting for baseline imbalances, none of these differences was significant, and no adverse effects were identified. CONCLUSIONS: Neither magnesium nor diazepam significantly improved neurologic outcome from cardiac arrest.

Payer status: the unspoken triage criterion.

BACKGROUND: The cost of uncompensated trauma care is a significant barrier to trauma system development. Trauma center designation may burden an institution with an unprofitable mix of underinsured, severely injured patients. Concerns about inadequate reimbursement may motivate interhospital transfers on the basis of insurance status rather than medical necessity, potentially undermining the effectiveness of the system. We set out to explore whether this phenomenon exists in a mature trauma system. METHODS: Trauma patients receiving definitive care at Level III or IV trauma centers were compared with patients transferred from these centers to the only Level I regional center. Insurance status was classified as either commercial or noncommercial. Logistic regression was used to determine the independent predictors of transfer after adjusting for differences in injury severity. RESULTS: Only 12% of 2,008 patients initially evaluated at Level III/IV centers were transferred to the Level I center, an indicator of the effectiveness of prehospital triage protocols in the region. The presence of specific complex injuries, younger age, male gender, and insurance status were all associated with an increased likelihood of transfer. Insurance status was an independent predictor of transfer: patients without commercial insurance were 2.4 (95% confidence interval, 1.6-3.6) times more likely to be transferred to a Level I facility than patients with commercial insurance after adjusting for differences in injury severity. CONCLUSION: Insurance status influences the decision to transfer to higher levels of care. These findings suggest that the financial burden of a trauma system may be inequitably distributed. This inequitable distribution may be necessary for trauma system sustainability and calls for the development of disproportionate reimbursement strategies to support regional referral centers.

Cardiopulmonary resuscitation by chest compression alone or with mouth-to-mouth ventilation.

BACKGROUND: Despite extensive training of citizens of Seattle in cardiopulmonary resuscitation (CPR), bystanders do not perform CPR in almost half of witnessed cardiac arrests. Instructions in chest compression plus mouth-to-mouth ventilation given by dispatchers over the telephone can require 2.4 minutes. In experimental studies, chest compression alone is associated with survival rates similar to those with chest compression plus mouth-to-mouth ventilation. We conducted a randomized study to compare CPR by chest compression alone with CPR by chest compression plus mouth-to-mouth ventilation. METHODS: The setting of the trial was an urban, fire-department-based, emergency-medical-care system with central dispatching. In a randomized manner, telephone dispatchers gave bystanders at the scene of apparent cardiac arrest instructions in either chest compression alone or chest compression plus mouth-to-mouth ventilation. The primary end point was survival to hospital discharge. RESULTS: Data were analyzed for 241 patients randomly assigned to receive chest compression alone and 279 assigned to chest compression plus mouth-to-mouth ventilation. Complete instructions were delivered in 62 percent of episodes for the group receiving chest compression plus mouth-to-mouth ventilation and 81 percent of episodes for the group receiving chest compression alone (P=0.005). Instructions for compression required 1.4 minutes less to complete than instructions for compression plus mouth-to-mouth ventilation. Survival to hospital discharge was better among patients assigned to chest compression alone than among those assigned to chest compression plus mouth-to-mouth ventilation (14.6 percent vs. 10.4 percent), but the difference was not statistically significant (P=0.18). CONCLUSIONS: The outcome after CPR with chest compression alone is similar to that after chest compression with mouth-to-mouth ventilation, and chest compression alone may be the preferred approach for bystanders inexperienced in CPR.

Dietary intake of long-chain n-3 polyunsaturated fatty acids and the risk of primary cardiac arrest.

Whether the dietary intake of long-chain n-3 polyunsaturated fatty acids (PUFAs) from seafood reduces the risk of ischemic heart disease remains a source of controversy, in part because studies have yielded inconsistent findings. Results from experimental studies in animals suggest that recent dietary intake of long-chain n-3 PUFAs, compared with saturated and monounsaturated fats, reduces vulnerability to ventricular fibrillation, a life-threatening cardiac arrhythmia that is a major cause of ischemic heart disease mortality. Until recently, whether a similar effect of long-chain n-3 PUFAs from seafood occurred in humans was unknown. We summarize the findings from a population-based case-control study that showed that the dietary intake of long-chain n-3 PUFAs from seafood, measured both directly with a questionnaire and indirectly with a biomarker, is associated with a reduced risk of primary cardiac arrest in humans. The findings also suggest that 1) compared with no seafood intake, modest dietary intake of long-chain n-3 PUFAs from seafood (equivalent to 1 fatty fish meal/wk) is associated with a reduction in the risk of primary cardiac arrest; 2) compared with modest intake, higher intakes of these fatty acids are not associated with a further reduction in such risk; and 3) the reduced risk of primary cardiac arrest may be mediated, at least in part, by the effect of dietary n-3 PUFA intake on cell membrane fatty acid composition. These findings also may help to explain the apparent inconsistencies in earlier studies of long-chain n-3 PUFA intake and ischemic heart disease.

Detection of acute alcohol intoxication and chronic alcohol dependence by trauma center staff.

BACKGROUND: Trauma patients with acute alcohol intoxication or chronic alcohol dependence are at greater risk for morbidity and mortality. We hypothesized that relying on clinical suspicion to detect acute alcohol intoxication and chronic alcohol dependence in trauma patients is inaccurate, influenced by injury factors, and biased by race, gender, age, and socioeconomic status. METHODS: Trauma patients were screened with a blood alcohol concentration and with the Short Michigan Alcohol Screening Test and CAGE questionnaire. Before screening, physicians and emergency department nurses were asked whether the patient was acutely intoxicated (blood alcohol concentration > 100 mg/ dL) or had a chronic alcohol problem. Sensitivity, specificity, positive, and negative predictive values were determined by comparing responses with blood alcohol concentration, Short Michigan Alcohol Screening Test, and CAGE questionnaire results, stratified by injury and demographic factors. RESULTS: Clinical evaluations were obtained on 462 patients. Overall, 23% of acutely intoxicated patients were not identified by physicians. The miss rate increased to one third in severely injured, chemically paralyzed, or intubated patients. Specificity was also poor. Patients with a negative blood alcohol concentration were more likely to be falsely suspected of intoxication if they were either young, male, perceived as disheveled, uninsured, or having a low income (p < 0.05). Staff identified < 50% of patients with a positive Short Michigan Alcohol Screening Test or CAGE, and falsely identified 26% of patients as alcoholic. CONCLUSIONS: Formal alcohol screening should be routine because clinical detection of acute alcohol intoxication and dependence is inaccurate. Screening should also be routine to avoid discriminatory bias attributable to patient characteristics.

Alcohol interventions in a trauma center as a means of reducing the risk of injury recurrence.

OBJECTIVE: Alcoholism is the leading risk factor for injury. The authors hypothesized that providing brief alcohol interventions as a routine component of trauma care would significantly reduce alcohol consumption and would decrease the rate of trauma recidivism. METHODS: This study was a randomized, prospective controlled trial in a level 1 trauma center. Patients were screened using a blood alcohol concentration, gamma glutamyl transpeptidase level, and short Michigan Alcoholism Screening Test (SMAST). Those with positive results were randomized to a brief intervention or control group. Reinjury was detected by a computerized search of emergency department and statewide hospital discharge records, and 6- and 12-month interviews were conducted to assess alcohol use. RESULTS: A total of 2524 patients were screened; 1153 screened positive (46%). Three hundred sixty-six were randomized to the intervention group, and 396 to controls. At 12 months, the intervention group decreased alcohol consumption by 21.8+/-3.7 drinks per week; in the control group, the decrease was 6.7+/-5.8 (p = 0.03). The reduction was most apparent in patients with mild to moderate alcohol problems (SMAST score 3 to 8); they had 21.6+/-4.2 fewer drinks per week, compared to an increase of 2.3+/-8.3 drinks per week in controls (p < 0.01). There was a 47% reduction in injuries requiring either emergency department or trauma center admission (hazard ratio 0.53, 95% confidence interval 0.26 to 1.07, p = 0.07) and a 48% reduction in injuries requiring hospital admission (3 years follow-up). CONCLUSION: Alcohol interventions are associated with a reduction in alcohol intake and a reduced risk of trauma recidivism. Given the prevalence of alcohol problems in trauma centers, screening, intervention, and counseling for alcohol problems should be routine.

Amiodarone for resuscitation after out-of-hospital cardiac arrest due to ventricular fibrillation.

BACKGROUND: Whether antiarrhythmic drugs improve the rate of successful resuscitation after out-of-hospital cardiac arrest has not been determined in randomized clinical trials. METHODS: We conducted a randomized, double-blind, placebo-controlled study of intravenous amiodarone in patients with out-of-hospital cardiac arrest. Patients who had cardiac arrest with ventricular fibrillation (or pulseless ventricular tachycardia) and who had not been resuscitated after receiving three or more precordial shocks were randomly assigned to receive 300 mg of intravenous amiodarone (246 patients) or placebo (258 patients). RESULTS: The treatment groups had similar clinical profiles. There was no significant difference between the amiodarone and placebo groups in the duration of the resuscitation attempt (42+/-16.4 and 43+/-16.3 minutes, respectively), the number of shocks delivered (4+/-3 and 6+/-5), or the proportion of patients who required additional antiarrhythmic drugs after the administration of the study drug (66 percent and 73 percent). More patients in the amiodarone group than in the placebo group had hypotension (59 percent vs. 48 percent, P=0.04) or bradycardia (41 percent vs. 25 percent, P=0.004) after receiving the study drug. Recipients of amiodarone were more likely to survive to be admitted to the hospital (44 percent, vs. 34 percent of the placebo group; P=0.03). The benefit of amiodarone was consistent among all subgroups and at all times of drug administration. The adjusted odds ratio for survival to admission to the hospital in the amiodarone group as compared with the placebo group was 1.6 (95 percent confidence interval, 1.1 to 2.4; P=0.02). The trial did not have sufficient statistical power to detect differences in survival to hospital discharge, which differed only slightly between the two groups. CONCLUSIONS: In patients with out-of-hospital cardiac arrest due to refractory ventricular arrhythmias, treatment with amiodarone resulted in a higher rate of survival to hospital admission. Whether this benefit extends to survival to discharge from the hospital merits further investigation.

Influence of cardiopulmonary resuscitation prior to defibrillation in patients with out-of-hospital ventricular fibrillation.

CONTEXT: Use of automated external defibrillators (AEDs) by first arriving emergency medical technicians (EMTs) is advocated to improve the outcome for out-of-hospital ventricular fibrillation (VF). However, adding AEDs to the emergency medical system in Seattle, Wash, did not improve survival. Studies in animals have shown improved outcomes when cardiopulmonary resuscitation (CPR) was administered prior to an initial shock for VF of several minutes' duration. OBJECTIVE: To evaluate the effects of providing 90 seconds of CPR to persons with out-of-hospital VF prior to delivery of a shock by first-arriving EMTs. DESIGN: Observational, prospectively defined, population-based study with 42 months of preintervention analysis (July 1, 1990-December 31, 1993) and 36 months of post-intervention analysis (January 1, 1994-December 31, 1996). SETTING: Seattle fire department-based, 2-tiered emergency medical system. PARTICIPANTS: A total of 639 patients treated for out-of-hospital VF before the intervention and 478 after the intervention. INTERVENTION: Modification of the protocol for use of AEDs, emphasizing approximately 90 seconds of CPR prior to delivery of a shock. MAIN OUTCOME MEASURES: Survival and neurologic status at hospital discharge determined by retrospective chart review as a function of early (<4 minutes) and later (> or =4 minutes) response intervals. RESULTS: Survival improved from 24% (155/639) to 30% (142/478) (P=.04). That benefit was predominantly in patients for whom the initial response interval was 4 minutes or longer (survival, 17% [56/321] before vs 27% [60/220] after; P = .01). In a multivariate logistic model, adjusting for differences in patient and resuscitation factors between the periods, the protocol intervention was estimated to improve survival significantly (odds ratio, 1.42; 95% confidence interval, 1.07-1.90; P = .02). Overall, the proportion of victims who survived with favorable neurologic recovery increased from 17% (106/634) to 23% (109/474) (P = .01). Among survivors, the proportion having favorable neurologic function at hospital discharge increased from 71% (106/150) to 79% (109/138) (P<.11). CONCLUSION: The routine provision of approximately 90 seconds of CPR prior to use of AED was associated with increased survival when response intervals were 4 minutes or longer.

Ten-year trend in survival and resource utilization at a level I trauma center.

OBJECTIVE: To determine the impact of increasing trauma center experience over time on survival and resource utilization. METHODS: The authors studied a retrospective cohort at a single level I trauma center over a 10-year period, from 1986 to 1995. Patients included all hospital admissions and emergency department deaths. The main outcome measures were the case-fatality rate adjusted for injury severity, hospital length of stay, and costs. RESULTS: A total of 25,979 patients were admitted or died. The number of patients per year increased, from 2063 in 1986 to 3313 in 1995. The proportion of patients transferred from another institution increased from 16.2% to 34.4%. Although mean length of stay declined by 28.4%, from 9.5 to 6.8 days, costs increased by 16.7%, from $14,174 to $16,547. The use of specific radiologic investigations increased; the frequency of operative procedures either remained unchanged (craniotomy, fracture fixation) or decreased (celiotomy). After adjusting for injury severity and demographic factors, the mortality rate decreased over 10 years. The improvement in survival was confined to patients with an injury severity score > or =16. CONCLUSION: Over a 10-year period, the case-fatality rate declined in patients with severe injuries. Overall acute care costs increased, partially because of the increased use of radiologic investigations. Even in otherwise established trauma centers, increasing cumulative experience results in improved survival rates in the most severely injured patients. These data suggest that experience contributes to a decrease in mortality rate after severe trauma and that developing trauma systems should consider this factor and limit the number of designated centers to maximize cumulative experience at individual centers.

Outcome after hemorrhagic shock in trauma patients.

BACKGROUND: It is essential to identify patients at high risk of death and complications for future studies of interventions to decrease reperfusion injury. METHODS: We conducted an inception cohort study at a Level I trauma center to determine the rates and predictors of death, organ failure, and infection in trauma patients with systolic blood pressure < or = 90 mm Hg in the field or in the emergency department. RESULTS: Among the 208 patients with hemorrhagic shock (blood pressure < or = 90 mm Hg), 31% died within 2 hours of emergency department arrival, 12% died between 2 and 24 hours, 11% died after 24 hours, and 46% survived. Among those who survived > or = 24 hours, 39% developed infection and 24% developed organ failure. Increasing volume of crystalloid in the first 24 hours was strongly associated with increased mortality (p = 0.00001). CONCLUSION: Hemorrhage-induced hypotension in trauma patients is predictive of high mortality (54%) and morbidity. The requirement for large volumes of crystalloid was associated with increased mortality.

Human immunodeficiency virus seroprevalence among adults treated for out-of-hospital cardiac arrest in Seattle, Washington, 1989-1993.

STUDY OBJECTIVE: To estimate the potential risk of HIV exposure for those providing emergency care for out-of-hospital cardiac arrest in Seattle, Washington, by surveying the seroprevalence of HIV in the patient population. METHODS: We surveyed the seroprevalence of HIV among 1,474 persons treated for out-of-hospital cardiac arrest by paramedics during the years 1989 through 1993. Blood specimens were obtained at the site of cardiac arrest, stripped of personal identifiers, and tested for HIV-1 and HIV-2 by enzyme immunoassay and Western blot. RESULTS: Among the 1,011 men, 8 (.8%, 95 percent confidence interval .3% to 1.4%) were seropositive for HIV-1 during this 5-year period; all 8 were younger than age 55. No serologic evidence of HIV infection was detected among the 463 women. CONCLUSION: The seroprevalence of HIV in this population was relatively low. Risk of possible HIV transmission during paramedic treatment was remote.

Progressive multifocal leukoencephalopathy presenting as human immunodeficiency virus type 1 (HIV)-associated dementia.

Progressive multifocal leukoencephalopathy (PML) is a demyelinating disorder of the CNS that usually causes hemiparesis or hemianopsia. Dementia occurs in combination with other neurologic abnormalities. We report a human immunodeficiency virus type 1 (HIV)-infected man whose only manifestation of proven PML was dementia that was clinically indistinguishable from HIV-associated dementia.

Introduction of unmarked mutations in the Helicobacter pylori vacA gene with a sucrose sensitivity marker.

Research on Helicobacter pylori has been hindered by the lack of useful genetic tools. Using the sacB gene of Bacillus subtilis, we developed a sucrose-based counterselection system that allows introduction of unmarked mutations in H. pylori. A kan-sacB cassette, consisting of the sacB gene expressed from the H. pylori flagellin promoter and the kanamycin resistance module, was introduced by homologous recombination into a target H. pylori gene. The resultant strains were sucrose sensitive and kanamycin resistant. Following transformation with a mutated allele, growth in sucrose-containing medium allowed the selection of strains that had lost the kan-sacB module and had integrated the unmarked allele. We have used this cassette to perform a site-directed modification of two histidine residues encoded by the vacA gene in a two-step procedure. This system should prove useful in the site-directed mutagenesis of H. pylori genes.

Urban-rural differences in prehospital care of major trauma.

OBJECTIVES: To compare differences in response times, scene times, and transport times by advanced life-support-trained paramedics to trauma incidents in urban and rural locations. METHODS: This report was a prospective cohort study of professional emergency medical services conducted in a five-county area in the state of Washington. Ninety-eight percent of trauma transports are provided by professional paramedics trained in advanced life support. Subjects were included in this study if they qualified as a major trauma victim and were transported or found dead at the scene by one of the region's advanced life support transport agencies between August 1, 1991, and January 31, 1992. The severity of injury was rated using the Prehospital Index. Incident locations were defined as "rural" if they occurred in a US Census division (a geographic area) in which more than 50% of the residents resided in a rural location. RESULTS: During the 6-month data collection period, advanced life support agencies responded to a total of 459 major trauma victims in the region. A geographic locations was determined for 452 of these subjects. Of these, 42% of subjects were injured in urban areas and the remainder in rural areas. The severity of injuries, as determined both by the triage classification (p = 0.17) and the distribution of Prehospital Index scores (p = 0.92), was similar for urban and rural major trauma patients. Twenty-six (5.7%) subjects died at the scene. About one quarter of both groups had a severe injury, as indicated by Prehospital Index score of more than 3. The mean response time for urban locations was 7.0 minutes (median = 6 minutes) compared with 13.6 minutes (median = 12 minutes) for rural locations (p < 0.0001). The mean scene time in rural areas was slightly longer than in urban areas (21.7 vs. 18.7 minutes, p = 0.015). Mean transport times from the scene to the hospital were also significantly longer for rural incidents (17.2 minutes vs. 8.2 minutes, p < 0.0001). Rural victims were over seven times more likely to die before arrival (relative risk = 7.4, 95% confidence interval 2.4-22.8) if the emergency medical services' response time was more than 30 minutes. CONCLUSIONS: Response and transport times among professional, advanced life-support-trained paramedics responding to major trauma incidents are longer in rural areas, compared with urban areas.

Diagnosis of injuries after stab wounds to the back and flank.

BACKGROUND: Historically, patients with deep posterior wounds underwent a formal celiotomy to rule out injury. Currently, we use a policy of selective management. The purpose of this review is to evaluate our experience with selective management to identify potential areas of further improvement. METHODS AND RESULTS: This study includes 203 patients over a 10-year period. By changing from a policy of mandatory exploration to selective management the total celiotomy rate decreased from 100 to 24% and the therapeutic celiotomy rate increased from 15 to 80%. CONCLUSIONS: In stable patients, a diagnostic peritoneal lavage should be performed as the initial diagnostic study. When diagnostic peritoneal lavage is negative, triple contrast computed tomography should be performed to evaluate the remaining retroperitoneal structures. Any suggestion of pericolonic extravasation of contrast or air, edema, or hemorrhage must be interpreted as a positive study and prompt consideration for operative exploration.

Pressure sores in the acute trauma patient: incidence and causes.

OBJECTIVES: The purpose of this review was to determine the incidence of pressure sores in acute trauma patients and to identify the causes. DESIGN AND MATERIALS AND METHODS: This study is a retrospective chart review of all acute trauma patients admitted to Harborview Medical Center between January of 1991 and December of 1993 who were discharged with an ICD-9 diagnosis of acute pressure sore. RESULTS: A total of 7,492 trauma patients were admitted. Thirty-two patients developed 44 pressure sores for an incidence of 0.4%. All of the patients were severely injured (mean Injury Severity Score, 21). Eighteen (41%) of the pressure sores developed as a result of unrelieved positional pressure. Thirteen (30%) of the pressure sores were equipment induced. CONCLUSION: Our incidence of pressure sores in acute trauma victims is very low and occurs in those most seriously injured. The majority of the wounds are due to unrelieved pressure from body positioning or equipment failure.

Cerebrospinal fluid creatine kinase BB isoenzyme activity and neurologic prognosis after cardiac arrest.

OBJECTIVE: To assess the relationship between CSF creatine kinase BB isoenzyme activity (CSF CKBB) and neurologic outcome after cardiac arrest in clinical practice. BACKGROUND: CSF CKBB reflects the extent of brain damage following cardiac arrest. METHODS: To help with prognosis, treating physicians ordered CSF CKBB tests on 474 patients over 7.5 years; 351 of these patients had experienced a cardiac arrest. Assays were performed in one laboratory using agarose electrophoresis. By chart review, we determined awakening status for all patients, defined as the patient having comprehensible speech or following commands. RESULTS: CSF CKBB was usually sampled 48 to 72 hours after cardiac arrest and was strongly associated with awakening (p < < 0.001). The median was 4 U/l for 61 patients who awakened and 191 U/l for 290 who never awakened. For those who awakened, 75% of CKBB levels were < 24 U/l, and for those who never awakened, 75% were > 86 U/l. The highest value in a patient who awakened was 204 U/l, a cutoff that yielded a specificity of 100% of never awakening but a sensitivity of forty-eight percent. Only nine patients who awakened had CSF CKBB values greater than 50 U/l, and none regained independence in activities of daily living. Only three unconscious patients were still alive at last contact, with follow-up of 63, 107, and 109 months. Using logistic regression, the probability of never awakening given a CSF CKBB result can be estimated as: 1/(1 + L), where L = e raised to (0.1267 - 0.0211 x CSF CKBB [U/l]). CONCLUSION: CSF CKBB measurement helps to estimate degree of brain damage and thus neurologic prognosis after cardiac arrest. However, results of this retrospective study could reflect in part a self-fulfilling prophecy.