3500 people suffer from out of hospital cardiac arrest each year in Denmark. Therapeutic hypothermia to 33 degrees celsius is now standard treatment of comatose cardiac arrest patients. The investigators are challenged in our attempt to predict outcome of these patients by both low body temperature in itself and the sedative and relaxing drugs used to keep the patient in a coma.

This study is a substudy in a large international multicenter randomized trial that investigates the possible benefit of 48 hours of therapeutic hypothermia versus todays standard of 24 hours.

In this substudy the investigators will approach early prediction of neurological outcome using a combination of new examinations backed up by well established ones to engage this challenge from different angles and find a battery of combined examinations, that together will enable us to accurately predict outcome at an earlier stage.

Our examinations have been chosen from the three fields of neurophysiology, biochemistry and neurology and consist of electroencephalography (EEG), somatosensory evoked potentials (SEP), biomarkers and clinical examinations such as brain stem reflexes chosen for their prognostic reliability.

Neurological prognostication of patients in therapeutic hypothermia after cardiac arrest

Background: In Denmark, approximately 3000 people suffer from cardiac arrest (CA) each year. After CA with no blood flow for more than five minutes, cerebral ischemia, free radicals along with other mediators creates chemical cascades during reperfusion that result in cerebral injury and often severe neurologic impairment.

Therapeutic hypothermia (TH) has become standard treatment of comatose CA patients and this has led to new challenges in the area of prognostication, because both the low temperature in itself and the sedative and relaxing drugs used to keep the patient in a coma, affect the examinations performed by the clinician.

This calls for a reevaluation of the existing prognostic examinations and the invention and implementation of new ones.

A new awareness towards a prognostic model involving several different modalities including SSEP, EEG, biomarkers and clinical examinations is thus awakening.

The aim of this study is to evaluate the use of EEG, SSEP, four biomarkers and five clinical examinations and observations. We will evaluate our patients using the Glasgow-Pittsburgh Cerebral Performance Categories (CPC).

EEG The EEG analysis is today a qualitative analysis based on the assessment of a trained specialist. New methods of EEG analysis are based on digital quantitative analyses and therefore less sensitive to interpersonal differences and the uncertainty of subjective assessments.

EEG reactivity is the observed change in EEG linked to a stimulation (fx noxious or auditory) and is a promising predictor of outcome concerning CA patient with anoxic coma (1, 9). In this study we will use power-spectral analyses in the assessment of EEG reactivity and investigate the prognostic value of power analyses in the four frequency bands delta, theta, alpha and beta.

Certain EEG-patterns, the malignant patterns, are predictors of bad neurological outcome:

1. burst-suppression pattern with discharges of epileptiform activity. (1).
2. Generalized suppression (17).
3. Isoelectric EEG pattern (17).
4. Electroencephalographic status epilepticus (ESE) on a flat background.

Hypotheses:

1. It is possible to measure reactivity quantitatively in the EEG using power spectrum analyses and it can be correlated to outcome after 6 months using the CPC scale dichotomized in good (CPC 1-2) and poor (CPC 3-5) outcome.
2. EEG has the same predictive value for good and bad outcome in the 48h group compared to the 24h group using the dichotomized CPC-scale at 6 months.

SSEP: Somatosensory Evoked Potentials is an examination where the median nerve is stimulated and cortical response (N20) is assessed. Bilateral absence of the N20 component of a SSEP is, with one exception, invariably associated with bad outcome, CPC 4-5 (1, 2, 7). SSEP maintains its high predictive value for bad outcome during hypothermia (7, 15) and only one case is described where a patient has regained the N20 response after an initial bilaterally absent response measured after 24 hours of hypothermia (106).

SSEP is a more hardy examination than the EEG concerning medication and metabolic imbalances and therefore more prognostically reliable when the response is absent (1).

Hypothesis: The predictive value of bilaterally absence of SEP response for death or permanent vegetative state (CPC4-5) is 100% at 24, 48 and 72 hours in both groups (24h and 48h hypothermia).

Neurological examination: The brainstem is more resistant to hypoxia than the cerebral cortex which is why an absence of brainstem reflexes is a good predictor of bad outcome in comatose patients. Several studies have shown that presence of myoclonias or seizures and absence of corneal-, light- and oculocephalic reflexes along with absence of motor response to noxious stimulation is strongly associated with severe neurological deficits, permanent vegetative state or death after CA (1, 12, 15). Our aim is to validate the use of these neurological examinations in prolonged hypothermia of 48 hours and to investigate the difference in incidences in the two groups.

Hypothesis: There is a higher incidence of absent brainstem reflexes in the 24h group compared to the 48h group evaluated at 72 hours.

Biochemical markers:

Specific biochemical markers have been demonstrated to have better predictive value than many clinical examinations in prediction of poor outcome in patients who remain unconscious after cardiopulmonary resuscitation (CPR). Especially neuron-specific enolase (NSE) and S-100B have been used in prognostication of the comatose CA patient (3, 4, 8, 74, 14). Cut-off values with a predictive value for bad outcome of 100% have been suggested, but because of the small number of patients in most studies and failure to agree on which biochemical marker is the best, new studies including NSE and S-100B along with testing of new biochemical markers are needed. In this study we wish to compare the two new biochemical markers Copeptin and NT-proBNP with NSE and S-100B.

Copeptin is released in response to stress and is a sensitive and stable surrogate marker for arginine vasopressin release. Measurement of copeptin levels has been shown to be promising as a prognostic marker in patients with CA and therapeutic hypothermia (16).

B-type Natriuretic peptide (BNP) is a cardiac neurohormone synthesized in the ventricular myocardium and secreted as a response to ventricular expansion and pressure overload (11). The BNP level on arrival at the emergency room has been shown to be significantly higher among patients who die of post-resuscitation syndrome, which is when reperfusion of ischemic tissue occurs and cytokines and endotoxines are spread throughout the organism resulting in a drop in cardiac output and cerebral blood flow. The ability of BNP to predict the degree of post-resuscitation syndrome therefore makes it a promising neurologic biochemical marker (5).

NSE is released because of neuronal damage and impairment of the blood-brain barrier integrity can be detected by the release of NSE into cerebrospinal fluid (CSF) and eventually into the blood. Increases in CSF and serum NSE levels have been reported after stroke, brain injury, and CA (80). When used with GCS and SEP, NSE has been demonstrated to augment the predictive value for neurological outcome (8).

Protein S100 beta (S100B) is expressed mainly in human astroglial cells and is a surrogate marker for neuronal damage (13). S-100B rises to its peak-concentration quickly after the CA thus, constant elevation of S100B level in serum reflects its continuous release from damaged tissue and predicts bad outcome (13). S-100B serves as a prognostic predictor within 24 hours after CA, and thus at an earlier stage than other factors like NSE (4).

Our aim is to investigate the new prognostic markers Copeptin and NT-proBNP compared to NSE and S-100B in patients treated with therapeutic hypothermia after CA.

Hypotheses:

1. 48 hours of therapeutic hypothermia reduces the total release of NSE and S-100B compared to 24 hours measured as area under the curve from 24 to 72 hours.
2. The total release of NT-proBNP and Copeptin measured as area under the curve (AUC) is significantly correlated to neurological outcome (dichotomized CPC score) after 6 months.
3. To find a cut-off-value for copeptin and NT-proBNP from samples taken at arrival to the hospital that can predict good (CPC1-2) or bad (CPC 3-5) outcome with a false positive rate of 0%.

Materials and methods: EEG and SEP is performed on patients at 1: 12-24h after reaching of target temperature (T0, 33 degrees celsius), 2: 36-48h and 3: 72-96h after T0 only on comatose patients. During the EEG recording three sequences of 30 seconds with noxious stimulation are marked and will be used later in the quantitative power-spectral analysis to assess reactivity. All EEG's are assessed by a neurophysiologist as soon after the recording as possible in order to see if ESE is present and if so, the treating physician will be notified at once. The final interpretation of the EEG will be conducted blinded at a later stage.

When analysing the EEG we will catalog benign and malignant patterns partly by using Rossettis (9) simple and operational method and assessing the following parameters: Reactivity to noxious stimulation, continuous background activity and ESE.

After 6 months our patients will be categorized with the Glasgow-Pittsburgh Cerebral Performance categories (CPC), in which 1 = good recovery, 2 = moderate disability, 3 = severe disability with dependency for daily-life activity, 4 = vegetative state, and 5 = death.

SEP: Cortical N20 responses of median nerve SEPs are recorded with standard procedures during hypothermia. SEP is performed maximum 3 times on each patient as described above. Recordings are assessed by a neurophysiologist and documented as ''absent'' (bilaterally absent cortical N20 responses after left and right median nerve stimulation, in the presence of a cervical potential), ''present'' (cortical N20 response present on at least 1 side), or ''undeterminable'' (technically insufficient recording). Undeterminable SEP results are considered as present

The chosen examinations and analyses, apart from power-spectral analysis, are well known standard examinations and are performed routinely in hospitals all over the world. EEG and SEP are performed by medical student Kristoffer Nøhr and physician Christophe Duez. The interpretation is done by a neurophysiologist and Christophe Duez will participate increasingly.

Biochemical markers: NT-proBNP, NSE, S-100B and Copeptin are all taken at hospital admission, 24h, 48h and 72h. NT-proBNP is furthermore taken at 96h.

Blinding: EEG is blinded to the treating physician unless ESE is present. All primary SEP are blinded to the treating physician. If SEP is performed at 72h it is unblinded. All biochemical markers are blinded.

Statistical analysis: The QEEG study is a study of a new method in the context of CA and data is therefore not available for statistical analysis. If there is no difference in outcome between the 24h and the 48h group we will pool our data and if not we will analyse each group separately.

The aim of the study on clinical neurological examinations, classical EEG and biochemical samples will be to generate new hypotheses as our power assessments are not strong enough when we assess each examination one by one. We still believe however that the examinations are highly relevant when investigating the additive effect of a multimodal examination and not many studies have investigated such an important body of examinations within the same study. Furthermore we have no data of this kind on patients treated with prolonged hypothermia of 48 hours since it has never been done in a randomized trial.

It will be possible to enrol 100 patients in 2 years in the study based on the patient flow of Aarhus University Hospital and we believe that 50 patients in either group will be sufficient for a basic validation of well established examinations as well as new ones and to form a basis for generating new hypotheses.

Prognostic studies of this kind have not been performed previously in Denmark and we see a unique opportunity in this patient group to create a basis for new studies in the area of neuro-prognostics in Emergency Medicine and Intensive Care.

This protocol is written with biostatistical assistance.

Perspective: The aim of this study is to contribute to the understanding of how a longer period of therapeutic hypothermia affects some of the examinations used in neurological prognostic after CA.

In the future we expect EEG, SEP clinical examination and biochemical markers to become a more integrated part of a faster and more precise neurologic prognostication to guide the physician in his decisions concerning discontinuation of intensive therapy. This will be of benefit on many levels because financial implications of caring for patients in a vegetative state or prolonged impaired consciousness are substantial, family members yearn to know what functional outcome they can expect from their loved ones and maybe most important; the patients themselves can be spared suffering and unfruitful continuation of therapy.

Inclusion Criteria:

• Out-of-hospital cardiac arrest with presumable cardiac origin
• Restored spontaneous circulation after resuscitation (No need for cardiac compressions during 20 min, and clinical sign of circulation)
• Glasgow Coma Score (GCS) < 8
• Age >= 18, < 80 years

Exclusion Criteria:

• Estimated time interval from collapse to return of spontaneous circulation>60 min
• Cardiac arrest with presumable non-cardiac origin (hypoxia etc.).
• In hospital cardiac arrest
• Terminal disease
• Coagulopathy (anticoagulation treatment including thrombolysis is not an exclusion criteria)
• Unwitnessed asystolia
• Time from cardiac arrest to initiation of cooling>240 min.
• Pregnancy
• Persistent cardiogenic shock, systolic blood pressure<80 mmHg despite vasoactive treatment
• Cerebral performance category (CPC) 3-4 before cardiac arrest.
• Suspicious or confirmed intracerebral bleeding
• Suspicious or confirmed acute stroke
• Acute CABG
• No informed consent from relatives
• No informed consent from general practitioner
• Patient unstable
• Other reason