Therapeutic hypothermia following resuscitation after out of hospital cardiac arrest is cost effective and supported by medical evidence [1-6]. Unfortunately some patients emerge from therapeutic hypothermia with devastating neurological complications. The Cerebral Performance Category (CPC) is an established classification scheme that uses neurologic outcome to group patients into one of five categories [7,8]. Patients in CPC of 1 or 2 are able to independently perform activities of daily living and return to some form of work; patients discharged from the ICU in CPC 1 or 2 are designated neurologic survivors. By contrast, patients in CPC 3-5 exhibit at least severe cerebral disability but may be comatose or brain dead; patients discharged from the ICU with a CPC or 3-5 are designated neurologic non-survivors. The dichotomization of outcomes into neurologic survivorship and non-survivorship based on the CPC is clinically useful for reporting outcomes. Predictors of neurologic survival may serve to identify patients who would benefit from intensive therapy thus promoting appropriate resource utilization. Preliminary evidence suggests the bispectral index (BIS) has some prognostic value particularly when the BIS is zero (100% sensitive for neurologic nonsurvivorship), however BIS alone is insufficient to predict neurologic survivorship [7,8]. The investigators report initial clinical evidence that in the first 24 hours following ICU admission a multiple risk factor model including BIS, echocardiographic, and metabolic parameters may predict neurologic survivorship or non-survivorship based on the CPC.

After obtaining Institutional Review Board approval, we retrospectively reviewed the medical records of 13 consecutive patients being treated with therapeutic hypothermia following out of hospital cardiac arrest at tertiary care university hospital in which a novel hemodynamic monitor was used by the investigators during a product quality assessment trial. This novel hemodynamic monitor, a miniaturized disposable monoplane transesophageal echocardiography (TEE) probeClariTEE™ (Imacor Inc., Garden City, NY, USA), was used for assessment of myocardial function and ventricular filling. Qualitative assessment of the left ventricle (LV) was performed using the mid-esophageal four-chamber and transgastric mid short axis views. Quantitative assessment of LV function was performed using a simple fractional of change (FAC) calculation utilizing the transgastric mid short axis view. Qualitative assessment of right ventricular (RV) function was performed using the mid-esophageal four-chamber view. RV function was elucidated from the apical displacement of the lateral tricuspid annulus, right ventricular free wall thickening, motion of the inter-ventricular septum, and right ventricular diameter relative to the left ventricle. An individual certified in Advanced Perioperative Transesophageal Echocardiography by the National Board of Echocardiography confirmed all echocardiographic assessments. There were no complications from utilizing the disposable TEE probe.

Therapeutic hypothermia was performed according to an institutional protocol including standardized ventilation and sedation strategies. All patients were sedated with midazolam and fentanyl infusions titrated to BIS 40-60. Infusions were never lower than 1 milligram per hour of midazolam and 50 micrograms per hour of fentanyl, even when BIS was <40. All patients were paralyzed with cisatracurium using a train of four technique titrated to goal of a two out of four twitches. All BIS values analyzed were obtained while the patient was paralyzed.

The investigators retrospectively dichotomized patients into two groups based on neurologic survivorship (CPC 1,2) or neurologic nonsurvivorship (CPC 3-5) and examined the medical records seeking to identify parameters that distinguished the two groups. Since the data showed a clear cut point between neurological survivors (0 or 1 risk factors) and neurological non-survivors (2 or more risk factors), sensitivity and specificity were calculated from a 2×2 contingency table with this cut point. Statistical significance was calculated with the most conservative appropriate test, Fisher’s two-tailed exact test (GraphPad Software, La Jolla CA, USA).

Risk factors for neurologic non-survival (CPC 3-5) present and identifiable within 24 hours after ICU admission are BIS < 40 (8-24 hours), arterial pH < 7.34 (12- 20 hours), best left ventricular fractional area of change < 25% (0-24 hours), and right ventricular dysfunction of at least moderate severity (0-24 hours). Tables one and two list the demographic and risk factors for the neurologic survivors and nonsurvivors respectfully (Table 1,2) Neurologic survivors exhibited 0.67 ± 0.52 risk factors while neurologic non-survivors exhibited 2.57 ± 0.79 risk factors (p = 0.0004). The presence of two or more risk factors predicts neurologic non- survival (CPC 3-5) with 100% sensitivity (95% CI: 0.56-1 due to small sample size) and 100% specificity (95% CI: 0.52- 1 due to small sample size), p=0.0006, Fisher’s two-tailed exact test.

Table 1: Neurologic Survivor Demographics and Risk Factors.

Table 2: Neurologic Non-Survivor Demographics and Risk Factors.

Our observations, limited by a small sample size and by the high percentage of drowning victims in our population, must be interpreted with caution. Whether or not our results can be reproduced there is little likelihood that BIS alone will have the sensitivity and specificity to be truly prognostic for all patients undergoing therapeutic hypothermia following out of hospital cardiac arrest. The authors postulate that an index consisting of neurologic, echocardiographic, and metabolic parameters collected within the first 24 hours following ICU admission may be useful to differentiate neurologic survivors and non-survivors. If these results can be reproduced in larger investigations, a similar scoring system may serve to identify patients who would most benefit from intensive therapy. Such a prognostic scoring system would also serve to address family angst and guide appropriate resource utilization.

In 13 patients retrospectively reviewed who underwent post cardiac arrest therapeutic hypothermia we found that within the first 24 hours BIS < 40 (8-24 hours), arterial pH < 7.34 (12- 20 hours), best left ventricular fractional area of change < 25% (0-24 hours), and right ventricular dysfunction of at least moderate severity (0-24 hours) were significant risk factors for neurologic prognostication. Future studies should delve into the value of a multiple risk factor model in neurologic prognostication post therapeutic hypothermia from cardiac arrest.

• Preliminary evidence suggests that a prognostic scoring system consisting of neurologic, cardiac, and metabolic parameters in the first 24 hours of ICU admission may predict neurologic non-survival in patients undergoing therapeutic hypothermia following out of hospital cardiac arrest.

• Echocardiographic data was obtained in all patients with the disposable TEE probe without complication.

The authors wish to thank Ronan Walker for formatting the tables.