Research ethics approval was granted by the Queen Elizabeth II Health Sciences Centre (QEII) Research Ethics Board.
This is a retrospective, medical record review of patients admitted to the QEII HSC, Halifax, Nova Scotia, Canada. The QEII HSC is a tertiary care cardiac referral center for the Canadian provinces of Nova Scotia and Prince Edward Island.
We compared two cohorts of patients admitted with out-of-hospital cardiac arrest. First, a prospectively evaluated intervention group of patients managed with TH was assessed. This group included patients who suffered an out-of-hospital VF/VT cardiac arrest, had a ROSC, were admitted to our coronary care unit (CCU) from December 2006 to January 2008 and were managed with TH. A TH protocol, based on published algorithms and available evidence, was introduced to the QEII HSC in December 2006.[15
] Education sessions were provided to attending and resident physicians, and to nursing staff. Nurse educators repeatedly reviewed and provided both formal and informal feedback on the use of TH. Data was collected prospectively and recorded in the QEII TH quality control registry.
As part of this protocol, TH was considered in patients with a primary cardiac arrest rhythm of VF or pulseless VT, time to advanced life support less than 15 minutes, ROSC within 60 minutes, and persistent coma after resuscitation. Exclusion criteria included immediately improving neurological status, arrest due to noncardiac factors, persistent shock despite the use of vasopressor medications, or a history of terminal illness.
TH was initiated in the emergency department (ED) or CCU utilizing a pre-printed order sheet. Cooling options included the application of ice packs, and/or a cooling blanket, and/or cold normal saline (4 degrees Celsius) to a target temperature between 32-34 degrees Celsius. Medications were used to maintain a mean arterial pressure (MAP) greater than 75 and included norepinephrine, phenylephrine, and dopamine. Sedation was provided with infusions of propofol and midazalam, analgesia with fentanyl or morphine, and neuromuscular blockade with pancuronium or rocuronium.
Our comparative cohort consisted of out-of-hospital cardiac arrest patients admitted to our center in the 5 years preceding the implementation of the TH protocol. Patients were identified from a manual search of the admission logs of all admissions to the CCU and medical surgical intensive care units of the QEII HSC from January 2001 to December 2006. During this time, patients were not routinely cooled at our institution. Patients were included for analysis if they would have been an appropriate patient for TH had our protocol been in place at that time. Namely, we included patients with a primary cardiac arrest rhythm of VF or pulseless VT, time to advanced life support less than 15 minutes, ROSC within 60 minutes, and persistent coma after resuscitation. We also excluded patients who had an improving neurological status, arrest due to noncardiac factors, persistent shock despite the use of vasopressor medications, or a history of terminal illness.
The primary outcomes assessed were in-hospital mortality rate and cerebral outcome. Patients’ Glasgow-Pittsburg Cerebral Performance Category (GPCPC) was determined based on available documentation. A GPCPC of I or II was considered a "good" neurologic outcome in that a GPCPC I score indicates good cerebral performance and a GPCPC II score indicates moderate cerebral disability but able to live independently. A score of GPCPC III indicates severe cerebral disability with dependence on others for daily support. GPCPC score IV indicates coma or vegetative state and a patient unaware of their surroundings. GPCPC V score indicates brain death or death.
Other data included the CCU/ICU and hospital length of stay, cardiac interventions, cardiac arrest variables, and demographic data. We also recorded specific adverse events (seizures, bleeding, sepsis and acute kidney injury).
Data was collected by 2 independent persons and included patient demographics, pre-arrest status, resuscitation details, course in hospital, and discharge status. Data was analyzed with Chi-Squared analysis for discrete endpoints and Student t-test and Nonparametric Wilcoxon rank sum test for continuous variables. A Univariate analysis was performed to calculate Odds Ratios.