Study to Evaluate the Effectiveness of SGB in Preventing Post-op Atrial Fibrillation

The investigator's hypothesis is that perioperative SGB in cardiac surgery will decrease POAF. To address this, the investigator will:

1. Recruit 707 patients to undergo perioperative SGB prior to cardiac surgery (CABG, Aortic Valve Replacement, CABG/Aortic Valve Replacement).
2. Perform a double blinded post-induction/pre-operative SGB using either 10 mL 0.5% bupivacaine or 10 mL 0.9% normal saline
3. Assess the effects on POAF as defined by the Northern New England (NNE) Cardiovascular Disease Study Group.

Significance: If the investigator is able to demonstrate that perioperative SGB in cardiac surgery affects POAF, this simple, low cost and low risk procedure may significantly impact POAF incidence and thereby its attendant morbidity, mortality and costs in this clinical population.

Study subjects 707 patients electively scheduled for aortic valve, coronary revascularization surgery (CABG) or a combination of the two (CABG/AVR) will be recruited and enrolled to receive SGB at the time of surgery. Patients will be introduced to the study by their surgical team and only those providing informed consent to participate will be enrolled. The informed consent document (attached) clearly states that SGB is not currently standard of care.

Randomization The investigator will use NQuery software (Statistical Solutions Ltd., Boston, MA) to create a randomization scheme stratified by the three types of elective surgery: CABG, aortic valve replacement, and CABG plus aortic valve replacement. Patients will be randomized in a 1:1 ratio to receive either SGB (case, injection of 10 mL 0.5% bupivacaine injection) or sham SGB (control, injection of 10 mL 0.9% saline). For each of the three surgery types, treatment assignments will be placed in opaque envelopes, numbered sequentially; these three series of assignments will be maintained in separate surgery-specific boxes. Medications will be unidentified at the time of injection and the anesthesiologists and cardiac surgeons will be blinded to sham versus treatment group.

Power analysis:

Sample size was calculated based on a group sequential test of two proportions (NQuery, Statistical Solutions Ltd., Boston, MA), which takes into account a single interim analysis of SGB efficacy once 50% of the proposed study group have been enrolled. At 80% power, cumulative alpha = 0.05, with a two-sided test with continuity correction, and using the O'Brien-Fleming spending function, a total study group of n=632 (n=316/ treatment group) would be required to detect a decrease in the rate post-operative atrial fibrillation from 30% to 20% with stellate ganglion block. After taking the interim analysis (performed when n=350; significance accepted at p\<0.003) into account, significance will be accepted at p\<0.047 for the overall study. In our preliminary feasibility study, the investigator established that stellate ganglion block was successful in 88% of cases; thus, the sample size was increased by 12%, from n=632 to n=707.

Data analysis. Descriptive statistics (mean (SD), median, frequency as appropriate) will be used to summarize the demographic, procedural and clinical characteristics of the study group. To evaluate the success of randomization, the investigator will describe demographic and clinical data both overall and after stratification by surgery type and by treatment type. As intra-operative conditions can direct modification in the actual surgery performed (e.g. CABG may become CABG plus valve replacement), the investigator will also describe the proportion of subjects who were misclassified by surgery type during randomization and if necessary will assess randomization balance according to the actual surgery type received.

The frequency of atrial fibrillation (AF) will be compared between treatment groups by chi square test; this primary analysis will be performed first using an intention to treat model. In post hoc analysis the investigator will explore treatment outcomes after stratification by type of surgery and established clinical risk factors for POAF (age, sex, smoking, obesity, hypertension, diabetes mellitus, myocardial infarction, and BMI). If descriptive analyses suggest imbalance in clinical or demographic characteristics between treatment groups, these will be explored using t tests or their non-parametric equivalent (continuous data) or chi square test (categorical data), as appropriate, and logistic regression will be used to evaluate differences in AF rate between treatment groups, while taking identified covariates into account.