Perioperative mortality in diabetic patients undergoing coronary artery bypass graft surgery

Objective: Investigate risk factors for in-hospital death in diabetic patients undergoing isolated CABG. Methods: Retrospective study of 305 consecutive diabetic patients undergoing CABG in the Division of Cardiovascular Surgery of our institution from April 2004 to April 2010. Univariate analysis for categorical variables was performed with the chi-square test or Fisher’s exact as appropriate. Potential risk factors with P <0.05 in univariate analysis were included in multivariate analysis, which was performed by backward logistic regression. P values <0.05 were considered statistically significant. Results: The study population had a mean age of 61.44 years (± 9.81) and 65.6% (n=200) were male. The in-hospital mortality rate was 11.8% (n=36). The following independent risk factors for death were identified: on-pump CABG (OR 6.15, 95% CI 1 0.57 to 24, 03, P=0.009) and low cardiac output in the postoperative period (OR 34.17, 95% CI 10.46 to 111.62, P <0.001). The use of internnal thoracic artery (ITA) was an independent protective factor for death (OR 0.27, 95% CI 0.08 to 0.093, P=0.038). Conclusion: This study identified the following independent risk factors for death after CABG: on-pump CABG and low cardiac output syndrome. The use of ITA was an independent protective


INTRODUCTION INTRODUCTION INTRODUCTION INTRODUCTION INTRODUCTION
T he prevalence of diabetes mellitus (DM) throughout the western world has been increasing at an alarming rate in recent years 1 .Coronary artery disease (CAD) is often an associated condition 2 .Diabetic patients have a worse prognosis when compared to non-diabetics in relation to coronary heart disease and display different evolutions when treated by percutaneous intervention with catheter or by surgery 3 .Studies show that the presence of DM is an independent risk factor for postoperative mortality of coronary artery bypass grafting (CABG), with an odds ratio of 1.73 for death from cardiovascular causes 4 and 2.94 for overall mortality 5 .
Medical evidence leads to a greater tendency of indicating CABG in diabetics with multivessel disease 2 .In such patients CABG should always be considered in view of the benefits in the medium and long term when compared to medical and interventional treatments.However, when indicating surgery in the presence of DM, one should consider the potential increased surgical risk and special care in pre-operative and postoperative handling 2 .
The search for factors that increase surgical risk, especially modifiable, is essential in order to decrease operative mortality.
The aim of this study was to investigate risk factors for in-hospital deaths of diabetic patients undergoing CABG at our local institution.The dependent variable was the in-hospital outcome (survival or death).The independent variables were divided into three categories: The data storage was done in double-entry to validation and conduction of data consistency analysis, in order to ensure minimal error in recording information in the software.

METHODS METHODS METHODS METHODS
Univariate analysis for categorical variables was performed with the chi-square test or Fishers exact test, as appropriate.Potential risk factors with P<0.05 in univariate analysis were included in multivariate analysis, which was performed by backward logistic regression.P values <0.05 were considered statistically significant.

Description of Population and Mortality Description of Population and Mortality Description of Population and Mortality Description of Population and Mortality Description of Population and Mortality
The study population had a mean age of 61.44 years (± 9.81), 65.6% (n = 200) were male and 34.4% (n = 105) were female.The mortality rate was 11.8% (n = 36).The study population was identified among 849 coronary artery bypass surgeries, showing a prevalence of 35.9% (n = 305) of diabetes among patients undergoing this type of surgical procedure in our local institution.
The use of ITA was a protective factor for inhospital death (OR 0.26, 95% CI 0.12 to 0.54, p <0.001), see table 2.

Multivariate Logistic Regression Analysis Multivariate Logistic Regression Analysis Multivariate Logistic Regression Analysis Multivariate Logistic Regression Analysis Multivariate Logistic Regression Analysis
We identified the following independent risk factors for in-hospital deaths: CPB (OR 6.15, 95% CI 1.57 to 24.03, p = 0.009) and low cardiac output in the postoperative period (OR 34.17, 95% CI 10.46 to 111.62, p <0.001).The use of ITA was an independent protective factor for in-hospital death (OR 0.27, 95% CI 0.08 to 0.093, p = 0.038).
Table 4 shows the data from multiple logistic regression analysis.This study showed 35.9% prevalence of DM in patients undergoing CABG at our institution during the reference period.This rate was 33.4% higher than the one reported in the study of Lauruschkat et al. .6 involving 7310 patients who consecutively underwent CABG, which observed a prevalence of 29.6% of DM diagnosed preoperatively.
The observed in-hospital post-CABG mortality rate of 11.8% is considered high.We should take into account that this study deals with a population under additional surgical risk, with a greater tendency to complications that can lead to death in the postoperative period, since all the individuals are diabetics 2 .Another aspect is the fact that we are studying a population operated at a public institution.Moraes et al. 7 conducted a study involving 752 patients undergoing CABG in a private institution, showing a 1.7% mortality.Moreover, Oliver et al. 8 recently published a work involving public hospitals and showed in-hospital mortality ranging from 7.0% to 14.3%.Another recent work 9 involving 600 patients undergoing CABG in public  Table 3 Table 3 Table 3 Table 3 Table 3   ones.This may be related to the probable difference between the population assisted by private institutions (population that has better access to basic and complex health services) and the population assisted by public institutions (population that has restricted access to basic health services and even more restricted access to high-tech services).Lima et al. 11 identified an important strategy to decrease surgical morbidity and mortality: off-pump CABG.In this study, which addressed specifically the diabetic population, it was observed that the use of CPB was an independent risk factor for in-hospital deaths.Taking into account that diabetes is a systemic disease with an important inflammatory component 2 , it is assumed that there is a significant disarray after CPB.Offpump CABG eliminates the non-pulsing flow and hypothermic myocardial ischemia, decreases release of inflammatory cytokines (tumor necrosis factor alpha, interleukins) and free radicals that are associated with cardiopulmonary bypass 12 .It has been noted that CPB is associated with higher levels of activated complement factors and markers of endothelial injury 12 .These effects are expressed in the clinical arena with a decrease of complications that increase mortality, such as renal failure, stroke, infections, atrial fibrillation, need for blood transfusions and low cardiac output 13 .
The latter complication, low cardiac output, occurs in 9.1% of CABGs 14 .Our study found a 14.8% occurrence of this complication in the postoperative period (n = 45), which means an increase of 62.6% in the incidence of low cardiac output compared with that described in the literature.This probably occurred because we studied a population with diabetes, a condition associated with an increase of 1.6 times the risk of low cardiac output postoperatively 14 .Rao et al, in a study involving 4558 consecutive CABGs, observed that operative mortality was higher in patients who developed low cardiac output in comparison to those who did not (16.9%versus 0.9%, p <0.001) 14 .Oliveira et al. 8 and Pivatto et al. 15 also identified low cardiac output as a risk factor for increased operative mortality.Our study also found that low cardiac output was an independent risk factor for in-hospital deaths.The low cardiac output syndrome is a clinical outcome that may result from inadequate myocardial protection or perioperative ischemia.Patients at high risk for low cardiac output should be the focus of trials of new techniques of myocardial protection to resuscitate the ischemic myocardium.
Despite the later impairment caused by diabetes, surgical treatment of diabetic patients (especially those with multivessel disease) is associated with significant improvement in event-free survival when compared to those undergoing medical treatment and percutaneous angioplasty, as reported in the BARI study 16 and ratified by the BARI 2D study 17 .The better survival in the BARI study (mortality of 5.8% in the surgical group compared with 20.6% in the percutaneous group, with average follow up 5.4 years) was related to the implantation of at least one ITA, emphasizing the importance of such graft in improving late prognosis.It is known that implantation of ITA in the left anterior descending artery coronary constitutes an independent factor of improved survival in the long term.In some situations, surgeons are afraid to use the ITA in diabetics, especially if they are elderly and/or obese and/or present a poor quality sternum, because of the risk of a catastrophic infectious event secondary to postoperative sternal ischemia by the artery harvest of its original bed: the mediastinitis 18,19 .In other situations ITA is not used due to the discovery, during the operation, that its flow is inadequate.However, our study showed that the use of ITA in the making of a coronary bypass was an independent protective factor for death, showing that the benefit of using this type of graft in diabetics may already be initiated during hospitalization.
Being a retrospective analysis of medical records is this study's major limitation, leaving it at the mercy of all the biases associated with this type of study and also the quality of records' filling.

CONCLUSION CONCLUSION CONCLUSION CONCLUSION CONCLUSION
This study identified the following independent risk factors for in-hospital deaths after CABG in diabetics: CPB and low cardiac output syndrome.The use of ITA was an independent protective factor for death.
Perioperative mortality in diabetic patients undergoing coronary artery bypass graft surgery DISCUSSION DISCUSSION DISCUSSION DISCUSSION DISCUSSION
10: odds ratio; CI: confidence interval; COPD: chronic obstructive pulmonary disease; NYHA: New York Heart Association; PCC: previous cardiac surgery; EF: ejection fraction Rev. Col. Bras.Cir.2012;39(1):022-027 hospitals showed a mortality rate during hospital stay of 12.2%.Sá et al.10showed a mortality rate during hospital stay of 13% involving 500 patients undergoing CABG in a public institution.Apparently there is some influence of the institutional factor (public versus private), with in-hospital mortality of public institutions being higher than in private

Table 4 Table 4 Table 4 Table 4 Table 4 -
Independent risk factors for death (multivariate logistic regression analysis).