Effect of Age on Postoperative Results in Elderly Coronary By-pass Patients

Results: Of these 510 patients, 382 (74.9%) were male and 89 (17.5%) were older than 70 years of age. Post-operative mortality, pulmonary and neurologic morbidity occurred in 27 (5.3%), 78 (15.3%) and 26 (5.1%) patients, respectively. The comparison of patients with ≤ 70 and > 70 years of age showed that the differences were signifi cant (mortality 4.0% vs. 11.2%, p= 0.012; pulmonary morbidity 13.5% vs. 23.6%, p= 0.022; neurologic morbidity 3.3% vs. 13.5%, p= 0.0001). Presence of cerebrovascular disease (odds ratio= 4.72; p= 0.014) and advanced age (odds ratio= 3.55; p= 0.016) were independent predictors of neurologic complications. Every year increase above 51.5 years of age was associated with 5.8% (p= 0.049) increased neurologic complication risk.


INTRODUCTION
The number of elderly patients having coronary bypass operations (CABG) are steadily increasing (1,2) .These patients are regarded as high risk group for the higher surgical risks because of the associated high mortality and morbidity rates along with greater length of hospital stay (2,3) .Postoperative neurologic events and neuropsychological disorders constitute signifi cant sources of mortality and morbidity after coronary by-pass operations (4,5) .
Age is a known risk factor for many postoperative complications in coronary surgery and the outcome results of CABG operations in the elderly have been studied extensively (2,3,(6)(7)(8) .The purpose of this study is to analyze the outcome results in the elderly in a large number of patients from a single institution.The postoperative neurologic complications have been analyzed in detail in order to fi nd out the risk factors and the relation of age and occurrence of neurologic complications has been assessed.

PATIENTS and METHODS
The rationale for this analysis is the view of our patient sample as a representative cohort of individuals for preoperative assessment units admitting for coronary bypass surgery.Age of the patients was evaluated as an independent risk factor for postoperative mortality and neurologic morbidity.
We have analyzed the hospital records of the patients and our prospectively collected coronary by-pass database after the study was approved by the institutional ethics board and all patient consents were completed.The outcome data of coronary by-pass operations performed from December 1, 2010 to July 31, 2012 were included in the analysis.
We have obtained the preoperative demographic characteristics, operation details and postoperative outcome results from the hospital records and the prospectively collected database.Neurological complication was defi ned as presence of any cerebrovascular event (stroke, transient ischemic attack or reversible ischemic neurologic defi cit) documented by tomography and clinical examination; encephalopathy (onset after 4 th postoperative day or lasting more than four days postoperatively in order to rule out effect of anesthesia) and neurologic complications of other causes (9) .As a routine preoperative work-up, all patients were evaluated with carotid duplex ultrasonography for presence of any extracranial cerebrovascular disease.In case a hemodynamically signifi cant cerebrovascular disease detected, computerized tomographic (CTA) or magnetic resonance angiography (MRA) was performed.Presence of any cerebrovascular disease was used as a binary variable in the analysis regardless of the hemodynamic signifi cance.Patient demographic characteristics and intraoperative measurements were recorded to the coronary by-pass database by the anesthesiologist.
All patients were operated under general anesthesia and conventional cardiac surgery via median sternotomy was performed.In patients operated for concurrent carotid procedures, carotid procedure was performed prior to sternotomy.Carotid by-pass with saphenous vein graft was used in patients with localized stenosis of internal carotid artery.In carotid endarterectomies, carotid shunts were used in each case.Seventy seven (15.1%) patients were operated off-pump.Cardiopulmonary by-pass was established using a non-pulsatile hypothermic fl ow of 2.0-2.4L/ min/m 2 , with mean arterial pressure maintained from 50-70 mmHg and α-stat blood gas management.Pulsatile perfusion was used in 33 patients.Total circulatory arrest was utilized in 3 (0.6%) cases.One of them had graft interposition in the ascending aorta and the other two were operated for aortic root replacement with modifi ed Bentall procedure.They had ascending aortic aneurysms along with coronary artery disease.Axillary arterial cannulation was utilized in these cases and cerebral protection was maintained with selective antegrad cerebroplegia.Arterial cannulation was made on the ascending aorta in the remaining cases.Since we do not have epiaortic ultrasonography available, ascending aorta was assessed intraoperatively by palpation and cannulation was made in the atherosclerosis free areas.According to the procedural needs, myocardial protection was modifi ed as antegrade, retrograde or both using cold or isothermic blood cardioplegia.As a standard cardioplegia strategy we preferred antegrade induction with 1000 mL at the time of the aortic cross-clamping and afterwards either continuous retrograde or intermittent antegrade (500 mL/20 min) infusion was used until aortic cross-clamp removal.Antifi brinolytic therapy was administered with aminocaproic acid as institutional routine.The intraoperative blood glucose management was made either with bolus or continuous infusions of insulin were administered to obtain a target serum glucose < 200 mg/dL (11.1 mmol/L).In the postoperative intensive care unit, normoglycemia was achieved following the Portland cardiac surgery insulin protocol (10) .
The primary clinical endpoint of this study was postoperative mortality and secondary end point is the occurrence of any neurologic complication after coronary bypass.Patients with diabetes were defi ned as those with a diagnosis of type 1 or type 2 diabetes mellitus before hospital admission and who were on oral or parenteral hypoglycemic drugs.Renal function was analyzed as a binary variable in the database.We used Acute Kidney Injury Network criteria that was ≥ 50% postoperative increase from baseline creatinine to peak postoperative creatinine level in the fi rst ten postoperative days (11,12) .Pulmonary complication was defi ned as presence of postoperative respiratory problems necessitating intensive physiotherapy, re-intubation or prolonged mechanical ventilation and presence of pneumothorax or pulmonary effusion.Infectious complications were defi ned as positive blood, urine, sputum or wound cultures postoperatively, requiring dressings and intravenous antibiotics, requiring revision surgery (like mediastinal infection) or presence of radiographic infi ltrate.Gastrointestinal complication was defi ned as presence of any postoperative complication requiring medical or surgical intervention.
Continuous variables were described as means (± standard deviation) or median (interquartile range) as appropriate; categorical variables were described as a percentage.For descriptive purposes, patients were grouped according to age: ≤ 70 and > 70.Logistic regression model was constructed twice with age either as a continuous variable and a binary variable.At the end of the analysis a ROC curve was obtained to assess how well the variable could be discriminated and the area under curve (AUC) was calculated.A logistic regression model for the outcome neurologic morbidity and mortality was constructed with the following dependent predictors: gender, age, previous cerebrovascular event, previous myocardial infarction, hypertension, known diabetes mellitus at the time of admission, hypercholesterolemia, chronic obstructive pulmonary disease (COPD), cerebrovascular disease, family history of coronary disease, peripheral arterial disease (PAD), tobacco use, left ventricular dysfunction (ejection fraction < 40%), chronic renal failure at the time of admission, emergency operation, use of pulsatile cardiopulmonary by-pass, off-pump procedure, any concomitant procedure and degree of hypothermia during the operation.The data management and analysis was made with SPSS version 11.0 (SPSS Inc, statistical software package).A p value of 0.05 was considered statistically signifi cant.

RESULTS
Five hundred and ten patients were operated for CABG during the study period.Of these 510 patients, 382 (74.9%) were male and 128 (25.1%) female.The preoperative characteristics were outlined in Table 1 and procedural characteristics were outlined in Table 2.
Postoperative mortality was seen in 27 (5.3%) of patients and mortality rate was signifi cantly higher in elderly patients (Table 3).It can be seen that one patient had delayed awakening (Table 3).This patient was considered to have an encephalopathy since no tomographic changes were present.He had pulmonary and infectious complications and he was discharged on postoperative 52 nd day without neurologic sequel.It can be seen that elder patients were extubated later than the younger patients and the incidence of pulmonary complications were signifi -cantly higher (Table 3).However, the incidences of pulmonary complications were not signifi cantly different among groups in patients with neurologic morbidity (Table 4).
The data of the patients with neurologic complications were summarized in Table 4. Facial paralysis was seen in two patients over 70 years of age.They had carotid endarterectomy concurrent with coronary by-pass.The facial paralysis was attributed to the intraoperative retraction of the facial nerve and the patient had complete recovery during their follow-up in the outpatient clinic.None of the patients who had valvular operations or operated with total circulatory arrest had any neurologic complication.None of the patients with total circulatory arrest had any neurological morbidity but the number is too small to draw signifi cant conclusions.Of the 23 patients who had a history of previous cerebrovascular event, only 3 of them had postoperative neurologic complications.All three were over 70 years of age and none of them had stroke.Two had facial paralysis and one had encephalopathy.The patient with encephalopathy died because of pneumonia.Of the 45 patients with cerebrovascular disease, seven patients had postoperative neurologic complication.Three of them had postoperative cerebrovascular event, two had facial paralysis, one had encephalopathy and one showed delayed awakening.Two of them (with stroke) had mortality.The patients with mortality had isolated CABG procedure and the etiology for mortality was massive stroke in one and low cardiac output in the other.The logistic regression analysis was run twice; one for age as a continuous variable and the other for age as a binary variable (≤ 70 and > 70).Advanced age was signifi cantly associated with postoperative neurologic complication in each run.Advanced age was associated with neurologic complications with an odds ratio (OR) of 3.55 [95% confi dence interval (CI)= 1.27-9.90;p= 0.016].Every year increase above 51.5 is associated with 5.8% (95% CI= 0.1%-11.9%;p= 0.049) increased neurologic complication risk.The ROC curve also showed 92.0% sensitivity with statistical signifi cance (area under curve = 0.702, 95% CI= 0.599-0.804;p= 0.001) (Figure 1).The other significant associations with postoperative neurologic complica-tion was presence of cerebrovascular disease (OR= 4.72) (95% CI= 1.38-16.13;p= 0.014).

DISCUSSION
In this study, we evaluated the outcome data of heterogeneous group of patients operated for CABG for comparison between elderly and younger patients.The result of the analysis revealed marked difference of neurologic complications in elderly patients (p= 0.0001).The main difference between the groups was related to the more frequent occurrence of encephalopathy in elderly patients.The signifi cance of age persisted in the multivariate analysis and 5.8% risk increase was found for every year increase in patients older than 51.5 years of age.Aortic atheroma is a well known risk factor for stroke and it has been postulated that it may have signifi cant correlation with postoperative neurocognitive dysfunction (13) .We do not analyze the cognitive functions routinely in our daily practice.Patients were evaluated for their mental status as a part of routine daily examination and mainly the distinctive alterations detected in the physical exami-nations were noted.A detailed analysis of neurocognitive function may reveal higher incidences in both groups.Cognitive dysfunction was reported to be decreasing in the early postoperative period and the amount of decline was shown to be correlated with delayed or absence recovery of such functions (14) .The incidence of cerebrovascular event was 2% in the overall cohort and groups did  not show signifi cant difference.The results were similar to previous reports (2,15) .Goto and colleagues analyzed the relation of arteriosclerosis severity and postoperative neurologic morbidity in patients older than 60 years of age (16) .
The grouped the patients according to the degree of arteriosclerosis of the ascending aorta.The incidences of stroke in normal or mild, moderate and severe arteriosclerosis were 10.5%, 7% and 1.8%, respectively (16) .It was previously stated that (in the methods section) we did not use epiaortic ultrasound intraoperatively.However, the incidence of cerebrovascular event in our study group s compatible with the normal or mild arteriosclerosis group in the aforementioned study (16) .This may be related to the fact that our study group is a heterogeneous group of patients admitted for CABG in our clinic.
Apart from cerebrovascular event, other neurologic complications constitute important sources of postoperative morbidity.A recent analysis showed that low mini mental state examination score and high creatinine values preoperatively and prolonged cardiopulmonary by-pass durations were predictors of postoperative delirium in the elderly patients (17) .Bucerius and colleagues analyzed the predictors of postoperative delirium in cardiac surgery patients (5) .In the aforementioned analysis of more than 16.000 patients, the prevalence of delirium was 8.4% and 16.2% of these cases had postoperative stroke.The prevalence of postoperative delirium was reported to be higher in patients over 70 years of age (12.9%) and lower after off-pump coronary by-pass operations (5.1%).Unlike their analysis, we did not detect signifi cant associations with off-pump operations in our series.No-touch strategies have been postulated to be more important in occurrence of neurologic complications (16) .The use of pulsatile cardiopulmonary by-pass also did not have any signifi cant effect.So, this analysis does not add any additional information on the contradictory results of pulsatile cardiopulmonary by-pass literature, however, the number of patients is small draw meaningful conclusions (18) .The patients with postoperative delirium had higher rates of pulmonary prob-  lems, sternum instability and longer duration of intensive care stays.However, as indicated in Table 4, the extraneurologic morbidity was similar in elderly and younger patients in our study group.So, the latter associations may be independent of patient's age unlike the occurrence of neurologic morbidity.The higher prevalence of encephalopathy constitutes the major difference between the age groups (Table 3) in our study.Excluding the two patients with hepatic encephalopathy in the younger patients, the signifi cance will be greater.In our analysis, we analyzed the patients with neurologic morbidities according to the age groups (Table 4).Elderly patients did not show signifi cant differences of morbidity (other than neurologic) than the younger counterparts.The overall group comparisons (Table 3), however, revealed signifi cant differences in terms of mortality, extubation times, neurologic and pulmonary morbidity.Considering the fact that age persisted as a signifi cant risk factor in the multivariate analysis, the higher neurologic morbidity should be explained by some other mechanisms.
The preoperative characteristics were summarized in Table 1.It can be seen that the prevalence of a previous cerebrovascular event and presence of extracranial cerebrovascular disease was higher in the elderly patients.Presence of cerebrovascular disease was also significantly associated with postoperative neurologic morbidity in regression analysis unlike presence of a history of cerebrovascular event.This may be due to the fact that, lower perfusion may be a more important factor.The association of age and neurologic morbidity may be due to the lower reserve in elderly patients.Although there was lack of association of atherosclerotic parameters with neurologic morbidity (peripheral arterial disease, diabetes mellitus, hypertension, and hypercholesterolemia), the higher incidence of cerebrovascular disease and signifi cant association in regression analysis may still favor the importance of atherosclerotic process.The prevalence of hemodynamically signifi cant cerebrovascular disease was reported to be between 12% and 17% (19) .Our patient cohort had lower rates overall but the patients over 70 years of age had compatible rates.The risk of perioperative stroke was reported to be lower than 2% in the absence of cerebrovascular disease (19) .This is compatible with the cerebrovascular event rate in the overall group cerebrovascular event prevalence.Although more up to 10% stroke rates were reported for the concomitant CABG and carotid procedures, we did not have perioperative cerebrovascular event in these concomitant cases (19,20) .Still, the expected stroke rates are higher than 2.4% in combined procedures (19) .Combined procedures did not have signifi cant associations with mortality or neurologic morbidity in our study group (Table 2).Another importance of cerebrovascular disease may be the importance of lower perfusion and lower reserve in elderly patients.Another important factor was the blood transfusion.It can be seen that elderly patients had lower preoperative hemoglobin values (Table 1) and higher need of erythrocyte transfusions (Table 2).Bucerius et al. also reported increased risk of delirium with high transfusion requirements in cardiac surgery patients (5) .
Another important fi nding was the higher prevalence of pulmonary complications in elderly patients.It could be objected that the pulmonary complications were too high.This difference was mainly due to the defi nition of pulmonary complications.Many studies reporting lower numbers consider mostly respiratory failure patients.A more strict defi nition of pulmonary morbidity was given in the methods section.Even the patients requiring prolonged respiratory physiotherapy were accepted morbid in this analysis.Again, with any defi nition, older patients were reported to have higher pulmonary problems (1,21) .In spite of this high morbidity, satisfactory mid and long results have been reported (2,7) .
The patient population that was analyzed is a heterogeneous group of patients.Patients who had concomitant procedures were included in the analysis in order to simulate a real-world situation.It can be seen from our data that a signifi cant number of patients (about 8%) had concomitant procedures.As it may be a limiting factor in interpreting the results, it has other advantages.First, we tried to make a risk assessment for neurologic complications in CABG patients.It can be seen that, elderly patients had concomitant procedures signifi cantly more often (Table 2).So, we used this fact intentionally in order to analyze this variable as a separate variable in the regression analysis.Our results confi rm our hypothesis that concomitant procedures do not pose an additional risk on the neurologic outcome.
In conclusion, elderly patients have higher rates of mortality, pulmonary and neurologic morbidity after CABG operations.Although no signifi cant associations were found between age and mortality, age constitutes an independent risk factor for neurologic morbidity along with presence cerebrovascular disease.

CONFLICT of INTEREST
None declared.

Figure 1 .
Figure 1.ROC curve graph for age cut-off point.