| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a Clinical Research Unit, Division of Cardiothoracic Surgery, Rollins School of Public Health, Emory University School of Medicine, Atlanta, Georgia
b Department of Biostatistics, Rollins School of Public Health, Emory University School of Medicine, Atlanta, Georgia
Accepted for publication June 18, 2008.
* Address correspondence to Dr Thourani, 550 Peachtree St, Crawford Long Hospital, 6th Floor, Medical Office Tower, Cardiothoracic Surgery, Atlanta, GA 30308 (Email: vthoura{at}emory.edu).
Presented at the Poster Session of the Forty-fourth Annual Meeting of The Society of Thoracic Surgeons, Fort Lauderdale, FL, Jan 28–30, 2008.
| Drs Puskas and Guyton disclose that they have a financial relationship with Medtronic and Maquet Cardiovascular; Dr Lattouf with Medtronic.
|
 |
Abstract |
|---|
 Top Abstract IntroductionMaterial and MethodsResultsCommentAcknowledgmentsReferences |
|---|
Methods: Preoperative HbA1c levels were obtained in 3,201 patients undergoing primary, elective coronary artery bypass surgery at Emory Healthcare Hospitals from January 2002 to December 2006 and entered prospectively into a computerized database. Long-term survival status was determined by cross-referencing patient records with the Social Security Death Index. Log-rank (unadjusted) and Cox proportional hazards regression models (adjusted) were employed to determine whether HbA1c and diabetes mellitus were independent risk factors for reduced long-term survival, adjusted for 29 covariates. Hazard ratios for each unit increase in continuous HbA1c were calculated.
Results: Patients with HbA1c of 7% or greater had lower unadjusted 5-year survival compared with patients with HbA1c less than 7% (p = 0.001). Similarly, patients with diabetes mellitus had lower unadjusted 5-year survival compared with patients without diabetes (p < 0.001). After multivariable adjustment, higher HbA1c (measured as a continuous variable) was associated with reduced long-term survival for each unit increase in HbA1c (hazard ratio 1.15, p < 0.001), but preoperative diagnosis of diabetes was not associated with reduced long-term survival after coronary artery bypass surgery (p = 0.41). Other multivariable predictors of reduced long-term survival included age, cerebrovascular disease, elevated serum creatinine, renal insufficiency, congestive heart failure, previous myocardial infarction, chronic lung disease, and peripheral vascular disease.
Conclusions: Poor preoperative glycemic control, as measured by an elevated HbA1c, is associated with reduced long-term survival after coronary artery bypass surgery. Optimizing glucose control in these patients may improve long-term survival.
Coronary artery bypass surgery (CABG) is the preferred method of revascularization in diabetic patients with multivessel coronary disease [1]. Although diabetes mellitus has historically been associated with increased in-hospital mortality after CABG [2–4], recent reports have documented dramatic reductions in hospital mortality with the use of a continuous perioperative insulin infusion to achieve strict glycemic control [5, 6]. Depending on the severity of diabetes and its associated complications and comorbidities, long-term survival is lower in diabetic compared with nondiabetic patients after CABG [3, 7–10]. Moreover, diabetic patients treated with insulin appear to have worse outcomes compared with those managed with diet or oral hypoglycemic medications [3, 9, 10]. For these patients, aggressive treatment to achieve glycemic control is associated with a reduced incidence of diabetes-related complications [11, 12]. Therefore, treatment aimed at achieving long-term glycemic control may improve long-term survival.
Hemoglobin A1c (HbA1c) is the recommended method of monitoring long-term glycemic control in patients with diabetes mellitus [13]. The American Diabetes Association currently recommends that patients with diabetes achieve HbA1c levels less than 7%, which is associated with a lower risk of diabetes-associated complications [11, 12]. Although patients with diabetes appear to have lower long-term survival after CABG, there is no objective risk factor that quantifies the risk of long-term mortality among these patients. We previously reported that elevated preoperative HbA1c was a powerful predictor of adverse in-hospital events after CABG [14]. Therefore, the purpose of this study was to determine the predictive role of preoperative HbA1c on long-term survival after CABG.
|
Material and Methods |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentAcknowledgmentsReferences |
|---|
Perioperative Glucose Control
All patients were treated with a uniform perioperative intravenous insulin protocol. In the operating room, an insulin infusion was premixed with 125 units of insulin in 250 cc 0.9% normal saline. Routine measurement of blood glucose was obtained from serial arterial blood gases measured every 30 minutes. In the intensive care unit, insulin infusion was continued and glucose levels were obtained from arterial blood gas samples or finger-stick samples every 2 hours. The insulin infusion was initiated for blood glucose greater than 120 mg/dL and adjusted to target intraoperative blood glucose between 80 and 110 mg/dL according to the discretion of the attending cardiac anesthesiologist. In the intensive care unit, patients received a continuous insulin infusion that was adjusted to maintain blood glucose between 80 and 110 mg/dL according to a sliding scale (blood glucose – 60 x 0.04 = Units of insulin per hour). Upon transfer out of the intensive care unit, most patients were monitored with blood glucose monitoring every 4 to 6 hours (goal blood glucose < 150 mg/dL) and maintained according to the sliding scale with subcutaneous insulin in addition to their preoperative subcutaneous regimens. If necessary to achieve glucose control, continuous insulin infusion was continued after transfer out of the intensive care unit. Newly diagnosed or poorly controlled diabetic patients also received endocrinology consultation for better control in the postoperative period.
Outcomes
The primary aim of this study was to determine whether HbA1c was an independent risk factor for all-cause long-term mortality. To this end, a total of 29 covariates were collected to use as risk adjustors to ensure that the effect of HbA1c was not confounded by their influence. These covariates are listed in Table 1
by HbA1c groupings (<7% and 
7%). The HbA1c classifications were chosen according to recommendations of the American Diabetes Association, which recommends a target HbA1c of less than 7% [11, 12]. Standard STS definitions of each risk factor and outcome were used. Diabetes control method (diet, oral, insulin, none) was a nominal variable that was analyzed as separate dichotomous variables. Canadian Cardiovascular Society (CCS) classification and New York Heart Association (NYHA) classification are ordinal variables that were modeled as discrete numerical variables.
|
Long-term survival data was obtained using the Social Security Death Index, which is a public use national database of death records extracted from the United States Social Security Administration's Death Master File Extract. Persons who have died since 1963 who had a social security number and whose death has been reported to the Social Security Administration will be listed in the Social Security Death Index. The sensitivity of the Social Security Death Index (92.2%) is comparable with that of the National Death Index among American-born persons (87% to 98%) [15]. Arrangements were made with the Social Security Death Index whereby individual death records for all patients in the study cohort were purchased in bulk. Thus, for each patient who died before the cutoff date of March 31, 2007, a mortality date was provided, allowing construction of Kaplan-Meier curves and product-limit estimates of survival time to measure 1-, 3-, and 5-year survival. Cause of death was neither considered nor available; the purpose of this study was to compare all-cause mortality according to diabetes and HbA1c classification.
Statistical Analysis
Checks for data quality are employed both at the institutional level and before final entry into the STS National Adult Cardiac Database. To statistically evaluate the main effects of HbA1c in the presence of 29 potential confounders, a Cox proportional hazards regression model was constructed that modeled the mortality hazard as a function of HbA1c and 29 other risk factor covariates. Of interest was whether HbA1c, modeled as a continuous factor, was predictive of long-term mortality in the presence of diabetes status and other potential confounders. Adjusted odds ratios were reported and measured the additional increase in odds of outcome per unit increase in HbA1c. Because of the inherent inefficiencies of such a large model (eg, artificially inflated standard errors, collinearity), a backward elimination algorithm was performed to determine which risk factors, considered together, were significant predictors of long-term survival. The backwards elimination approach begins with the full (saturated) model estimates (after multiple imputation and parameter estimate combination), removes the least significant risk factor and fits the model again. This process is sequentially repeated until every model term left in the model is significant at the 
= 0.05 level. Hazard ratios were generated for each significant model term along with 95% confidence intervals. Kaplan-Meier curves were generated to determine observed differences in HbA1c groups (<7.0, 7.0), diabetic (yes/no), HbA1c among diabetic patients, and insulin dependence among diabetic patients. The data were managed and analyzed using SAS version 9.1 (SAS Institute, Cary, North Carolina) and STATA 9.0 (Stata Corp, College Station, Texas). All statistical tests were two-sided using an = 0.05 level of significance.
|
Results |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentAcknowledgmentsReferences |
|---|
7%). Preoperative demographics, clinical variables, and perioperative factors comparing patients with HbA1c greater than or less than 7 are listed in Table 1. Forty-two percent of patients (538 of 1,285) with diabetes were well controlled with HbA1c levels less than 7%. Of patients without an established diagnosis of diabetes, 4.9% (94 of 1,916) had HbA1c levels of 7% or greater, suggesting undiagnosed and untreated diabetes mellitus. Similarly, 11.2% of patients (94 of 841) with HbA1c of 7% or greater had no previous history of diabetes. Most patients in this study had multivessel coronary artery disease, more than 95% of patients received arterial grafting with the left internal mammary artery or bilateral internal mammary artery grafting, and the majority of cases were performed without the use of cardiopulmonary bypass.
Prevalence of Comorbidities
Patients with a preoperative HbA1c of 7% or greater were more likely to have comorbid conditions and complications associated with diabetes, even if a prior diagnosis of diabetes had not been made. Patients with HbA1c of 7% or greater were more likely to have preoperative renal insufficiency, renal failure requiring dialysis support, and a higher baseline creatinine level. In addition, they were more likely to have a history of cerebrovascular disease, cerebrovascular accident, hypertension, congestive heart failure, and peripheral vascular disease. These patients also had a significantly higher number of diseased vessels and received more bypass grafts. In addition, despite the same protocols in place for perioperative glucose control, patients with HbA1c of 7% or greater had higher mean blood glucose levels on postoperative day 0 and postoperative days 1 through 3.
Five-Year Survival Outcomes
Kaplan-Meier survival curves were constructed to provide estimates of 5-year survival. The mean follow-up time for patients in this study was 2.81 ± 1.40 years. Diabetic patients had a significantly lower 5-year survival compared with nondiabetic patients (81.1% versus 89.8%; Fig 1,
Table 2). Similarly, patients with a preoperative HbA1c of 7% or greater had a significantly lower 5-year survival compared with patients with a preoperative HbA1c less than 7% (82.3% versus 87.6%; Fig 2,
Table 2). In this unadjusted analysis, there was no 5-year mortality difference observed among diabetic patients according to HbA1c level of 7% or greater versus less than 7% (Fig 3). Among diabetic patients treated with insulin, 5-year survival was significantly worse compared with patients treated with diet, oral hypoglycemic medication, or nothing (78.3% versus 82.4%; Fig 4).
|
|
|
|
|
|
|
Comment |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentAcknowledgmentsReferences |
|---|
The American Diabetes Association has recommended the use of blood HbA1c as a method of assessing long-term glycemic control in diabetic patients [3]. Hemoglobin A1c is formed when glucose in the blood binds irreversibly to hemoglobin to form a stable glycated hemoglobin complex. Because red cell turnover is continuous (life span 90 to 120 days), HbA1c is not affected by short-term glycemic lability, and thus allows better assessment of glucose control over a 3 to 4 month period. The American Diabetes Association currently recommends that patients with diabetes aim for a target HbA1c of less than 7% [13], which is associated with a reduced incidence of macrovascular and microvascular complications [11, 12].
In this study, we examined the impact of preoperative HbA1c as a risk factor for 5-year mortality after CABG. Among patients with diabetes and with HbA1c of 7% or greater, 5-year survival was significantly lower compared to patients without diabetes or with HbA1c less than 7%. Among diabetic patients, those treated with insulin had a significantly lower 5-year survival compared to patients treated without insulin. However, there was no difference in unadjusted 5-year survival among diabetic patients according to HbA1c stratification (7% versus <7%). Using a Cox proportional hazard regression model, HbA1c (as a continuous variable) emerged as a significant predictor of reduced long-term survival (odds ratio 1.15, p < 0.001). This corresponds with a 15% reduction in 5-year survival for each unit increase in HbA1c. Other well-known risk factors were also significant for reduced 5-year survival including renal insufficiency, peripheral vascular disease, age, congestive heart failure, and chronic lung disease. Diabetes and insulin treatment were not significant predictors of reduced long-term survival in the logistic regression model.
These findings have several implications. First, these data suggest that it may be the associated comorbidities and complications of diabetes that are associated with reduced long-term survival, not simply the diagnosis of diabetes or treatment methods. As seen in Table 1, patients with HbA1c of 7% or greater had significantly more preoperative comorbidities than did patients with HbA1c less than 7%. Although Figure 3 suggests that there is no difference in survival among diabetic patients according to HbA1c level, this unadjusted Kaplan-Meier analysis does not take into account the impact of the higher incidence of comorbidities in patients with HbA1c levels of 7% or greater nor does it account for patients with undiagnosed diabetes. Therefore, these results may be misleading, as 11.2% of patients (94 of 841) in this study with preoperative HbA1c of 7% or greater had no previous or preoperative history of diabetes and were therefore not included in this subgroup analysis. We conclude that patients with good preoperative glucose control, reflected by HbA1c less than 7%, may have a higher 5-year survival compared with patients having suboptimal or poor glucose control (HbA1c 7%). The improvement in survival may be due to a reduction in diabetes-associated complications and comorbidities. Specifically, the adverse metabolic effects of prolonged hyperglycemia as well as the associated arteriopathy and vasculopathy may be attenuated in patients with well-controlled diabetes [5, 11].
Long-term survival after CABG has been the subject of several investigations. Mohammadi and colleagues [9] examined long-term cardiac-specific mortality on 9,125 survivors of CABG and found that cardiac survival was adversely affected by the need for insulin therapy. Diabetic patients not treated with insulin before surgery had long-term cardiac-related survival comparable with that of nondiabetic patients. In a previous report from our institution [3], patients with diabetes had a significantly reduced 5- and 10-year survival compared with patients without diabetes. Among the diabetic cohort, insulin-treated diabetes had a significantly higher 10-year mortality compared with patients treated with oral medication or diet [3]. Similarly, in a report from Leavitt and colleagues [10], diabetic patients with peripheral vascular disease and renal insufficiency had markedly worse long-term survival compared with diabetic patients without these associated comorbidities. Diabetic patients with retinopathy also have reduced long-term survival after CABG [19].
The results from our study are similar to the aforementioned studies with the exception of examining preoperative HbA1c as a risk factor in an adjusted multivariable analysis. In the present study, we have provided a quantifiable risk factor, HbA1c, after adjusting for other well-known risk factors, which can provide 5-year survival estimates. Using the method of treatment or complications associated with diabetes to determine the risk of long-term mortality has limitations. First, the treatment of diabetes has evolved substantially, so that many patients are treated with different and combined regimens in an attempt to reach euglycemia [20]. The combination of oral hypoglycemics with short- and long-term insulin analogues as well as diet and lifestyle modification is more common in the current era. Thus, many patients without diabetes-related complications may still be treated aggressively with insulin and insulin analogues in addition to oral medication and lifestyle modification in an effort to prevent these well-known complications. Finally, patients who achieve euglycemia have a reduced incidence of diabetes-related complications, regardless of their treatment modality [11–13, 16, 20]. Therefore, using a marker for long-term glycemic control such as HbA1c may provide a more accurate reflection of long-term risk after CABG.
Limitations
The retrospective nature of this study does not permit complete accounting for all sources of bias, despite advanced statistical methodology designed to correct for both treatment selection bias and potential confounders of outcomes in the preplanned analyses. In this study, patients were classified as having a diagnosis of diabetes if they presented to the hospital with a history of diabetes or if an established diagnosis of diabetes was made preoperatively. Patients with no known history of diabetes were classified as "nondiabetic" even if they were determined to be diabetic in the postoperative period. The definitions in this study were made to reflect actual practice patterns where the preoperative diagnosis of diabetes was made according to the patient's history upon presentation. Importantly, in this study as well as other retrospective analyses, how the presence or absence of diabetes is determined has far-reaching implications with regard to outcomes. This is relevant because 11.2% of patients (94 of 841) in this study with preoperative HbA1c of 7% or greater had no previous history of diabetes. With these patients, their diagnosis would have been overlooked without a preoperative HbA1c level. Currently, HbA1c is not recommended as a modality to diagnose diabetes, although it is apparent that patients with an elevated HbA1c may in fact have diabetes [21]. Other than preoperative HbA1c, routine confirmatory testing in all preoperative patients to establish a diagnosis of diabetes was not performed. Therefore, the method in which diabetics were defined may have influenced the outcomes in this study. An additional limitation of this study is that more than 70% patients underwent off-pump coronary artery bypass surgery, which does not reflect national practice patterns for surgical revascularization, although this variable was accounted for in the adjusted analysis. Finally, although we had preoperative and perioperative glucose control data, we do not have glucose control data after hospital discharge nor do we know if there were treatment changes (ie, oral medication to insulin) or changes in comorbidity status.
In conclusion, patients with diabetes and diabetes-related complications appear to have worse long-term survival after CABG. These outcomes may be more related to complications associated with diabetes such as renal insufficiency and peripheral vascular disease than to simply a diagnosis of diabetes or insulin-treatment. Using a marker of glucose control such as a preoperative HbA1c may provide more accurate risk stratification to predict long-term outcomes of patients with diabetes mellitus.
|
Acknowledgments |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentAcknowledgmentsReferences |
|---|
|
References |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentAcknowledgmentsReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  G. Nicholson and G. M. Hall Diabetes mellitus: new drugs for a new epidemic Br. J. Anaesth., July 1, 2011; 107(1): 65 - 73. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. Knapik, D. Ciesla, K. Filipiak, M. Knapik, and M. Zembala Prevalence and clinical significance of elevated preoperative glycosylated hemoglobin in diabetic patients scheduled for coronary artery surgery Eur J Cardiothorac Surg, April 1, 2011; 39(4): 484 - 489. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. Sato, G. Carvalho, T. Sato, R. Lattermann, T. Matsukawa, and T. Schricker The Association of Preoperative Glycemic Control, Intraoperative Insulin Sensitivity, and Outcomes after Cardiac Surgery J. Clin. Endocrinol. Metab., September 1, 2010; 95(9): 4338 - 4344. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. E. Halkos, P. Kilgo, O. M. Lattouf, J. D. Puskas, W. A. Cooper, R. A. Guyton, and V. H. Thourani The Effect of Diabetes Mellitus on In-Hospital and Long-Term Outcomes After Heart Valve Operations Ann. Thorac. Surg., July 1, 2010; 90(1): 124 - 130. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. Tekumit, A. R. Cenal, A. Polat, K. Uzun, C. Tataroglu, and E. Akinci Diagnostic Value of Hemoglobin A1c and Fasting Plasma Glucose Levels in Coronary Artery Bypass Grafting Patients With Undiagnosed Diabetes Mellitus Ann. Thorac. Surg., May 1, 2010; 89(5): 1482 - 1487. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ANN THORAC SURG | ASIAN CARDIOVASC THORAC ANN | EUR J CARDIOTHORAC SURG |
| J THORAC CARDIOVASC SURG | ICVTS | ALL CTSNet JOURNALS |
