- Email: [email protected]
The Influence of Type 2 Diabetes Mellitus in Patients Undergoing Coronary Artery Bypass Graft Surgery
The Influence of Type 2 Diabetes Mellitus in Patients Undergoing Coronary Artery Bypass Graft Surgery* An 8-Year Prospective Cohort Study Scott E. Woods, MD, MPH; J. Michael Smith, MD, FCCP; Samina Sohail, MD; Amal Sarah, MD; and Amy Engle, MA
Objective: To prospectively assess whether there are any outcome differences between patients with and without type 2 diabetes mellitus undergoing coronary artery bypass graft (CABG) surgery. Study design: This was an 8-year, prospective hospitalization cohort study. Data were collected on 225 variables concurrently with hospital admission. The main outcome was total operative mortality. In addition, we evaluated 12 morbidity outcomes. To minimize confounding, we controlled for 16 other variables. Results: A total of 6,711 patients were available for our analysis (diabetic patients, 2,178; and nondiabetic patients, 4,533). The diabetic patients were significantly more likely to be women, to have more left ventricular hypertrophy, to have a history of cerebrovascular disease, hypertension, and COPD, to have a greater body surface area, to have higher creatinine levels, to be African-American, to have undergone more elective procedures, to have a shorter pump time, and to have less of a history of tobacco use compared to nondiabetic patients (p < 0.05). Multiple regression analysis found no significant difference between the two groups for all 12 morbidity outcomes of interest. Diabetic patients experienced significantly more mortality than nondiabetic patients (relative risk, 1.67; 95% confidence interval, 1.20 to 2.30; p < 0.004). Conclusion: Patients with type 2 diabetes who are undergoing CABG surgery experience significantly more total operative mortality compared to nondiabetic patients, even after controlling for multiple variables. There was no difference between the groups for 12 morbidity outcomes. (CHEST 2004; 126:1789 –1795) Key words: cohort study; coronary artery bypass graft; diabetes mellitus Abbreviations: BSA ⫽ body surface area; CABG ⫽ coronary artery bypass graft; CAD ⫽ coronary artery disease; CPB ⫽ cardiopulmonary bypass; IABP ⫽ intraaortic balloon pump; MI ⫽ myocardial infarction; NYHA ⫽ New York Heart Association; OPCAB ⫽ off-pump coronary bypass grafting; PTCA ⫽ percutaneous transluminal coronary angioplasty; RR ⫽ relative risk
mellitus, which affects at least 17 million D iabetes Americans or approximately 6% of the population has many associated cardiovascular comorbidities. For example, patients with diabetes have the same risk for myocardial infarction (MI) as nondia*From the Bethesda Family Medicine Residency Program (Drs. Woods, Sohail, and Sarah), Cincinnati, OH; and Cardiovascular Thoracic Surgery Group (Dr. Smith), and E. Kenneth Hatton Research Center (Ms. Engle), Good Samaritan Hospital, Cincinnati, OH. Manuscript received January 28, 2004; revision accepted July 21, 2004. Reproduction of this article is prohibited without written permission from the American College of Chest Physicians (e-mail: [email protected]). Correspondence to: Scott E. Woods, MD, MPH, Director of Epidemiology, Bethesda Family Residency Program, 4411 Montgomery Rd, Suite 200, Cincinnati, OH 45212; e-mail: [email protected] www.chestjournal.org
betic patients who have had an MI.1 In addition, diabetic patients have higher fatality rates from first MIs, secondary to accelerated atherosclerosis and more extensive disease.2 Since diabetes confers a significant increased risk of death from coronary artery disease (CAD), aggressive management of risk factors and appropriate treatment of established CAD is essential. The Bypass Angioplasty Revascularization Investigation3 demonstrated the long-term survival benefit of coronary artery bypass graft (CABG) surgery over percutaneous transluminal coronary angioplasty (PTCA) in diabetic patients with multivessel CAD. This trial found that the 5-year, all-cause mortality rate for diabetic patients undergoing PTCA was 34.7%, compared to 19% for diabetic patients undergoing CHEST / 126 / 6 / DECEMBER, 2004
1789
CABG surgery. Comparatively, nondiabetic patients had 5-year mortality rates of 9.5% and 10.3%, respectively, for PTCA and CABG surgery. Conversely, other studies4,5 have suggested that very careful patient selection for each revascularization modality can result in similar outcomes between diabetic patients and nondiabetic patients. Although CABG surgery has generally become the preferred procedure for initial revascularization in diabetic patients with multivessel CAD, it is not without risk. Several studies6 –14 have examined the role of diabetes and outcomes in CABG patients, and the results have been varied. While most studies have shown that diabetic patients, in general, experience more postoperative morbidity and long-term mortality, the results of short-term mortality have been less conclusive. The purpose of this large, prospective study was to further clarify the morbidity and total operative mortality risk of diabetic patients undergoing CABG surgery.
Materials and Methods We conducted a prospective hospitalization cohort study. Patients entered the cohort on admission to the hospital and exited the cohort 30 days after hospital discharge. The study population consisted of patients undergoing CABG surgery by the Cardiovascular Thoracic Surgery Group, located in Cincinnati, OH. The inclusion criteria included CAGB surgery between October 1, 1993, and July 1, 2002, and age ⬎ 18 years. The exclusion criteria included any other surgery performed simultaneously with the CABG surgery (eg, CABG and valve surgery) and repeat procedures. In addition, patients with type 1 diabetes were eliminated from the study due to small numbers. Institutional review board approval was obtained prior to conducting this study. Data Collection Data were collected on 225 variables concurrently with hospital admission by physicians, nurses, and perfusionists for all patients undergoing cardiac surgery (partial listing, Table 1). The data were grouped into demographic, medical history, postoperative, perfusion, and procedure sections. All data forms were audited for completeness and consistency with a series of crosschecking questions answered by two different individuals. Additionally, another 10% of patient forms were randomly audited by a physician for accuracy and consistency. The data were stored in an interactive multi-institutional database (Patient Analysis and Tracking System; Axis Clinical Systems; Portland, OR). Control Variables To minimize confounding, we controlled for 16 variables. These included the demographic variables of age, gender, and race. We also controlled for the comorbidities of hypertension, body surface area (BSA), history of tobacco use, COPD, serum creatinine level, left ventricular hypertrophy, hypercholesterolemia, cerebrovascular disease history, urgency of the operation, New York Heart Association (NYHA) functional category, cardiac 1790
pump time, previous intervention in the last 30 days, and the number of venous and arterial grafts. Outcomes The primary outcome was total operative mortality, which includes deaths occurring within 30 days of the procedure and deaths occurring during the hospitalization in which the procedure was performed. The other 12 outcomes that were analyzed included the following: total length of stay; time in the ICU; intraoperative complications; return to the ICU; GI complications; pulmonary complications; neurologic complications; renal complications; arrhythmias; reoperation for bleeding; wound infections; and low cardiac output requiring assistance. Statistical Analysis Univariate analysis using 2 and t tests was performed comparing diabetes status with each of the three demographic variables and with each of the 13 comorbidities. To generate the unadjusted risks for each outcome, 2 and t tests were performed comparing diabetes with each of the 13 outcomes of interest. Then, multiple logistic regression analysis for dichotomous variables and multiple linear regression for continuous variables investigated the adjusted risk between diabetes and each of the 13 outcomes of interest, each controlling for the 16 potential confounding variables. The significance of each adjusted analysis was set at p ⫽ 0.004 (0.05/13 ⫽ 0.004) based on a complete Bonferroni adjustment for multiple comparisons. Data analysis was conducted using a statistical software package (SPSS software; SPSS Inc; Chicago, IL). Utilizing a two-tailed ␣ of 0.004 and a  of 0.10, it was estimated that approximately 6,500 patients would be required for this study to have a 90% power to find a significant difference between diabetic patients and nondiabetic patients for any outcome if they differed by 5%.
Results During the study period, 9,551 patients had CABG surgery. A total of 6,711 patients met our inclusion criteria and were available for our analysis. This consisted of 2,178 patients with type 2 diabetes and 4,533 nondiabetic patients. Among the diabetic patients, diabetes was controlled by diet in 247 patients and by oral medication in 1,207 patients, and 724 patients were insulin-dependent. Univariate analysis comparing diabetic status with the three demographic variables and the 13 comorbidities (Table 2) revealed that there was no significant difference between the diabetic patients and nondiabetic patients in terms of age, hypercholesterolemia, NYHA functional class, graft number and type, and having undergone a previous intervention in the last 30 days. The diabetic patients were significantly more likely to be women, with more patients having left ventricular hypertrophy, a history of cerebrovascular disease, hypertension, COPD, a higher BSA, higher creatinine levels, more likely to be African-American, more elective CABG procedures, shorter pump time, and less of a history of tobacco use compared to nondiabetic patients (p ⬍ 0.05). Clinical Investigations
Table 1—Explanation of Comorbidities and Risk Factors* Variables Continuous variables Age Creatinine level Pump time BSA Number of vessels bypassed Categoric variables Reported history of MI Diabetes Hypertension History of tobacco use Cerebrovascular diagnosis history Hypercholesterolemia Medicaid patient Left ventricular hypertrophy NYHA functional class Character of operation Elective Urgent Emergent Desperate Race Bleeding history COPD IMA used in graft Previous intervention ⬍ 30 d Reoperation for bleeding Arrhythmia requiring treatment Positive culture (postoperative) Renal complications Neurologic complications Pulmonary complications GI complications Low cardiac output Return to ICU Mortality Intraoperative complications Hours in ICU Hospitalization
Description Age in years at time of surgery Preoperative creatinine level Operative time patient was receiving mechanical perfusion in minutes BSA in meters squared Number of coronary arteries surgically bypassed with venous or arterial graft Patient’s self-report of prior MI in preoperative history and physical examination Includes patients with dietary as well a medical control Diastolic BP ⬎ 90 mm Hg The pack-years of tobacco use in 10-year strata Reported history of TIA, CVA, carotid bruit, and abnormal carotid pulse Yes, if total cholesterol was ⬎ 200 mg/dL or receiving medical therapy Primary insurance carrier listed on inpatient chart Mild, moderate, or severe (measured by echocardiogram) Class I, class II, class III, class IV Yes or no Next available OR within 24 h OR within 1–2 h On CPB or mechanical support White, black, Hispanic, oriental, middle-eastern, other Bleeding complications at prior surgery or coagulation disorder (eg, hemophilia, ITP, Von Willebrand factor) Mild (no medications), moderate (symptoms on exertion), and severe (symptoms at rest) Use of any part of IMA in any part of grafting process Cardiac catheterization, PTCA, atherectomy, stent placement Yes or no Any arrhythmia requiring medical therapy or pacemaker Including blood, urine (⬎ 10,000 colony count), arterial or venous lines (⬎ 15,000 colony count) Mild (double preoperative creatinine), moderate (creatinine, ⬎ 4.0 mg/dL), severe (dialysis) Any neurologic complication except mild mental status changes (severe mental status change, CVA, peripheral nerve, seizure, TIA) Any pulmonary complication with the exception of mild atelectasis Any, including severe GI bleeding, perforated ulcer, cholecystitis, hepatitis, pancreatitis, bowel obstruction, ileus, ischemic bowel Mild (dopamine, ⱕ 2 g/kg/min), moderate (pressor ⫻ 1), severe (pressor ⫻ 2), IABP, IABP ⫹ pressor, IABP ⫹ 2 pressors Yes or no Patient expired within 30 d of hospitalization Included dissection, hemorrhage, arrhythmia, cardiac laceration, cardiac dilation, air embolism, MI, aortic tear, cardiac arrest, valve trauma, unsatisfactory graft harvest Time in hours Duration in days from operation to hospital discharge
*OR ⫽ operating room; TIA ⫽ transient ischemic attack; CVA ⫽ cerebrovascular accident; ITP ⫽ idiopathic thrombocytopenia; IMA ⫽ internal mammary artery.
The unadjusted, univariate analysis comparing diabetes and the 13 outcomes of interest (Table 3) revealed no significant difference between diabetic patients and nondiabetic patients for GI complications, pulmonary complications, neurologic complications, arrhythmias, return to the ICU, intraoperative complications, ICU length of stay, and reoperation for bleeding. Diabetic patients had a significantly increased total operative mortality rate, total time in the hospital, number of renal complications and wound complications, and a significantly more frewww.chestjournal.org
quent need for therapy with pressors and intraaortic balloon pump (IABP) procedures due to low cardiac output (p ⬍ 0.05). The adjusted relative risks (RRs) comparing diabetic patients and nondiabetic patients and the 13 outcomes of interest are displayed in Table 4. There was no significant difference between the two groups for all 12 morbidity outcomes of interest. Diabetic patients experienced significantly more mortality than nondiabetic patients (RR, 1.67; 95% confidence interval, 1.20 to 2.30; p ⬍ 0.004). The confounding CHEST / 126 / 6 / DECEMBER, 2004
1791
Table 2—CABG Cohort Characteristics* Variables Continuous variables Age Creatinine level Pump time BSA Categoric variables Graft type IMA (ⱖ 1) Veins only Other Vessels bypassed, No. History of tobacco use Cerebrovascular history Hypercholesterolemia Previous intervention ⬍ 30 d Left ventricular hypertrophy Hypertension NYHA functional class I II III IV Urgency of procedure Elective Nonelective OPCAB grafting COPD Race White Black Gender Female Male
Diabetic Patients (n ⫽ 2,178)
Nondiabetic Patients (n ⫽ 4,533)
64.6 ⫾ 10.36 1.28 ⫾ 2.54 93.8 ⫾ 43.2 2.00 ⫾ .25
64.95 ⫾ 10.99 1.11 ⫾ 0.76 91.0 ⫾ 47.1 1.96 ⫾ .23
1,376 (68.3) 633 (31.4) 7 (0.3) 3.09 ⫾ 1.37 1,102 (58.5) 296 (13.6) 1,331 (64.8) 2,091 (97.1) 1,285 (62.5) 1,658 (76.3)
2,933 (70.4) 1,212 (29.1) 19 (0.5) 2.91 ⫾ 1.41 2,413 (62.9) 352 (7.8) 2,748 (64.1) 4,357 (97.0) 2,276 (53.0) 2,740 (60.5)
288 (14.3) 499 (24.8) 698 (34.6) 530 (26.3)
565 (13.5) 1,155 (27.6) 1,349 (32.2) 1,123 (26.8)
1,507 (72.5) 571 (27.5) 377 (17.3) 379 (17.4)
3,032 (69.8) 1,309 (30.2) 800 (17.6) 703 (15.5)
1,976 (90.7) 202 (9.3)
4,298 (94.8) 235 (5.2)
830 (38.1) 1,348 (61.9)
1,265 (27.9) 3,268 (72.1)
p Value 0.261 0.002 0.020 ⬍ 0.001 0.0157
⬍ 0.001 ⬍ 0.001 ⬍ 0.001 0.577 0.782 ⬍ 0.001 ⬍ 0.001 0.064
0.028
0.732 0.047 ⬍ 0.001 ⬍ 0.001
*Values given as mean ⫾ SD or No. (%), unless otherwise indicated. See Table 1 for abbreviation not used in text.
variables that had the most impact on the regression model, in order, were BSA, gender, and age. After controlling for these variables and changing our significance level to 0.004 due to the Bonferroni adjustment for multiple comparisons, the morbidity outcomes were no longer significant. Discussion In this study, we determined the impact of diabetes mellitus on morbidity and total operative mortality after coronary bypass surgery. This study has the advantage of being prospective and including a relatively high prevalence of diabetic patients (32%). Univariate analysis revealed that the diabetic patients had more cardiovascular comorbidities at the time of surgery compared to nondiabetic patients, even though there was no significant difference in age. This is similar to the findings of previous research in this area.6 –9,12,14 Our diabetic patients 1792
had experienced significantly more left ventricular hypertrophy, had more history of cerebrovascular disease, hypertension, and COPD, and were more likely to be African-American. It is noteworthy that the prevalence of hypertension in our diabetic population was high (76.3% vs 60.5%, respectively; p ⬍ 0.001) compared to the nondiabetic population. The presence of both hypertension and diabetes accelerates the development of atherosclerosis more than either comorbidity alone.15,16 Although we controlled for the presence of diabetes in our regression model, residual confounding could still be present. Our data indicate that diabetic patients are more likely to experience short-term mortality compared to nondiabetic patients (3.9% vs 2.6%, respectively; adjusted odds ratio, 1.67) even after controlling for multiple confounding variables. Our total operative mortality results were very similar to the results found in two large previous studies.6,7 Thourani et al6 reviewed the data on 12,198 CABG patients over a Clinical Investigations
Table 3—CABG Cohort: Unadjusted Hospitalized Outcomes by Diabetes* Diabetic (n ⫽ 2,178)
Outcomes Reoperation for bleeding Arrhythmia requiring treatment Renal complications Neurologic complications Intraoperative complications Pulmonary complications GI complications Low cardiac output Pressors IABP Wound complications Return to ICU Mortality Total operative Admission to 30 d postsurgery Admission to ⬎ 30 d postsurgery Hours in ICU Duration of hospitalization, d postsurgery
Nondiabetic (n ⫽ 4,533)
p Value
39 (1.8) 857 (39.5) 141 (6.5) 77 (3.5) 31 (1.5) 107 (4.9) 14 (0.6)
81 (1.8) 1,839 (40.7) 175 (3.9) 153 (3.4) 81 (1.9) 255 (5.6) 36 (0.8)
0.990 0.367 ⬍ 0.001 0.744 0.286 0.225 0.501 ⬍ 0.001
1,294 (59.7) 73 (3.4) 32 (1.5) 46 (2.1)
2,443 (54.1) 147 (3.3) 37 (0.8) 80 (1.8)
0.013 0.329
85 (3.9) 81 (3.7) 4 (0.2) 51.48 ⫾ 120 8.04 ⫾ 7.5
116 (2.6) 110 (2.4) 6 (0.1) 47.19 ⫾ 191 7.33 ⫾ 5.7
0.002 0.003 0.64 0.265 ⬍ 0.001
*Values given as No. (%) or mean ⫾ SD.
16-year period and found that diabetic patients had an increased short-term mortality rate compared to nondiabetic patients (3.9% vs 1.6%, respectively; p ⬍ 0.0001). Carson et al7 reviewed the data on 146,786 patients from the 1997 database of the Society of Thoracic Surgeons and found diabetic patients to have had an increased total operative mortality rate (3.74% vs 2.7%, respectively; adjusted odds ratio, 1.23). Conversely, other authors in two smaller studies8,9 showed no significant difference in the 30-day mortality rate. Table 4 —CABG Hospitalization Cohort: Adjusted RR Comparing Diabetics to Nondiabetics for 13 Outcomes* Outcome
RR†‡
95% CI
p Value§
Mortality Intraoperative complications Neurologic complications Wound complications Arrhythmias Low-output conditions Pulmonary complications GI complications Reoperation/bleeding ICU length of stay Return to the ICU Renal complications Length of stay
1.67 0.46 0.83 1.72 0.85 0.93 0.76 0.57 0.90 0.99 0.98 1.24 1.00
1.20–2.30 0.26–0.80 0.58–1.17 0.96–3.08 0.75–0.96 0.65–1.31 0.57–1.00 0.25–1.27 0.56–1.44 0.99–1.00 0.63–1.50 0.96–1.67 0.99–1.01
0.002 0.006 0.280 0.066 0.013 0.680 0.052 0.170 0.660 0.530 0.920 0.140 0.655
*CI ⫽ confidence interval. †Adjusted for age, gender, COPD, hypertension, hypercholesterolemia, BSA, race, serum creatinine level, left ventricular, hypertrophy, pump time, type and number of grafts, NYHA functional class, history of tobacco use, cerebrovascular disease history, previous intervention in the last 30 d and character of operation. ‡Diabetic patients compared to nondiabetic patients. §Significance level p ⫽ 0.004. www.chestjournal.org
The existing morbidity data in the literature on diabetic patients undergoing CABG is somewhat limited compared to the available mortality data. We found the diabetic patients to have more unadjusted morbidity compared to nondiabetic patients; however, after adjustment there was no significant difference. Only one previous author has found no difference in morbidity between diabetic patients and nondiabetic patients.10 Previous researchers have generally found diabetic patients to have significantly more morbidity.6 –9 Thourani et al6 found that diabetic patients had increased rates of postoperative stroke (2.9% vs 1.4%, respectively; p ⬍ 0.0001), angina, and hospital length of stay compared to nondiabetic patients. There was also a significantly reduced long-term survival rate in diabetic patients vs nondiabetic patients (5 years, 78% vs 88%, respectively; 10 years, 50% vs 71%, respectively). Similarly, Carson et al7 found diabetic patients to have a 35% increased risk of postoperative morbidity, secondary to infection, renal failure, MI, and multisystem failure. The population for our study was primarily white and did not have Medicaid insurance. It is possible that a more diverse population, similar to those in the studies conducted by Thourani et al6 and Carson et al,7 may have resulted in a different outcome. There are a few potential explanations for the increased total operative mortality demonstrated in diabetic patients undergoing CABG surgery. In our study, and in the majority of the research published in the literature on this subject, diabetic patients tend to have more comorbidities and more advanced disease compared to nondiabetic patients. Although CHEST / 126 / 6 / DECEMBER, 2004
1793
projects like ours can control for multiple known confounders using regression, there could easily be unknown confounders or even residual confounding from the controlled variables. Second, the metabolic effects from hyperglycemia and increased free fatty acids in diabetic patients may influence their clinical course. Hyperglycemia is associated with decreased function of leukocytes, increased platelet activity, and decreased endothelial function. Clinical trials of more intensive insulin therapy in hospitalized patients have suggested improved outcomes.17,18 There has been an increasing effort to find alternatives to the traditional method of performing CABG surgery. Historically, this procedure has been conducted with the assistance of complete cardiopulmonary bypass (CPB). Cardiac surgeons are now turning toward off-pump coronary bypass (OPCAB) grafting, as well as other minimally invasive techniques, in an effort to avoid the complications associated with CPB. Randomized trials comparing OPCAB grafting to CABG surgery with CPB have demonstrated that OPCAB grafting produces less renal dysfunction, a decreased release of troponin, fewer cardiac arrhythmias, the suppression of the perioperative inflammatory response, a lower infection rate, and a decreased length of hospital stay.19 It is possible that OPCAB surgery could help to reduce short-term mortality and morbidity for diabetic patients that is frequently demonstrated in the literature. Several randomized controlled studies with extended follow-up data are required to truly know whether OPCAB grafting produces better outcomes for certain populations. In our study, 17% of the surgeries were performed off-pump. The percentage of CABG and OPCAB procedures was evenly distributed in both the diabetic and nondiabetic groups (17.3% vs 17.6%, respectively; p ⫽ 0.732) [Table 2]. This project has some limitations that need to be considered while contemplating the results. First, this is a hospital cohort, and outcomes that occurred after 30 days following hospital discharge were not recorded. It is possible that diabetic patients had different results after a longer period of time. Second, the population in this study was primarily white, with few having Medicaid insurance. It is possible that a population that includes a more diverse racial distribution and socioeconomic status could have different results. And finally, there were likely some patients classified as nondiabetic patients who were actually undiagnosed diabetic patients. This misclassification would make the mortality results stronger since a difference between the two was still demonstrated. That same misclassification could be a partial explanation for the insignificant findings for some of the morbidity outcomes. 1794
Conclusion Patients with type 2 diabetes who are undergoing CABG surgery experience significantly more total operative mortality compared to nondiabetic patients, even after controlling for multiple variables. There was no difference between diabetic patients and nondiabetic patients for 12 morbidity outcomes. ACKNOWLEDGMENT: The authors would like to acknowledge the assistance of Dr. Loren F. Hiratzka in the collection and quality control of the data for this publication.
References 1 Haffner SM, Lehto S, Ronnemaa T, et al. Mortality from coronary heart disease in subjects with type 2 diabetes and in nondiabetic subjects with and without prior myocardial infarction. N Engl J Med 1998; 339:229 –234 2 Wilson P. Diabetes and coronary heart disease. Endocrinol Metab Clin North Am 2001; 30:857– 881 3 BARI Investigators. Influence of diabetes on 5-year mortality and morbidity in a randomized trial comparing CABG and PTCA in patients with multivessel disease: the Bypass Angioplasty Revascularization Investigation (BARI). Circulation 1997; 96:1761–1769 4 Barsness G, Peterson E, Ohman E, et al. Relationship between diabetes mellitus and long-term survival after coronary bypass and angioplasty. Circulation 1997; 96:633– 640 5 Niles N, McGrath P, Malenka D, et al. Northern New England Study Group: survival of patients with diabetes and multivessel coronary artery disease after surgical or percutaneous coronary revascularization: results of a large regional prospective study. J Am Coll Cardiol 2001; 37:1008 –1015 6 Thourani VH, Weintraub WS, Stein B, et al. Influence of diabetes mellitus on early and late outcome after coronary artery bypass grafting. Ann Thorac Surg 1999; 67:1045–1052 7 Carson JL, Scholz PM, Chen AY, et al. Diabetes mellitus increases short-term mortality and morbidity in patients undergoing coronary artery bypass graft surgery. J Am Coll Cardiol 2002; 40:418 – 423 8 Szabo Z, Hakason E, Svedjeholm R. Early postoperative outcome and medium-term survival in 540 diabetic and 2239 nondiabetic patients undergoing coronary artery bypass grafting. Ann Thorac Surg 2002; 74:712–719 9 Yamamoto T, Hosoda Y, Takazawa K, et al. Is diabetes mellitus a major risk factor in coronary artery bypass grafting? The influence of internal thoracic artery grafting on late survival in diabetic patients. Jpn J Thorac Cardiovasc Surg 2000; 48:344 –352 10 Risum M, Abdelnoor JL, Svennevig K, et al. Diabetes mellitus and morbidity and mortality risks after coronary artery bypass surgery. Scand J Thorac Cardiovasc Surg 1996; 30:70 –75 11 Lawrie GM, Morris GC, Glaeser DH. Influence of diabetes mellitus on the results of coronary bypass surgery: follow-up of 212 diabetic patients ten to 15 years after surgery. JAMA 1986; 256:2967–2971 12 Morris JJ, Smith LR, Jones RH, et al. Influence of diabetes and mammary artery grafting on survival after coronary bypass. Circulation 1991; 84:275–283 13 Koike Y, Nakagawa S, Kimura M. Influence of diabetes mellitus and complications on long-term outcome of coronary artery bypass surgery. J Cardiol 2000; 35:9 –17 14 Herlitz J, Wognsen GB, Karlson BW, et al. Mortality, mode of death and risk indicators for death during 5 years after coronary artery bypass grafting among patients with and Clinical Investigations
without a history of diabetes mellitus. Coron Artery Dis 2000; 11:339 –346 15 Assmann G, Schulte H. The Prospective Cardiovascular Munster Study: prevalence of hyperlipidemia in persons with hypertension and/or diabetes mellitus and the relationship to coronary heart disease. Am Heart J 1988; 116: 1713–1724 16 Fernandez-Britto JE, Bacallao J, Castillo JA, et al. Athersclerosis in diabetes and hypertension: a comparative morphometric study if their progression using an atherometric system. Zentralbl Pathol 1991; 137:487– 491
www.chestjournal.org
17 Fath-Ordoubadi F, Beatt KJ. Glucose-insulin-potassium therapy for treatment of acute myocardial infarction: an overview of randomized placebo-controlled trials. Circulation 1997; 96:1152–1156 18 Van den Berghe G, Wouters P, Weekers F, et al. Intensive insulin therapy in critically ill patients. N Engl J Med 2002; 345:1359 –1367 19 Bowles BJ, Lee JD, Dang CR, et al. Coronary artery bypass performed without the use of cardiopulmonary bypass is associated with reduced cerebral microemboli and improved clinical results. Chest 2001; 119:25–30
CHEST / 126 / 6 / DECEMBER, 2004
1795
Objective: To prospectively assess whether there are any outcome differences between patients with and without type 2 diabetes mellitus undergoing coronary artery bypass graft (CABG) surgery. Study design: This was an 8-year, prospective hospitalization cohort study. Data were collected on 225 variables concurrently with hospital admission. The main outcome was total operative mortality. In addition, we evaluated 12 morbidity outcomes. To minimize confounding, we controlled for 16 other variables. Results: A total of 6,711 patients were available for our analysis (diabetic patients, 2,178; and nondiabetic patients, 4,533). The diabetic patients were significantly more likely to be women, to have more left ventricular hypertrophy, to have a history of cerebrovascular disease, hypertension, and COPD, to have a greater body surface area, to have higher creatinine levels, to be African-American, to have undergone more elective procedures, to have a shorter pump time, and to have less of a history of tobacco use compared to nondiabetic patients (p < 0.05). Multiple regression analysis found no significant difference between the two groups for all 12 morbidity outcomes of interest. Diabetic patients experienced significantly more mortality than nondiabetic patients (relative risk, 1.67; 95% confidence interval, 1.20 to 2.30; p < 0.004). Conclusion: Patients with type 2 diabetes who are undergoing CABG surgery experience significantly more total operative mortality compared to nondiabetic patients, even after controlling for multiple variables. There was no difference between the groups for 12 morbidity outcomes. (CHEST 2004; 126:1789 –1795) Key words: cohort study; coronary artery bypass graft; diabetes mellitus Abbreviations: BSA ⫽ body surface area; CABG ⫽ coronary artery bypass graft; CAD ⫽ coronary artery disease; CPB ⫽ cardiopulmonary bypass; IABP ⫽ intraaortic balloon pump; MI ⫽ myocardial infarction; NYHA ⫽ New York Heart Association; OPCAB ⫽ off-pump coronary bypass grafting; PTCA ⫽ percutaneous transluminal coronary angioplasty; RR ⫽ relative risk
mellitus, which affects at least 17 million D iabetes Americans or approximately 6% of the population has many associated cardiovascular comorbidities. For example, patients with diabetes have the same risk for myocardial infarction (MI) as nondia*From the Bethesda Family Medicine Residency Program (Drs. Woods, Sohail, and Sarah), Cincinnati, OH; and Cardiovascular Thoracic Surgery Group (Dr. Smith), and E. Kenneth Hatton Research Center (Ms. Engle), Good Samaritan Hospital, Cincinnati, OH. Manuscript received January 28, 2004; revision accepted July 21, 2004. Reproduction of this article is prohibited without written permission from the American College of Chest Physicians (e-mail: [email protected]). Correspondence to: Scott E. Woods, MD, MPH, Director of Epidemiology, Bethesda Family Residency Program, 4411 Montgomery Rd, Suite 200, Cincinnati, OH 45212; e-mail: [email protected] www.chestjournal.org
betic patients who have had an MI.1 In addition, diabetic patients have higher fatality rates from first MIs, secondary to accelerated atherosclerosis and more extensive disease.2 Since diabetes confers a significant increased risk of death from coronary artery disease (CAD), aggressive management of risk factors and appropriate treatment of established CAD is essential. The Bypass Angioplasty Revascularization Investigation3 demonstrated the long-term survival benefit of coronary artery bypass graft (CABG) surgery over percutaneous transluminal coronary angioplasty (PTCA) in diabetic patients with multivessel CAD. This trial found that the 5-year, all-cause mortality rate for diabetic patients undergoing PTCA was 34.7%, compared to 19% for diabetic patients undergoing CHEST / 126 / 6 / DECEMBER, 2004
1789
CABG surgery. Comparatively, nondiabetic patients had 5-year mortality rates of 9.5% and 10.3%, respectively, for PTCA and CABG surgery. Conversely, other studies4,5 have suggested that very careful patient selection for each revascularization modality can result in similar outcomes between diabetic patients and nondiabetic patients. Although CABG surgery has generally become the preferred procedure for initial revascularization in diabetic patients with multivessel CAD, it is not without risk. Several studies6 –14 have examined the role of diabetes and outcomes in CABG patients, and the results have been varied. While most studies have shown that diabetic patients, in general, experience more postoperative morbidity and long-term mortality, the results of short-term mortality have been less conclusive. The purpose of this large, prospective study was to further clarify the morbidity and total operative mortality risk of diabetic patients undergoing CABG surgery.
Materials and Methods We conducted a prospective hospitalization cohort study. Patients entered the cohort on admission to the hospital and exited the cohort 30 days after hospital discharge. The study population consisted of patients undergoing CABG surgery by the Cardiovascular Thoracic Surgery Group, located in Cincinnati, OH. The inclusion criteria included CAGB surgery between October 1, 1993, and July 1, 2002, and age ⬎ 18 years. The exclusion criteria included any other surgery performed simultaneously with the CABG surgery (eg, CABG and valve surgery) and repeat procedures. In addition, patients with type 1 diabetes were eliminated from the study due to small numbers. Institutional review board approval was obtained prior to conducting this study. Data Collection Data were collected on 225 variables concurrently with hospital admission by physicians, nurses, and perfusionists for all patients undergoing cardiac surgery (partial listing, Table 1). The data were grouped into demographic, medical history, postoperative, perfusion, and procedure sections. All data forms were audited for completeness and consistency with a series of crosschecking questions answered by two different individuals. Additionally, another 10% of patient forms were randomly audited by a physician for accuracy and consistency. The data were stored in an interactive multi-institutional database (Patient Analysis and Tracking System; Axis Clinical Systems; Portland, OR). Control Variables To minimize confounding, we controlled for 16 variables. These included the demographic variables of age, gender, and race. We also controlled for the comorbidities of hypertension, body surface area (BSA), history of tobacco use, COPD, serum creatinine level, left ventricular hypertrophy, hypercholesterolemia, cerebrovascular disease history, urgency of the operation, New York Heart Association (NYHA) functional category, cardiac 1790
pump time, previous intervention in the last 30 days, and the number of venous and arterial grafts. Outcomes The primary outcome was total operative mortality, which includes deaths occurring within 30 days of the procedure and deaths occurring during the hospitalization in which the procedure was performed. The other 12 outcomes that were analyzed included the following: total length of stay; time in the ICU; intraoperative complications; return to the ICU; GI complications; pulmonary complications; neurologic complications; renal complications; arrhythmias; reoperation for bleeding; wound infections; and low cardiac output requiring assistance. Statistical Analysis Univariate analysis using 2 and t tests was performed comparing diabetes status with each of the three demographic variables and with each of the 13 comorbidities. To generate the unadjusted risks for each outcome, 2 and t tests were performed comparing diabetes with each of the 13 outcomes of interest. Then, multiple logistic regression analysis for dichotomous variables and multiple linear regression for continuous variables investigated the adjusted risk between diabetes and each of the 13 outcomes of interest, each controlling for the 16 potential confounding variables. The significance of each adjusted analysis was set at p ⫽ 0.004 (0.05/13 ⫽ 0.004) based on a complete Bonferroni adjustment for multiple comparisons. Data analysis was conducted using a statistical software package (SPSS software; SPSS Inc; Chicago, IL). Utilizing a two-tailed ␣ of 0.004 and a  of 0.10, it was estimated that approximately 6,500 patients would be required for this study to have a 90% power to find a significant difference between diabetic patients and nondiabetic patients for any outcome if they differed by 5%.
Results During the study period, 9,551 patients had CABG surgery. A total of 6,711 patients met our inclusion criteria and were available for our analysis. This consisted of 2,178 patients with type 2 diabetes and 4,533 nondiabetic patients. Among the diabetic patients, diabetes was controlled by diet in 247 patients and by oral medication in 1,207 patients, and 724 patients were insulin-dependent. Univariate analysis comparing diabetic status with the three demographic variables and the 13 comorbidities (Table 2) revealed that there was no significant difference between the diabetic patients and nondiabetic patients in terms of age, hypercholesterolemia, NYHA functional class, graft number and type, and having undergone a previous intervention in the last 30 days. The diabetic patients were significantly more likely to be women, with more patients having left ventricular hypertrophy, a history of cerebrovascular disease, hypertension, COPD, a higher BSA, higher creatinine levels, more likely to be African-American, more elective CABG procedures, shorter pump time, and less of a history of tobacco use compared to nondiabetic patients (p ⬍ 0.05). Clinical Investigations
Table 1—Explanation of Comorbidities and Risk Factors* Variables Continuous variables Age Creatinine level Pump time BSA Number of vessels bypassed Categoric variables Reported history of MI Diabetes Hypertension History of tobacco use Cerebrovascular diagnosis history Hypercholesterolemia Medicaid patient Left ventricular hypertrophy NYHA functional class Character of operation Elective Urgent Emergent Desperate Race Bleeding history COPD IMA used in graft Previous intervention ⬍ 30 d Reoperation for bleeding Arrhythmia requiring treatment Positive culture (postoperative) Renal complications Neurologic complications Pulmonary complications GI complications Low cardiac output Return to ICU Mortality Intraoperative complications Hours in ICU Hospitalization
Description Age in years at time of surgery Preoperative creatinine level Operative time patient was receiving mechanical perfusion in minutes BSA in meters squared Number of coronary arteries surgically bypassed with venous or arterial graft Patient’s self-report of prior MI in preoperative history and physical examination Includes patients with dietary as well a medical control Diastolic BP ⬎ 90 mm Hg The pack-years of tobacco use in 10-year strata Reported history of TIA, CVA, carotid bruit, and abnormal carotid pulse Yes, if total cholesterol was ⬎ 200 mg/dL or receiving medical therapy Primary insurance carrier listed on inpatient chart Mild, moderate, or severe (measured by echocardiogram) Class I, class II, class III, class IV Yes or no Next available OR within 24 h OR within 1–2 h On CPB or mechanical support White, black, Hispanic, oriental, middle-eastern, other Bleeding complications at prior surgery or coagulation disorder (eg, hemophilia, ITP, Von Willebrand factor) Mild (no medications), moderate (symptoms on exertion), and severe (symptoms at rest) Use of any part of IMA in any part of grafting process Cardiac catheterization, PTCA, atherectomy, stent placement Yes or no Any arrhythmia requiring medical therapy or pacemaker Including blood, urine (⬎ 10,000 colony count), arterial or venous lines (⬎ 15,000 colony count) Mild (double preoperative creatinine), moderate (creatinine, ⬎ 4.0 mg/dL), severe (dialysis) Any neurologic complication except mild mental status changes (severe mental status change, CVA, peripheral nerve, seizure, TIA) Any pulmonary complication with the exception of mild atelectasis Any, including severe GI bleeding, perforated ulcer, cholecystitis, hepatitis, pancreatitis, bowel obstruction, ileus, ischemic bowel Mild (dopamine, ⱕ 2 g/kg/min), moderate (pressor ⫻ 1), severe (pressor ⫻ 2), IABP, IABP ⫹ pressor, IABP ⫹ 2 pressors Yes or no Patient expired within 30 d of hospitalization Included dissection, hemorrhage, arrhythmia, cardiac laceration, cardiac dilation, air embolism, MI, aortic tear, cardiac arrest, valve trauma, unsatisfactory graft harvest Time in hours Duration in days from operation to hospital discharge
*OR ⫽ operating room; TIA ⫽ transient ischemic attack; CVA ⫽ cerebrovascular accident; ITP ⫽ idiopathic thrombocytopenia; IMA ⫽ internal mammary artery.
The unadjusted, univariate analysis comparing diabetes and the 13 outcomes of interest (Table 3) revealed no significant difference between diabetic patients and nondiabetic patients for GI complications, pulmonary complications, neurologic complications, arrhythmias, return to the ICU, intraoperative complications, ICU length of stay, and reoperation for bleeding. Diabetic patients had a significantly increased total operative mortality rate, total time in the hospital, number of renal complications and wound complications, and a significantly more frewww.chestjournal.org
quent need for therapy with pressors and intraaortic balloon pump (IABP) procedures due to low cardiac output (p ⬍ 0.05). The adjusted relative risks (RRs) comparing diabetic patients and nondiabetic patients and the 13 outcomes of interest are displayed in Table 4. There was no significant difference between the two groups for all 12 morbidity outcomes of interest. Diabetic patients experienced significantly more mortality than nondiabetic patients (RR, 1.67; 95% confidence interval, 1.20 to 2.30; p ⬍ 0.004). The confounding CHEST / 126 / 6 / DECEMBER, 2004
1791
Table 2—CABG Cohort Characteristics* Variables Continuous variables Age Creatinine level Pump time BSA Categoric variables Graft type IMA (ⱖ 1) Veins only Other Vessels bypassed, No. History of tobacco use Cerebrovascular history Hypercholesterolemia Previous intervention ⬍ 30 d Left ventricular hypertrophy Hypertension NYHA functional class I II III IV Urgency of procedure Elective Nonelective OPCAB grafting COPD Race White Black Gender Female Male
Diabetic Patients (n ⫽ 2,178)
Nondiabetic Patients (n ⫽ 4,533)
64.6 ⫾ 10.36 1.28 ⫾ 2.54 93.8 ⫾ 43.2 2.00 ⫾ .25
64.95 ⫾ 10.99 1.11 ⫾ 0.76 91.0 ⫾ 47.1 1.96 ⫾ .23
1,376 (68.3) 633 (31.4) 7 (0.3) 3.09 ⫾ 1.37 1,102 (58.5) 296 (13.6) 1,331 (64.8) 2,091 (97.1) 1,285 (62.5) 1,658 (76.3)
2,933 (70.4) 1,212 (29.1) 19 (0.5) 2.91 ⫾ 1.41 2,413 (62.9) 352 (7.8) 2,748 (64.1) 4,357 (97.0) 2,276 (53.0) 2,740 (60.5)
288 (14.3) 499 (24.8) 698 (34.6) 530 (26.3)
565 (13.5) 1,155 (27.6) 1,349 (32.2) 1,123 (26.8)
1,507 (72.5) 571 (27.5) 377 (17.3) 379 (17.4)
3,032 (69.8) 1,309 (30.2) 800 (17.6) 703 (15.5)
1,976 (90.7) 202 (9.3)
4,298 (94.8) 235 (5.2)
830 (38.1) 1,348 (61.9)
1,265 (27.9) 3,268 (72.1)
p Value 0.261 0.002 0.020 ⬍ 0.001 0.0157
⬍ 0.001 ⬍ 0.001 ⬍ 0.001 0.577 0.782 ⬍ 0.001 ⬍ 0.001 0.064
0.028
0.732 0.047 ⬍ 0.001 ⬍ 0.001
*Values given as mean ⫾ SD or No. (%), unless otherwise indicated. See Table 1 for abbreviation not used in text.
variables that had the most impact on the regression model, in order, were BSA, gender, and age. After controlling for these variables and changing our significance level to 0.004 due to the Bonferroni adjustment for multiple comparisons, the morbidity outcomes were no longer significant. Discussion In this study, we determined the impact of diabetes mellitus on morbidity and total operative mortality after coronary bypass surgery. This study has the advantage of being prospective and including a relatively high prevalence of diabetic patients (32%). Univariate analysis revealed that the diabetic patients had more cardiovascular comorbidities at the time of surgery compared to nondiabetic patients, even though there was no significant difference in age. This is similar to the findings of previous research in this area.6 –9,12,14 Our diabetic patients 1792
had experienced significantly more left ventricular hypertrophy, had more history of cerebrovascular disease, hypertension, and COPD, and were more likely to be African-American. It is noteworthy that the prevalence of hypertension in our diabetic population was high (76.3% vs 60.5%, respectively; p ⬍ 0.001) compared to the nondiabetic population. The presence of both hypertension and diabetes accelerates the development of atherosclerosis more than either comorbidity alone.15,16 Although we controlled for the presence of diabetes in our regression model, residual confounding could still be present. Our data indicate that diabetic patients are more likely to experience short-term mortality compared to nondiabetic patients (3.9% vs 2.6%, respectively; adjusted odds ratio, 1.67) even after controlling for multiple confounding variables. Our total operative mortality results were very similar to the results found in two large previous studies.6,7 Thourani et al6 reviewed the data on 12,198 CABG patients over a Clinical Investigations
Table 3—CABG Cohort: Unadjusted Hospitalized Outcomes by Diabetes* Diabetic (n ⫽ 2,178)
Outcomes Reoperation for bleeding Arrhythmia requiring treatment Renal complications Neurologic complications Intraoperative complications Pulmonary complications GI complications Low cardiac output Pressors IABP Wound complications Return to ICU Mortality Total operative Admission to 30 d postsurgery Admission to ⬎ 30 d postsurgery Hours in ICU Duration of hospitalization, d postsurgery
Nondiabetic (n ⫽ 4,533)
p Value
39 (1.8) 857 (39.5) 141 (6.5) 77 (3.5) 31 (1.5) 107 (4.9) 14 (0.6)
81 (1.8) 1,839 (40.7) 175 (3.9) 153 (3.4) 81 (1.9) 255 (5.6) 36 (0.8)
0.990 0.367 ⬍ 0.001 0.744 0.286 0.225 0.501 ⬍ 0.001
1,294 (59.7) 73 (3.4) 32 (1.5) 46 (2.1)
2,443 (54.1) 147 (3.3) 37 (0.8) 80 (1.8)
0.013 0.329
85 (3.9) 81 (3.7) 4 (0.2) 51.48 ⫾ 120 8.04 ⫾ 7.5
116 (2.6) 110 (2.4) 6 (0.1) 47.19 ⫾ 191 7.33 ⫾ 5.7
0.002 0.003 0.64 0.265 ⬍ 0.001
*Values given as No. (%) or mean ⫾ SD.
16-year period and found that diabetic patients had an increased short-term mortality rate compared to nondiabetic patients (3.9% vs 1.6%, respectively; p ⬍ 0.0001). Carson et al7 reviewed the data on 146,786 patients from the 1997 database of the Society of Thoracic Surgeons and found diabetic patients to have had an increased total operative mortality rate (3.74% vs 2.7%, respectively; adjusted odds ratio, 1.23). Conversely, other authors in two smaller studies8,9 showed no significant difference in the 30-day mortality rate. Table 4 —CABG Hospitalization Cohort: Adjusted RR Comparing Diabetics to Nondiabetics for 13 Outcomes* Outcome
RR†‡
95% CI
p Value§
Mortality Intraoperative complications Neurologic complications Wound complications Arrhythmias Low-output conditions Pulmonary complications GI complications Reoperation/bleeding ICU length of stay Return to the ICU Renal complications Length of stay
1.67 0.46 0.83 1.72 0.85 0.93 0.76 0.57 0.90 0.99 0.98 1.24 1.00
1.20–2.30 0.26–0.80 0.58–1.17 0.96–3.08 0.75–0.96 0.65–1.31 0.57–1.00 0.25–1.27 0.56–1.44 0.99–1.00 0.63–1.50 0.96–1.67 0.99–1.01
0.002 0.006 0.280 0.066 0.013 0.680 0.052 0.170 0.660 0.530 0.920 0.140 0.655
*CI ⫽ confidence interval. †Adjusted for age, gender, COPD, hypertension, hypercholesterolemia, BSA, race, serum creatinine level, left ventricular, hypertrophy, pump time, type and number of grafts, NYHA functional class, history of tobacco use, cerebrovascular disease history, previous intervention in the last 30 d and character of operation. ‡Diabetic patients compared to nondiabetic patients. §Significance level p ⫽ 0.004. www.chestjournal.org
The existing morbidity data in the literature on diabetic patients undergoing CABG is somewhat limited compared to the available mortality data. We found the diabetic patients to have more unadjusted morbidity compared to nondiabetic patients; however, after adjustment there was no significant difference. Only one previous author has found no difference in morbidity between diabetic patients and nondiabetic patients.10 Previous researchers have generally found diabetic patients to have significantly more morbidity.6 –9 Thourani et al6 found that diabetic patients had increased rates of postoperative stroke (2.9% vs 1.4%, respectively; p ⬍ 0.0001), angina, and hospital length of stay compared to nondiabetic patients. There was also a significantly reduced long-term survival rate in diabetic patients vs nondiabetic patients (5 years, 78% vs 88%, respectively; 10 years, 50% vs 71%, respectively). Similarly, Carson et al7 found diabetic patients to have a 35% increased risk of postoperative morbidity, secondary to infection, renal failure, MI, and multisystem failure. The population for our study was primarily white and did not have Medicaid insurance. It is possible that a more diverse population, similar to those in the studies conducted by Thourani et al6 and Carson et al,7 may have resulted in a different outcome. There are a few potential explanations for the increased total operative mortality demonstrated in diabetic patients undergoing CABG surgery. In our study, and in the majority of the research published in the literature on this subject, diabetic patients tend to have more comorbidities and more advanced disease compared to nondiabetic patients. Although CHEST / 126 / 6 / DECEMBER, 2004
1793
projects like ours can control for multiple known confounders using regression, there could easily be unknown confounders or even residual confounding from the controlled variables. Second, the metabolic effects from hyperglycemia and increased free fatty acids in diabetic patients may influence their clinical course. Hyperglycemia is associated with decreased function of leukocytes, increased platelet activity, and decreased endothelial function. Clinical trials of more intensive insulin therapy in hospitalized patients have suggested improved outcomes.17,18 There has been an increasing effort to find alternatives to the traditional method of performing CABG surgery. Historically, this procedure has been conducted with the assistance of complete cardiopulmonary bypass (CPB). Cardiac surgeons are now turning toward off-pump coronary bypass (OPCAB) grafting, as well as other minimally invasive techniques, in an effort to avoid the complications associated with CPB. Randomized trials comparing OPCAB grafting to CABG surgery with CPB have demonstrated that OPCAB grafting produces less renal dysfunction, a decreased release of troponin, fewer cardiac arrhythmias, the suppression of the perioperative inflammatory response, a lower infection rate, and a decreased length of hospital stay.19 It is possible that OPCAB surgery could help to reduce short-term mortality and morbidity for diabetic patients that is frequently demonstrated in the literature. Several randomized controlled studies with extended follow-up data are required to truly know whether OPCAB grafting produces better outcomes for certain populations. In our study, 17% of the surgeries were performed off-pump. The percentage of CABG and OPCAB procedures was evenly distributed in both the diabetic and nondiabetic groups (17.3% vs 17.6%, respectively; p ⫽ 0.732) [Table 2]. This project has some limitations that need to be considered while contemplating the results. First, this is a hospital cohort, and outcomes that occurred after 30 days following hospital discharge were not recorded. It is possible that diabetic patients had different results after a longer period of time. Second, the population in this study was primarily white, with few having Medicaid insurance. It is possible that a population that includes a more diverse racial distribution and socioeconomic status could have different results. And finally, there were likely some patients classified as nondiabetic patients who were actually undiagnosed diabetic patients. This misclassification would make the mortality results stronger since a difference between the two was still demonstrated. That same misclassification could be a partial explanation for the insignificant findings for some of the morbidity outcomes. 1794
Conclusion Patients with type 2 diabetes who are undergoing CABG surgery experience significantly more total operative mortality compared to nondiabetic patients, even after controlling for multiple variables. There was no difference between diabetic patients and nondiabetic patients for 12 morbidity outcomes. ACKNOWLEDGMENT: The authors would like to acknowledge the assistance of Dr. Loren F. Hiratzka in the collection and quality control of the data for this publication.
References 1 Haffner SM, Lehto S, Ronnemaa T, et al. Mortality from coronary heart disease in subjects with type 2 diabetes and in nondiabetic subjects with and without prior myocardial infarction. N Engl J Med 1998; 339:229 –234 2 Wilson P. Diabetes and coronary heart disease. Endocrinol Metab Clin North Am 2001; 30:857– 881 3 BARI Investigators. Influence of diabetes on 5-year mortality and morbidity in a randomized trial comparing CABG and PTCA in patients with multivessel disease: the Bypass Angioplasty Revascularization Investigation (BARI). Circulation 1997; 96:1761–1769 4 Barsness G, Peterson E, Ohman E, et al. Relationship between diabetes mellitus and long-term survival after coronary bypass and angioplasty. Circulation 1997; 96:633– 640 5 Niles N, McGrath P, Malenka D, et al. Northern New England Study Group: survival of patients with diabetes and multivessel coronary artery disease after surgical or percutaneous coronary revascularization: results of a large regional prospective study. J Am Coll Cardiol 2001; 37:1008 –1015 6 Thourani VH, Weintraub WS, Stein B, et al. Influence of diabetes mellitus on early and late outcome after coronary artery bypass grafting. Ann Thorac Surg 1999; 67:1045–1052 7 Carson JL, Scholz PM, Chen AY, et al. Diabetes mellitus increases short-term mortality and morbidity in patients undergoing coronary artery bypass graft surgery. J Am Coll Cardiol 2002; 40:418 – 423 8 Szabo Z, Hakason E, Svedjeholm R. Early postoperative outcome and medium-term survival in 540 diabetic and 2239 nondiabetic patients undergoing coronary artery bypass grafting. Ann Thorac Surg 2002; 74:712–719 9 Yamamoto T, Hosoda Y, Takazawa K, et al. Is diabetes mellitus a major risk factor in coronary artery bypass grafting? The influence of internal thoracic artery grafting on late survival in diabetic patients. Jpn J Thorac Cardiovasc Surg 2000; 48:344 –352 10 Risum M, Abdelnoor JL, Svennevig K, et al. Diabetes mellitus and morbidity and mortality risks after coronary artery bypass surgery. Scand J Thorac Cardiovasc Surg 1996; 30:70 –75 11 Lawrie GM, Morris GC, Glaeser DH. Influence of diabetes mellitus on the results of coronary bypass surgery: follow-up of 212 diabetic patients ten to 15 years after surgery. JAMA 1986; 256:2967–2971 12 Morris JJ, Smith LR, Jones RH, et al. Influence of diabetes and mammary artery grafting on survival after coronary bypass. Circulation 1991; 84:275–283 13 Koike Y, Nakagawa S, Kimura M. Influence of diabetes mellitus and complications on long-term outcome of coronary artery bypass surgery. J Cardiol 2000; 35:9 –17 14 Herlitz J, Wognsen GB, Karlson BW, et al. Mortality, mode of death and risk indicators for death during 5 years after coronary artery bypass grafting among patients with and Clinical Investigations
without a history of diabetes mellitus. Coron Artery Dis 2000; 11:339 –346 15 Assmann G, Schulte H. The Prospective Cardiovascular Munster Study: prevalence of hyperlipidemia in persons with hypertension and/or diabetes mellitus and the relationship to coronary heart disease. Am Heart J 1988; 116: 1713–1724 16 Fernandez-Britto JE, Bacallao J, Castillo JA, et al. Athersclerosis in diabetes and hypertension: a comparative morphometric study if their progression using an atherometric system. Zentralbl Pathol 1991; 137:487– 491
www.chestjournal.org
17 Fath-Ordoubadi F, Beatt KJ. Glucose-insulin-potassium therapy for treatment of acute myocardial infarction: an overview of randomized placebo-controlled trials. Circulation 1997; 96:1152–1156 18 Van den Berghe G, Wouters P, Weekers F, et al. Intensive insulin therapy in critically ill patients. N Engl J Med 2002; 345:1359 –1367 19 Bowles BJ, Lee JD, Dang CR, et al. Coronary artery bypass performed without the use of cardiopulmonary bypass is associated with reduced cerebral microemboli and improved clinical results. Chest 2001; 119:25–30
CHEST / 126 / 6 / DECEMBER, 2004
1795