Original Cardiovascular
Race and Survival among Diabetic Patients after Coronary Artery Bypass Grafting Wesley T. O’Neal1 Jimmy T. Efird2,3 Stephen W. Davies4 Jason B. O’Neal5 W. Randolph Chitwood2 T. Bruce Ferguson2 Alan P. Kypson2 1 Department of Internal Medicine, Wake Forest University School of
Medicine, Winston-Salem, North Carolina 2 Department of Cardiovascular Sciences, East Carolina Heart Institute, Brody School of Medicine, East Carolina University, Greenville, North Carolina 3 Center for Health Disparities, Brody School of Medicine, East Carolina University, Greenville, North Carolina 4 Department of General Surgery, University of Virginia School of Medicine, Charlottesville, Virginia 5 Department of Anesthesia, Critical Care, and Pain Medicine, Beth Israel Deaconess Medical Center, Harvard Medical School, Boston, Massachusetts
Curtis A. Anderson2
Address for correspondence Wesley T. O’Neal, MD, Department of Internal Medicine, Wake Forest University School of Medicine, Medical Center Boulevard, Winston-Salem, NC (e-mail: [email protected]).
Thorac Cardiovasc Surg 2014;62:308–316.
Abstract
Keywords
► ► ► ►
diabetes CABG outcomes epidemiology
Background Diabetes is a known predictor of decreased long-term survival after coronary artery bypass grafting (CABG). Differences in survival by race have not been examined. Methods A retrospective cohort study was conducted for CABG patients between 1992 and 2011. Long-term survival was compared in patients with and without diabetes and stratified by race. Hazard ratios (HR) and 95% confidence intervals (CI) were computed using a Cox regression model. Results Out of the 13,053 patients undergoing CABG, 35% (black n ¼ 1,655; white n ¼ 2,884) had diabetes at the time of surgery. The median follow-up for study participants was 8.2 years. Long-term survival after CABG was similar between black and white diabetic patients (no diabetes, HR ¼ 1.0; white diabetic patients, adjusted HR ¼ 1.5, 95%CI ¼ 1.4– 1.6; black diabetic patients, adjusted HR ¼ 1.5, 95%CI ¼ 1.4–1.7). Conclusion A survival disadvantage after CABG was not observed among black versus white diabetic patients in our study.
Introduction The prevalence of diabetes in the United States has increased by 82% between 1995 and 2010.1 Diabetes is a known risk factor for coronary artery disease (CAD) development, and diabetic patients are more likely to have multivessel disease.2,3 Also, an increased prevalence of diabetes and subsequent risk of mortality has been observed among blacks compared with white patients.4,5
received March 15, 2013 accepted after revision August 22, 2013 published online October 25, 2013
Diabetes is an important predictor of decreased long-term survival after coronary artery bypass grafting (CABG).6–11 In addition, the influence of insulin dependency at the time of surgery has been shown to negatively influence survival compared with patients managed by diet or oral hypoglycemic agents.7–9 Limited information is available regarding how race may affect survival of diabetic patients undergoing CABG. The current study was designed to determine and compare
© 2014 Georg Thieme Verlag KG Stuttgart · New York
DOI http://dx.doi.org/ 10.1055/s-0033-1357297. ISSN 0171-6425.
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
308
Diabetes and Survival
Patients and Methods Study Design This was a retrospective cohort study of patients undergoing isolated (e.g., no concomitant valve procedures) CABG for the first time at the East Carolina Heart Institute during 1992 and 2011. Patients receiving single or multiple bypass grafts were included in this study. Demographic data, comorbid conditions, CAD severity, and surgical data were collected at the time of surgery. Patients with diabetes were compared with those without diabetes. Only black and white patients were included to minimize the potential for residual confounding (1% other races). Racial identity was self-reported. Emergency cases were considered a clinically different population with a different etiology following surgery and were excluded in our analysis (n ¼ 420). Patients who died during the inhospital postoperative period also were excluded (n ¼ 301). The study was approved by the Institutional Review Board at the Brody School of Medicine, East Carolina University.
Operative Procedure In most cases, the left internal mammary artery was used for left anterior descending revascularization. Cardiopulmonary bypass or off-pump coronary artery bypass was selected depending upon patient presentation and surgeon preference. Cold-blood cardioplegia was used to achieve cardiac standstill. Distal anastomoses were performed first followed by proximal anastomoses. If off-pump coronary artery bypass was performed, left internal mammary artery to left anterior descending artery anastomosis was performed first, followed by the remaining distal anastomoses. Proximal anastomoses of the saphenous vein conduits were sewn directly to the ascending aorta.
Definitions Diabetes was defined as having a history of a diagnosis and/or treatment by a physician that was documented in the medical record. Diabetes management was recorded and indicated by medical therapies that were present at the time of admission. Insulin-dependent patients required insulin and insulinindependent patients were managed with diet or oral hypoglycemic agents. Mortality was defined as any cause of death postoperatively. CAD was defined as at least 50% stenosis and was confirmed by angiography before surgery.
Setting The East Carolina Heart Institute is a 120-bed cardiovascular hospital located in the center of eastern North Carolina, a rural region with a large black population. The institute is the largest stand-alone, tertiary referral hospital focusing on cardiovascular care in the state of North Carolina. Cardiovascular disease is the number one cause of death in North Carolina with an increased prevalence occurring in eastern North Carolina.12 Nearly all patients treated at the East
309
Carolina Heart Institute live and remain within a 240-km radius of the medical center.
Data Collection and Follow-Up The primary sources of data extraction were the Society of Thoracic Surgery (STS) Adult Cardiac Surgery Database and the electronic medical record at the Brody School of Medicine. Cardiovascular surgery information at our facility has been reported to the STS since 1989. Data quality and cross-field validation are routinely performed by the Epidemiology and Outcomes Research Unit at the East Carolina Heart Institute. An electronic medical record was introduced at the Brody School of Medicine in 1997. Patient information from 1989 to 1997 was retrospectively scanned into the electronic medical record, with complete conversion beginning in 1992. Local and regional clinics were consolidated under a single electronic medical record in 2005, which allowed efficient patient follow-up. The electronic medical record system applies multiple logical comparisons to reliably reduce mismatching of patient data across clinics and follow-up visits. The STS database is linked to the electronic medical record through a unique patient medical record number. The National Death Index was used to obtain death dates for patients lost to follow-up and also used to validate death information captured in our electronic medical record. Linkage with the National Death Index was based on a multiple criteria, deterministic matching algorithm.13 In our database, less than 5% of validated deaths failed to correctly match with the National Death Index.
Statistical Analysis Categorical variables were reported as frequency and percentage, while continuous variables were reported as mean standard deviation, median, and range. Variables not previously categorized were divided into quartiles before the statistical analysis. Follow-up time was measured from the date of surgery to the date of death or censoring. Survival probabilities were computed using the Kaplan-Meier product-limit method and stratified by diabetes and race. The log-rank test was used to compare survival between patients with and without diabetes and among diabetic patients by race. Cox proportional hazard regression models were used to compute hazard ratios (HR) and 95% confidence intervals (CI) for long-term mortality. The initial multivariable models included variables that have been previously reported to be associated with cardiovascularrelated mortality, regardless of their statistical significance in our dataset.14–16 These included age, sex, race, hypertension, CAD severity, heart failure, and previous stroke. The post hoc addition of other variables into the model was performed in a pairwise fashion. The test statistic of Grambsch and Therneau was used to check the proportional hazards assumption.17 Statistical significance for categorical variables was tested using the Chi-square (χ2) method. Differences in central tendency between or among groups were tested using the DeuchlerWilcoxon and Kruskal-Wallis procedures. Ptrend was computed using a likelihood ratio test. Temporality during the study period was assessed by stratifying the analysis by three time periods. Thoracic and Cardiovascular Surgeon
Vol. 62
No. 4/2014
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
survival between black and white diabetic patients following CABG.
O’Neal et al.
Diabetes and Survival
O’Neal et al.
Few values were missing (< 1% for included variables). However, when values were missing they were entered into the regression models as a separate category. A sensitivity analysis with missing values excluded also was performed to confirm that model β-coefficients did not substantively differ from the above results. Statistical significance was defined as p < 0.05. SAS Version 9.3 (SAS Institute Inc., Cary, NC) was used for all analyses.
different between black and white patients with diabetes (no diabetes, HR ¼ 1.0; white diabetic patients, adjusted HR ¼ 1.5, 95%CI ¼ 1.4–1.6; black diabetic patients, adjusted HR ¼ 1.5, 95%CI ¼ 1.4–1.7) (►Table 5). The multivariable results did not substantively change with the pairwise addition of other variables listed in ►Table 1.
Results
Consistent with previous reports, diabetes was a statistically significant predictor of decreased long-term survival after CABG in our study.6–11 However, a long-term survival disadvantage after CABG was not observed among black compared with white diabetic patients. To the best of our knowledge, we are the first to report this unique finding among diabetic patients undergoing CABG. Recent studies that have examined the influence of diabetes and survival after CABG have been limited to homogenous populations with little patient diversity.6–11 In the United States, black patients comprise approximately 4% of the STS database and this limits many centers from investigating racial differences regarding survival.18 The lack of a survival difference by race among diabetic patients in our study is contrary to several population-based reports that have shown a survival disadvantage for black patients with diabetes.4,5 A recent analysis of Medicare beneficiaries found that black diabetic patients had a higher death rate than white diabetic patients.4 Also, a national cohort study of adults aged 25 to 74 from the first National Health and Nutrition Examination Survey found that black patients with diabetes had a higher mortality rate over a 22year period than white patients with diabetes.5 However, our results suggest that survival related to diabetes may differ from survival with diabetes after CABG. Several studies also have reported black race to be an independent predictor of decreased long-term survival after CABG compared with white patients.19–23 However, survival paradoxes in black patients are well known in the literature.24–27 For example, hypertension and obesity are associated with increased survival among black hemodialysis patients.28 Black patients with diabetes in our study were more likely to be obese and hypertensive at the time of surgery than white patients. Conceivably, diabetes confers a survival advantage among blacks after CABG similar to that observed among hemodialysis patients. While this is intriguing, we cannot rule out chance as an explanation of our findings. Insulin dependency has been shown to negatively impact survival among diabetics after CABG.7–9 In the current study, no difference in survival between black and white patients with diabetes was observed when stratified by insulin therapy. In addition, males have been observed to have a higher incidence rate of diabetes than females and differences in survival may exist after CABG.29 While sex differences at presentation were noted in our sample, survival differences between black and white patients were not observed when the analysis was stratified by sex (female: no diabetes HR ¼ 1.0, black adjusted HR ¼ 1.4, 95%CI ¼ 1.2–1.6, white
Out of the 13,053 patients undergoing CABG, 35% (black n ¼ 1,655; white n ¼ 2,884) had diabetes at the time of surgery. The prevalence of diabetes was higher in black compared with white patients (47 vs. 32%; p < 0.01). Patient characteristics, preoperative medications, and operative characteristics are shown in ►Tables 1–3, respectively. The median follow-up for study participants was 8.2 years. Patients with diabetes had a higher body mass index (mean ¼ 31 5.9 vs. 28 5.1) and were more likely to be black (24 vs. 14%) and female (36 vs. 25%) than patients without diabetes (p < 0.01). On admission, diabetic patients were more likely to have three-vessel CAD (71 vs. 64%), hypertension (81 vs. 67%), peripheral arterial disease (14 vs. 10%), heart failure (21 vs. 11%), chronic obstructive pulmonary disease (9 vs. 7%), previous myocardial infarction (41 vs. 39%), and previous stroke (10 vs. 6%) (p < 0.01). Diabetic patients also were more likely to require preoperative dialysis (3 vs. 1%) but less likely to smoke (20 vs. 27%) and have left main disease (19 vs. 21%) (p < 0.01). Compared with white diabetics, black diabetic patients were more likely to be younger, female, obese, hypertensive, and to have heart failure and hemodialysis (p < 0.01; ►Table 1). In addition, they were more likely to receive preoperative diuretics, angiotensin converting enzyme inhibitors and angiotensin receptor blockers, and β-blockers (p < 0.01; ►Table 2). Intraoperative differences included shorter cardiopulmonary bypass and aortic cross-clamp times and greater likelihood for off-pump coronary artery bypass among black diabetic patients (p < 0.01; ►Table 3). The median survival for patients with and without diabetes was 11 and 16 years, respectively (p < 0.0001; ►Fig. 1). Among black and white patients with diabetes the median survival was 11 years (p ¼ 0.69). The unadjusted HR for diabetes was 1.6 (95%CI ¼ 1.5–1.7) compared with nondiabetic patients. Statistically significant survival differences also were observed for age, body mass index, CAD severity, peripheral arterial disease, heart failure, hypertension, hemodialysis, chronic obstructive pulmonary disease, and previous stroke (►Table 1). Although more black diabetic patients were female, a survival difference was not observed for gender. Similarly, a survival difference was not observed between black and white patients with diabetes when stratified by insulin dependency (►Table 4). Controlling for age, sex, race, hypertension, CAD severity, heart failure, and previous stroke, the HR for long-term mortality for diabetic patients after CABG decreased to 1.5 (95%CI ¼ 1.4–1.6). Long-term survival after CABG was not Thoracic and Cardiovascular Surgeon
Vol. 62
No. 4/2014
Discussion
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
310
Diabetes and Survival
O’Neal et al.
311
Table 1 Patient characteristics and univariable survival (N ¼ 13,053) Characteristic
No diabetes n (%)
Diabetes
Univariable HR (95%CI)
Black n (%)
White n (%)
p-Value
8,514 (65)
1,075 (8)
3,464 (27)
1.6 (1.5–1.7)b
Q1 ( 56)
2,326 (27)
316 (29)
785 (23)
< 0.0001
1.0 Referent
Q2 (> 56–64)
2,090 (25)
305 (28)
900 (26)
1.6 (1.5–1.8)
Q3 (> 64–71)
2,050 (24)
252 (23)
1,004 (29)
2.6 (2.4–2.8)
Q4 (> 71)c
2,048 (24)
202 (19)
775 (22)
4.3 (3.9–4.7)
63 11
62 10
64 10
P trend < 0.0001
64 (24–94)
62 (32–85)
65 (26–89)
6,361 (75)
560 (52)
2,324 (67)
2,153 (25)
515 (48)
1,140 (33)
7,319 (86)
1,195 (14)
Obese ( 30)
2,945 (35)
649 (60)
1,736 (50)
Overweight (25–29.9)
3,636 (43)
313 (29)
1,252 (36)
Overall
a
Mean SD Median (range)
c
Sex Male Female
c
< 0.0001
1.0 Referent 1.1 (1.02–1.2)
Race White Black
1.0 Referent 1.1 (1.06–1.2)
2 d
BMI (kg/m )
< 0.0001
1.0 Referent 1.1 (1.03–1.2)
Normal (18.5–24.9)
1,807 (21)
104 (10)
444 (13)
1.5 (1.4–1.6)
Underweight (< 18.5)c
83 (1)
3 (< 1)
12 (< 1)
2.0 (1.5–2.6)
Mean SD
28 5.1
32 6.2
30 5.8
28 (13–70)
31 (17–64)
30 (14–66)
Elective
3,484 (41)
443 (41)
1,448 (42)
Nonelective
5,030 (59)
632 (59)
2,016 (58)
653 (8)
52 (5)
175 (5)
2,379 (28)
266 (25)
823 (24)
1.6 (1.4–1.9)
5,482 (64)
757 (70)
2,466 (71)
2.0 (1.7–2.3) P trend < 0.0001
No
6,687 (79)
861 (80)
2,823 (82)
Yese
1,827 (21)
214 (20)
641 (18)
No
6,236 (73)
860 (80)
2,758 (80)
Yesc
2,278 (27)
215 (20)
706 (20)
2,771 (33)
105 (10)
740 (21)
5,743 (67)
970 (90)
2,724 (79)
7,642 (90)
902 (84)
3,008 (87)
872 (10)
173 (16)
456 (13)
Median (range)
c
P trend < 0.0001
Status 0.73
1.0 Referent 1.2 (1.1–1.3)
CAD severity 1 Vessel 2 Vessel 3 Vessel
c
0.79
1.0 Referent
Left main disease 0.30
1.0 Referent 1.2 (1.1–1.3)
Recent smoker 0.79
1.0 Referent 0.90 (0.84–0.96)
Hypertension No Yes
c
< 0.0001
1.0 Referent 1.3 (1.2–1.35)
Peripheral arterial disease No Yes
c
0.015
1.0 Referent 2.0 (1.8–2.2) (Continued)
Thoracic and Cardiovascular Surgeon
Vol. 62
No. 4/2014
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
Age (y)
Diabetes and Survival
O’Neal et al.
Table 1 (Continued) Characteristic
No diabetes n (%)
Diabetes Black n (%)
White n (%)
p-Value
7,609 (89)
790 (73)
2,799 (81)
< 0.0001
905 (11)
285 (27)
665 (19)
No
8,433 (99)
1,002 (93)
3,412 (99)
Yesc
81 (1)
73 (7)
52 (1)
No
7,951 (93)
990 (92)
3,159 (91)
Yesc
563 (7)
85 (8)
305 (9)
No
5,208 (61)
609 (57)
2,085 (60)
Yese
3,306 (39)
466 (43)
1,379 (40)
7,961 (94)
954 (89)
3,143 (91)
553 (6)
121 (11)
321 (9)
No
6,926 (81)
854 (79)
2,792 (81)
Yes
1,588 (19)
221 (21)
672 (19)
a
Univariable HR (95%CI)
Heart failure No Yes
c
1.0 Referent 2.1 (2.0–2.3)
Dialysis < 0.0001
1.0 Referent 5.2 (4.4–6.2)
COPD 0.36
1.0 Referent 1.8 (1.6–2.2)
Previous MI 0.039
1.0 Referent 1.2 (1.1–1.3)
Previous stroke No Yes
c
0.055
1.0 Referent 2.0 (1.8–2.2)
Previous PCI 0.40
1.0 Referent 0.85 (0.79–0.92)
Abbreviations: BMI, body mass index; CAD, coronary artery disease; CI, confidence interval; COPD, chronic obstructive pulmonary disease; HR, hazard ratio; MI, myocardial infarction; PCI, percutaneous coronary intervention; Q1, quartile 1; Q2, quartile 2; Q3, quartile 3; Q4, quartile 4; SD, standard deviation; y, years. a Black versus white diabetics (Chi-square). b Diabetes versus no diabetes. c p < 0.01, Chi-square across all columns (χ2) (categorical variables), Kruskal-Wallis test across all columns (continuous variables). d Missing category not shown. e p < 0.05, Chi-square across all columns (χ2) (categorical variables), Kruskal-Wallis test across all columns (continuous variables).
Table 2 Preoperative medications (N ¼ 13,053) Medication
No diabetes n (%)
Aspirin Lipid lowering agents
b
Diabetes Black n (%)
White n (%)
p-Valuea
6,057 (71)
741 (69)
2,407 (69)
0.73
3,230 (38)
514 (48)
1,585 (46)
0.24
Anticoagulants
2,800 (33)
328 (31)
1,113 (32)
0.32
Antiplatelet agentsb
4,585 (54)
479 (45)
1,717 (50)
0.0041
β-Blockersb
4,799 (57)
668 (62)
1,925 (56)
0.0001
Calcium channel blockers
2,608 (31)
358 (33)
1,056 (30)
0.081
Diureticsb
1,420 (17)
383 (36)
965 (28)
< 0.0001
ACE inhibitors/ARBsb
2,176 (26)
512 (48)
1,336 (39)
< 0.0001
Digitalis
b
438 (5)
73 (7)
298 (9)
0.058
Nitrates
1,337 (16)
158 (15)
531 (15)
0.61
Inotropic agentsc
62 (1)
16 (1)
32 (1)
0.11
Abbreviations: ACE, angiotensin converting enzyme; ARB, angiotensin receptor blocker. a Black versus white diabetics (Chi-square). b p < 0.01, Chi-square across all columns (χ2) (categorical variables), Kruskal-Wallis test across all columns (continuous variables). c p < 0.05, Chi-square across all columns (χ2) (categorical variables), Kruskal-Wallis test across all columns (continuous variables). Thoracic and Cardiovascular Surgeon
Vol. 62
No. 4/2014
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
312
Diabetes and Survival
O’Neal et al.
313
Table 3 Operative characteristics (N ¼ 13,053) Characteristic
No diabetes n (%)
Diabetes Black n (%)
White n (%)
p-Valuea
Left internal mammary arteryb,c
5,792 (68)
848 (79)
2,527 (73)
0.035
Intra-aortic balloon pumpb
543 (6)
46 (4)
188 (5)
0.20
827 (10)
112 (10)
265 (8)
0.0041 < 0.0001
Off-pump CABG
b
Cardiopulmonary bypass time (min)
b
Mean SD
98 34
93 30
101 36
Median (range)
95 (5–416)
90 (20–283)
97 (2–941)
Mean SD
62 23
60 21
64 23
Median (range)
60 (5–260)
58 (14–195)
61 (11–486)
< 0.0001
Abbreviations: CABG, coronary artery bypass grafting; min, minutes; SD, standard deviation. a Black versus white diabetics (Chi-square for categorical and Deuchler-Wilcoxon for continuous variables). b p < 0.01, Chi-square across all columns (χ2) (categorical variables), Kruskal-Wallis test across all columns (continuous variables). c Data available from 1995 to 2011.
adjusted HR ¼ 1.4, 1.3–1.6; male: no diabetes HR ¼ 1.0, black adjusted HR ¼ 1.6, 95%CI ¼ 1.4–1.9, white adjusted HR ¼ 1.5, 95%CI ¼ 1.3–1.6). Diabetes in the general population is more prevalent among black than white patients and we observed a similar finding in our analysis.4,5 Consistent with other studies, diabetics were more likely to have multivessel CAD.3 Furthermore, black patients with diabetes in our CABG population were more likely to have left main disease than white diabetic patients. These findings suggest that racial differences exist in the progression of CAD among diabetics.30
Strengths and Limitations Our study is strengthened by its large sample size and longterm follow-up. Furthermore, we were able to accurately determine time of death using a combination of the National Death Index and our comprehensive electronic medical record. An additional strength of this study is its target base. A large priority population in eastern North Carolina allowed
for us to report on a group that has experienced historical differences in socioeconomic position and discrimination. A total of 28 (97%) of the 29 counties in eastern North Carolina fall below the national per capita income of $27,915, with half reporting a value less than $20,000.31 Similarly, 90% of the counties have a higher percentage of blacks than the national value of 13.1%.31 Our results are generalizable to other lowincome, rural, and racially dichotomous populations. Diabetes diagnosis was confirmed through patient history and information recorded in our medical record. Hemoglobin A1C values were not recorded in our database before 2009 and survival may have differed by A1C. The diagnostic criteria for diabetes changed over the study period. However, few patients would have been misclassified because we relied on history and medication use recorded in our electronic medical record. Data regarding socioeconomic position, education, and income were not collected and these factors may have influenced survival.32 Payer status, which has been shown in some studies to predict survival independent of race, was
Fig. 1 Unadjusted Kaplan-Meier survival after CABG. CABG, coronary artery bypass grafting.
Thoracic and Cardiovascular Surgeon
Vol. 62
No. 4/2014
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
Aortic cross-clamp time (min)b
Diabetes and Survival
O’Neal et al.
Table 4 Unadjusted survival by insulin dependency (N ¼ 4,492) Diabetes management
Black
White
n (%)
5-, 10-, and 15-y survival
n (%)
5-, 10-, and 15-y survival
Insulin dependenta
490 (46)
79, 50, 31
1,067 (31)
79, 50, 29
Noninsulin dependent
581 (54)
86, 58, 41
2,354 (69)
83, 59, 40
Abbreviation: y, years. Note: Diabetes management was unknown for 47 patients. a Black diabetic patients were more likely to require insulin at the time of CABG than white diabetic patients (p < 0.0001).
Table 5 Survival by race (N ¼ 13,053) Comparison
Univariable HR (95%CI)
Multivariable HR (95%CI)
No diabetes
1.0 Referent
1.0 Referent
White diabetes
1.6 (1.5–1.7)
1.5 (1.4–1.7)
Black diabetes
1.6 (1.5–1.8)
1.5 (1.4–1.6)
Abbreviations: CI, confidence interval; HR, hazard ratio.
not consistently collected and consequently was not used in our analysis.33 Furthermore, we were unable to reliably estimate socioeconomic position using zip codes, because a large percentage of patients in our catchment area live in rural areas with postal box addresses. Patients in this study were recruited over a relatively long period (20 y), over which practice methods and clinical care may have changed considerably. However, results were consistent throughout the study after stratifying by three time periods, indicating the robustness of the data to temporal changes. Also, the status of several variables in our analysis may have changed over time. We did not adjust for these variables in a time-dependent manner due to their potential to be in the causal pathway. Similarly, surgical complications and medication use were not included in our analysis because of their time-dependent status. Cause of death is not recorded in the National Death Index and diabetic status may have been unrelated to their mortality. Although, we adjusted for known clinically relevant variables, we acknowledge that other unmeasured factors could have influenced our results due to the retrospective nature of this study. We considered missing values to be a distinct category and they were entered into the regression models as a separate category. However, we cannot rule out misclassification bias due to grouping missing values into a distinct category although such bias likely is trivial given the small number of missing values. Furthermore, we performed a complete case analysis with missing values removed and also a separate analysis imputing missing values using the expectationmaximization algorithm to account for missing values.34–36 The results of our sensitivity analysis suggest that missing values did not significantly affect out study findings. While quartile categorization is advantageous because it limits the influence of outliers and allows for the assessment of trend across categories, their use may have yielded overly broad categories and the potential for residual confounding. Thoracic and Cardiovascular Surgeon
Vol. 62
No. 4/2014
However, the substitution of continuous variables in our models yielded similar results. Except for race, we did not examine interactions among other clinically relevant variables in our dataset. Given the large number of potential multilevel interactions in our data, it is difficult to interpret such effects. Furthermore, we did not use regression-based tests for interaction because they are known to have weak power and often fail to detect interactions when they exist.37 Multivariable Cox regression models, rather than propensity score matching, were used to control for confounding because of potential “noncollapsibility bias” inherent to logistic regression-based propensity scores and the possible loss of power due to incomplete matching.38 Alternative methods such as machine learning (e.g., random forest algorithm) may introduce misspecification into the propensity score model due to the “black box” nature of the algorithm that obscures the etiological relationship between predictors and outcome and were not used in the current analysis.39 While in some cases our Cox proportional hazard model diverged from the proportional hazards assumption, there was no effect modification by time and results remained clinically interpretable in terms of the average relative hazard over the observation period.
Conclusion While black diabetic patients in the general population historically have worse survival than white patients, we did not observe a long-term survival difference between black and white diabetics following CABG in our rural, tertiary referral medical center. In the United States, race often is an important surrogate marker for other risk factors, such as socioeconomic position, access, and utilization of care that may not be adequately captured in clinical cardiovascular databases. However, these factors likely are important in cardiovascular disease, independent of geographical boundaries. Our unexpected finding provides important etiological
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
314
clues regarding heart disease progression following CABG among diabetic patients. Further research is needed to determine if the timely identification of CAD in black diabetic patients and appropriate access to CABG surgery will reduce racial differences in survival observed among diabetic patients.
16
17
Acknowledgment The authors would like to thank the East Carolina Heart Institute for providing resources to conduct this study.
References
18
19
1 Centers for Disease Control and Prevention (CDC). Increasing
2
3
4
5
6
7
8
9
10
11
12 13
14
15
prevalence of diagnosed diabetes—United States and Puerto Rico, 1995-2010. MMWR Morb Mortal Wkly Rep 2012;61(45): 918–921 Koskinen P, Mänttäri M, Manninen V, Huttunen JK, Heinonen OP, Frick MH. Coronary heart disease incidence in NIDDM patients in the Helsinki Heart Study. Diabetes Care 1992;15(7):820–825 Melidonis A, Dimopoulos V, Lempidakis E, et al. Angiographic study of coronary artery disease in diabetic patients in comparison with nondiabetic patients. Angiology 1999;50(12):997–1006 McBean AM, Li S, Gilbertson DT, Collins AJ. Differences in diabetes prevalence, incidence, and mortality among the elderly of four racial/ethnic groups: whites, blacks, hispanics, and asians. Diabetes Care 2004;27(10):2317–2324 Gu K, Cowie CC, Harris MI. Mortality in adults with and without diabetes in a national cohort of the U.S. population, 1971-1993. Diabetes Care 1998;21(7):1138–1145 Leavitt BJ, Sheppard L, Maloney C, et al; Northern New England Cardiovascular Disease Study Group. Effect of diabetes and associated conditions on long-term survival after coronary artery bypass graft surgery. Circulation 2004;110(11, Suppl 1):II41–II44 Luciani N, Nasso G, Gaudino M, et al. Coronary artery bypass grafting in type II diabetic patients: a comparison between insulin-dependent and non-insulin-dependent patients at short- and mid-term follow-up. Ann Thorac Surg 2003;76(4):1149–1154 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(4):1045–1052 Szabó Z, Håkanson 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(3):712–719 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 without a history of diabetes mellitus. Coron Artery Dis 2000;11(4):339–346 van Straten AH, Soliman Hamad MA, van Zundert AA, et al. Diabetes and survival after coronary artery bypass grafting: comparison with an age- and sex-matched population. Eur J Cardiothorac Surg 2010;37(5):1068–1074 Morris PJ. Heart disease and stroke in North Carolina. N C Med J 2012;73(6):448–449 Morales DL, McClellan AJ, Jacobs JP. Empowering a database with national long-term data about mortality: the use of national death registries. Cardiol Young 2008;18(Suppl 2):188–195 Efird JT, O’Neal WT, Anderson CA, et al. The effect of race and chronic obstructive pulmonary disease on long-term survival after coronary artery bypass grafting. Front Public Health Serv Syst Res 2013:1 O’Neal WT, Efird JT, Anderson CA, et al. The impact of prior percutaneous coronary intervention on long-term survival after
20
21
22
23
24
25
26
27
28
29
30
31
32
33
O’Neal et al.
coronary artery bypass grafting. Heart Lung Circ 2013 doi:pii: S1443-9506(13)00320-X. 10.1016/j.hlc.2013.04.110. [Epub ahead of print] PMID: 23683716 Efird JT, O’Neal WT, O’Neal JB, Ferguson TB, Chitwood WR, Kypson AP. Effect of peripheral arterial disease and race on survival after coronary artery bypass grafting. Ann Thorac Surg 2013;96(1): 112–118 Grambsch P, Therneau T. Proportional hazards tests and diagnostics based on weighted residuals. Biometrika 1994;81(3): 515–526 Shahian DM, O’Brien SM, Sheng S, et al. Predictors of long-term survival after coronary artery bypass grafting surgery: results from the Society of Thoracic Surgeons Adult Cardiac Surgery Database (the ASCERT study). Circulation 2012;125(12): 1491–1500 Gray RJ, Nessim S, Khan SS, Denton T, Matloff JM. Adverse 5-year outcome after coronary artery bypass surgery in blacks. Arch Intern Med 1996;156(7):769–773 Cooper WA, Thourani VH, Guyton RA, et al. Racial disparity persists after on-pump and off-pump coronary artery bypass grafting. Circulation 2009;120(11, Suppl):S59–S64 Brooks MM, Jones RH, Bach RG, et al. For the BARI Investigators. Predictors of mortality and mortality from cardiac causes in the bypass angioplasty revascularization investigation (BARI) randomized trial and registry. Circulation 2000;101(23): 2682–2689 Maynard C, Fisher LD, Passamani ER. Survival of black persons compared with white persons in the Coronary Artery Surgery Study (CASS). Am J Cardiol 1987;60(7):513–518 Taylor HA Jr, Mickel MC, Chaitman BR, Sopko G, Cutter GR, Rogers WJ. Long-term survival of African Americans in the Coronary Artery Surgery Study (CASS). J Am Coll Cardiol 1997;29(2): 358–364 Bloembergen WE, Port FK, Mauger EA, Wolfe RA. Causes of death in dialysis patients: racial and gender differences. J Am Soc Nephrol 1994;5(5):1231–1242 Kucirka LM, Grams ME, Lessler J, et al. Association of race and age with survival among patients undergoing dialysis. JAMA 2011; 306(6):620–626 Owen WF Jr, Chertow GM, Lazarus JM, Lowrie EG. Dose of hemodialysis and survival: differences by race and sex. JAMA 1998;280(20):1764–1768 Agodoa L, Eggers P. Racial and ethnic disparities in end-stage kidney failure-survival paradoxes in African-Americans. Semin Dial 2007;20(6):577–585 Kalantar-Zadeh K, Kovesdy CP, Derose SF, Horwich TB, Fonarow GC. Racial and survival paradoxes in chronic kidney disease. Nat Clin Pract Nephrol 2007;3(9):493–506 Centers for Disease Control. Age-adjusted incidence of diagnosed diabetes per 1,000 population aged 18–79 years, by sex, United States, 1980–2011. http://www.cdc.gov/diabetes/statistics/incidence/fig4.htm. Accessed May 23, 2013 McClelland RL, Chung H, Detrano R, Post W, Kronmal RA. Distribution of coronary artery calcium by race, gender, and age: results from the Multi-Ethnic Study of Atherosclerosis (MESA). Circulation 2006;113(1):30–37 United States Census Bureau. State & County Quick Facts2010. http://quickfacts.census.gov/qfd/index.html. Accessed June 1, 2013 Koch CG, Li L, Kaplan GA, et al. Socioeconomic position, not race, is linked to death after cardiac surgery. Circ Cardiovasc Qual Outcomes 2010;3(3):267–276 Zacharias A, Schwann TA, Riordan CJ, Durham SJ, Shah A, Habib RH. Operative and late coronary artery bypass grafting outcomes in matched African-American versus Caucasian patients: evidence of a late survival-Medicaid association. J Am Coll Cardiol 2005;46(8): 1526–1535 Thoracic and Cardiovascular Surgeon
Vol. 62
No. 4/2014
315
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
Diabetes and Survival
Diabetes and Survival
O’Neal et al.
34 Dempster AP, Laird NM, Rubin DB. Maximum likelihood from
37 Aiken LS, West SG, Reno RR. Multiple Regression: Testing and
incomplete data via the EM algorithm. J R Stat Soc Series B Stat Methodol 1977;39:1–38 35 Ware JH, Harrington D, Hunter DJ, D’Agostino R. Missing data. N Engl J Med 2012;367:1353–1354 36 Little RJ, D’Agostino R, Cohen ML, et al. The prevention and treatment of missing data in clinical trials. N Engl J Med 2012; 367(14):1355–1360
Interpreting Interactions. Thousand Oaks, CA: SAGE Publications; 1991 38 Efird JT, Lea S, Toland A, Phillips CJ. Informational odds ratio: a useful measure of epidemiologic association in environment exposure studies. Environ Health Insights 2012;6: 17–25 39 Breiman L. Random Forests. Mach Learn 2001;45:5–32
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
316
Thoracic and Cardiovascular Surgeon
Vol. 62
No. 4/2014
Copyright of Thoracic & Cardiovascular Surgeon is the property of Georg Thieme Verlag Stuttgart and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.
Race and Survival among Diabetic Patients after Coronary Artery Bypass Grafting Wesley T. O’Neal1 Jimmy T. Efird2,3 Stephen W. Davies4 Jason B. O’Neal5 W. Randolph Chitwood2 T. Bruce Ferguson2 Alan P. Kypson2 1 Department of Internal Medicine, Wake Forest University School of
Medicine, Winston-Salem, North Carolina 2 Department of Cardiovascular Sciences, East Carolina Heart Institute, Brody School of Medicine, East Carolina University, Greenville, North Carolina 3 Center for Health Disparities, Brody School of Medicine, East Carolina University, Greenville, North Carolina 4 Department of General Surgery, University of Virginia School of Medicine, Charlottesville, Virginia 5 Department of Anesthesia, Critical Care, and Pain Medicine, Beth Israel Deaconess Medical Center, Harvard Medical School, Boston, Massachusetts
Curtis A. Anderson2
Address for correspondence Wesley T. O’Neal, MD, Department of Internal Medicine, Wake Forest University School of Medicine, Medical Center Boulevard, Winston-Salem, NC (e-mail: [email protected]).
Thorac Cardiovasc Surg 2014;62:308–316.
Abstract
Keywords
► ► ► ►
diabetes CABG outcomes epidemiology
Background Diabetes is a known predictor of decreased long-term survival after coronary artery bypass grafting (CABG). Differences in survival by race have not been examined. Methods A retrospective cohort study was conducted for CABG patients between 1992 and 2011. Long-term survival was compared in patients with and without diabetes and stratified by race. Hazard ratios (HR) and 95% confidence intervals (CI) were computed using a Cox regression model. Results Out of the 13,053 patients undergoing CABG, 35% (black n ¼ 1,655; white n ¼ 2,884) had diabetes at the time of surgery. The median follow-up for study participants was 8.2 years. Long-term survival after CABG was similar between black and white diabetic patients (no diabetes, HR ¼ 1.0; white diabetic patients, adjusted HR ¼ 1.5, 95%CI ¼ 1.4– 1.6; black diabetic patients, adjusted HR ¼ 1.5, 95%CI ¼ 1.4–1.7). Conclusion A survival disadvantage after CABG was not observed among black versus white diabetic patients in our study.
Introduction The prevalence of diabetes in the United States has increased by 82% between 1995 and 2010.1 Diabetes is a known risk factor for coronary artery disease (CAD) development, and diabetic patients are more likely to have multivessel disease.2,3 Also, an increased prevalence of diabetes and subsequent risk of mortality has been observed among blacks compared with white patients.4,5
received March 15, 2013 accepted after revision August 22, 2013 published online October 25, 2013
Diabetes is an important predictor of decreased long-term survival after coronary artery bypass grafting (CABG).6–11 In addition, the influence of insulin dependency at the time of surgery has been shown to negatively influence survival compared with patients managed by diet or oral hypoglycemic agents.7–9 Limited information is available regarding how race may affect survival of diabetic patients undergoing CABG. The current study was designed to determine and compare
© 2014 Georg Thieme Verlag KG Stuttgart · New York
DOI http://dx.doi.org/ 10.1055/s-0033-1357297. ISSN 0171-6425.
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
308
Diabetes and Survival
Patients and Methods Study Design This was a retrospective cohort study of patients undergoing isolated (e.g., no concomitant valve procedures) CABG for the first time at the East Carolina Heart Institute during 1992 and 2011. Patients receiving single or multiple bypass grafts were included in this study. Demographic data, comorbid conditions, CAD severity, and surgical data were collected at the time of surgery. Patients with diabetes were compared with those without diabetes. Only black and white patients were included to minimize the potential for residual confounding (1% other races). Racial identity was self-reported. Emergency cases were considered a clinically different population with a different etiology following surgery and were excluded in our analysis (n ¼ 420). Patients who died during the inhospital postoperative period also were excluded (n ¼ 301). The study was approved by the Institutional Review Board at the Brody School of Medicine, East Carolina University.
Operative Procedure In most cases, the left internal mammary artery was used for left anterior descending revascularization. Cardiopulmonary bypass or off-pump coronary artery bypass was selected depending upon patient presentation and surgeon preference. Cold-blood cardioplegia was used to achieve cardiac standstill. Distal anastomoses were performed first followed by proximal anastomoses. If off-pump coronary artery bypass was performed, left internal mammary artery to left anterior descending artery anastomosis was performed first, followed by the remaining distal anastomoses. Proximal anastomoses of the saphenous vein conduits were sewn directly to the ascending aorta.
Definitions Diabetes was defined as having a history of a diagnosis and/or treatment by a physician that was documented in the medical record. Diabetes management was recorded and indicated by medical therapies that were present at the time of admission. Insulin-dependent patients required insulin and insulinindependent patients were managed with diet or oral hypoglycemic agents. Mortality was defined as any cause of death postoperatively. CAD was defined as at least 50% stenosis and was confirmed by angiography before surgery.
Setting The East Carolina Heart Institute is a 120-bed cardiovascular hospital located in the center of eastern North Carolina, a rural region with a large black population. The institute is the largest stand-alone, tertiary referral hospital focusing on cardiovascular care in the state of North Carolina. Cardiovascular disease is the number one cause of death in North Carolina with an increased prevalence occurring in eastern North Carolina.12 Nearly all patients treated at the East
309
Carolina Heart Institute live and remain within a 240-km radius of the medical center.
Data Collection and Follow-Up The primary sources of data extraction were the Society of Thoracic Surgery (STS) Adult Cardiac Surgery Database and the electronic medical record at the Brody School of Medicine. Cardiovascular surgery information at our facility has been reported to the STS since 1989. Data quality and cross-field validation are routinely performed by the Epidemiology and Outcomes Research Unit at the East Carolina Heart Institute. An electronic medical record was introduced at the Brody School of Medicine in 1997. Patient information from 1989 to 1997 was retrospectively scanned into the electronic medical record, with complete conversion beginning in 1992. Local and regional clinics were consolidated under a single electronic medical record in 2005, which allowed efficient patient follow-up. The electronic medical record system applies multiple logical comparisons to reliably reduce mismatching of patient data across clinics and follow-up visits. The STS database is linked to the electronic medical record through a unique patient medical record number. The National Death Index was used to obtain death dates for patients lost to follow-up and also used to validate death information captured in our electronic medical record. Linkage with the National Death Index was based on a multiple criteria, deterministic matching algorithm.13 In our database, less than 5% of validated deaths failed to correctly match with the National Death Index.
Statistical Analysis Categorical variables were reported as frequency and percentage, while continuous variables were reported as mean standard deviation, median, and range. Variables not previously categorized were divided into quartiles before the statistical analysis. Follow-up time was measured from the date of surgery to the date of death or censoring. Survival probabilities were computed using the Kaplan-Meier product-limit method and stratified by diabetes and race. The log-rank test was used to compare survival between patients with and without diabetes and among diabetic patients by race. Cox proportional hazard regression models were used to compute hazard ratios (HR) and 95% confidence intervals (CI) for long-term mortality. The initial multivariable models included variables that have been previously reported to be associated with cardiovascularrelated mortality, regardless of their statistical significance in our dataset.14–16 These included age, sex, race, hypertension, CAD severity, heart failure, and previous stroke. The post hoc addition of other variables into the model was performed in a pairwise fashion. The test statistic of Grambsch and Therneau was used to check the proportional hazards assumption.17 Statistical significance for categorical variables was tested using the Chi-square (χ2) method. Differences in central tendency between or among groups were tested using the DeuchlerWilcoxon and Kruskal-Wallis procedures. Ptrend was computed using a likelihood ratio test. Temporality during the study period was assessed by stratifying the analysis by three time periods. Thoracic and Cardiovascular Surgeon
Vol. 62
No. 4/2014
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
survival between black and white diabetic patients following CABG.
O’Neal et al.
Diabetes and Survival
O’Neal et al.
Few values were missing (< 1% for included variables). However, when values were missing they were entered into the regression models as a separate category. A sensitivity analysis with missing values excluded also was performed to confirm that model β-coefficients did not substantively differ from the above results. Statistical significance was defined as p < 0.05. SAS Version 9.3 (SAS Institute Inc., Cary, NC) was used for all analyses.
different between black and white patients with diabetes (no diabetes, HR ¼ 1.0; white diabetic patients, adjusted HR ¼ 1.5, 95%CI ¼ 1.4–1.6; black diabetic patients, adjusted HR ¼ 1.5, 95%CI ¼ 1.4–1.7) (►Table 5). The multivariable results did not substantively change with the pairwise addition of other variables listed in ►Table 1.
Results
Consistent with previous reports, diabetes was a statistically significant predictor of decreased long-term survival after CABG in our study.6–11 However, a long-term survival disadvantage after CABG was not observed among black compared with white diabetic patients. To the best of our knowledge, we are the first to report this unique finding among diabetic patients undergoing CABG. Recent studies that have examined the influence of diabetes and survival after CABG have been limited to homogenous populations with little patient diversity.6–11 In the United States, black patients comprise approximately 4% of the STS database and this limits many centers from investigating racial differences regarding survival.18 The lack of a survival difference by race among diabetic patients in our study is contrary to several population-based reports that have shown a survival disadvantage for black patients with diabetes.4,5 A recent analysis of Medicare beneficiaries found that black diabetic patients had a higher death rate than white diabetic patients.4 Also, a national cohort study of adults aged 25 to 74 from the first National Health and Nutrition Examination Survey found that black patients with diabetes had a higher mortality rate over a 22year period than white patients with diabetes.5 However, our results suggest that survival related to diabetes may differ from survival with diabetes after CABG. Several studies also have reported black race to be an independent predictor of decreased long-term survival after CABG compared with white patients.19–23 However, survival paradoxes in black patients are well known in the literature.24–27 For example, hypertension and obesity are associated with increased survival among black hemodialysis patients.28 Black patients with diabetes in our study were more likely to be obese and hypertensive at the time of surgery than white patients. Conceivably, diabetes confers a survival advantage among blacks after CABG similar to that observed among hemodialysis patients. While this is intriguing, we cannot rule out chance as an explanation of our findings. Insulin dependency has been shown to negatively impact survival among diabetics after CABG.7–9 In the current study, no difference in survival between black and white patients with diabetes was observed when stratified by insulin therapy. In addition, males have been observed to have a higher incidence rate of diabetes than females and differences in survival may exist after CABG.29 While sex differences at presentation were noted in our sample, survival differences between black and white patients were not observed when the analysis was stratified by sex (female: no diabetes HR ¼ 1.0, black adjusted HR ¼ 1.4, 95%CI ¼ 1.2–1.6, white
Out of the 13,053 patients undergoing CABG, 35% (black n ¼ 1,655; white n ¼ 2,884) had diabetes at the time of surgery. The prevalence of diabetes was higher in black compared with white patients (47 vs. 32%; p < 0.01). Patient characteristics, preoperative medications, and operative characteristics are shown in ►Tables 1–3, respectively. The median follow-up for study participants was 8.2 years. Patients with diabetes had a higher body mass index (mean ¼ 31 5.9 vs. 28 5.1) and were more likely to be black (24 vs. 14%) and female (36 vs. 25%) than patients without diabetes (p < 0.01). On admission, diabetic patients were more likely to have three-vessel CAD (71 vs. 64%), hypertension (81 vs. 67%), peripheral arterial disease (14 vs. 10%), heart failure (21 vs. 11%), chronic obstructive pulmonary disease (9 vs. 7%), previous myocardial infarction (41 vs. 39%), and previous stroke (10 vs. 6%) (p < 0.01). Diabetic patients also were more likely to require preoperative dialysis (3 vs. 1%) but less likely to smoke (20 vs. 27%) and have left main disease (19 vs. 21%) (p < 0.01). Compared with white diabetics, black diabetic patients were more likely to be younger, female, obese, hypertensive, and to have heart failure and hemodialysis (p < 0.01; ►Table 1). In addition, they were more likely to receive preoperative diuretics, angiotensin converting enzyme inhibitors and angiotensin receptor blockers, and β-blockers (p < 0.01; ►Table 2). Intraoperative differences included shorter cardiopulmonary bypass and aortic cross-clamp times and greater likelihood for off-pump coronary artery bypass among black diabetic patients (p < 0.01; ►Table 3). The median survival for patients with and without diabetes was 11 and 16 years, respectively (p < 0.0001; ►Fig. 1). Among black and white patients with diabetes the median survival was 11 years (p ¼ 0.69). The unadjusted HR for diabetes was 1.6 (95%CI ¼ 1.5–1.7) compared with nondiabetic patients. Statistically significant survival differences also were observed for age, body mass index, CAD severity, peripheral arterial disease, heart failure, hypertension, hemodialysis, chronic obstructive pulmonary disease, and previous stroke (►Table 1). Although more black diabetic patients were female, a survival difference was not observed for gender. Similarly, a survival difference was not observed between black and white patients with diabetes when stratified by insulin dependency (►Table 4). Controlling for age, sex, race, hypertension, CAD severity, heart failure, and previous stroke, the HR for long-term mortality for diabetic patients after CABG decreased to 1.5 (95%CI ¼ 1.4–1.6). Long-term survival after CABG was not Thoracic and Cardiovascular Surgeon
Vol. 62
No. 4/2014
Discussion
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
310
Diabetes and Survival
O’Neal et al.
311
Table 1 Patient characteristics and univariable survival (N ¼ 13,053) Characteristic
No diabetes n (%)
Diabetes
Univariable HR (95%CI)
Black n (%)
White n (%)
p-Value
8,514 (65)
1,075 (8)
3,464 (27)
1.6 (1.5–1.7)b
Q1 ( 56)
2,326 (27)
316 (29)
785 (23)
< 0.0001
1.0 Referent
Q2 (> 56–64)
2,090 (25)
305 (28)
900 (26)
1.6 (1.5–1.8)
Q3 (> 64–71)
2,050 (24)
252 (23)
1,004 (29)
2.6 (2.4–2.8)
Q4 (> 71)c
2,048 (24)
202 (19)
775 (22)
4.3 (3.9–4.7)
63 11
62 10
64 10
P trend < 0.0001
64 (24–94)
62 (32–85)
65 (26–89)
6,361 (75)
560 (52)
2,324 (67)
2,153 (25)
515 (48)
1,140 (33)
7,319 (86)
1,195 (14)
Obese ( 30)
2,945 (35)
649 (60)
1,736 (50)
Overweight (25–29.9)
3,636 (43)
313 (29)
1,252 (36)
Overall
a
Mean SD Median (range)
c
Sex Male Female
c
< 0.0001
1.0 Referent 1.1 (1.02–1.2)
Race White Black
1.0 Referent 1.1 (1.06–1.2)
2 d
BMI (kg/m )
< 0.0001
1.0 Referent 1.1 (1.03–1.2)
Normal (18.5–24.9)
1,807 (21)
104 (10)
444 (13)
1.5 (1.4–1.6)
Underweight (< 18.5)c
83 (1)
3 (< 1)
12 (< 1)
2.0 (1.5–2.6)
Mean SD
28 5.1
32 6.2
30 5.8
28 (13–70)
31 (17–64)
30 (14–66)
Elective
3,484 (41)
443 (41)
1,448 (42)
Nonelective
5,030 (59)
632 (59)
2,016 (58)
653 (8)
52 (5)
175 (5)
2,379 (28)
266 (25)
823 (24)
1.6 (1.4–1.9)
5,482 (64)
757 (70)
2,466 (71)
2.0 (1.7–2.3) P trend < 0.0001
No
6,687 (79)
861 (80)
2,823 (82)
Yese
1,827 (21)
214 (20)
641 (18)
No
6,236 (73)
860 (80)
2,758 (80)
Yesc
2,278 (27)
215 (20)
706 (20)
2,771 (33)
105 (10)
740 (21)
5,743 (67)
970 (90)
2,724 (79)
7,642 (90)
902 (84)
3,008 (87)
872 (10)
173 (16)
456 (13)
Median (range)
c
P trend < 0.0001
Status 0.73
1.0 Referent 1.2 (1.1–1.3)
CAD severity 1 Vessel 2 Vessel 3 Vessel
c
0.79
1.0 Referent
Left main disease 0.30
1.0 Referent 1.2 (1.1–1.3)
Recent smoker 0.79
1.0 Referent 0.90 (0.84–0.96)
Hypertension No Yes
c
< 0.0001
1.0 Referent 1.3 (1.2–1.35)
Peripheral arterial disease No Yes
c
0.015
1.0 Referent 2.0 (1.8–2.2) (Continued)
Thoracic and Cardiovascular Surgeon
Vol. 62
No. 4/2014
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
Age (y)
Diabetes and Survival
O’Neal et al.
Table 1 (Continued) Characteristic
No diabetes n (%)
Diabetes Black n (%)
White n (%)
p-Value
7,609 (89)
790 (73)
2,799 (81)
< 0.0001
905 (11)
285 (27)
665 (19)
No
8,433 (99)
1,002 (93)
3,412 (99)
Yesc
81 (1)
73 (7)
52 (1)
No
7,951 (93)
990 (92)
3,159 (91)
Yesc
563 (7)
85 (8)
305 (9)
No
5,208 (61)
609 (57)
2,085 (60)
Yese
3,306 (39)
466 (43)
1,379 (40)
7,961 (94)
954 (89)
3,143 (91)
553 (6)
121 (11)
321 (9)
No
6,926 (81)
854 (79)
2,792 (81)
Yes
1,588 (19)
221 (21)
672 (19)
a
Univariable HR (95%CI)
Heart failure No Yes
c
1.0 Referent 2.1 (2.0–2.3)
Dialysis < 0.0001
1.0 Referent 5.2 (4.4–6.2)
COPD 0.36
1.0 Referent 1.8 (1.6–2.2)
Previous MI 0.039
1.0 Referent 1.2 (1.1–1.3)
Previous stroke No Yes
c
0.055
1.0 Referent 2.0 (1.8–2.2)
Previous PCI 0.40
1.0 Referent 0.85 (0.79–0.92)
Abbreviations: BMI, body mass index; CAD, coronary artery disease; CI, confidence interval; COPD, chronic obstructive pulmonary disease; HR, hazard ratio; MI, myocardial infarction; PCI, percutaneous coronary intervention; Q1, quartile 1; Q2, quartile 2; Q3, quartile 3; Q4, quartile 4; SD, standard deviation; y, years. a Black versus white diabetics (Chi-square). b Diabetes versus no diabetes. c p < 0.01, Chi-square across all columns (χ2) (categorical variables), Kruskal-Wallis test across all columns (continuous variables). d Missing category not shown. e p < 0.05, Chi-square across all columns (χ2) (categorical variables), Kruskal-Wallis test across all columns (continuous variables).
Table 2 Preoperative medications (N ¼ 13,053) Medication
No diabetes n (%)
Aspirin Lipid lowering agents
b
Diabetes Black n (%)
White n (%)
p-Valuea
6,057 (71)
741 (69)
2,407 (69)
0.73
3,230 (38)
514 (48)
1,585 (46)
0.24
Anticoagulants
2,800 (33)
328 (31)
1,113 (32)
0.32
Antiplatelet agentsb
4,585 (54)
479 (45)
1,717 (50)
0.0041
β-Blockersb
4,799 (57)
668 (62)
1,925 (56)
0.0001
Calcium channel blockers
2,608 (31)
358 (33)
1,056 (30)
0.081
Diureticsb
1,420 (17)
383 (36)
965 (28)
< 0.0001
ACE inhibitors/ARBsb
2,176 (26)
512 (48)
1,336 (39)
< 0.0001
Digitalis
b
438 (5)
73 (7)
298 (9)
0.058
Nitrates
1,337 (16)
158 (15)
531 (15)
0.61
Inotropic agentsc
62 (1)
16 (1)
32 (1)
0.11
Abbreviations: ACE, angiotensin converting enzyme; ARB, angiotensin receptor blocker. a Black versus white diabetics (Chi-square). b p < 0.01, Chi-square across all columns (χ2) (categorical variables), Kruskal-Wallis test across all columns (continuous variables). c p < 0.05, Chi-square across all columns (χ2) (categorical variables), Kruskal-Wallis test across all columns (continuous variables). Thoracic and Cardiovascular Surgeon
Vol. 62
No. 4/2014
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
312
Diabetes and Survival
O’Neal et al.
313
Table 3 Operative characteristics (N ¼ 13,053) Characteristic
No diabetes n (%)
Diabetes Black n (%)
White n (%)
p-Valuea
Left internal mammary arteryb,c
5,792 (68)
848 (79)
2,527 (73)
0.035
Intra-aortic balloon pumpb
543 (6)
46 (4)
188 (5)
0.20
827 (10)
112 (10)
265 (8)
0.0041 < 0.0001
Off-pump CABG
b
Cardiopulmonary bypass time (min)
b
Mean SD
98 34
93 30
101 36
Median (range)
95 (5–416)
90 (20–283)
97 (2–941)
Mean SD
62 23
60 21
64 23
Median (range)
60 (5–260)
58 (14–195)
61 (11–486)
< 0.0001
Abbreviations: CABG, coronary artery bypass grafting; min, minutes; SD, standard deviation. a Black versus white diabetics (Chi-square for categorical and Deuchler-Wilcoxon for continuous variables). b p < 0.01, Chi-square across all columns (χ2) (categorical variables), Kruskal-Wallis test across all columns (continuous variables). c Data available from 1995 to 2011.
adjusted HR ¼ 1.4, 1.3–1.6; male: no diabetes HR ¼ 1.0, black adjusted HR ¼ 1.6, 95%CI ¼ 1.4–1.9, white adjusted HR ¼ 1.5, 95%CI ¼ 1.3–1.6). Diabetes in the general population is more prevalent among black than white patients and we observed a similar finding in our analysis.4,5 Consistent with other studies, diabetics were more likely to have multivessel CAD.3 Furthermore, black patients with diabetes in our CABG population were more likely to have left main disease than white diabetic patients. These findings suggest that racial differences exist in the progression of CAD among diabetics.30
Strengths and Limitations Our study is strengthened by its large sample size and longterm follow-up. Furthermore, we were able to accurately determine time of death using a combination of the National Death Index and our comprehensive electronic medical record. An additional strength of this study is its target base. A large priority population in eastern North Carolina allowed
for us to report on a group that has experienced historical differences in socioeconomic position and discrimination. A total of 28 (97%) of the 29 counties in eastern North Carolina fall below the national per capita income of $27,915, with half reporting a value less than $20,000.31 Similarly, 90% of the counties have a higher percentage of blacks than the national value of 13.1%.31 Our results are generalizable to other lowincome, rural, and racially dichotomous populations. Diabetes diagnosis was confirmed through patient history and information recorded in our medical record. Hemoglobin A1C values were not recorded in our database before 2009 and survival may have differed by A1C. The diagnostic criteria for diabetes changed over the study period. However, few patients would have been misclassified because we relied on history and medication use recorded in our electronic medical record. Data regarding socioeconomic position, education, and income were not collected and these factors may have influenced survival.32 Payer status, which has been shown in some studies to predict survival independent of race, was
Fig. 1 Unadjusted Kaplan-Meier survival after CABG. CABG, coronary artery bypass grafting.
Thoracic and Cardiovascular Surgeon
Vol. 62
No. 4/2014
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
Aortic cross-clamp time (min)b
Diabetes and Survival
O’Neal et al.
Table 4 Unadjusted survival by insulin dependency (N ¼ 4,492) Diabetes management
Black
White
n (%)
5-, 10-, and 15-y survival
n (%)
5-, 10-, and 15-y survival
Insulin dependenta
490 (46)
79, 50, 31
1,067 (31)
79, 50, 29
Noninsulin dependent
581 (54)
86, 58, 41
2,354 (69)
83, 59, 40
Abbreviation: y, years. Note: Diabetes management was unknown for 47 patients. a Black diabetic patients were more likely to require insulin at the time of CABG than white diabetic patients (p < 0.0001).
Table 5 Survival by race (N ¼ 13,053) Comparison
Univariable HR (95%CI)
Multivariable HR (95%CI)
No diabetes
1.0 Referent
1.0 Referent
White diabetes
1.6 (1.5–1.7)
1.5 (1.4–1.7)
Black diabetes
1.6 (1.5–1.8)
1.5 (1.4–1.6)
Abbreviations: CI, confidence interval; HR, hazard ratio.
not consistently collected and consequently was not used in our analysis.33 Furthermore, we were unable to reliably estimate socioeconomic position using zip codes, because a large percentage of patients in our catchment area live in rural areas with postal box addresses. Patients in this study were recruited over a relatively long period (20 y), over which practice methods and clinical care may have changed considerably. However, results were consistent throughout the study after stratifying by three time periods, indicating the robustness of the data to temporal changes. Also, the status of several variables in our analysis may have changed over time. We did not adjust for these variables in a time-dependent manner due to their potential to be in the causal pathway. Similarly, surgical complications and medication use were not included in our analysis because of their time-dependent status. Cause of death is not recorded in the National Death Index and diabetic status may have been unrelated to their mortality. Although, we adjusted for known clinically relevant variables, we acknowledge that other unmeasured factors could have influenced our results due to the retrospective nature of this study. We considered missing values to be a distinct category and they were entered into the regression models as a separate category. However, we cannot rule out misclassification bias due to grouping missing values into a distinct category although such bias likely is trivial given the small number of missing values. Furthermore, we performed a complete case analysis with missing values removed and also a separate analysis imputing missing values using the expectationmaximization algorithm to account for missing values.34–36 The results of our sensitivity analysis suggest that missing values did not significantly affect out study findings. While quartile categorization is advantageous because it limits the influence of outliers and allows for the assessment of trend across categories, their use may have yielded overly broad categories and the potential for residual confounding. Thoracic and Cardiovascular Surgeon
Vol. 62
No. 4/2014
However, the substitution of continuous variables in our models yielded similar results. Except for race, we did not examine interactions among other clinically relevant variables in our dataset. Given the large number of potential multilevel interactions in our data, it is difficult to interpret such effects. Furthermore, we did not use regression-based tests for interaction because they are known to have weak power and often fail to detect interactions when they exist.37 Multivariable Cox regression models, rather than propensity score matching, were used to control for confounding because of potential “noncollapsibility bias” inherent to logistic regression-based propensity scores and the possible loss of power due to incomplete matching.38 Alternative methods such as machine learning (e.g., random forest algorithm) may introduce misspecification into the propensity score model due to the “black box” nature of the algorithm that obscures the etiological relationship between predictors and outcome and were not used in the current analysis.39 While in some cases our Cox proportional hazard model diverged from the proportional hazards assumption, there was no effect modification by time and results remained clinically interpretable in terms of the average relative hazard over the observation period.
Conclusion While black diabetic patients in the general population historically have worse survival than white patients, we did not observe a long-term survival difference between black and white diabetics following CABG in our rural, tertiary referral medical center. In the United States, race often is an important surrogate marker for other risk factors, such as socioeconomic position, access, and utilization of care that may not be adequately captured in clinical cardiovascular databases. However, these factors likely are important in cardiovascular disease, independent of geographical boundaries. Our unexpected finding provides important etiological
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
314
clues regarding heart disease progression following CABG among diabetic patients. Further research is needed to determine if the timely identification of CAD in black diabetic patients and appropriate access to CABG surgery will reduce racial differences in survival observed among diabetic patients.
16
17
Acknowledgment The authors would like to thank the East Carolina Heart Institute for providing resources to conduct this study.
References
18
19
1 Centers for Disease Control and Prevention (CDC). Increasing
2
3
4
5
6
7
8
9
10
11
12 13
14
15
prevalence of diagnosed diabetes—United States and Puerto Rico, 1995-2010. MMWR Morb Mortal Wkly Rep 2012;61(45): 918–921 Koskinen P, Mänttäri M, Manninen V, Huttunen JK, Heinonen OP, Frick MH. Coronary heart disease incidence in NIDDM patients in the Helsinki Heart Study. Diabetes Care 1992;15(7):820–825 Melidonis A, Dimopoulos V, Lempidakis E, et al. Angiographic study of coronary artery disease in diabetic patients in comparison with nondiabetic patients. Angiology 1999;50(12):997–1006 McBean AM, Li S, Gilbertson DT, Collins AJ. Differences in diabetes prevalence, incidence, and mortality among the elderly of four racial/ethnic groups: whites, blacks, hispanics, and asians. Diabetes Care 2004;27(10):2317–2324 Gu K, Cowie CC, Harris MI. Mortality in adults with and without diabetes in a national cohort of the U.S. population, 1971-1993. Diabetes Care 1998;21(7):1138–1145 Leavitt BJ, Sheppard L, Maloney C, et al; Northern New England Cardiovascular Disease Study Group. Effect of diabetes and associated conditions on long-term survival after coronary artery bypass graft surgery. Circulation 2004;110(11, Suppl 1):II41–II44 Luciani N, Nasso G, Gaudino M, et al. Coronary artery bypass grafting in type II diabetic patients: a comparison between insulin-dependent and non-insulin-dependent patients at short- and mid-term follow-up. Ann Thorac Surg 2003;76(4):1149–1154 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(4):1045–1052 Szabó Z, Håkanson 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(3):712–719 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 without a history of diabetes mellitus. Coron Artery Dis 2000;11(4):339–346 van Straten AH, Soliman Hamad MA, van Zundert AA, et al. Diabetes and survival after coronary artery bypass grafting: comparison with an age- and sex-matched population. Eur J Cardiothorac Surg 2010;37(5):1068–1074 Morris PJ. Heart disease and stroke in North Carolina. N C Med J 2012;73(6):448–449 Morales DL, McClellan AJ, Jacobs JP. Empowering a database with national long-term data about mortality: the use of national death registries. Cardiol Young 2008;18(Suppl 2):188–195 Efird JT, O’Neal WT, Anderson CA, et al. The effect of race and chronic obstructive pulmonary disease on long-term survival after coronary artery bypass grafting. Front Public Health Serv Syst Res 2013:1 O’Neal WT, Efird JT, Anderson CA, et al. The impact of prior percutaneous coronary intervention on long-term survival after
20
21
22
23
24
25
26
27
28
29
30
31
32
33
O’Neal et al.
coronary artery bypass grafting. Heart Lung Circ 2013 doi:pii: S1443-9506(13)00320-X. 10.1016/j.hlc.2013.04.110. [Epub ahead of print] PMID: 23683716 Efird JT, O’Neal WT, O’Neal JB, Ferguson TB, Chitwood WR, Kypson AP. Effect of peripheral arterial disease and race on survival after coronary artery bypass grafting. Ann Thorac Surg 2013;96(1): 112–118 Grambsch P, Therneau T. Proportional hazards tests and diagnostics based on weighted residuals. Biometrika 1994;81(3): 515–526 Shahian DM, O’Brien SM, Sheng S, et al. Predictors of long-term survival after coronary artery bypass grafting surgery: results from the Society of Thoracic Surgeons Adult Cardiac Surgery Database (the ASCERT study). Circulation 2012;125(12): 1491–1500 Gray RJ, Nessim S, Khan SS, Denton T, Matloff JM. Adverse 5-year outcome after coronary artery bypass surgery in blacks. Arch Intern Med 1996;156(7):769–773 Cooper WA, Thourani VH, Guyton RA, et al. Racial disparity persists after on-pump and off-pump coronary artery bypass grafting. Circulation 2009;120(11, Suppl):S59–S64 Brooks MM, Jones RH, Bach RG, et al. For the BARI Investigators. Predictors of mortality and mortality from cardiac causes in the bypass angioplasty revascularization investigation (BARI) randomized trial and registry. Circulation 2000;101(23): 2682–2689 Maynard C, Fisher LD, Passamani ER. Survival of black persons compared with white persons in the Coronary Artery Surgery Study (CASS). Am J Cardiol 1987;60(7):513–518 Taylor HA Jr, Mickel MC, Chaitman BR, Sopko G, Cutter GR, Rogers WJ. Long-term survival of African Americans in the Coronary Artery Surgery Study (CASS). J Am Coll Cardiol 1997;29(2): 358–364 Bloembergen WE, Port FK, Mauger EA, Wolfe RA. Causes of death in dialysis patients: racial and gender differences. J Am Soc Nephrol 1994;5(5):1231–1242 Kucirka LM, Grams ME, Lessler J, et al. Association of race and age with survival among patients undergoing dialysis. JAMA 2011; 306(6):620–626 Owen WF Jr, Chertow GM, Lazarus JM, Lowrie EG. Dose of hemodialysis and survival: differences by race and sex. JAMA 1998;280(20):1764–1768 Agodoa L, Eggers P. Racial and ethnic disparities in end-stage kidney failure-survival paradoxes in African-Americans. Semin Dial 2007;20(6):577–585 Kalantar-Zadeh K, Kovesdy CP, Derose SF, Horwich TB, Fonarow GC. Racial and survival paradoxes in chronic kidney disease. Nat Clin Pract Nephrol 2007;3(9):493–506 Centers for Disease Control. Age-adjusted incidence of diagnosed diabetes per 1,000 population aged 18–79 years, by sex, United States, 1980–2011. http://www.cdc.gov/diabetes/statistics/incidence/fig4.htm. Accessed May 23, 2013 McClelland RL, Chung H, Detrano R, Post W, Kronmal RA. Distribution of coronary artery calcium by race, gender, and age: results from the Multi-Ethnic Study of Atherosclerosis (MESA). Circulation 2006;113(1):30–37 United States Census Bureau. State & County Quick Facts2010. http://quickfacts.census.gov/qfd/index.html. Accessed June 1, 2013 Koch CG, Li L, Kaplan GA, et al. Socioeconomic position, not race, is linked to death after cardiac surgery. Circ Cardiovasc Qual Outcomes 2010;3(3):267–276 Zacharias A, Schwann TA, Riordan CJ, Durham SJ, Shah A, Habib RH. Operative and late coronary artery bypass grafting outcomes in matched African-American versus Caucasian patients: evidence of a late survival-Medicaid association. J Am Coll Cardiol 2005;46(8): 1526–1535 Thoracic and Cardiovascular Surgeon
Vol. 62
No. 4/2014
315
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
Diabetes and Survival
Diabetes and Survival
O’Neal et al.
34 Dempster AP, Laird NM, Rubin DB. Maximum likelihood from
37 Aiken LS, West SG, Reno RR. Multiple Regression: Testing and
incomplete data via the EM algorithm. J R Stat Soc Series B Stat Methodol 1977;39:1–38 35 Ware JH, Harrington D, Hunter DJ, D’Agostino R. Missing data. N Engl J Med 2012;367:1353–1354 36 Little RJ, D’Agostino R, Cohen ML, et al. The prevention and treatment of missing data in clinical trials. N Engl J Med 2012; 367(14):1355–1360
Interpreting Interactions. Thousand Oaks, CA: SAGE Publications; 1991 38 Efird JT, Lea S, Toland A, Phillips CJ. Informational odds ratio: a useful measure of epidemiologic association in environment exposure studies. Environ Health Insights 2012;6: 17–25 39 Breiman L. Random Forests. Mach Learn 2001;45:5–32
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
316
Thoracic and Cardiovascular Surgeon
Vol. 62
No. 4/2014
Copyright of Thoracic & Cardiovascular Surgeon is the property of Georg Thieme Verlag Stuttgart and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.
