Kogan et al.

Cardiovasc Diabetol (2018) 17:151

https://doi.org/10.1186/s12933-018-0796-7 Cardiovascular Diabetology

ORIGINAL INVESTIGATION Open Access

Impact of type 2 diabetes mellitus

on short‑ and long‑term mortality
after coronary artery bypass surgery
Alexander Kogan1,2,3*, Eilon Ram1,3, Shany Levin1, Enrique Z. Fisman3, Alexander Tenenbaum3, Ehud Raanani1,3
and Leonid Sternik1,3

Abstract 
Background:  Type 2 diabetes mellitus (DM) is a frequent co-morbidity among patients undergoing coronary artery
bypass grafting (CABG) surgery. The aim of this study was to evaluate the impact of DM on the early- and long-term
outcomes of patients who underwent isolated CABG.
Methods:  We performed an observational cohort study in a large tertiary medical center over a period of 11 years.
All data from patients who had undergone isolated CABG surgery between 2004 and 2014 were obtained from our
departmental database. The study population included 2766 patients who were divided into two groups: Group
I (1553 non-diabetic patients), and Group II (1213 patients suffering from type 2 DM). Group II patients were then
divided into two subgroups: subgroup IIA (981 patients treated with oral antihyperglycemic medications) and sub-
group IIB (232 insulin-treated patients with or without additional oral antihyperglycemic drugs). In-hospital, 1-, 3-, 5-
and 10-year mortality outcome variables were evaluated. Mean follow-up was 97 ± 41 months.
Results:  In-hospital mortality was similar between Group I and Group II patients (1.87% vs. 2.31%, p = 0.422) and
between the subgroups IIA and IIB (2.14% vs. 3.02%, p = 0.464). Long-term mortality (1, 3, 5 and 10 years) was higher
in Group II (DM type 2) compared with Group I (non-diabetic patients) (5.3% vs. 3.6%, p = 0.038; 9.3% vs. 5.6%,
p < 0.001; 15.3% vs. 9.3%, p < 0.001 and 47.3% vs. 29.6% p < 0.001). Kaplan–Meier analysis demonstrated that all-cause
mortality was higher in Group II compared with Group I (p < 0.001) and in subgroup IIB compared with subgroup IIA
(p = 0.001). Multivariable analysis showed that DM increased the mortality hazard by twofold, and among diabetic
patients, insulin treatment increased the mortality hazard by twofold.
Conclusions:  Diabetic and non-diabetic patients have similar in-hospital mortality rates. Survival rates of diabetic
patients start to deteriorate 3 year after surgery. Type 2 DM is an independent predictor for long-term mortality after
isolated CABG surgery. Mortality is even higher when the diabetes treatment strategy included insulin.
Keywords:  Diabetes mellitus, Coronary artery bypass grafting, Revascularization, Insulin

*Correspondence: alexanderkogan140@hotmail.com;
Alexander.Kogan@sheba.health.gov.il
1
Department of Cardiac Surgery, Sheba Medical Center, Tel Hashomer,
Affiliated to the Sackler School of Medicine, Tel Aviv University, 52621 Tel
Aviv, Israel
Full list of author information is available at the end of the article

© The Author(s) 2018. This article is distributed under the terms of the Creative Commons Attribution 4.0 International License
(http://creat​iveco​mmons​.org/licen​ses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium,
provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license,
and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creat​iveco​mmons​.org/
publi​cdoma​in/zero/1.0/) applies to the data made available in this article, unless otherwise stated.
Kogan et al. Cardiovasc Diabetol (2018) 17:151 Page 2 of 8

Introduction they were discharged either to their home or to a reha-

Diabetes mellitus (DM) doubles the risk of cardiovascular bilitation facility. In the operating room and in the ICU,
disease [1] and about 75% of deaths in diabetic patients patients from both groups received intravenous continu-
are due to coronary artery disease [2]. Studies performed ous infusion of regular insulin according to the Society
during the 1980’s and 90’s demonstrated increasing of Thoracic Surgeons practice guideline series [8]. After
short- [3] and long-term mortality in diabetic patients discharge from the ICU, patients from Group I (non-
undergoing CABG compared with non-diabetic patients diabetic) did not receive insulin or any other hypoglyce-
[4, 5]. However, more recently, reports have shown a sig- mic medication, while patients from Group II continued
nificant reduction in mortality among patients with dia- to receive preoperative antiglycemic treatment (per oral
betes [6]. The aim of the current study was to investigate drugs and insulin), as soon as they began to eat. We eval-
the impact of DM type 2 on short- and long-term mortal- uated the following outcome variables: in-hospital, 1-,
ity. The study was conducted on a large cohort of patients 3-, 5- and 10-year mortality. Mortality was ascertained
undergoing CABG. from the Israeli Ministry of Interior Population Register.
Mean follow-up was 98 ± 41  months and completed for
Methods all patients.
The study was approved by the Sheba Medical Center Throughout the study no major changes in hospital pol-
Institutional Ethical Committee (Protocol No 4257). The icy, surgical or anesthesiological techniques were intro-
requirement for informed consent was waived because duced. All patients received standardized anesthesia.
of the retrospective nature of the study. We performed
a retrospective, observational study that included pro- Database management and statistical analysis
spectively-collected data from consecutive patients who At discharge, all patient data were checked, corrected
had undergone a first isolated CABG at a large tertiary and entered into a database. Descriptive statistics were
university hospital over an 11-year period, between used to summarize data, and numerical data were
01.04.2004 and 31.03.2015. Patients were divided into expressed as means (SD). Normality of the distribu-
two groups: Group I (non-diabetic patients), and Group tion of continuous data variables was analyzed using the
II (patients suffering from type 2 DM). DM type 2 was Kolmogorov–Smirnov test. Since not all numerical data
defined according to the American Diabetes Association were distributed normally, Mann–Whitney U-tests were
as: (a) hemoglobin A1C ≥ 6.5%; (b) fasting plasma glucose used to evaluate differences between groups. Differences
levels ≥ 126 mg/day; (c) classic symptoms of hyperglyce- between the frequencies of categorical variables were
mia or a hyperglycemic crisis, a random plasma glucose also estimated using Fisher’s exact test. All outcome vari-
level ≥ 200 mg/dL (11.1 mmol/L) [7]; or (d) currently on ables were compared between whole groups of patients.
pharmacologic treatment (oral antihyperglycemic drugs Variables that had a possible influence on patient
and/or insulin). Group II patients were divided into two mortality (univariate predictors) were evaluated using
subgroups: subgroup IIA, treated only with oral antihy- Fisher’s exact test. Suspected predictors for mortality
perglycemics and subgroup IIB, treated with insulin or a by univariable analysis were analyzed by a multivariable
combination of oral antihyperglycemic drugs and insulin. logistic regression analysis in order to find independ-
Using non-identifiable patient data from our depart- ent predictors of mortality. The variables included in the
ment’s database, we evaluated the following variables: final model among the entire cohort were: DM, Euro-
gender, age, chronic obstructive pulmonary disease, score, NYHA functional class, and hyperlipidemia. The
congestive heart failure [New York Heart Association variables included in the final model among the diabetic
(NYHA) functional class], DM, dialysis-dependent renal patients were: insulin therapy, Euroscore, NYHA func-
failure, peripheral vascular disease, a previous cardiovas- tional class, and hyperlipidemia. p values of 0.05 and less
cular accident (CVA)/transient ischemic attack (TIA), were considered as statistically significant. All statistical
systemic and pulmonary hypertension, left ventricular analyses were performed using SPSS 11.5 for Windows.
function, priority of surgery (elective, urgent, or emer-
gent), and logistic EuroSCORE I. Perioperative variables Results
included using left and right internal mammary arteries, Baseline characteristics
the number of anastomosis, time of cross-clamping and During the 11-year study period, 2972 patients under-
duration of cardiopulmonary bypass. went isolated CABG, of whom 36 patients were excluded
After surgery all patients were admitted to the intensive from the study due to incomplete data, and a further 170
care unit (ICU) directly from the operating room. After patients were excluded since they suffered from DM type
discharge from the ICU, patients were transferred either 1. The study population included 2766 patients (Table 1).
to a step-down unit or directly to the floor, from where Group I (non-diabetic patients), comprised 1553
Kogan et al. Cardiovasc Diabetol (2018) 17:151 Page 3 of 8

Table 1  Patients’ data (Groups I and II)

Non-DM Group I DM Group II p value
No. of patients (1553) No. of patients (1213)

Age (years) 63 ± 11 65 ± 10 0.000

Males 1308 (84%) 947 (78%) 0.000
Elective 978 (63%) 788 (65%) 0.378
NYHA FC III–IV 373 (20%) 376 (27%) 0.001
Previous operation 32 (2%) 37 (3%) 0.110
Ejection fraction (%) 51 ± 11 50 ± 11 0.001
Logistic EuroSCORE 5.6 ± 8.2 6.4 ± 9.6 0.020
Hypertension 1065 (69%) 1052 (87%) 0.000
COPD 70 (4%) 61 (5%) 0.529
Dialysis 17 (1%) 24 (2%) 0.059
Hyperlipidemia 1097 (71%) 1019 (84%) 0.000
PVD 111 (7%) 185 (15%) 0.000
CVA/TIA 100 (6%) 130 (11%) 0.000
Pulmonary hypertension 3 (0.2%) 16 (1%) 0.001
Arrhythmia 8 (0.5%) 6 (0.5%) 1.000
LIMA 1509 (97%) 1171 (97%) 0.378
No mammary 26 (2%) 28 (2%) 1.000
Double mammary 781 (50%) 468 (39%) 0.000
Number of grafts 3.1 ± 1.1 3.1 ± 1.0 0.411
Bypass 80 ± 27 82 ± 42 0.133
Clamp 56 ± 20 56 ± 29 0.701
Mini invasive 33 (2%) 21 (2%) 0.491
DM Diabetes mellitus, NYHA FC New York Heart Association functional class, COPD chronic obstruction pulmonary disease, PVD peripheral vascular disease, CVA
cerebral vascular accident, TIA transient ischemic attack, LIMA left internal mammary artery

patients, and Group II (patients suffering from DM type significantly higher rate of prolonged mechanical ventila-
2), comprised 1213 patients. Of them, 981 patients were tion time (over 72  h) (5.6% vs. 3.1%, p = 0.001 and 9.7%
treated with oral antihyperglycemic medications (Group vs. 4.7%, p = 0.006), > 7  days in the ICU (5.7% vs. 2.7%,
IIA) and 232 patients were treated with insulin (Group p < 0.001 and 9.7% vs. 4.8%, p = 0.007) and a significantly
IIB) (Table  2). Compared with the non-diabetic group, longer hospital duration (7.6 ± 7.9 vs. 6.7 ± 5.2, p < 0.001
the diabetic group of patients were older, more frequently and 9 ± 8 vs. 7 ± 8, p = 0.042).
women, had a higher mean logistic EuroSCORE, and had
a higher prevalence of systemic and pulmonary hyper- Follow‑up mortality
tension, peripheral vascular disease and previous CVA. Long-term mortality (1, 3, 5 and 10  years) was higher
in Group II (DM type 2) compared with Group I (non-
Early outcomes diabetic patients) (5.3% vs. 3.6%, p = 0.038; 9.3% vs. 5.6%,
Overall in-hospital mortality was similar between p < 0.001; 15.3% vs. 9.3%, p < 0.001 and 47.3% vs. 29.6%,
Groups I and II: 1.87% vs. 2.31%, p = 0.422, and between p < 0.001) (Fig. 1; Additional file 1: Table S1). These results
subgroups IIA and IIB: 2.14% vs. 3.01%, p = 0.464. Other were also consistent among the subgroup of patients with
in-hospital major events were similar between Groups NYHA functional class I–II and III–IV (Additional file 2:
I and II, such as CVA (0.3% vs. 0.5%, p = 0.350), use of Figure S1 and Additional file 3: Figure S2).
intra-aortic balloon pump (5% vs. 5%, p = 0.859) and Furthermore, long-term mortality was higher in sub-
renal replacement therapy (0.7% vs. 1.5%, p = 0.059). group IIB (insulin-treated patients) compared to sub-
Major events were also similar between the non-insulin group IIA (non-insulin treated patients) with 1-, 3-,
dependent and insulin-dependent DM patients: CVA 5- and 10-year mortality rates of 6.5% vs. 5%, p = 0.413;
(0.5% vs. 0.4%, p = 1.000), use of intra-aortic balloon 14.2% vs. 8.2%, p = 0.008; 22.1% vs. 13.6%, p = 0.002
pump (5% vs. 6%, p = 0.239) and renal replacement ther- and 59.3% vs. 44.3%, p = 0.002 (Fig.  2; Additional file  1:
apy (1% vs. 2%, p = 0.364). Group II compared to Group Table  S2). These results were also consistent among the
I, and Group IIB compared to Group IIA experienced a subgroup of patients with NYHA functional class I–II
Kogan et al. Cardiovasc Diabetol (2018) 17:151 Page 4 of 8

Table 2  Patients’ data (subgroups IIA and IIB)

Non-insulin Insulin p value
No. of patients = 981 No. of patients = 232

Age (years) 65 ± 10 64 ± 10 0.023

Males 775 (79%) 172 (74%) 0.113
Elective 627 (65%) 149 (65%) 1.000
NYHA FC III–IV 284 (26%) 88 (31%) 0.003
Previous operation 28 (3%) 9 (4%) 0.399
Ejection fraction (%) 51 ± 11 47 ± 12 0.000
Logistic EuroSCORE 6 ± 9 8 ± 11 0.012
Hypertension 847 (86%) 205 (88%) 0.453
COPD 51 (5%) 10 (4%) 0.738
Dialysis 12 (1%) 12 (5%) 0.001
Hyperlipidemia 818 (83%) 201 (87%) 0.273
PVD 125 (13%) 60 (26%) 0.000
CVA/TIA 106 (11%) 24 (10%) 0.906
Pulmonary hypertension 10 (1%) 6 (3%) 0.099
Arrhythmia 2 (0.2%) 4 (2%) 0.014
LIMA 949 (97%) 222 (96%) 0.426
Double mammary 396 (40%) 72 (31%) 0.032
Number of grafts 3 ± 1 3 ± 1 0.231
Bypass 83 ± 42 81 ± 40 0.471
Clamp 57 ± 31 53 ± 17 0.014
Mini invasive 18 (2%) 3 (1%) 0.781
NYHA FC New York Heart Association functional class, COPD chronic obstruction pulmonary disease, PVD peripheral vascular disease, CVA cerebral vascular accident,
TIA transient ischemic attack, LIMA left internal mammary artery

and III–IV (Additional file  4: Figure S3 and Additional patients was higher than that of non-diabetic patients.
file 5: Figure S4). Furthermore, mortality was even higher when the dia-
Multivariable analysis demonstrated that predictors betic treatment strategy included insulin compared to
for 10  years’ mortality were higher on EuroSCORE (HR treatment without insulin. Macrovascular disease (coro-
1.38 95% CI 1.31–1.46, p < 0.001), hyperlipidemia (HR nary artery disease, stroke, and peripheral vascular dis-
1.39 95% CI 1.02–1.92, p = 0.037) and DM (HR 2.08 95% ease) was responsible for the majority of morbidity and
CI 1.57–2.74, p < 0.001). Among the diabetic patients, mortality associated with type 2 DM [9]. It should be
predictors for 10  years’ mortality were higher on Euro- pointed out that incident diabetes in adults is associated
SCORE (HR 1.37 95% CI 1.3–1.45, p < 0.001), hyper- with a substantial risk for mortality, especially in younger
lipidemia (HR 1.45 95% CI 1.06–1.97, p = 0.021) and adults [10], and in the presence of coronary artery dis-
insulin treatment (HR 2.11 95% CI 1.34–3.3, p = 0.001). ease (CAD) early revascularization is overwhelmingly
Although higher NYHA functional class was a risk fac- important, since a deferred revascularization procedure
tor for late mortality by univariable analysis (p = 0.001), in patients with and without DM shows that the former
it was not associated with late mortality after adjustment are associated with a significantly higher target lesion
for confounders by multivariable analysis, both in the failure rate [11].
entire cohort (p = 0.834) and among the diabetic patients The impact of diabetes on short-term mortality after
(p = 0.831). CABG was insignificant. Abizaid et al. [12] reported sim-
ilar in-hospital and 1-year mortality rates for diabetic and
Discussion non-diabetic patients: 2.1% vs. 1.2% and 3.1% vs. 2.8%. In
Our study investigated the impact of DM type 2 on in- addition, Marui et al. [13] found no differences in 30-day
hospital and long-term mortality in patients after their and 1-year mortality between diabetic and non-diabetic
first isolated CABG. First, we found that among patients patients: 0.9% vs. 1.2% and 4.4% vs. 4.5%. Carson et  al.
undergoing CABG, diabetic and non-diabetic patients [3] reported a 30-day mortality rate of 3.7% in patients
had similar in-hospital mortality. Second, our princi- with DM and 2.7% in those without DM. While Li et al.
pal finding was that the long-term mortality of diabetic [14] reported similar post-CABG mortality rates for
Kogan et al. Cardiovasc Diabetol (2018) 17:151 Page 5 of 8

non-diabetic and diabetic patients treated with oral anti- is associated with increased long-term mortality after
hyperglycemic medication (1.88% vs. 2.01%), they showed CABG, while CABG was associated with significantly
a significant increase in mortality in a subgroup of dia- lower long-term adverse clinical outcomes compared to
betic patients treated with insulin (3.08%). Zalewska- percutaneous coronary intervention (PCI) in patients
Adamiec et  al. [15] found no differences in 30-day, with insulin-treated type 2 DM [24].
1- and 2-year mortality rates (11.6% vs. 11.1%) in dia- While in the general population DM is associated with
betic and non-diabetic patients with left main coronary excess mortality, compared with the general population
artery disease. Whang et al. [16] found no differences in without DM, with a hazard ratio of 1.15 at 5-years [25],
2-year mortality rates between diabetic and non-diabetic we reported a greater impact of DM on patients who
patients: 26% vs. 24%, and concluded that diabetes was underwent CABG (HR of 1.68 at 5  years). Although we
not a predictor of mortality after CABG among patients have no data regarding the cause of death, myocardial
with left ventricular dysfunction. Although diabetic infarction, and graft patency during the follow-up period,
patients in our series were older and had more comor- we assume that this higher impact of DM in our cohort,
bidities, differences in early mortality did not reach sta- compared to the natural history of DM in the general
tistical significance. We report here that in-hospital population, is due to accelerated CAD, based on previous
mortality among diabetic (N = 1553) and non- diabetic studies that support this theory [26].
patients (N = 1213), was 2.3% and 1.9%. With these rates In our report the 5-year mortality rate was 15.3%
of mortality in a group the size of ours, we achieved a among diabetic patients and 9.3% among non-diabetic
power of 13% to detect differences between the groups. patients. Although we did not compare our results to
Even if we would have reached the statistical signifi- a percutaneous approach, based on previous reports
cance, the clinical implications of these differences would among diabetic and non-diabetic patients with multi-
have been questionable, since it is obvious that the most vessel CAD who were treated by PCI, our results were
important impact of DM is on long-term results. In the somewhat better. Among diabetic patients, Farkouh et al.
non-insulin compared with the insulin-treated subgroup [27] reported 5-year mortality rates of 26.6% in the PCI
of patients, mortality was 2.14% vs. 3.01%, p = 0.464, and group. Contini et al. [28] reported a 5-year mortality rate
at 1  year, mortality was 5% vs. 6.5%, p = 0.413. In this of 24.5% in diabetic patients who underwent PCI. Kap-
context, it should be stressed that among hospitalized petein et al. [20], throughout a 5-year follow-up, showed
heart failure patients with no pre-existing DM there is a that PCI revealed a mortality rate of 19.5% in diabetic
linear relationship between admission blood glucose and patients, and 12.0% in non-diabetic patients.
long-term mortality, whereas among patients with DM Only 18.5% of the patients in the current study were
only an admission level of > 200 mg/dL is associated with female, as reported in previous studies where the patients
increased mortality risk [17]. were also predominantly male [13, 18, 19]. The explana-
However, the impact of diabetes on long-term mortal- tion for this observation could be that cardiovascular dis-
ity after CABG has been controversial. Marui et al. [13] ease develops in an older age in females than in males,
reported an increase in 3- and 5-years’ mortality in dia- and older age increases the surgical risk. The differences
betic compared to non-diabetic patients, 11% vs. 9.7% in clinical presentation in women lead to less aggressive
and 19.6% vs. 16.2%. Mohamadi et  al. [18] investigated treatment strategies with less referral for surgical revas-
only cardiac-specific survival rates, but not all-cause cularization [29].
mortality, and postulated that DM type 2 diabetes is not We reported that 1213 out of the 2766 patients (44%)
an independent predictor of late cardiac death 6  years who underwent isolated CABG in our Institute had dia-
after CABG. Furthermore, Onuma et  al. [19] reported betes, more than the average in previous reports [18–20].
slightly increasing mortality rates in diabetic compared Blumenfeld et al. [30] reported that in Israel, the thresh-
with non-diabetic patients 5  years after CABG: 8.6% vs. old of patient referral for surgical revascularization has
7.1%. Kappetein et al. [20] also investigated 5 years’ mor- risen in the last two decades, and that those patients have
tality after CABG and reported non-significant differ- more comorbidities than in the past. Since the advan-
ences between diabetic and non-diabetic patients: 12.9% tage of CABG over PCI is seen mostly in patients with
vs. 10.9%. However, Koshizaka et al. [21] reported signifi- complex CAD, such as left-main stenosis or high SYN-
cant differences in 5 years’ mortality: 15.5% vs. 8.5%. Wit TAX score, and in diabetic patients, the characteristics
et  al. [22] reported significantly higher 3-year mortality of patients who are referred to CABG has changed dur-
rates in patients with insulin-treated compared with non- ing the last decade toward patients with a more com-
insulin-treated DM (16.7% vs. 8.7%) and non-diabetic plex anatomy and more diabetes. Recent reports based
patients (6.3%). In their systematic reviews and meta- on National Registries have shown that 43–49% of
analyses Bundhun et  al. [23] postulated that diabetes the patients who were referred to CABG in Israel had
Kogan et al. Cardiovasc Diabetol (2018) 17:151 Page 6 of 8

1.0
0.8
0.6
Survival
0.4

Diabetes

No−Diabetes
0.2

p<0.001
0.0

0 12 24 36 48 60 72 84 96 108 120 132 144 156

Time in months

# at risk
Diabetes 1213 1152 1129 1101 1075 969 829 728 602 493 365 262 157 60
No−Diabetes 1553 1499 1480 1467 1438 1326 1181 1062 906 776 603 444 267 111
Fig. 1  Survival rate by DM groups. DM Diabetes mellitus

diabetes [31, 32]. The current study is based on the pop- department. Third, we had no information regarding
ulation of a tertial referral center in Israel, and thus our the main cause of death, the rate of cardiac events and
report is consistent with real-life situations in Israel. data regarding graft patency during the follow-up period.
Diabetes mellitus is a chief cause of heart failure, either Analysis of cardiac events could reinforce the conclusion
secondary to CAD or secondary to diabetic cardiomyopa- that DM provides less favorable results after CABG. The
thy [33]. In general, insulin-dependent diabetic patients lack of information regarding the main cause of death
have more comorbidities than non-insulin dependent weakens the conclusions of this study.
diabetic patients. Although the presence of insulin treat-
ment is indeed a marker for more advanced disease, its Conclusions
underlying biological mechanism has not been fully elu- DM type-2 is an independent predictor of long-term
cidated. It may be related to the impact of a procoagu- mortality after CABG surgery. Mortality rates increased
lant imbalance, chronic exposure to high glucose levels, significantly when the diabetic treatment strategy
and direct effects of hyperinsulinemia. Interestingly, included insulin. The high-risk population of insulin-
endogenous hyperinsulinemia has been associated with dependent DM may require specific and/or more intense
increased long-term mortality following myocardial cardiovascular protective therapies after CABG. Further
infarction in patients without diabetes [34]. Further stud- studies are needed to examine whether novel interven-
ies are needed to examine whether insulin-dependent tions, such as GLP-1 analogues or SGLT2 inhibitors, can
diabetic patients should be included in risk stratification improve their long-term outcomes. Since the current
algorithms for patients who undergo CABG. study is underpowered to detect reduced early survival
in diabetic patients a larger study is warranted in order
Limitations to reach significant differences in early mortality, which
There are a few limitations in our study. First, while it is would enable us to conclude that DM may impact early
retrospective in design, data were collected prospectively results after CABG.
and recorded in a well-defined database. Second, our
study was conducted in a single-center cardiac surgery
Kogan et al. Cardiovasc Diabetol (2018) 17:151 Page 7 of 8

1.0
0.8
0.6
Survival
0.4

Insulin

No−Insulin
0.2

p<0.001
0.0

0 12 24 36 48 60 72 84 96 108 120 132 144 156

Time in months

# at risk

Insulin 232 218 209 199 194 173 139 119 90 72 57 40 23 9

No−Insulin 981 934 920 902 881 796 690 609 512 421 308 222 134 51
Fig. 2  Survival rate in the DM group by insulin treatment. DM Diabetes mellitus

Additional files editing. ER: study revision. LS: wrote the research project and study revision. All
authors read and approved the final manuscript.

Additional file 1: Table S1. Life table for mortality in the entire cohort Author details
1
by patients with and without diabetes mellitus. A. Patients with diabetes  Department of Cardiac Surgery, Sheba Medical Center, Tel Hashomer,
mellitus. B. Patients without diabetes mellitus. Table S2. Life table for mor- Affiliated to the Sackler School of Medicine, Tel Aviv University, 52621 Tel
tality among diabetic patients on insulin. A. Patients treated by insulin. B. Aviv, Israel. 2 Cardiac Surgery Intensive Care Unit, Sheba Medical Center, Tel
Patients without insulin treatment. Hashomer, Affiliated to the Sackler School of Medicine, Tel Aviv University, Tel
Aviv, Israel. 3 Tel Aviv University, Tel Aviv, Israel.
Additional file 2: Figure S1. Survival rate by DM groups among patients
with NYHA functional class I–II. DM = Diabetes mellitus; NYHA = New-
Acknowledgements
York Heart Association.
Not applicable.
Additional file 3: Figure S2. Survival rate by DM groups among patients
with NYHA functional class III–IV. DM = Diabetes mellitus; NYHA = New- Competing interests
York Heart Association. The authors declare that they have no competing interests.
Additional file 4: Figure S3. Survival rate in the DM group receiving
insulin treatment among patients with NYHA functional class I–II. DM = Availability of data and materials
Diabetes mellitus; NYHA = New-York Heart Association. Data collected from a departmental database.

Additional file 5: Figure S4. Survival rate in the DM group receiving Consent for publication
insulin treatment among patients with NYHA functional class III–IV. DM = Not applicable.
Diabetes mellitus; NYHA = New-York Heart Association.
Ethics approval and consent to participate
Institutional Ethical Committee (Protocol No 4257).
Abbreviations
CABG: coronary artery by-pass graft; CAD: coronary artery disease; CVA: Financial disclosures
cerebrovascular accident; DM: diabetes mellitus; ICU: intensive care unit; PCI: None to declare.
percutaneous coronary intervention; TIA: transient ischemic attack.
Funding
Authors’ contributions None.
AK: wrote the research project and the full manuscript. ER: study revision
and editing. SL: data collection and interpretation. EZF, AT: study revision and
Kogan et al. Cardiovasc Diabetol (2018) 17:151 Page 8 of 8

Publisher’s Note patients with severe left ventricular dysfunction: results from The CABG
Springer Nature remains neutral with regard to jurisdictional claims in pub- Patch Trial database. J Am Coll Cardiol. 2000;36:1166–72.
lished maps and institutional affiliations. 17. Zadok OIB, Kornowski R, Goldenberg I, Klempfner R, Toledano Y, Biton Y,
et al. Admission blood glucose and 10-year mortality among patients
Received: 4 November 2018 Accepted: 26 November 2018 with or without pre-existing diabetes mellitus hospitalized with heart
failure. Cardiovasc Diabetol. 2017;16:102.
18. Mohammadi S, Dagenais F, Mathieu P, Kingma JG, Doyle D, Lopez S, et al.
Long-term impact of diabetes and its comorbidities in patients undergo-
ing isolated primary coronary artery bypass graft surgery. Circulation.
References 2007;116:I220–5.
1. Sarwar N, Gao P, Seshasai SR, Gobin R, Kaptoge S, Di Angelantonio E, 19. Onuma Y, Wykrzykowska JJ, Garg S, Vranckx P, Serruys PW, ARTS I and II
et al. Diabetes mellitus, fasting blood glucose concentration, and risk of Investigators. 5-Year follow-up of coronary revascularization in diabetic
vascular disease: a collaborative meta-analysis of 102 prospective studies. patients with multivessel coronary artery disease: insights from ARTS
Lancet. 2010;375:2215–22. (arterial revascularization therapy study)-II and ARTS-I trials. J Am Coll
2. O’Gara PT, Kushner FG, Ascheim DD, Casey DE Jr, Chung MK, de Lemos Cardiol Cardiovasc Interv. 2011;4:317–23.
JA, et al. 2013 ACCF/AHA guideline for the management of ST-elevation 20. Kappetein AP, Head SJ, Morice MC, Banning AP, Serruys PW, Mohr FW,
myocardial infarction: a report of the American College of Cardiology et al. Treatment of complex coronary artery disease in patients with
Foundation/American Heart Association Task Force on Practice Guide- diabetes: 5-year results comparing outcomes of bypass surgery and per-
lines. Circulation. 2013;127:e362–425. cutaneous coronary intervention in the SYNTAX trial. Eur J Cardiothorac
3. Carson JL, Scholz PM, Chen AY, Peterson ED, Gold J, Schneider SH. Dia- Surg. 2013;43:1006–13.
betes mellitus increases short-term mortality and morbidity in patients 21. Koshizaka M, Lopes RD, Reyes EM, Gibson CM, Schulte PJ, Hafley GE, et al.
undergoing coronary artery bypass graft surgery. J Am Coll Cardiol. Long-term clinical and angiographic outcomes in patients with diabetes
2002;40:418–23. undergoing coronary artery bypass graft surgery: results from the Project
4. Adler DS, Goldman L, O’Neil A, Cook EF, Mudge GH Jr, Shemin RJ, et al. of Ex-vivo Vein Graft Engineering via Transfection IV trial. Am Heart J.
Long-term survival of more than 2,000 patients after coronary artery 2015;169:175–84.
bypass grafting. Am J Cardiol. 1986;58:195–202. 22. Wit MA, de Mulder M, Jansen EK, Umans VA. Diabetes mellitus and its
5. Barsness GW, Peterson ED, Ohman EM, Nelson CL, DeLong ER, Reves JG, impact on long-term outcomes after coronary artery bypass graft sur-
et al. Relationship between diabetes mellitus and long-term survival after gery. Acta Diabetol. 2013;50:123–8.
coronary bypass and angioplasty. Circulation. 1997;96:2551–6. 23. Bundhun PK, Bhurtu A, Yuan J. Impact of type 2 diabetes mellitus on the
6. Rawshani A, Rawshani A, Franzén S, Eliasson B, Svensson AM, Miftaraj M, long-term mortality in patients who were treated by coronary artery
et al. Mortality and cardiovascular disease in type 1 and type 2 diabetes. bypass surgery: a systematic review and meta-analysis. Medicine (Balti-
N Engl J Med. 2017;376:1407–18. more). 2017;96:e7022.
7. American Diabetes Association. Classification and diagnosis of diabetes. 24. Bundhun PK, Wu ZJ, Chen MH. Coronary artery bypass surgery compared
Diabetes Care. 2015;38(Suppl):S8–16. with percutaneous coronary interventions in patients with insulin-treated
8. Lazar HL, McDonnell M, Chipkin SR, Furnary AP, Engelman RM, Sadhu type 2 diabetes mellitus: a systematic review and meta-analysis of 6
AR, Society of Thoracic Surgeons Blood Glucose Guideline Task Force, randomized controlled trials. Cardiovasc Diabetol. 2016;15:2.
et al. The Society of Thoracic Surgeons practice guideline series: blood 25. Tancredi M, Rosengren A, Svensson AM, Kosiborod M, Pivodic A, Gudb-
glucose management during adult cardiac surgery. Ann Thorac Surg. jörnsdottir S, et al. Excess mortality among persons with type 2 diabetes.
2009;87:663–9. N Engl J Med. 2015;373:1720–32.
9. Monteiro P, Gonçalves L, Providência LA. Diabetes and cardiovascular 26. Lathief S, Inzucchi SE. Approach to diabetes management in patients
disease: the road to cardioprotection. Heart. 2005;91:1621–5. with CVD. Trends Cardiovasc Med. 2016;26:165–79.
10. Zucker I, Shohat T, Dankner R, Chodick G. New onset diabetes in adult- 27. Farkouh ME, Domanski M, Sleeper LA, Siami FS, Dangas G, Mack M, et al.
hood is associated with a substantial risk for mortality at all ages: a Strategies for multivessel revascularization in patients with diabetes. N
population based historical cohort study with a decade-long follow-up. Engl J Med. 2012;367:2375–84.
Cardiovasc Diabetol. 2017;16:105. 28. Contini GA, Nicolini F, Fortuna D, Pacini D, Gabbieri D, Vignali L, et al. Five-
11. Kennedy MW, Kaplan E, Hermanides RS, Fabris E, Hemradj V, Koopmans year outcomes of surgical or percutaneous myocardial revascularization
PC, et al. Clinical outcomes of deferred revascularisation using fractional in diabetic patients. Int J Cardiol. 2013;168:1028–33.
flow reserve in patients with and without diabetes mellitus. Cardiovasc 29. Maas AH, Appelman YE. Gender differences in coronary heart disease.
Diabetol. 2016;15:100. Neth Heart J. 2010;18:598–602.
12. Abizaid A, Costa MA, Centemero M, Abizaid AS, Legrand VM, Limet RV, 30. Blumenfeld O, Na’amnih W, Shapira-Daniels A, Lotan C, Shohat T, Shapira
Arterial Revascularization Therapy Study Group, et al. Clinical and eco- OM. Trends in coronary revascularization and ischemic heart disease-
nomic impact of diabetes mellitus on percutaneous and surgical treat- related mortality in Israel. J Am Heart Assoc. 2017;6:e004734.
ment of multivessel coronary disease patients: insights from the Arterial 31. Fink N, Nikolsky E, Assali A, Shapira O, Kassif Y, Barac YD, et al. Revasculari-
Revascularization Therapy Study (ARTS) trial. Circulation. 2001;104:533–8. zation strategies and survival in patients with multi-vessel coronary artery
13. Marui A, Kimura T, Nishiwaki N, Mitsudo K, Komiya T, Hanyu M, CREDO- disease. Ann Thorac Surg. 2018. https​://doi.org/10.1016/j.athor​acsur​
Kyoto PCI/CABG Registry Cohort-2 Investigators, et al. Five-year outcomes .2018.07.070.
of percutaneous versus surgical coronary revascularization in patients 32. Ram E, Goldenberg I, Kassif Y, Segev A, Lavee J, Shlomo N, et al.
with diabetes mellitus (from the CREDO-Kyoto PCI/CABG Registry Comparison of patients with multivessel disease treated at centers
Cohort-2). Am J Cardiol. 2015;115:1063–72. with and without on-site cardiac surgery. J Thorac Cardiovasc Surg.
14. Li Z, Amsterdam EA, Young JN, Hoegh H, Armstrong EJ. Contemporary 2018;155(865–873):e3.
outcomes of coronary artery bypass grafting among patients with 33. Meagher P, Adam M, Civitarese R, Bugyei-Twum A, Connelly KA. Heart
insulin-treated and non-insulin-treated diabetes. Ann Thorac Surg. failure with preserved ejection fraction in diabetes: mechanisms and
2015;100:2262–9. management. Can J Cardiol. 2018;34:632–43.
15. Zalewska-Adamiec M, Bachorzewska-Gajewska H, Malyszko J, Tomaszuk- 34. Kragelund C, Snorgaard O, Køber L, Bengtsson B, Ottesen M, Højberg S,
Kazberuk A, Nowak K, Hirnle T, et al. Impact of diabetes on mortality and et al. Hyperinsulinaemia is associated with increased long-term mortality
complications after coronary artery by-pass graft operation in patients following acute myocardial infarction in non-diabetic patients. Eur Heart
with left main coronary artery disease. Adv Med Sci. 2014;59:250–5. J. 2004;25:1891–7.
16. Whang W, Bigger JT Jr, The CABG Patch Trial Investigators and Coordina-
tors. Diabetes and outcomes of coronary artery bypass graft surgery in