There are conflicting data on whether patients with insulin-treated diabetes mellitus (ITDM) have poorer outcomes compared with non-insulin treated diabetic (non-ITDM) patients following percutaneous coronary intervention (PCI). We therefore compared clinical outcomes following PCI in ITDM versus non-ITDM patients. We prospectively collected data on 4,579 patients with diabetes underwent PCI between 2005 and 2014 in a large multicenter registry and dichotomized them as having ITDM (n = 1,111) or non-ITDM (n = 3,468). The non-ITDM group was further divided into diet control only (diet-DM; n = 786) and those taking oral hypoglycemic agents (OHG-DM; n = 2,639), and clinical outcomes were compared with ITDM patients. Median follow-up for long-term mortality was 4.2 years (IQR 2.0 to 6.6 years). ITDM patients were more likely to be female, obese, and have severe renal impairment (all p <0.001). Procedural characteristics were similar other than a greater use of drug-eluting stents in ITDM patients. On multivariable analysis, ITDM was an independent predictor of 12-month major adverse cardiovascular and cerebrovascular events (MACCE; OR 1.26, 95% CI 1.02 to1.55, p = 0.03). Dividing the non-ITDM group further by treatment, a progressively higher rate of 12-month MACCE across the 3 groups was observed (13.5% vs 17.9% vs 21.8%; p <0.001). Long-term mortality was similar in the diet-DM and OHG-DM groups, but significantly higher in the ITDM group on Kaplan-Meier analysis (log-rank p <0.001). In conclusion, there is a clear gradient of adverse outcomes with escalation of therapy from diet control to OHGs to insulin.
Diabetes mellitus (DM) is an important risk factor for the development and progression of coronary artery disease. Coronary artery lesions in patients with diabetes are more often diffuse and involve multiple vessels with small luminal diameters. Patients with DM have also been consistently shown to have worse cardiovascular outcomes after percutaneous coronary intervention (PCI) compared with nondiabetics. , , However, the results of studies comparing clinical outcomes in patients with insulin-treated DM (ITDM) with patients with non-insulin treated DM (non-ITDM) have been inconsistent. Patients with ITDM often have had a more prolonged duration of disease, a greater burden of co-morbidities, as well as poorer glycemic control, and therefore may be expected to have worse outcomes. A recent meta-analysis comparing PCI outcomes in ITDM with non-ITDM patients showed that both short- and long-term mortality was higher in the ITDM group. However, data from a large German drug-eluting stent (DES) registry and from the secondary analysis of the Taxus Element versus Xience Prime in a Diabetic Population trial, both showed a similar incidence of major adverse cardiovascular events (MACE) at 12 months following PCI regardless of insulin treatment. , Given the conflicting data in the published literature, we sought to compare clinical outcomes following PCI in patients with diabetes mellitus according to their treatment status in a large, multicenter Australian registry.
Methods
We analyzed data from consecutive patients with diabetes underwent PCI from January 1, 2005 to December 31, 2014 enrolled prospectively in the Melbourne Interventional Group (MIG) registry. Patients were divided into two groups based on whether they were on insulin (ITDM group) or not on insulin (non-ITDM group). Diabetic status and treatment were determined at the time of PCI by the interventional cardiologist through assessment of medical records and medication charts, and recorded on a prespecified registry case report form.
The MIG registry is a multicenter Australian PCI registry and has been previously described in detail. Briefly, it collects data from 6 participating hospitals located in metropolitan Melbourne and regional Victoria, that all have 24-hour cardiac catheterization laboratory services. The registry is coordinated by the Centre of Cardiovascular Research and Education in Therapeutics; an independent research body within the School of Public Health and Preventive Medicine at Monash University (Melbourne, Australia). Demographic, clinical, procedural and in-hospital outcome data are prospectively recorded on case-report forms using standardized definitions for all fields. Thirty-day and 12-month outcomes are obtained through telephone follow-up and medical records were reviewed to verify events. Long-term mortality data were obtained by linkage to the Australian National Death Index (NDI), a database housed at the Australian Institute of Health and Welfare that contains records of all deaths occurring in Australia since 1980. The censoring date for linkage with the NDI in this study was 30 July, 2014. Successful matching of patients through this linkage process was achieved in 99.4% of all patients in the MIG registry. The primary ethics approval has been granted by the ethics committee at The Alfred Hospital (approval number 92/04), and also approved by each participating hospital, including the use of “opt-out” consent as previously described. ,
Baseline and procedural characteristics, as well as in-hospital, 30-day and 12-month clinical outcomes were compared between the groups. The primary end point was a composite end point of major adverse cardiovascular and cerebrovascular events (including all-cause mortality, myocardial infarction (MI), target vessel revascularization (TVR) and stroke; MACCE) at 12 months. Secondary end points included 30-day and 12-month mortality, MI, stroke and TVR, as well as long-term NDI-linked mortality. MI was defined using the Third Universal Definition of Myocardial Infarction. Post-PCI major bleeding was defined as any bleeding requiring a transfusion and/or prolonging the hospital stay and/or causing a fall in hemoglobin > 3.0 g/dl. We also performed a sensitivity analysis of outcomes by diabetes treatment status in patients according to the generation of drug-eluting stent (DES) received (first vs second generation DES). First generation DES were defined as Taxus (Boston Scientific Corporation, Natick, Massachusetts) and Cypher (Cordis Corporation, Miami Lakes, Florida) stents while all other DES were defined as second generation DES. Patients who received >1 DES from >1 generation were excluded from this analysis. Further subgroup analyses were also performed comparing 12-month outcomes in patients treated with diet-control only (diet-DM), oral hypoglycemic agents only (OHG-DM) and insulin (ITDM).
Continuous variables are expressed as mean ± standard deviation and were compared using Kruskal-Wallis equality-of-populations rank test. Categorical data are expressed as numbers and percentages and were compared using Pearson’s Chi-square test or Fisher’s exact test as appropriate. The Kaplan-Meier method was used to evaluate 12-month MACCE-free survival rates and long-term NDI-linked mortality rates, while the log-rank test was used for survival comparisons. Multiple logistic regression analysis was used to identify independent predictors of 12-month MACCE. In this model, in addition to diabetes treatment status, 26 other clinically relevant variables were considered (Supplementary Table 1). Those with a p value of <0.1 on univariate analysis that were not co-linear were entered into a stepwise backward selection modelling process for multivariable assessment. Complete case analysis was performed for purposes of multivariable modelling (i.e., patients with missing values were excluded). The proportion of missing data were <1% for all variables. However, out of the non-ITDM group, 43 patients (1.3%) did not have their specific treatment for diabetes (i.e., whether on oral hypoglycemic agents or diet-control alone) recorded and therefore were excluded from analysis. All statistical analyses were performed using Stata 13.1 software (StataCorp LP, College Station, Texas). p values of <0.05 were considered to be statistically significant.
Results
In total, 4,579 patients were included in this study, of which 1,111 patients (24.3%) were in the ITDM group, and 3,468 (75.7%) were in the non-ITDM group.
Table 1 shows the baseline characteristics of the 2 groups. ITDM patients tended to be slightly younger and were more likely to be female and obese (defined by body mass index ≥30 kg/m 2 ) than non-ITDM patients (all p <0.001). They were also more likely to have stage 4 to 5 chronic kidney disease (defined as estimated glomerular filtration rate <30 ml/min/1.73m 2 ), peripheral vascular disease, and a history of MI or coronary artery bypass graft surgery (CABG) (all p <0.001).
| Variable | Non-ITDM (n = 3,468) | ITDM (n = 1,111) | p value |
|---|---|---|---|
| Mean age ± SD (years) | 67.1 ± 11.2 | 65.2 ± 11.3 | < 0.001 |
| Male | 2,531 (73.0%) | 741 (66.7%) | < 0.001 |
| Mean body mass index (kg/m 2 ) ± SD | 29.7 ± 5.5 | 30.8 ± 6.2 | < 0.001 |
| Body mass index ≥ 30 kg/m 2 | 1,305 (42.0%) | 522 (51.0%) | < 0.001 |
| Hypertension | 2,875 (83.0%) | 947 (85.3%) | 0.06 |
| Hypercholesterolemia | 2,833 (81.8%) | 930 (83.8%) | 0.13 |
| Current smoker | 578 (16.9%) | 179 (16.6%) | 0.79 |
| Peripheral vascular disease | 307 (8.9%) | 185 (16.7%) | < 0.001 |
| Previous myocardial infarction | 1,161 (33.5%) | 485 (43.7%) | < 0.001 |
| Previous coronary artery bypass graft surgery | 431 (12.4%) | 199 (17.9%) | < 0.001 |
| Previous stroke | 261 (7.5%) | 134 (12.1%) | <0.001 |
| Family history of coronary artery disease | 1,234 (37.3%) | 374 (35.9%) | 0.42 |
| eGFR (ml/min/1.73m 2 ) | |||
| > 60 | 2,404 (71.1%) | 595 (54.7%) | < 0.001 |
| 30 – 60 | 852 (25.2%) | 344 (31.6%) | |
| < 30 | 124 (3.7%) | 149 (13.7%) |
Presentation and procedural characteristics of the 2 groups are shown in Table 2 . ITDM patients were more likely to have PCI to the left main coronary artery (2.3% vs 1.2%; p = 0.004), receive a drug-eluting stent (64.5% vs 60.9%; p=0.002) and require rotational atherectomy (2.9% vs 1.9%; p = 0.03) than their non-ITDM counterparts. Mean total stent length was also slightly longer in the ITDM group (20.0 ± 9.6 vs 19.2 ± 8.7 mm; p = 0.04). ITDM patients were also more likely to have moderate-severe or severe left ventricular dysfunction (defined as a left ventricular ejection fraction 30% to 45% and <30%; 27.7% vs 21.6% and 4.0% vs 2.7% respectively; both p <0.001)
| Variable | Non-ITDM (n = 3,468) | ITDM (n=1,111) | p value |
|---|---|---|---|
| Clinical presentation | |||
| ST elevation myocardial infarction | 751 (21.7%) | 221 (19.9%) | 0.17 |
| Non-ST elevation myocardial infarction | 1,046 (30.2%) | 334 (30.1%) | |
| Unstable angina pectoris | 375 (10.8%) | 105 (9.5%) | |
| Stable angina pectoris | 1,293 (37.3%) | 450 (40.5%) | |
| Cardiogenic shock at presentation | 102 (2.9%) | 41 (3.7%) | 0.21 |
| Post-out-of-hospital cardiac arrest at presentation | 59 (1.7%) | 18 (1.6%) | 0.86 |
| Left ventricular ejection fraction | |||
| >45% | 2,252 (75.7%) | 640 (68.4%) | < 0.001 |
| 30-45% | 643 (21.6%) | 259 (27.7%) | |
| <30% | 81 (2.7%) | 37 (4.0%) | |
| Radial access | 658 (19.1%) | 222 (20.0%) | 0.49 |
| Femoral access | 2,793 (80.9%) | 887 (80.0%) | |
| Single vessel coronary disease | 1,131 (32.7%) | 298 (26.9%) | <0.001 |
| Multi-vessel coronary disease | 2,328 (67.3%) | 809 (73.1%) | |
| Coronary Vessel treated | |||
| Left main | 51 (1.2%) | 31 (2.3%) | 0.004 |
| Left anterior descending | 1,427 (34.0%) | 426 (31.8%) | 0.15 |
| Left circumflex | 533 (12.7%) | 193 (14.4%) | 0.10 |
| Right | 1,303 (31.0%) | 417 (31.2%) | 0.93 |
| Bypass graft | 191 (4.6%) | 65 (4.9%) | 0.64 |
| AHA/ACC B2/C lesion | 2,318 (55.2%) | 759 (56.7%) | 0.33 |
| PCI to chronic total occlusion | 185 (4.4%) | 62 (4.6%) | 0.73 |
| PCI to bifurcation lesion | 483 (11.5%) | 141 (10.5%) | 0.33 |
| PCI to in-stent restenosis | 253 (6.0%) | 98 (7.3%) | 0.09 |
| Glycoprotein IIb/IIIa inhibitor used | 842 (24.3%) | 226 (20.4%) | 0.007 |
| Rotational atherectomy used | 80 (1.9%) | 39 (2.9%) | 0.03 |
| Drug eluting stent implanted | 2,113 (60.9%) | 717 (64.5%) | 0.002 |
| Bare-metal stent implanted | 1,114 (32.1%) | 299 (26.9%) | |
| Balloon angioplasty only | 241 (7.0%) | 95 (8.6%) | |
| Mean total stent length (mm) ± SD | 19.2 ± 8.7 | 20.0 ± 9.6 | 0.04 |
| Pre-PCI TIMI flow 0-1 | 941 (22.5%) | 255 (19.3%) | < 0.001 |
| Post PCI TIMI flow 3 | 4,019 (95.8%) | 1,266 (94.8%) | 0.21 |
| PCI complications: | |||
| Acute closure | 17 (0.4%) | 6 (0.5%) | 0.83 |
| Transient no-reflow | 89 (2.3%) | 26 (2.1%) | 0.59 |
| Persistent no-reflow | 27 (0.7%) | 12 (1.0%) | |
| Unsuccessful PCI | 194 (5.6%) | 73 (6.6%) | 0.23 |
A comparison of clinical outcomes between the 2 groups is shown in Table 3 . In-hospital and 30-day MACCE rates were both significantly higher in the ITDM group compared with the non-ITDM group (6.8% vs 4.7%; p = 0.005 and 9.2% vs 7.2%; p = 0.04 respectively). There were no differences in rates of death, MI or stroke during the index admission or at 30-day follow-up, between the two groups (all p >0.05). However, by 12-month follow-up, mortality (8.6% vs 6.4%; p = 0.01), MI (8.2% vs 5.9%; p = 0.006) and MACCE rates (21.8% vs 16.8%; p <0.001) were all significantly higher in the ITDM group compared with the non-ITDM group. Similarly, long-term NDI-linked mortality was higher in the ITDM group (27.7% vs 18.7%; p<0.001). These differences in 12-month and long-term outcomes between the ITDM and non-ITDM groups persisted regardless of the type of initial clinical presentation (stable angina vs acute coronary syndrome) (all p <0.05) (Supplementary Table 2).
