Published reports describe a strong association between plasma glucose levels on admission and mortality in patients who undergo primary percutaneous coronary intervention for ST-segment elevation myocardial infarction. The aim of this study was to assess the predictive value of admission glucose levels for early and late mortality. From 2005 to 2007, 1,646 patients underwent primary percutaneous coronary intervention for ST-segment elevation myocardial infarction and were stratified according to admission plasma glucose level in category 1 (<7.8 mmol/L; n = 747), category 2 (7.8 to 11.0 mmol/L; n = 620), or category 3 (>11 mmol/L; n = 279). Event rates were estimated using the Kaplan-Meier method. A landmark survival analysis to 3-year follow-up was performed, with a landmark set at 30 days. Time-extended Cox regression was used to assess the predictive value of admission glucose levels. Furthermore, a stratified analysis was performed for known diabetes mellitus status at admission. Thirty-day mortality was 2.4% in category 1, 6% in category 2, and 22% in category 3 (p <0.01). Three-year mortality in 30-day survivors was 5.9% in category 1, 8.2% in category 2, and 7.1% in category 3 (p = 0.27). Glucose level on admission was a strong predictor of 30-day mortality: for every 1 mmol/L increase, the hazard increased by 14% (hazard ratio 1.14, 95% confidence interval 1.09 to 1.19, p <0.01) in patients without diabetes, by 12% (hazard ratio 1.12, 95% confidence interval 1.05 to 1.19, p <0.01) in those with diabetes, and by 13% (hazard ratio 1.13, 95% confidence interval 1.09 to 1.17, p <0.01) in the total cohort. After 30 days, glucose level at admission lost its predictive value. In conclusion, in patients with and those without diabetes, glucose level at admission is an independent predictor of early but not late mortality.
Published reports describe an association between hyperglycemia and reduced long-term survival. However, careful analysis of these reports actually suggests that most patients die within the first month. To improve further insight into the role of admission glucose level as a marker for adverse outcomes, it is important to explore the differences in early and late mortality in patients with ST-segment elevation myocardial infarction (STEMI) with different plasma glucose levels at admission. To the best of our knowledge, this is the first study using a landmark survival analysis in a true STEMI cohort treated with primary percutaneous coronary intervention (PCI) to differentiate the association of admission glucose levels with early and late mortality. The landmark was set at 30 days, with a total follow-up period of 3 years, which enabled us to explore mortality in 2 time periods: from 0 to 30 days and from 30 days to 3 years in 30-day survivors. Furthermore, we performed a stratified analysis for known diabetes mellitus (DM) status on admission, because biologically, a significant correlation with glucose levels can be expected.
Methods
From March 2005 to December 2007, a total of 1,965 consecutive and unselected patients were admitted to our hospital with STEMI. Data from these patients were checked for consistency and completeness. For patients who underwent multiple primary PCIs during the study period (n = 49), only the first primary PCI was counted as a qualifying event for this analysis. Patients treated with rescue PCI for failed intravenous thrombolysis (n = 16) were excluded. During this 3-year period, glucose levels at admission were not available for 254 patients, so the present study cohort consisted of 1,646 patients.
Acute STEMI was diagnosed when patients had symptoms of acute myocardial infarction lasting 30 minutes to 12 hours, accompanied by electrocardiographic ST-segment elevation >1 mm (0.1 mV) in ≥2 contiguous leads. Patients were immediately transported to the cardiac catheterization laboratory and underwent immediate coronary angiography with a view to performing primary PCI. PCI was performed using standard techniques if the coronary anatomy was suitable. All procedural decisions, including device selection and adjunctive pharmacotherapy, such as glycoprotein IIb/IIIa inhibitor use, were made at the discretion of the operator. All patients were treated with heparin (5,000 IU) and aspirin (500 mg) before PCI. If a coronary stent was implanted, ticlopidine or clopidogrel was prescribed according to guidelines.
Patients were stratified according to their histories of DM and their plasma glucose levels at admission into category 1 (<7.8 mmol/L; n = 712 without DM, n = 35 with DM), category 2 (7.8 to 11.0 mmol/L; n = 550 without DM, n = 70 with DM), or category 3 (>11 mmol/L; n = 175 without DM, n = 104 with DM). These cutoff levels were based on diagnostic criteria for impaired glucose tolerance and DM suggested by the American Diabetes Association. Category 1 consisted of patients with no evidence of impaired glucose metabolism, category 2 consisted of patients with possible impaired glucose metabolism, and category 3 consisted of patients with profound impaired glucose metabolism.
All patients who underwent PCI at our institution were prospectively followed. Baseline clinical, angiographic, and procedural information was entered by qualified cardiac catheterization laboratory personnel and interventional cardiologists in a dedicated electronic database. Information on mortality was synchronized with the computerized records from the national population registry (Statistics Netherlands, Voorburg, The Netherlands) and was verified through August 15, 2010. Blood samples before primary PCI were obtained as part of routine clinical care. Blood samples were drawn immediately after the insertion of the arterial sheath before primary PCI for the assessment of plasma glucose. Blood samples were centrifuged without undue delay and analyzed. Glucose was measured with an enzymatic assay on a Hitachi modular P-800 analyzer (Roche Diagnostics GmbH, Mannheim, Germany).
The primary outcome for the present analysis was the predictive value of admission glucose level for all-cause mortality at 30 days and at 3 years in 30-day survivors.
Statistical analysis was performed using SPSS version 16.0 (SPSS, Inc., Chicago, Illinois). Discrete variables are summarized as frequencies and percentages. Differences in baseline characteristics among groups were tested for significance using chi-square tests. Skewed-distributed continuous variables were compared using Kruskal-Wallis tests. Statistical significance was defined as a p value <0.05. Cumulative event rates of all-cause death were estimated using the Kaplan-Meier method and compared using the log-rank statistic. Follow-up for mortality was censored at the date of last follow-up by checking vital status in the Dutch population registry, or at 3 years, whichever came first. A landmark survival analysis was performed, with a landmark set at 30 days to provide insight into the predictive value of admission glucose on early and late mortality. Hazard ratios for all-cause mortality were calculated using time-extended Cox proportional-hazard regression analyses after verification of the proportional-hazards assumption. A covariate was allowed in the model when in univariate analyses, its p value was <0.10. The following variables were included in the model: glucose level at admission, age, male gender, body mass index, history of DM, hypertension, smoking status, hypercholesterolemia, previous myocardial infarction, shock, time to treatment (symptom onset to first balloon inflation), left anterior descending coronary artery–related myocardial infarction, and multivessel disease. It is likely that glucose on admission is correlated with known DM status, so we first stratified for this variable before analyzing the entire cohort.
Results
Of the 1,645 patients with STEMI, 209 (13%) had known diagnoses of DM. At admission, the median glucose levels were 7.8 mmol/L (interquartile range 6.7 to 9.4) for patients without DM and 11 mmol/L (interquartile range 9 to 14.4) for those with DM. Patients were subsequently stratified according to their DM status and their plasma glucose levels at admission, as detailed previously.
Table 1 lists baseline clinical angiographic and procedural characteristics. In the DM subgroups, there were no significant differences among the 3 glucose categories. In the non-DM subgroup, patients in categories 1 and 2 in comparison to those in category 3 were more often male and had a higher incidence of hypertension, hypercholesterolemia, and postprocedural Thrombolysis In Myocardial Infarction (TIMI) flow grade 3. Moreover, they had a lower incidence of cardiogenic shock at presentation and concomitant lower use of intra-aortic balloon pumps. Patients in categories 2 and 3 compared to those in category 1 were older and more frequently had anterior myocardial infarction, stent placement, and glycoprotein IIb/IIIa inhibitor use.
| Variable | DM (n = 209 [13%]) | No DM (n = 1,437 [87%]) | ||||||
|---|---|---|---|---|---|---|---|---|
| Glucose Category 1 (n = 35) | Glucose Category 2 (n = 70) | Glucose Category 3 (n = 104) | p Value | Glucose Category 1 (n = 712) | Glucose Category 2 (n = 550) | Glucose Category 3 (n = 175) | p Value | |
| Men | 19 (54%) | 49 (70%) | 65 (63%) | 0.27 | 530 (74%) | 393 (72%) | 114 (65%) | 0.01 |
| Age (years) | 69 (58–73) | 71 (61–78) | 66 (57–74) | 0.06 | 59 (49–69) | 63 (55–74) | 62 (52–74) | <0.01 |
| Body mass index (kg/m 2 ) | 28 (25–31) | 29 (26–31) | 29 (24–32) | 0.59 | 26 (24–28) | 26 (24–28) | 26 (24–29) | 0.06 |
| Hypertension | 17 (49%) | 42 (60%) | 45 (43%) | 0.10 | 210 (30%) | 166 (30%) | 32 (18%) | <0.01 |
| Smokers | 12 (34%) | 16 (23%) | 31 (30%) | 0.42 | 336 (47%) | 220 (40%) | 46 (26%) | <0.01 |
| Hypercholesterolemia | 13 (37%) | 24 (34%) | 23 (22%) | 0.11 | 145 (20%) | 104 (19%) | 16 (99.1%) | <0.01 |
| Family history of cardiovascular disease | 12 (34%) | 23 (33%) | 34 (33%) | 0.99 | 322 (45%) | 193 (35%) | 37 (21%) | <0.01 |
| Previous myocardial infarction | 5 (14%) | 22 (31%) | 22 (21%) | 0.23 | 89 (13%) | 64 (12%) | 22 (13%) | 0.74 |
| Cardiogenic shock | 0 (0%) | 4 (5.7%) | 10 (9.6%) | 0.13 | 10 (1.4%) | 11 (2.0%) | 32 (18%) | <0.01 |
| Time to treatment (minutes) | 212 (108–268) | 199 (136–275) | 239 (149–368) | 0.34 | 195 (132–301) | 176 (134–249) | 159 (116–239) | <0.01 |
| Angiographic characteristics | ||||||||
| Left anterior descending infarct-related coronary artery | 12 (34%) | 21 (30%) | 47 (45%) | 0.11 | 251 (35%) | 236 (43%) | 79 (45%) | <0.01 |
| Multivessel disease | 16 (46%) | 32 (46%) | 52 (50%) | 0.83 | 215 (30%) | 186 (34%) | 69 (39%) | 0.05 |
| Postprocedural TIMI flow grade 3 | 34 (97%) | 58 (83%) | 93 (89%) | 0.09 | 663 (93%) | 496 (90%) | 147 (84%) | <0.01 |
| Procedural characteristics | ||||||||
| Thrombosuction | 9 (25%) | 25 (36%) | 41 (39%) | 0.34 | 261 (37%) | 245 (45%) | 72 (41%) | 0.02 |
| Use of intra-aortic balloon pump | 2 (5.7%) | 4 (5.7%) | 10 (9.6%) | 0.57 | 17 (2.4%) | 34 (6.2%) | 48 (27%) | <0.01 |
| Stent placement | 26 (74%) | 50 (71%) | 76 (73%) | 0.95 | 564 (79%) | 477 (87%) | 149 (85%) | <0.01 |
| Use of glycoprotein IIb/IIIa inhibitor | 7 (20%) | 21 (30%) | 30 (29%) | 0.53 | 168 (24%) | 160 (29%) | 53 (30%) | 0.04 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
