Prevalence and relevance of abnormal glucose metabolism in acute coronary syndromes: insights from the PLATelet inhibition and patient Outcomes (PLATO) trial

Diabetes mellitus (DM) and abnormal glucose metabolism are associated with cardiovascular (CV) disease. We investigated the prevalence and prognostic importance of dysglycaemia in patients with acute coronary syndromes (ACS) in the PLATelet inhibition and patient Outcomes (PLATO) trial. Diabetes was defined as known diabetes or HbA1c ≥ 6.5% or non-fasting glucose ≥ 11.1 mmol/L on admission, prediabetes as HbA1c ≥ 5.7% but < 6.5%, and no diabetes as HbA1c < 5.7%. The primary endpoint was the composite of CV death, spontaneous myocardial infarction type 1 (sMI) or stroke at 12 months. Multivariable Cox regression models, adjusting for baseline characteristics, and biomarkers NT-proBNP and troponin I, were used to explore the association between glycaemia and outcome. On admission, 16,007 (86.1%) patients had HbA1c and/or glucose levels available and were subdivided into DM 38.5% (6160) (1501 patients had no previous DM diagnosis), prediabetes 38.8% (6210), and no DM 22.7% (3637). Kaplan Meier event rates at 12 months for CV death, sMI or stroke per subgroups were 14.5% (832), 9.0% (522), and 8.5% (293), respectively with multivariable adjusted HRs, versus no diabetes, for diabetes: 1.71 (1.50–1.95) and for prediabetes 1.03 (0.90–1.19). Corresponding event rates for CV death were 6.9% (391), 3.4% (195) and 3.0% (102), respectively, with adjusted HRs for patients with DM of: 1.92 (1.42–2.60) and for prediabetes 1.02 (0.79–1.32). Abnormal glucose metabolism is common in ACS patients, but only patients with definite DM have an increased CV risk, indicating that prediabetes is not immediately associated with worse CV outcomes.


Introduction
Diabetes mellitus (DM) is associated with a two-to fourfold increased risk for cardiovascular (CV) events compared with non-diabetics [1]. In patients without diagnosed DM, but with borderline increased haemoglobin A1c (HbA1c), heightened fasting glucose or abnormal glucose tolerance test, an increased risk of CV events has been observed in a general population [2,3]. The impact of dysglycaemia on CV risk in patients with acute coronary syndrome (ACS) is less clear with partly conflicting results [2][3][4][5][6][7][8][9][10]. In the PLATelet inhibition and patient Outcomes (PLATO) trial, we evaluated the effect of the antiplatelet drug ticagrelor, compared to clopidogrel, in patients with ACS [11,12]. We have previously reported ticagrelor to be both more efficacious and of similar safety compared with clopidogrel, irrespective of diabetes status [8]. In the current substudy, we stratified patients into clinically accepted subgroups of glycaemia based on admission non-fasting glucose and HbA1c to explore the impact of dysglycaemia on CV outcomes [3,8].

Study population
The PLATO trial (NCT00391872) enrolled 18624 patients with ACS, who were randomised to either clopidogrel or ticagrelor treatment in addition to optimal medical therapy, including aspirin, and optional invasive strategy [11,12]. All patients were evaluated for diabetes at baseline, and baseline non-fasting glucose concentration and HbA1c were collected. All participants provided written informed consent and the study complied with the declaration of Helsinki. The study was approved by all local Ethics Committees and Institutional Review Boards.

Laboratory analysis
The details of laboratory sampling and handling have previously been published [11,12].
Plasma glucose and HbA1c concentrations were determined with sandwich immunoassays on the Cobas® Analytics e601 Immunoanalyzer (Roche Diagnostics, Mannheim, Germany) at the Uppsala Clinical Research Center laboratory, Uppsala, Sweden.

Definitions of outcome events and glycaemia subgroups
The primary endpoint of the present biomarker substudy was the composite of CV death, stroke or spontaneous myocardial infarction (sMI) within one year of follow up. sMI was defined in accordance with the universal definition of myocardial infarction (MI), hence a non-procedurerelated, non-fatal, MI type 1 [13].

Statistical analysis
Baseline characteristics were presented by subgroups of glycaemia, and compared using chi-square or Kruskal-Wallis tests. Kaplan Meier event rates at 1-year, overall and by glycaemic subgroups, were estimated. The associations between glycaemic subgroups, with the non-diabetes group as reference, and the outcomes were assessed by Cox proportional hazard models. Unadjusted and adjusted hazard ratios were derived. Adjustment variables included baseline characteristics (age, gender, smoking status, type of ACS, and history of angina, MI, hypercholesterolemia, diabetes, hypertension, chronic kidney disease, previous; percutaneous coronary intervention (PCI), coronary artery bypass grafting (CABG); stroke or transient ischemic attack (TIA), as well as log transformed biomarkers; NT-pro-BNP and hs-troponin I. p-values less than 0.05 were considered statistically significant results. Statistical analyses were performed with SAS 9.4 (SAS Institute, Cary, NC) ( Fig. 1).

Results
Information on previous DM, non-fasting glucose and/ or HbA1c, were available for 16,007 patients (85.9%). On admission, 6160 patients (38.5%) had DM according to our pre-specified criteria, of which 1501 patients (9.4%) had no previous DM diagnosis, consequently new-onset DM. In the new-onset DM subgroup, 1076 patients had an HbA1c ≥ 6.5%. Another 201 individuals had non-fasting blood glucose above 11.1 mmol/L, but with normal HbA1c. A total of 224 patients had both an elevated HbA1c and nonfasting blood glucose above 11.1 mmol/L.

Glycaemia subgroups and outcome
The Kaplan Meier event rates at 12 months for the combined endpoint of CV death, sMI or stroke per subgroups were 12.2% (694) for patients with DM, 7.2% (409) for pre-DM and 6.3% (213) for patients with no DM (Table 2 and Fig 1).
In a sensitivity analysis, all MIs (not only sMI) were evaluated in the composite endpoint (CV death, MI and stroke), with event rates at 12 months per subgroups of 14.5% (832) for patients with DM, 9.0% (522) for pre-DM and 8.5% (293) for patients with no DM.
The association between glycaemia subgroups and the combination of CV death, sMI or stroke is presented in Table 2.
The individual components of the composite endpoint were evaluated in secondary analysis and are presented in Table 2.

Discussion
DM is independently associated with CV events, but there are conflicting reports on the impact of pre-DM on outcomes [2][3][4][5][6][7]9]. This may be due to differences in the identification, classification, follow-up and, in some cases, a lack of other important prognostic variables in ACS, including hstroponin and NT-proBNP [5]. DM is a multifaceted syndrome, and the pathogenesis for CV disease may include macrovascular changes, endothelial cell dysfunction and accelerated atherosclerosis, with a subsequent increased risk for plaque rupture, thrombus formation and ultimately vessel occlusion [3]. In the current substudy, we found comorbidities to be more common in patients with DM, including disadvantageous biomarker levels such as elevated serum creatinine and NT-proBNP, indicating underlying organ dysfunction. We did not, however, observe any differences between patients with pre-DM and patients without DM.
Pre-DM, defined as blood glucose concentrations higher than normal but not meeting the definition of diabetes, is associated with an approximately 5-10% annual risk of developing type 2 DM [14]. When diagnosed with DM, the patient momentarily is considered to have heightened risk for adverse events, and the indication for cardiac revascularisation, medical treatments e.g. ACE inhibitors, and fluid therapy with angiography uses the presence of DM as a dichotomous condition [2,3,15]. It is, however, unknown if the observed heightened CV risk starts even at moderate levels of dysglycaemia or if abnormal glucose metabolism below a certain threshold may be non-pathogenic [3,14].
In the current cohort, the proportion of patients with DM and at the same time a normal HbA1c was 2.5%, indicating that medical and lifestyle treatments still are not able to maintain normal HbA1c in patients with diabetes.  The increased CV risk for patients with manifest DM was, however, not observed in patients with pre-DM during the first year after ACS, and we believe this highlights that a direct effect of hyperglycaemia on clinical outcome may not be present, rather the deleterious hyperglycaemic effects may arise after a prolonged and accumulated exposure to dysglycemia.

Limitations
Manifest DM was one of several enrichment criteria employed to enter the PLATO trial if presenting with a non-STEACS [11,12]. It is likely that the current population exhibited a larger proportion of patients with diagnosed DM than other ACS populations, and consequently proportionally smaller subgroups of patients with pre-DM and no DM. Although follow up was up to one year, the detrimental effect of DM was not observed within the first year of diagnosis in the current study, however this study was not designed to answer these important questions.

Conclusions
Abnormal glucose metabolism is common in patients with ACS and 9% of the PLATO study patients had undiagnosed diabetes. Although the overall event rate during the first year after ACS is high, we only observed patients with definite DM to have a further increased CV risk, indicating that pre-DM is not immediately associated with worse CV outcomes over the first year after ACS. Nonetheless, this finding does not diminish the need for close follow-up of patients with pre-DM in order to detect DM early, as recommended by international guidelines.  Open Access 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.