Diabetes mellitus (DM) is considered not only a risk factor for coronary artery disease but a coronary artery disease equivalent [1]. Actually patients with diabetes mellitus tend to have worse outcomes after percutaneous coronary intervention (PCI) than non diabetic patients [2-7], thus proving negative clinical relation between exposition to high level of glucose and cardiac ischemia [8-13]. Nonetheless, there is contrasting evidence about target level of glucose to pursue for patients undergoing PCI, without a clear benefit related to an aggressive glucose lowering strategy [14-18,38]. Counterintuitive experimental research proved that diabetes seems to be protective for ischemic heart in animal models, through increased expression of cell survival proteins and decreased infarct size [19], greater plasticity and cellular resistance to injury [20] and the enhanced myocardial activity against oxidative damage during both ischemia and reperfusion [21].

To our knowledge no observational studies performed on patients have investigated value and degree of interaction between diabetes mellitus and cardiac ischemia. In order to address this issue, we retrospectively analyzed an unselected population undergoing PCI in our center.

The present retrospective study included all consecutive patients undergoing PTCA in our centre between July 2002 and December 2004.These patients were first divided into 2 cohorts: in the first diabetic patients, both insulin and non insulin dependent, and in the second non diabetic patients. Moreover they were divided into 4 groups: patients with DM and a history of myocardial ischemia, that is a prior myocardial infarction (MI) or a prior percutaneous or surgical revascularization, were included in the diabetes and ischemia (DI) group, those with a diagnosis of DM without ischemia in the diabetes only (DO) group, those with previous ischemia but non diabetic in ischemic only group (IO) and non-diabetics without an history of ischemia in the ND group. All patients provided written informed consent for the procedure, and ethical approval was waived given the retrospective observational design.

All patients were pre-treated with aspirin, 100 mg once daily, and clopidogrel 75 mg once daily or ticlopidine 250 mg twice daily at least three days before the procedure. A loading dose of 300-600 mg of clopidogrel and 250-500 mg of aspirin were given to patients not pre-treated. At the start of the procedure, unfractionated heparin was administered at a dose of 70-100 UI/KG to achieve an activated clotting time ≥ 250 seconds. The use of glycoprotein IIb/IIIa was left to the discretion of the operators. Coronary vasodilators (nitroglycerin) have been routinely used both pre and post-procedure. Coronary angioplasty and stent implantation were performed according to the current practice and technical guidelines.

The diagnosis of diabetes mellitus was made according to the WHO Consultation [20]. The primary outcomes were rates of death, MI (defined as Q-wave or non Q-wave with elevation of total CK-MB 2 times above the upper limit of normal), or repeat revascularization. In order to assess all procedural and in-hospital outcomes, we consulted our institutional electronic database and individual patient charts. We recorded short term outcomes with at least 3 years of follow up, performed with phone calling, ambulatorial visits or or formal query of primary care physicians

Continuous variables were expressed as mean±standard deviation and were compared with ANOVA, with a provision to perform additional Gosset t tests with Bonferroni adjustment in case of overall p<0.05. Categorical variables were presented as counts and percentage and were compared with the chi-squared test, with a provision to perform additional chi-squared tests with Bonferroni adjustment in case of overall p<0.05. Statistical significance was set at the twotailed 0.05 level. Computations were performed with SPSS 11.0 (SPSS, Chicago, IL, USA).

A total of 1610 patients were identified: 403 had diabetes, and 1207 were non-diabetic. Thus 152 patients were assigned to the DI group, 173 to the DO, 449 patients to the DI group and 836 to the ND group. Clinical, angiographic and procedural characteristics were summarized, respectively, in and Table 2.

Table 1: Baseline features comparing patients with vs those without diabetes mellitus (DM).

Table 2: Angiographic features comparing patients with vs those without diabetes mellitus (DM).

While non diabetic patients were more likely to be male, diabetic patients more frequently reported in their history a diagnosis of hypertension, renal failure or heart failure in addition to higher rates of previous surgical revascularization. During coronary angiography small vessel disease was more frequently found in diabetic people (35.2% vs 22.9%, p<0.001): moreover the number of diseased vessels and number and length of stent length were higher in people with diabetes mellitus.

Short term outcomes are reported in Table 3 and Table 4. Both at short and long term follow up patients belonging to DI groups did not show higher rates of death and AMI and percutaneous revascularization among the four groups, also when stratified for admission diagnosis. The only significant difference was the lower incidence of short term incidence of AMI in patients presenting with unstable angina or NSTEMI between non-diabetic patients (1.8% vs 0.5% vs 5.4% and vs 3.8%; p=0.049), and of long term incidence of target vessel revascularization of patients with diabetes mellitus, or a previous ischemia or both (29.5% vs 25.3% vs 25.3% vs 17.5%; p<0.001)

Table 3: Thirty-day outcomes.

Table 4: Long term outcomes

Diabetes mellitus is an important risk factor burdened by higher rates of cardiovascular events but in our article its negative impact was attenuated after dividing patients according to the presence or not of previous cardiovascular ischemia.

Actually the negative effect of diabetes mellitus has been widely supported [2-11], as outlined in our registry. Death and myocardial infarction were more likely to be reported in patients with diabetes, probably both because of its proved negative effects on cell survival [23] and because of higher rates of important co morbidities like renal failure [25], previous coronary artery bypass graft [26], peripheral artery disease [27] and small vessels [28].

On the contrary a clinical endorsement to experimental positive interaction between exposure to both previous ischemia and high glucose levels could be elicited from our results. Actually while experimental data arise from biological background [19-21], their main limit is that they are performed in a non-human clinical setting. Moreover many studies performed in clinical setting have demonstrated a negative predictive value both for previous ischemia [25-28] and for diabetes mellitus in the setting of coronary revascularization [2-12]. Our article on the contrary demonstrated that the combination of both these risk factors did not play a negative role for our patients. This could be related to the protective effect of antischemical drugs [29-32], but also to the previous reported biological correlation [33].

One of the main clinical applications of all the previously cited work is to perceive the target level of glucose to be achieved in diabetic patients undergoing PCI [37]. Actually no evenness of results has been achieved, in different clinical settings. For example in the DIGAMI trial, who was the only experimental work to report a mortality reduction for patients with a tight glucose control [17], about 35% of patients had a history of previous cardiac ischemia, while in the DIGAMI 2, in which this benefit disappeared, more than 50% of patients reported previous cardiovascular events or interventions [18]. Many explanations have been reported [14-18,38], and a possible reason of this variability is that clinical history of the patients could have influenced these outcomes:these data and their interpretation should be evaluated in the ongoing trials [34-36].

This study has several founding limitations, including the retrospective nonrandomized design and the descriptive aim. Moreover information about drug taken from our patients and their glucose blood level are lacking, thus this article should be viewed as hypothesis-generating only.

Diabetes mellitus and previous cardiac ischemic injury interplay an indifferent interaction in our study, thus fostering further researches in order to establish if patients’ cardiovascular history could influence the effectiveness of different glycemia target level protocols.