Impaired glucose tolerance and endothelial damage, as assessed by levels of von Willebrand factor and circulating endothelial cells, following acute myocardial infarction

Abstract

Background. Impaired glucose tolerance (IGT) following acute myocardial infarction (AMI) increases the incidence of major adverse cardiac events. We hypothesized that endothelial damage following AMI, as assessed by levels of von Willebrand factor (vWF) and circulating endothelial cells (CECs), would be more pronounced in patients with IGT compared to those with normal glucose tolerance (NGT).

Methods. We studied non-diabetic patients with AMI (n=125; 107 (86%) male; mean age 59 years (SD 12.5)) who underwent oral glucose tolerance testing 3–5 days after admission. We measured vWF (enzyme-linked immunosorbent assay) and CECs (CD146 immunobead capture) in the fasting state and at 2 h post glucose load.

Results. Base-line vWF and CEC levels were higher in IGT patients versus those with NGT and healthy controls (HC) (P<0.001). The acute increase in vWF and CECs in response to the glucose load was significantly higher in the IGT group compared to those with NGT and HC (P<0.01)—an increase on a par with that seen in newly diagnosed diabetics.

Conclusion. The degree of endothelial damage post AMI in patients with IGT is greater than NGT, and comparable to that seen in frank diabetes mellitus. Subjects with IGT therefore need to be as actively sought and managed.

• Acute myocardial infarction
• cardiovascular disease
• circulating endothelial cells
• diabetes mellitus
• impaired glucose tolerance
• von Willebrand factor

Introduction

A high prevalence of impaired glucose tolerance (IGT) has recently been reported in patients without a previous diagnosis of diabetes mellitus, following acute myocardial infarction (AMI) [1–6]. Also, follow-up data from these studies not only demonstrate a significantly shorter survival in those with IGT compared to those with normal glucose tolerance (NGT) [2], but also a 2-fold increase in major adverse cardiac events (MACE) in IGT subjects, compared to the newly diagnosed diabetic group [7].

Similar effects of both constant and intermittent hyperglycaemia have been observed ex vivo on human umbilical vein endothelial cells (HUVECs). By enhancing free radical production, hyperglycaemia induces oxidative stress and pathways implicated in cell apoptosis [8] and endothelial cell damage [9].

Von Willebrand factor (vWF) is probably the most useful plasma marker of endothelial damage [10], being almost exclusively produced by vascular endothelial cells [11]. Also, vWF levels increase in response to various stimuli associated with acute ischaemia, including hypoxia and inflammatory cytokines [12]. More importantly vWF is associated with arterial thrombus formation. Hence, vWF is not only a marker of endothelial damage/dysfunction but plays a role in the actual pathogenesis and haemodynamic deterioration associated with myocardial infarction [12], [13]. Recently, the quantification of immunologically defined circulating endothelial cells (CECs) from peripheral blood has rapidly gained ground as a novel technique in the assessment of endothelial damage [14]. A strong correlation between vWF and CECs has previously been demonstrated with elevated levels being associated with a poor prognosis [15].

Key messages

• Impaired glycaemic tolerance (IGT) following acute myocardial infarction (AMI) increases the incidence of major adverse cardiac events.

• The degree of endothelial damage post AMI in patients with IGT is greater than that seen in normal glucose tolerance and is on a scale comparable to that seen in frank diabetes mellitus.

• Subjects with IGT therefore need to be as actively sought and managed as those with frank diabetes.

We hypothesized that levels of vWF would be higher in patients with IGT post AMI, when compared to patients with NGT post AMI or healthy controls (HC), and that such endothelial damage would persist for at least 6 months post AMI. Secondly, we sought to reproduce the in vivo endothelial damage associated with hyperglycaemia (as assessed by endothelial cell detachment) in an in vitro model by exposing human umbilical vein endothelial cells (HUVECs) to incubation media at different glucose concentrations.

Methods

We recruited 125 consecutive non-diabetic patients aged 18–80 years who were admitted to the coronary care unit of Sandwell and City Hospitals, Birmingham, UK, with AMI between January 2006 and March 2007. AMI was defined according to the European Society of Cardiology and the American College of Cardiology 2000 guide-lines [16]. Thus, patients were diagnosed as AMI if the troponin I level was ≥0.6 ng/mL and/or the total creatine kinase (CK) and creatine kinase MB fraction (CK-MB) was more than twice the upper normal limit, in the presence of either typical symptoms (chest pain or discomfort >20 minutes), or development of pathological ‘Q’ waves in at least 2 of the 12 standard electrocariogram (ECG) leads or ECG changes indicative of ischaemia, ST-segment elevation >2 mm in 2 or more contiguous chest leads or 1 mm in two or more limb leads, or with presumed new left bundle-branch block or ST-segment depression or T wave abnormalities [16]. Patients with admission plasma glucose of >11.1 mmol/L or a previous documented history of diabetes mellitus were excluded.

Abbreviations

Table

ACS	acute coronary syndrome
AMI	acute myocardial infarction
CECs	circulating endothelial cells
CVD	cardiovascular disease
DM	diabetes mellitus
FPG	fasting plasma glucose
IGT	impaired glucose tolerance
NGT	normal glucose tolerance
OGTT	oral glucose tolerance test
vWF	von Willebrand factor

Demographic details were recorded on a standard proforma. Risk factors for Coronary heart disease (CHD) like hypertension and hyperlipidaemia in the patient group were defined according to the Joint British Societies’ guide-lines on prevention of cardiovascular disease in clinical practice 2005 [17]. All patients received the standard therapy that included aspirin, clopidogrel (300 mg loading dose on admission), low-molecular-weight heparin, intravenous nitrates, statin, β-blocker, angiotensin-converting enzyme (ACE) inhibitor, thrombolysis, and percutaneous coronary intervention where appropriate.

Forty healthy controls were recruited from hospital staff and relatives of patients. Healthy controls had no clinical evidence of vascular, metabolic, neoplastic, or inflammatory disease, as assessed by careful history and examination. These subjects were normotensive and in sinus rhythm. All healthy controls had NGT on a standardized oral glucose tolerance test (OGTT). Informed consent was obtained from all patients and control subjects, and the protocol was approved by the local research ethics committee.

Base-line assessment

An OGTT with 75 g of glucose (dissolved in 200 mL water) was administered following a 12-hour overnight fast before discharge (between days 3 and 5) post AMI. Glucose tolerance status was defined according to the 1998 World Health organization (WHO) criteria [18] and the fasting plasma glucose (FPG) criteria adopted by the American Diabetes Association (ADA) [19]. Thus, patients were classified as having diabetes mellitus (DM) if their fasting plasma glucose exceeded 7.0 mmol/L or their 2-hour post load plasma glucose concentration exceeded 11.1 mmol/L, or both. IGT was defined as a fasting plasma glucose <6.1 mmol/L and 2-hour plasma glucose of between 7.8 and 11.1 mmol/L. NGT was defined as fasting plasma glucose <6.1 mmol/L and 2-hour plasma glucose <7.8 mmol/L. Newly diagnosed diabetic patients were referred to the diabetic nurse specialist and managed accordingly with insulin and/or oral hypoglycaemic agents.

Blood collection was performed by atraumatic puncture of the antecubital vein. The first 4 mL of blood was discarded. All patients and healthy controls had 25 mL of blood drawn in the fasting state (before the glucose load): 4 mL in an ethylene diamine tetra acetic acid (EDTA) tube for CEC quantification (see below), a further 4 mL in an EDTA tube for HbA_1c, 4 mL in a fluoridated tube for fasting plasma glucose; and the remaining blood in citrated tubes were taken for plasma separation and storage. An additional 25 mL of blood was collected at 2 hours post glucose challenge (4 mL of which was used for CEC quantification), 4 mL for 2-hour plasma glucose, and the remaining for plasma separation and storage. A urine sample was also collected in a fasting state to detect the presence of microalbuminuria. The fasting and 2-hour plasma glucose were analysed using the hexokinase catalytic reaction, whilst the HbA_1c was measured by the liquid chromatography (HPLC) TOSOH G7 in our biochemistry laboratory. All samples for plasma were placed immediately on ice and centrifuged at 3000 rpm/1000 g at 4°C for 20 minutes within 1 hour of collection. Plasma was divided into aliquots and stored at -70°C to allow batch analysis. All base-line and 2-hour post glucose challenge CEC quantification was performed in blinded fashion with respect to the glucose status of subjects and healthy controls.

Clinical follow-up

All patients were followed up at 2 months and at 6 months for the occurrence of MACE, defined as a composite end-point of cardiovascular death, stroke or transient ischaemic attack, heart failure, occurrence of new or worsening of existing peripheral arterial disease, non-fatal myocardial infarction (MI), or refractory angina with ECG ST–T-segment changes requiring hospitalization, or urgent coronary revascularization. Non-fatal MI was defined by a history of angina with ST–T-segment changes on ECG and elevated cardiac enzymes (total CK and CK-MB above twice the upper normal limit or troponin I levels ≥0.6 ng/mL). Those patients with no MACE at 2 and 6 months had further blood samples taken for quantification of CECs and vWF.

Laboratory

Levels of vWF were measured in duplicate by enzyme-linked immunosorbent assay (ELISA) using commercial reagents (Dako-Patts, Ely, United Kingdom). The lower limits of detection were 0.5 IU/dL. Intra- and inter-assay coefficients of variations (CVs) were less than 5% and less than 10%, respectively. The immunobead technique of CEC isolation (using CD146-coated immunomagnetic beads) with cellular counterstaining using fluorescein isothiocyanate (FITC)-stained endothelial-specific Ulex Europeus lectin was used [20]. This involved diluting 1 mL of venous blood (collected in EDTA tube after discarding at least the first 4 mL of blood) in 1 mL of phosphate-buffered saline (PBS) solution (containing 0.1% sodium azide and 0.6% sodium citrate). A total of 100 µL (5×10⁶) of a preparation of anti-CD146 M-450 immunobeads (Dynal, Norway) and 20 µL of fragment crystalizable receptor-blocking agent (FC-receptor) to reduce non-specific binding (Octagam, Octapharma Ltd, UK) were then added to the diluted blood [20]. This mixture was incubated at 14–16°C for ≥30 min whilst being gently rotated (30 rpm) to ensure continued mixing in a head-over-head mixer. The rosetted beads were then washed five times inside a magnet (Dynal, MPC-L) to remove all unbound blood cells. The resulting rosetted cells and beads were suspended in 1 mL of PBS, with 100 µL of a 2 mg/mL prepared solution of Ulex Europeus lectin (Vector Laboratories, UK) added to it, and gently mixed for ≥30 minutes at room temperature (22–24°C) in darkness. The solution was then washed twice using 1 mL of PBS in a magnet. Finally, the resulting rosetted cells and beads were suspended in 125 µL of PBS for counting under epifluorescence microscopy (Zeiss) in Nageotte double counting chamber at 20× objective. CECs were defined as CD146 rosetted cells, bearing ≥4 beads, sized approximately 10–50 µm in diameter and fluorescing green due to their expression of Ulex Europeus lectin (Figure 1). The intra-observer coefficient of variation (CV) was 15.7% (n=20) with an inter-observer CV of 20.2% (n=70). The intra-assay CV was 24.5% (n=40), and the inter-assay CV was 28.3% (n=40). The observer and assay variability is consistent with other work [21].

Figure 1.  Epifluorescent micrograph of a circulating endothelial cell (Zeiss; 20× objective) stained bright green with Ulex Europeus lectin and rosetted by several anti-human CD146-coated beads (4.5 µm each).

Ex-vivo model—human umbilical vein endothelial cells (HUVECs)

Cultured human umbilical vein endothelial cells (HUVECs) forming a confluent monolayer were used for this model. Cell viability was 90%–95%, as estimated by tryphan blue accumulation. These cells were exposed to glucose concentrations of 15 mmol/L (n=12), or 7.5 mmol/L (n=12), with 1 mL of media per well, mimicking glucose concentrations in vivo at diabetic levels and just marginally raised hyperglycaemia as the case in subjects with IGT. As controls, we used samples with glucose-free media (n=12). After 60 hours of exposure to glucose, the plate with HUVECs was gently shaken using vortex for 10 seconds. Media was then collected and centrifuged. Supernatant was used for vWF quantification (ELISA, as above) and the cell pellet processed for flow cytometry to determine the number of detached endothelial cells.

Power calculations

Leurs et al. [22] demonstrated a significant increase in the level of vWF in subjects with IGT compared to those with normal glucose tolerance in a population-based prospective cohort study. We based our sample size calculations on this study [22] and determined that a sample size of 40 patients was required for both the NGT and IGT groups, and 15 patients in newly detected diabetic patients, to give 80% power to detect a difference of at least 20% in the means of vWF between NGT and IGT at 1 standard deviation and a significance level (alpha) of 0.05 (2-tailed).

Statistical analysis

The Kolmogorov-Smirnov test was applied to continuous data to confirm or refute a normal distribution. Results are expressed as mean and standard deviation (SD) or as median with interquartile range (IQR), as appropriate. Between-group (HC, NGT, IGT, and DM) comparisons were made using the chi-square test for categorical variables. For some variables, for example type of MI and risk factors, comparisons were only made among the patient groups. One-way analysis of variance (ANOVA) and the Kruskal-Wallis test were utilized for comparisons of continuous data with more than two groups, with Tukey's or Dunn's post hoc test correction performed, where appropriate. For paired data, the t test or the Wilcoxon matched signed-ranks test was used accordingly. Repeated measures ANOVAs with Tukey's and the Friedman test with Dunn's post hoc tests were used to analyse change in continuous variables over time where appropriate. Potential associations between clinical and biological parameters and MACE were tested using the univariate Cox regression, and event-free survival by Cox regression. Correlations were performed by the Pearson and Spearman correlation methods. The ‘acute change’ of vWF and CECs was defined as the difference (Δ) between their plasma concentrations at 2 hours post glucose challenge and fasting (base-line—pre glucose load) levels. All tests were 2-tailed, and probability values were considered significant at the 0.05 level. All statistical analyses were performed using SPSS software, version 14.0 (SPSS, Chicago, IL) except the Dunn's post hoc test which was performed using Graph Pad InStat version 3.05 for Windows 95/NT (GraphPad Software, San Diego California USA; www.graphpad.com).

Results

Oral glucose tolerance testing divided the patients into three groups: NGT was present in 42% (95% CI 29–55), with IGT and newly diagnosed DM detected in 38% (95% CI 24–52) and 20% (95% CI 4–36), respectively (Table I). The entire control group had normal glucose tolerance by definition. The only significant differences between healthy controls and patients were in ethnicity and waist-hip ratio (see Table I).

Table I.  Base-line demographic and clinical characteristics of patients and healthy control subjects.

Clinical characteristics Mean (SD)	Healthy controls (n=40)	Normal glucose tolerance (n=52)	Impaired glucose tolerance (n=48)	Diabetes mellitus (n=25)	P-value
Age (years)	57.9 (11.2)	57.4 (12.4)	61.0 (12.9)	57.9 (11.9)	0.475
Male, n (%)	35 (87.5)	43 (82.7)	43 (84.0)	21 (84.0)	0.766
Ethnicity, n (%)
White Europeans	38 (95)	40 (77)	30 (63)	14 (56)	<0.001
South Asian	2 (5)	12 (23)	18 (37)	11 (44)
Type of AMI, n (%)
STEMI		30 (58)	21 (45)	11 (44)	0.346
NSTEMI		22 (42)	27 (55)	14 (56)
Cardiac enzymes
Creatine kinase^a (IU/L)		662.5 (261–1392)	935 (387–1510)	871 (493–1497)	0.331
Troponin I^a (ng/mL)		4.9 (1.9–20.1)	6.0 (1.85–16.2)	10.9 (1.7–31.0)	0.729
Body mass index (kg/m^2)	26.7 (3.4)	27.5 (4.4)	28.2 (4.8)	29.4 (4.8)	0.09
Waist/hip ratio	0.93 (0.07)	0.98 (0.12)	0.98 (0.09)	1.0 (0.07)	0.012
Microalbuminuria ^a (mg/mmol)	0.54 (0.21–1.17)	0.8 (0.26–1.43)	0.74 (0.35–1.98)	1.13 (0.27–4.2)	0.155
Random plasma glucose (mmol/L)	–	6.2 (1.5)	6.6 (1.4)	7.4 (2.0)	0.012
HbA1c (%)	5.4 (0.28)	5.6 (0.4)	5.8 (0.5)	6.3 (1.0)	<0.0001
Drug treatment, n (%)	–
Thrombolysis		10 (20)	8 (16.7)	4 (16)
Primary PCI		19 (36)	10 (20.8)	4 (16)
PCI		18 (34)	20 (40)	13 (52)	0.66
Medical treatment		4 (8)	8 (16.7)	2 (8)
CABG		1 (2)	3 (6.3)	2 (8)
Aspirin		50 (96)	47 (98)	25 (100)	0.77
Clopidogrel		50 (96)	45 (94)	24 (96)	0.55
β-Blockers		42 (81)	42 (88)	23 (92)	0.49
Statin		50 (96)	46 (96)	25 (100)	0.52
ACE-inhibitor		47 (90)	47 (92)	25 (100)	0.39

^a Median and interquartile range.

ACE = angiotensin-converting enzyme; AMI = acute myocardial infarction; MI = myocardial infarction; NSTEMI = non-ST-elevated acute MI; PCI = percutaneous coronary intervention; STEMI = ST-elevated acute MI; CABG = Coronary artery bypass graft.

Base-line changes in vWF and CECs

Fasting and absolute 2-h vWF levels (Table IIA) were significantly higher in the IGT and DM groups compared to those with NGT post AMI (P < 0.01) and HC (P < 0.001). Following OGTT, the 2-h vWF level increased significantly in all patient groups, versus the fasting level (P < 0.0001), except in the HCs.

Table IIA.  A: Plasma von Willebrand factor (vWF) levels by base-line glycaemic status post acute myocardial infarction.

vWF IU/dL Mean (SD)	Healthy controls (n=40)	Normal glucose tolerance (n=52)	Impaired glucose tolerance (n = 48)	Diabetes mellitus (n=25)
Fasting	103.3 (19.2)	118.2 (20.3) ^a	134.27 (27.4) ^a, b	134.1 (17.2) ^a, b
2-h post glucose challenge	103.7 (17.6)	125.4 (19.4) ^a, c	145.9 (29.4) ^a, b, c	145.4 (17.7) ^a, b, c

Between-group analyses by ANOVA with Tukey post hoc test.

Pre and post glucose challenge vWF analysis by paired t test.

^a Significantly higher (P<0.001) compared to healthy controls.

^b Significantly higher vWF compared to normal glucose tolerance group (P<0.01).

^c Significantly higher vWF at 2-hour post glucose challenge (P<0.0001).

vMF-von Willebrand factor; CEC-Circulating endothelial cells.

This rise in vWF (ΔvWF) was significantly greater in the IGT and DM groups compared to the NGT and HC groups, whilst those with NGT had a greater rise than the HC (P < 0.001) (Table IIA, Figure 2A). Similar results were observed for absolute CECs counts at fasting and 2-hour post glucose challenge (Table IIB), except that the circulating endothelial cells (CEC) count increase was also observed in the HC group. The ΔCECs was also significantly higher in the IGT and DM groups compared to those with NGT and HC (P < 0.001). Again, those with NGT following AMI had a significantly greater ΔCECs following OGTT than HC (P < 0.001) (see Figure 2B).

Figure 2.  A: Acute increases (Δ) in von Willebrand factor following an oral glucose tolerance test according to different glucose tolerance status. B: Acute increases (Δ) in circulating endothelial cells following an oral glucose tolerance test according to different glucose tolerance status.

Table IIB.  B: Circulating endothelial cells (CEC) by base-line glycaemic status post acute myocardial infarction.

CECs (Cells/mL) Median (IQR)	Healthy control (n=40)	Normal glucose tolerance (n=52)	Impaired glucose tolerance (n=48)	Diabetes mellitus (n=25)
Fasting	1 (0–1)	4 (3–6) ^a	7 (7–9) ^a, b	10 (9–12) ^a, b
2-h post glucose challenge	1 (0–3) ^c	6 (5–8) ^a, c	11 (9–15) ^a, b, c	14 (13–16) ^a, b, c

Between-group analyses by Kruskal-Wallis and Dunn's post hoc test.

Pre and post glucose challenge CEC analysis by Wilcoxon signed ranked test.

^a Significantly higher (P<0.001) compared to healthy controls.

^b Significantly higher CECs compared to normal glucose tolerance group (P<0.001).

^c Significantly higher CECs at 2-hour post glucose challenge (P<0.001).

In the patient groups (n=125), ΔvWF correlated significantly with ΔCECs (r=0.35, P < 0.0001). Modest correlations between fasting vWF and fasting CECs levels were also found within the IGT and DM groups (r=0.293, P = 0.046; and r=0.409, P = 0.042), respectively.

Follow-up data

Of the original cohort, 94 (75%) and 84 (67%) patients completed clinical review at 2 months and 6 months, respectively. Von Willebrand factor and CECs were measured if the patient had no MACE end-points. Over 2 months, 7 major adverse events occurred in total: 1 in the NGT group (heart failure), 6 in the IGT group (3 deaths, 1 angina, 1 stroke, and 1 heart failure), and 1 in the DM group (1 angina). At 6 months, 8 further events had occurred (all revascularizations): 2 in the NGT group, 4 in the IGT group, and 2 in the DM group.

Univariate Cox regression demonstrates that base-line glycaemic status (OGTT post AMI-derived glycaemic status ie. IGT + DM) was a significant predictor of MACE at 6 months (Hazard ratio (HR) 5.0 (95% CI 1.1–22.2); P = 0.034) but not at 2 months (P=0.19). Base-line glycaemic status was an independent predictor of MACE at 6 months, after adjustment for age, waist-hip ratio, microalbuminuria, fasting vWF, and CECs (HR 5.6 (95% CI 1.1–28.9); P=0.046). Kaplan-Meier event-free survival curves for MACE events at 6 months by base-line glycaemic status (NGT vs IGT + DM) are shown in Figure 3.

Figure 3.  Cumulative event-free survival at 6 months post acute myocardial infarction according to different base-line glycaemic status.

Von Willebrand factor and CECs levels of those who were event-free at 2 and 6 months are summarized in Table IV. As expected, the levels decreased over 2 to 6 months, except for vWF between base-line and 2 months and base-line and 6 months in the DM group (P>0.05), and for CECs levels between 2–6 months in the IGT and DM groups (P > 0.05) (Table IV). Post hoc analysis showed higher vWF and CECs levels in the IGT and DM at 2 and 6 months compared to those with NGT (P<0.01).

Table III.  Baseline demographic and clinical characteristics of patients with and without major adverse cardiac events at 2 and 6 months.

	Two months				Six months
Mean (SD)	Event free (n=87)	MACE (n=7)	Hazard ratio (95% CI)	P value	Event free (n=69)	MACE (n=15)	Hazard ratio (95% CI)	P value
Age (years)	58.9 (12.4)	68.4 (7.4)	1.07 (0.99–1.15)	0.06	60.6 (11.9)	64.8 (13.3)	1.03 (0.98–1.08)	0.21
Male (%)	76 (87)	4 (57)	0.22 (0.05–0.98)	0.05	61 (88.4)	11 (73.3)	0.4 (0.13–1.27)	0.12
Ethnicity: (%)
White Europeans	59 (68)	6 (86)	2.69 (0.32–22.3)	0.36	50 (73)	10(67)	0.82 (0.28–2.4)	0.73
South Asian	28 (32)	1 (14)			19 (27)	5(33)
Cardiac enzyme:
Creatine kinase §	871 (97–5464)	623.5 (134–2159)	1.0 (0.99–1.00)	0.42	871 (97–3016)	730 (134–2159)	1.0 (1.00–1.001)	0.56
Troponin I §	6.9 (0.8–120)	3.3 (1.3–12.9)	0.94 (0.84–1.05)	0.25	8.25 (0.8–108)	5.4 (1.3–51.9)	0.98 (0.94–1.02)	0.32
Fasting plasma glucose (mmol/L)	5.3 (0.8)	5.6 (0.4)	1.44 (0.7–3.0)	0.32	5.3 (0.8)	5.4 (0.78)	1.18 (0.65–2.1)	0.58
2-h plasma glucose (mmol/L)	8.4 (2.8)	9.7(1.7)	1.1 (0.94–1.3)	0.21	8.6 (2.9)	10.1 (3.2)	1.10 (0.98–1.24)	0.1
Fasting CECs §	6.0 (2–36)	7.0(6–14)	1.03 (0.91–1.17)	0.62	6.0 (2–36)	8.0 (3 ~ 14)	0.99 (0.89–1.11)	0.97
2-hr CEC§	9.0 (2–41)	10.0(2–13)	0.96 (0.81–1.13)	0.59	9.0 (3–41)	11.0 (2–18)	0.98 (0.89–1.08)	0.68
Fasting vWf	125.8 (26)	146.3 (16.5)	1.02 (1.0–1.03)	0.041	127.0 (27.4)	137.1 (19.5)	1.01 (0.99–1.02)	0.19
2-h vWF	135.7 (26.9)	159.0 (17.2)	1.02 (1.0–1.03)	0.027	136.5 (29.1)	148.1 (21.9)	1.01 (0.99–1.02)	0.14

CEC circulating endothelial cells; MACE major adverse cardiac events: vWF-von Willebrand factor. § Median and IQR.

Table IV.  vWF and CECs at 2 and 6 months in different glycaemic groups.

	Normal glucose tolerance	Impaired glucose tolerance	Diabetes mellitus
Base-line:
VWF mean (SD)	116.3 (21.6) (n=28)	142.2 (28.5) (n=25)^a	127.3 (21.1) (n=10)
CECs median (IQR)	5 (4–6) (n=30)	7 (6–9) (n=27)^c	11 (9–12) (n=10)^c
2 months:
VWF mean (SD)	104.0 (8.1) (n=28)	122.6 (3.8) (n=25)^b	120.5 (11.1) (n=10)^b
CECs median (IQR)	2 (2–3) (n=30)	5 (4–6) (n =27)^c	4 (4–5) (n=10)^d
6 months:
VWF mean (SD)	99.9 (7.8) (n= 28)	116.0 (5.3) (n=25)^b	113.8 (11.0) (n=10)^b
CECs median (IQR)	2 (–) (n=30)	4 (3–5) (n=27)^c	3 (3–4) (n=10)^d
Post hoc comparisons:
Base-line versus 2 months
VWF^e (IU/dL)	P=0.036	P=0.009	P=0.296
CECs^f (cells/ml)	P<0.001	P<0.01	P<0.05
Base-line versus 6 months
VWF^e (IU/dL)	P=0.002	P<0.0001	P = 0.053
CECs^f (cells/ml)	P<0.001	P < 0.001	P<0.01
2 months versus 6 months
VWF^e (IU/dL)	P < 0.0001	P < 0.0001	P<0.0001
CECs^f (cells/ml)	P<0.05	P > 0.05	P>0.05

^aBetween-group analyses by ANOVA with Tukey post hoc test significantly higher than NGT (P < 0.01).

^bBetween-group analyses by ANOVA with Tukey post hoc test significantly higher than NGT (P<0.0001).

^cBetween-group analyses by Kruskal-Wallis and Dunn's post hoc test significantly higher than NGT (P<0.001).

^dBetween-group analyses by Kruskal-Wallis and Dunn's post hoc test significantly higher than NGT (P<0.01).

^eRepeated measures ANOVA with Tukey post hoc test.

^fFriedman test with Dunn's post hoc test was used for repeated measures of CECs.

vMF-von Willebrand factor; CEC = circulating endothelial cells.

In vitro HUVECs model

There was a significant increase in detached HUVECs count with increasing glucose concentration (Table V, Figure 4). The increase in detached HUVECs was significantly correlated with increased levels of vWF (Pearson r=0.89, P<0.001).

Figure 4.  A: Detachment of human umbilical vein endothelial cells (HUVECs) according to different glucose concentrations in media. B: Supernatant von Willebrand factor levels according to different glucose concentrations in media.

Table V.  In vitro experiment showing HUVECs count and vWF levels at different glucose concentrations.

Glucose concentration in (mmol/L)
Mean (SD)	0	7.5	15.0	P-value
HUVECs/mL	309 (112)	1494 (228)^a	2778 (1141)^a	<0.001
vWF, IU/dL	0.27 (0.08)	0. 62 (0.15)^a	0.96 (0.52)^a	<0.05

^a15 mmol/L and 7.5 mmol/L versus glucose-free media.

Discussion

To our knowledge this is the first study to examine the degree of endothelial damage in subjects with IGT post AMI, and to further demonstrate that the degree of endothelial damage observed is of a similar magnitude to that seen in established diabetes. Of note, our study patient profile is similar to that seen in the other studies [1–6] as confirmed by the similar rates of IGT and newly diagnosed DM detected in this post AMI cohort. The significantly higher degree of endothelial damage in subjects with IGT compared to those with NGT is not only evident at base-line but is persistent up to 6 months post AMI. This probably explains the high MACE rate observed in this group, in our exploratory analysis.

Excessive postprandial hyperglycaemia in IGT initiates a cascade of pathophysiological abnormalities, such as increased generation of reactive oxygen species (ROS) [23], reduced availability of nitric oxide [24], activation of protein kinase-C [25], production of advanced glycation end-products [26], and inflammation [27]. These mechanisms are crucial in the pathogenesis of endothelial damage that precedes atherosclerotic changes of the vascular wall. This possibly explains the association between postprandial hyperglycaemia and surrogates of atherosclerosis, including carotid intima-media thickness (IMT) [28]. More importantly, IMT not only correlates with the postprandial glucose levels but to the actual hyperglycaemic spikes during the OGTT [29], with significant regression in carotid IMT following better management of postprandial hyperglycaemia [30].

Hence, the acute increase of vWF and CECs in response to the glucose challenge of OGTT in our patients indicates that a certain degree of vascular damage occurs in everyday life for some of our patients, in response to postprandial hyperglycaemia. More importantly, the actual increase (Δ) in vWF and CECs in the IGT group was once again comparable to that seen in those with frank diabetes mellitus. Similarly, markers of inflammation such as C-reactive protein and white blood cell counts have not only been shown to positively correlate with postprandial hyperglycaemia during an OGTT [31], [32], but also to deteriorate with increasing degrees of glucose intolerance.

We further demonstrate in vitro what we have observed in vivo by using our HUVECs model, where the number of detached/damaged HUVECs and vWF levels were significantly higher in the glucose media at 7.5 mmol/L compared to media with no glucose, and comparable to media at glucose concentrations of 15 mmol/L.

Clinical implications

Subjects with IGT post AMI are clearly at increased cardiovascular risk, with an increased rate of MACE in the IGT group, consistent with previous studies [2], [6], [7], even though the primary aim of our study was not to evaluate MACE rates. Such data place great emphasis on mandating OGTT in non-diabetic post AMI patients, as suggested by our group [33], and in the recently published European Society of Cardiology guide-lines for the management of AMI [34]. This helps risk stratification, giving an opportunity for early identification of dysglycaemic patients with AMI before they are discharged home. Even simple and cost-effective life-style intervention studies with personalized recommendations involving diet and exercise have been shown to reduce the transition of IGT to frank diabetes [35], [36], as have pharmacological interventions [37–39]. Agents such as semisynthetic insulin, disaccharidase inhibitors, thiazolidinediones, metformin, and meglitinides have all been shown to counteract postprandial glucose elevations successfully [40].

Limitations

The outcome analysis is purely exploratory observation (and not a primary objective of the study) since our study is relatively small with a short period of follow-up for MACE end-points. Some may consider our results in terms of glucose tolerance status to be a mere reflection of ‘stress’, even though previous studies have demonstrated a concordance of 50%–70% between OGTT done at discharge post AMI and after 3 months [3], [41]. Furthermore, HbA_1c levels on admission largely reflect glycaemic status by OGTT, confirming any dysglycaemia as chronic rather than acute. Many patients were also managed by percutaneous coronary intervention (PCI), which has been shown to influence vWF and CECs, although there was no difference between the study groups in such interventions, nor in other cardiovascular prevention drugs such as statins. The patients with IGT are not as aggressively managed as ‘frank diabetes’ post AMI, which may account for the MACE events, but larger cohorts are required to explore this. Finally, vWF levels and CEC counts in our study were lower than recently published data [42]. This could probably be explained by the fact that we measured these parameters between days 3 and 5 post acute MI, whilst they measured within hours of acute insult in other studies. However, this difference could also be due to different laboratory methods used in other studies. Finally, our study involved measuring vWF and CECs (as direct products of the endothelium) as markers of endothelial damage. Measuring a clinical functional marker of endothelial perturbation, such as flow-mediated dilatation (FMD) (an index of endothelial dysfunction), in addition to these, would have been useful. However, significant correlations between vWF, CECs, and FMD have previously been demonstrated [43].

Conclusion

The degree of endothelial damage (vWF, CECs) in subjects with impaired glucose tolerance is not only significantly higher than in those with normal glucose tolerance, but appears comparable to that observed in frank diabetes mellitus. Furthermore the pathological increase in vWF and CECs in response to glucose challenge in IGT (clinically mirrored by postprandial hyperglycaemia) may indicate a more ‘fragile’ endothelium perhaps contributing to the adverse prognosis observed in these patients.

The clinical significance of such data therefore mandates impaired glucose tolerance post acute myocardial infarction to be identified at index admission before discharge and to be managed as aggressively as frank diabetes.

Acknowledgements

Declaration of interest: The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.