Notwithstanding all the improvements seen in primary angioplasty programs, with dramatic reductions in the time between symptom onset and intervention, and although normal epicardial coronary artery flow is almost always achieved in a timely fashion after primary percutaneous coronary intervention (P-PCI) in patients with ST-elevation myocardial infarction (STEMI), a significant proportion of patients (between 20% and 60%) have a poor outcome because of microvascular coronary damage.1,2

The negative prognostic implications for the risk of left ventricular (LV) remodeling and of hard endpoints, including death, associated with coronary microvascular damage have been repeatedly confirmed with several invasive and non-invasive indicators.2–13 However, the precise mechanisms underlying coronary microcirculatory dysfunction before and after the restoration of epicardial blood flow are largely unknown and likely to be multifactorial. Traditionally, coronary microvascular dysfunction in this setting is seen as a consequence of the primary epicardial event and/or of coronary reperfusion, either pharmacological (thrombolysis) or mechanical (P-PCI). Mechanical obstruction (due to distal embolization of atherothrombotic debris14), coronary endothelial dysfunction (ED) (mediated by the release of vasoactive factors such as endothelin-1 and tissue factor15) and reperfusion injury (through various mechanisms including platelet and leukocyte accumulation leading to thrombosis, vasoconstriction and release of free oxygen radicals, proteases, and pro-inflammatory mediators, and complement activation leading to the release of histamine and to increased cell permeability, resulting in endothelial cell and myocyte swelling, interstitial edema, and further stimulation of leukocyte adherence16), have all been suggested as mechanisms involved in coronary microvascular dysfunction in STEMI patients.

An alternative explanation is that pre-existing and/or simultaneous coronary microvascular dysfunction may itself have pathophysiological importance and contribute to myocardial infarct extension, LV remodeling and future events.16 Whichever theory of microvascular coronary dysfunction in STEMI patients is correct (cause, consequence or both), endothelial function always seems to be at the core of the proposed mechanisms. There is accumulating evidence that ED is not simply a risk factor and precursor of coronary artery disease (CAD), but also plays a central role in processes leading to acute coronary syndrome (ACS) and STEMI.17 In patients with ACS, endothelial function, measured in the peripheral circulation, has been shown to be an independent predictor of events,18 and subsequent normalization of endothelial function in these patients predicts a lower risk.19,20 Similarly, in STEMI patients treated with P-PCI, peripheral ED was associated with larger infarctions as measured by peak troponin levels,21 and endothelial function improvement six months after the event also correlated with lower end-diastolic LV volumes.22

It is therefore plausible to speculate that previous ED or a blunted reaction of endothelial mechanisms during STEMI may be a central component of coronary microvascular dysfunction in these patients and, as a consequence, a determinant of microvascular reperfusion, infarction extension and LV remodeling. Ideally, this hypothesis would best be proved if endothelial function had been assessed before the myocardial infarction, which clearly is not possible. In fact, most studies assessing endothelial function in ACS patients were performed relatively late (several days or weeks) after the onset of the event.21,23–27 Consequently, the worse results documented in patients with ED may simply be a consequence of larger and more complicated infarcts. Taking peripheral endothelial function (which is much easier to measure) as a surrogate of coronary endothelial function, if the above theory were true we would expect worse acute ED, worse microvascular reperfusion and worse LV remodeling in patients with larger infarcts.

The purpose of this pilot study was to analyze the relation between peripheral endothelial function assessed early (24 h after reperfusion) by peripheral arterial tonometry (PAT), the presence of microvascular obstruction (MVO) and the extent of myocardial infarction.

Patients admitted with a first STEMI and treated with P-PCI were prospectively included. The inclusion criteria were chest pain lasting at least 20 min and ST-segment elevation of at least 1 mm in two or more contiguous leads; pain-to-balloon time less than 6 h (or 6-12 h if ongoing pain); P-PCI performed successfully on the culprit lesion in a native coronary artery; age over 18 years; and informed consent obtained. Exclusion criteria were previous history of myocardial infarction; left bundle branch block; implanted pacemaker; cardiogenic shock; known myocardial disease and severe LV hypertrophy (>15 mm); previous coronary artery bypass surgery; percutaneous revascularization in the previous three months; contraindication to adenosine; and any serious non-cardiac disease associated with a life expectancy of less than 12 months. The institutional ethics committee approved the study, which was conducted in compliance with the Declaration of Helsinki.

P-PCI and adjuvant treatment were performed according to standard of care for patients with acute STEMI. Endothelial function was assessed 24 h after P-PCI by PAT. Infarct extension was assessed by contrast-enhanced cardiac magnetic resonance (ceCMR) and by the release of troponin I. MVO was assessed by ceCMR and also by indirect indicators, including ST resolution on the electrocardiogram (ECG) and angiographic indicators (corrected TIMI frame count [cTFC], TIMI myocardial perfusion grade [TMPG] and index of microcirculatory resistance [IMR]). Additionally, an echocardiographic exam was performed in the first 12 h after P-PCI.

Thirty-eight patients were included (mean age 60.0±13.7 years, 29 male). Mean RHI was 1.87±0.60 and 16 patients (42.1%) had ED (RHI <1.67).

The main characteristics of the study population according to the presence of ED are presented in Table 1. The flowchart of the patient selection process and exams performed in the study is presented in Figure 1.

ceCMR: contrast-enhanced cardiac magnetic resonance; ECG: electrocardiographic; echo: echocardiographic; PAT: peripheral arterial tonometry.'> ceCMR: contrast-enhanced cardiac magnetic resonance; ECG: electrocardiographic; echo: echocardiographic; PAT: peripheral arterial tonometry.' title='Flowchart of the patient selection process and exams performed. cath lab: catheterization laboratory; ceCMR: contrast-enhanced cardiac magnetic resonance; ECG: electrocardiographic; echo: echocardiographic; PAT: peripheral arterial tonometry.'/>

Figure 1.

Flowchart of the patient selection process and exams performed. cath lab: catheterization laboratory; ceCMR: contrast-enhanced cardiac magnetic resonance; ECG: electrocardiographic; echo: echocardiographic; PAT: peripheral arterial tonometry.

(0.36MB).

There were no significant differences between the two groups in physical characteristics, risk factors or previous medication. Admission blood tests were also similar in both populations, except for high-sensitivity C-reactive protein, which showed a tendency for higher values in patients with ED. There were no significant differences in ischemic (pain-to-balloon) time between patients with and without ED, although there was a trend for lower door-to-balloon times in patients with ED. The prevalence of multivessel disease in patients with ED was almost twice that in patients without ED (68.8% vs. 36.4%), although this difference was not statistically significant (p=0.10). SYNTAX score also tended to be higher in patients with ED. There was a clear trend towards greater area at risk, measured by both APPROACH and BARI scores (p=0.08 and 0.07, respectively), and these two scores showed a weak but significant inverse correlation with RHI (r=-0.426, p=0.008 and r=-0.361, p=0.026, respectively). RHI values tended to be lower in patients in whom the left anterior descending artery (LAD) was the culprit vessel (1.74±0.65 vs. 2.02±0.51 in other vessels, p=0.15). Patients with ED also showed a trend for worse initial TIMI score (TIMI 0-1: 93.8% vs. 72.7% in patients without ED). Finally, treatment options (use of mechanical aspiration, stent implantation technique and use of abciximab) were similar in patients with or without RHI <1.67 (Table 1).

Extent of infarction

Both peak TnI and the AUC of TnI release were significantly higher in patients with ED (Figure 2). These differences were even more significant when TnI values were indexed to the area at risk, by both the APPROACH and BARI scores (Table 2).

P-PCI), according to the presence of endothelial dysfunction. RHI: reactive hyperemia index.'> P-PCI), according to the presence of endothelial dysfunction. RHI: reactive hyperemia index.' title='Peak troponin I and area under the curve of troponin I release (determined from seven scheduled blood tests in the first 48 h after P-PCI), according to the presence of endothelial dysfunction. RHI: reactive hyperemia index.'/>

Figure 2.

Peak troponin I and area under the curve of troponin I release (determined from seven scheduled blood tests in the first 48 h after P-PCI), according to the presence of endothelial dysfunction. RHI: reactive hyperemia index.

(0.12MB).

Table 2.

Troponin I release according to the presence of endothelial dysfunction.

Variablea,b	Total population (n=38)	Endothelial dysfunction (RHI <1.67)		pc
		No (n=22)	Yes (n=16)
TnI (ng/dl)
Peak TnI	95 (96)	67 (81)	118 (186)	0.024
Peak TnI (APPROACH)	25 (32)	17 (20)	34 (56)	0.009
Peak TnI (BARI)	24 (31)	17 (22)	33 (47)	0.008
Peak TnI (both scores)	24 (31)	17 (21)	33 (55)	0.009
AUC
AUC of TnI release	1293 (1580)	1076 (1042)	2305 (2486)	0.012
AUC of TnI release (APPROACH)	403 (522)	315 (303)	664 (1080)	0.008
AUC of TnI release (BARI)	383 (448)	314 (326)	618 (799)	0.007
AUC of TnI release (both scores)	393 (482)	314 (300)	641 (984)	0.007

APPROACH: modified Alberta Provincial Project for Outcome Assessment in Coronary Heart Disease; AUC: area under the curve; BARI: Bypass Angioplasty Revascularization Investigation; TnI: troponin I.

a

Median (IQR).

b

Peak value and area under the curve (AUC) of seven serial troponin I measurements performed in the first 48 h after primary angioplasty; total values and values indexed to the APPROACH, BARI or both scores are presented.

c

Mann-Whitney test.

LVEF by echocardiography was significantly lower and WMSI significantly higher in patients with ED (Figure 3). Likewise, GLS was significantly higher in patients with RHI <1.67, although there were no other significant differences in Doppler variables (Table 3).

RHI <1.67. *independent t test; **Mann-Whitney test; RHI: reactive hyperemia index.'> RHI <1.67. *independent t test; **Mann-Whitney test; RHI: reactive hyperemia index.' title='Left ventricular ejection fraction (mean ± standard deviation) and wall motion score index (median and interquartile range) assessed by echocardiography, according to the presence of endothelial dysfunction, defined as RHI <1.67. *independent t test; **Mann-Whitney test; RHI: reactive hyperemia index.'/>

Figure 3.

Left ventricular ejection fraction (mean ± standard deviation) and wall motion score index (median and interquartile range) assessed by echocardiography, according to the presence of endothelial dysfunction, defined as RHI <1.67. *independent t test; **Mann-Whitney test; RHI: reactive hyperemia index.

(0.12MB).

Table 3.

Echocardiographic results according to the presence of endothelial dysfunction.

Variable	Total population	Endothelial dysfunction (RHI <1.67)		pc
		No	Yes
2D measurements	(n=26)	(n=15)	(n=11)
LVEDV (ml)a	109.8±23.4	106.5±30.2	113.8±11.4	0.48
LVESV (ml)a	55.9±12.6	51.1±12.3	61.7±10.9	0.047
LVEF (%)a	48.6±7.1	51.4±4.7	45.3±8.3	0.045
WMSIb	1.44 (0.41)	1.35 (0.47)	1.77 (0.47)	0.006
LA volume (ml/m2)a	36.1±11.0	36.7±12.1	35.3±10.0	0.78
Doppler measurements	(n=21)	(n=11)	(n=10)
E/A ratioa	1.10±0.40	1.17±0.40	0.97±0.0.39	0.30
E/e′ ratioa	8.91±3.30	8.83±3.91	9.05±2.18	0.89
2D speckle tracking	(n=21)	(n=11)	(n=10)
GLSa	-13.16±2.35	-14.32±1.72	-11.89±2.35	0.014

2D: two-dimensional; GLS: global longitudinal strain; LA: left atrial; LVEDV: left ventricular end-diastolic volume; LVEF: left ventricular ejection fraction; LVESV: left ventricular end-systolic volume; WMSI: wall motion score index.

a

Mean ± standard deviation.

b

Median (interquartile range).

c

Independent t test for continuous variables with a normal distribution, Mann-Whitney test for continuous variables with a non-normal distribution.

On ceCMR, patients with ED had significantly lower LVEF and higher WMSI, confirming the echocardiographic results. There was a trend for more transmural necrosis (22.2% vs. 63.6%, p=0.06) and greater infarct mass (median value 10.1% vs. 17.5, p=0.08) in patients with RHI <1.67. The percentage of infarct mass indexed to area-at-risk scores (APPROACH and BARI) also tended to be higher in patients with ED (Table 4).

Table 4.

Contrast-enhanced cardiac magnetic resonance results according to the presence of endothelial dysfunction.

Variable	Total population (n=29)	Endothelial dysfunction (RHI <1.67)		pd
		No (n=18)	Yes (n=11)
LVEDV (ml)a	138.6±26.9	139.2±26.5	137.5±28.9	0.87
LVESV (ml)a	63.5±21.0	59.3±19.4	70.4±22.6	0.17
LVEF (%)a	53.9±8.4	56.6±8.1	49.5±7.2	0.025
WMSIa	1.37±0.33	1.28±0.31	1.53±0.32	0.05
Edema massb	19.1 (19.0)	17.2 (14.1)	21.2 (28.6)	0.28
Transmural necrosisc	12 (38.7)	4 (22.2)	7 (63.6)	0.06e
Infarct mass
Totalb	11.6 (9.3)	10.1 (10.3)	17.5 (15.4)	0.08
Percentageb	11.5 (13.7)	10.2 (7.6)	17.5 (21.8)	0.10
Indexed to APPROACH	3.2 (7.0)	2.7 (2.6)	4.9 (11.5)	0.10
Indexed to BARI	3.4 (5.8)	2.3 (2.7)	5.1 (11.5)	0.09
Salvaged massb	5.0 (14.0)	5.0 (8.8)	4.7 (27.6)	0.87
MVO
Presentc	8 (27.6)	2 (11.1)	6 (54.5)	0.03e
Massb	6.1 (15.5)	5.3 (-)	6.8 (20.8)	0.29

APPROACH: modified Alberta Provincial Project for Outcome Assessment in Coronary Heart Disease; BARI: Bypass Angioplasty Revascularization Investigation; LVEDV: left ventricular end-diastolic volume; LVEF: left ventricular ejection fraction; LVESV: left ventricular end-systolic volume; MVO: microvascular obstruction; WMSI: wall motion score index.

a

Mean ± standard deviation.

b

Median (interquartile range).

c

Number (%).

d

Independent t test for continuous variables with a normal distribution, Mann-Whitney test for continuous variables with a non-normal distribution, chi-square test for categorical variables.

e

Yates’ correction.

To our knowledge, this is the first study reporting the relation between peripheral ED measured early (24 h) and microvascular coronary damage and myocardial infarction extension in patients with STEMI treated with P-PCI.

The presence of ED in patients with ACS has been repeatedly demonstrated. Elbaz et al. invasively assessed endothelial function in non-culprit arteries in patients with non-ST elevation acute coronary syndrome (NSTE-ACS) using intracoronary acetylcholine infusion and found that 81% of patients had a blunted response to this endothelium-dependent vasodilator.40 Spiro et al. used flow-mediated dilation and low flow-mediated constriction to demonstrate that endothelial function is significantly more compromised in patients with NSTE-ACS compared to patients with stable CAD, both before and after angioplasty.41 Careri et al. assessed endothelial function in the first 12 h in both stable and NSTE-ACS patients, and showed that flow-mediated dilation is significantly lower in acute patients.20 Finally, we also demonstrated that patients with STEMI treated with P-PCI have a high prevalence of ED as assessed with PAT 3-5 days after the event.21

Importantly, peripheral endothelial function has been shown to relate closely with coronary microvascular dysfunction, using several different techniques42–44 including PAT.45 However, the exact role of peripheral ED (as a surrogate for coronary ED) in the early phase after P-PCI has never been investigated. Testing peripheral endothelial function is particularly challenging in acute patients, for several reasons: in addition to the unpredictable nature of the acute event onset, patients’ unstable condition and the pain, anxiety and fear they usually experience, it is generally impossible to comply with several recommendations for endothelial function assessment, including fasting state, morning exam, avoidance of caffeine or smoking, etc.46 Additionally, our patients were assessed after an arterial puncture (for urgent angiography) and several venous punctures (both for blood tests and for administration of drugs). Finally, the time of day was not the same for all measurements (since it depended on the time of the primary angioplasty), and therefore patients may or may not have been in a fasting state.

Notwithstanding all these potential pitfalls, the RHI results we describe in this population are similar to the values reported by Schoenenberger et al. (1.82±0.48) in the only published study that also measured RHI 24 h after P-PCI in STEMI patients.23 Furthermore, the percentage of patients with RHI <1.67 (42.1%) is consistent with the results observed in all studies that reported this information in the acute/subacute phase of ACS: 47.9% in the study by Yamomoto et al.,25 43.1% in our pilot study21 and 45.0% in the study by Schoenenberger et al.23

Patients with RHI <1.67 had significantly higher values of both peak TnI and AUC of TnI release, both total and indexed to the area-at-risk scores. Correspondingly, they had a three-fold higher risk of transmural necrosis and a clear trend to higher total and percentage infarct mass on ceCMR. In line with these results, they also had lower LVEF and higher (worse) WMSI, both on the initial echocardiogram and on ceCMR.

Similarly, patients with RHI <1.67 had a clear trend towards worse electrocardiographic (particularly residual ST elevation after P-PCI) and angiographic (IMR, cTFC and TMPG) indicators of microvascular reperfusion. Altogether, these data suggest greater damage to the coronary microvasculature in patients with peripheral ED, which is confirmed by the ceCMR results: patients with RHI <1.67 were almost five times more likely to have MVO on ceCMR, and tended to have greater MVO mass.

The fact that patients with ED tended to have more severe CAD (the prevalence of multivessel disease was almost twice as high in patients with ED, and SYNTAX scores also tended to be higher in these patients) is in agreement with our previous work24 and suggests that at least the previous severity of CAD is related to ED. We also found a trend for lower initial TIMI flow in patients with ED, which is in agreement with a recent study by Kandhai-Ragunat et al. These authors documented lower RHI values (although measured 4-6 weeks after STEMI) in patients with an occluded artery before P-PCI, compared with patients with a patent artery (2.08±0.34 vs. 1.75±0.35; p<0.007),26 suggesting that ED may be involved in microvascular reperfusion in STEMI patients.

However, we also found a significant negative correlation between both area-at-risk scores and RHI, and there was a trend for lower RHI in patients in whom the LAD was the culprit artery. Both larger areas at risk and the LAD as culprit artery are associated with larger infarctions, implying that the documented lower RHI values may reflect the extension of the infarction, as opposed to being part of the pathophysiological process that caused it. Indeed, our finding that patients with ED in the first 24 h after STEMI have larger infarcts and lower LVEF is in line with the evidence of neurohumoral activation in patients with heart failure. It has been demonstrated that reduction in stroke volume produces lower endothelial shear stress, which causes dysregulation in gene expression of nitric oxide (NO) synthase isoforms,47,48 leading to reduced NO bioavailability.49 Furthermore, there is an additional reduction in NO bioavailability caused by direct destruction of NO by reactive oxygen species, mainly driven by an increase in angiotensin II and aldosterone activity, and purine metabolism.50

While we cannot infer from our results whether ED was a cause or a consequence of the severity of myocardial infarction, it is reasonable to conclude that peripheral ED in these patients may be a marker of the severity of myocardial disease and therefore of its systemic consequences, which makes it useful to assess the effectiveness of therapy and to predict events, independently of its exact pathophysiological role.

In patients with acute myocardial infarction, early treatment with statins, which are known to reduce events, was initially assessed through their beneficial effect on endothelial function.51,52 Likewise, in heart failure patients, peripheral endothelial function has been described as a tool to predict response to cardiac resynchronization therapy (CRT), and ED improvement after CRT correlates with functional improvement,53 related to an increase in cardiac output,54 likely mediated by shear stress.55

In summary, although it is not possible at this point to determine the exact role of ED in STEMI patients, we demonstrate that its measurement early after P-PCI is feasible and related to infarct extent and MVO, which makes it a valuable tool for identifying high-risk patients and for testing new pharmacological or device-based strategies.

This study is limited by the small study sample, and larger studies will be necessary to confirm these results. Additionally, the EndoPAT exams were not performed in accordance with recommendations for assessing endothelial function: patients were not fasting, coffee or tea consumption was not excluded and the exam was not necessarily performed in the morning. However, measuring endothelial function in the acute phase was in fact the purpose of the study. Importantly, all the operators performing the EndoPAT assessment were highly trained in this procedure, having participated in several previous studies.

The presence of ED assessed by PAT at 24 h after P-PCI in patients with STEMI is associated with larger infarcts, lower LVEF, higher WMSI and higher prevalence of MVO.

The study was sponsored by unrestricted investigational grants from St. Jude Medical™, AstraZeneca, Merck Sharp & Dohme, Medtronic and Cordis.

Dr. Baptista has received consultancy and speaker fees from St. Jude Medical and AstraZeneca, not related to the present study. Dr. Pedro Farto e Abreu has received consultancy and speaker fees from AstraZeneca and Medtronic, not related to the present study. Prof. Victor Gil has received speaker fees from AstraZeneca and Merck Sharp & Dohme, not related to the present study. The other authors have no conflicts of interest to declare.