Article Text

Download PDFPDF

Original article
The index of microvascular resistance identifies patients with periprocedural myocardial infarction in elective percutaneous coronary intervention
  1. Jamie J Layland1,2,
  2. Robert J Whitbourn1,2,
  3. Andrew T Burns1,2,
  4. Jithendra Somaratne1,
  5. Georg Leitl1,
  6. Andrew I MacIsaac1,2,
  7. Andrew Wilson1,2
  1. 1Department of Cardiology, St Vincent's Hospital, Fitzroy, Victoria, Australia
  2. 2University of Melbourne, Department of Medicine, St Vincent's Hospital, Fitzroy, Victoria, Australia
  1. Correspondence to Dr Jamie Layland, St Vincent's Hospital, Fitzroy, Victoria 31010, Australia; jamie.layland{at}


Background This study was designed to assess whether measurement of the index of microvascular resistance (IMR) could help prospectively identify patients who develop periprocedural myocardial infarction (PPMI).

Methods and results IMR was measured in 54 patients before and following percutaneous coronary intervention (PCI) in a culprit vessel with a PressureWire using the equation IMR = PaHyp × TmnHyp (PdHyp−Pw/PaHyp−Pw). IMR was also measured in an angiographically normal reference vessel. The relative pre-IMR ratio (rPIMR) defined as IMR Culprit divided by IMR Non-Culprit was also calculated. Troponin was sequentially sampled up to 24 h following PCI. Mean troponin post-PCI was 0.37±0.8 ng/ml. 33 (61%) patients fulfilled the criteria for PPMI. IMR pre-PCI was the most significant correlate of post-PCI troponin (r=0.43 p=0.001), however, the number of balloon inflations (r=0.3, p=0.02) and rPIMR (r=0.33 p=0.017) were also correlated. IMR pre-PCI was higher in patients with periprocedural myocardial infarction compared with patients without PPMI (IMR pre-PCI 21.2±2.1 PPMI vs 15.6±1.8 No PPMI, p=0.02). The strongest predictor of troponin post-PCI was IMR pre-PCI (β 0.7, p=0.02). Both IMR pre- and rPIMR were predictive of PPMI (OR 11 (1.3 to 90.5) p=0.026, OR 1.09 (1 to 1.19) p=0.03, respectively).

Conclusion Microvascular function prior to PCI is an important determinant of PPMI. Measuring IMR pre-PCI and rPIMR may allow prospective identification of patients at risk of periprocedural myocardial infarction. Future studies in a larger cohort are required to establish the predictive ability of IMR in PPMI.

  • Microcirculation
  • PCI
  • troponin
  • elective
  • coronary physiology
  • cardiomyopathy
  • exercise echocardiography
  • cardiac ultrasound
  • angiography
  • myocardial function
View Full Text

Statistics from


The number of percutaneous coronary interventions (PCI) performed has risen exponentially over the past 2 decades. Figures from Europe suggest that the number of coronary stenting procedures has increased from 3000 in 1992 to 777 000 in 2004.1 Evidence of myocardial injury in patients undergoing PCI (periprocedural myocardial infarction: PPMI) has been associated with adverse in-hospital outcomes and a worse overall prognosis depending on criteria used.2–4 Contemporary guidelines provide a threshold of an increase in biomarkers >3×99th percentile upper reference limit for defining PPMI.5 Using such a definition, the number of patients experiencing PPMI in elective PCI is estimated at up to 30%.6

However, despite known risk factors for the development of PPMI, the majority of cases are clinically silent with detection of PPMI occurring sometime later on routine blood tests precluding an opportunity to intervene.6 ,7 Therefore, the ability to prospectively identify patients developing PPMI in the catheter laboratory would be advantageous, and perhaps provide a window of opportunity to adjust treatment strategies (eg, more aggressive antiplatelet therapy) for patients at risk of developing PPMI, and ultimately improve outcomes in elective PCI.

Occlusion of the coronary microvasculature by distal embolisation is a major cause of myocardial infarction post-PCI.8 However, studies in patients with acute coronary syndromes have suggested that the baseline status of the microvasculature is also an important determinant of cardiac injury following PCI.9 Recent work has highlighted the clinical utility of the index of microvascular resistance (IMR) in the evaluation of the coronary microvasculature.10 ,11 In patients with ST segment elevation myocardial infarction, IMR has shown a good correlation with areas of microvascular obstruction demonstrated on cardiac MRI .12 Furthermore, Cuisset et al have shown that in patients with stable angina randomised to either direct or a conventional stenting strategy, patients who had a troponin elevation following PCI had higher IMR values.13 Therefore, we hypothesised that evaluation of microvascular integrity using IMR before and after PCI would help in the identification of patients with PPMI.


Study population

The study population consisted of consecutive patients undergoing elective PCI for stable angina. Patients were excluded if they had a recent myocardial infarction (defined as a positive troponin >3 times the upper limit of normal within 7 days of PCI), or a history of myocardial infarction in the culprit territory in the previous 12 months, history of myocardial infarction in the non-culprit vessel, severe renal impairment (eGFR <30 ml/min), left ventricular hypertrophy on ECG, acute inflammatory illness, chronic atrial fibrillation, left ventricular dysfunction (ejection fraction ≤35%), previous coronary artery bypass surgery, contraindication to prolonged dual antiplatelet therapy, significant valvular heart disease or if the lesion involved a major side branch (>2.0 mm). The Human Research Ethics Committee at St Vincent's Hospital Melbourne approved the study protocol.

For the procedure, all patients received an initial bolus of 5000 units of intravenous heparin with additional bolus dosing to maintain an activated clotting time of 250 s, and were receiving aspirin and clopidogrel. A 6F coronary guiding catheter was used to engage the selected coronary artery. A 5F sheath was placed within the right femoral vein for drug delivery. All patients received 200 μg of intracoronary nitroglycerin in each study artery to minimise changes in coronary volume. A 0.014 coronary temperature and pressure-sensing guidewire was calibrated, and then equalised to the guiding catheter pressure with the distal sensor placed at the ostium of the coronary artery. The wire was then passed beyond the stenosis into the distal third of the vessel. Intravenous adenosine was administered via the right femoral vein (140 μg/kg/min) to achieve maximal hyperaemia. A physiological response to adenosine was observed in all patients.

Microvascular resistance was measured utilising the IMR as previously described10 in the culprit vessel. Three milliliters of room temperature saline was injected intracoronary to produce three reproducible and consistent thermodilution curves and derive the hyperaemic transit time (TmnHyp).

IMR calculation

In the culprit vessel, IMR was calculated as follows14:IMR=Pa×TmnHyp(PdPw/PaPw)where Pa and Pd were the mean hyperaemic aortic and distal coronary pressures, respectively, and Pw the coronary wedge pressure defined as the distal coronary pressure obtained during a 30 s balloon occlusion of the culprit vessel during the initial balloon inflation, and representing recruitable collateral vessels.15 A simplified method of IMR calculation was used for the non-culprit vessel (IMR =Pd × TmnHyp) as previously described.16 Fractional flow reserve (FFR) was defined as the mean distal coronary pressure divided by the mean aortic pressure during hyperaemia. Care was taken to ensure that the distal sensor was in the same position between measurements to avoid errors in transit time acquisition.

The change in IMR ratio (CIMRr) defined as the post-PCI IMR minus the pre-PCI IMR and divided by the pre-PCI IMR value was calculated to account for baseline levels of microvascular resistance pre-PCI.

In order to determine the relative contribution of the local microcirculation from the entire coronary tree, we measured IMR in an angiographically normal coronary artery without prior infarction. A ratio of the non-culprit to the culprit territory microvascular resistance was then calculated. The Relative pre-IMR ratio (rPIMR) was defined as the pre-PCI IMR in the culprit vessel divided by the IMR in an angiographically normal reference vessel.

Coronary flow reserve (CFR) was defined by dividing the baseline transit time by the hyperaemic transit time.

Troponin I, three times the 99th percentile (≥0.12 ng/ml) was defined as a periprocedural myocardial infarction.5 Endpoints for Troponin values five times and eight times the 99th percentile were also calculated.

Study protocol

IMR, FFR and CFR were measured in the culprit artery at baseline and in an angiographically normal reference vessel. Stenting of the culprit vessel was then performed and physiological measures were repeated. The decision to intervene and the use of glycoprotein IIb/IIIa inhibitors and direct thrombin inhibitors were at the operators' discretion.

Troponin I (Abbott-Architect) and creatine phosphokinase (CK) were sequentially measured every 6 h up to a maximum of 18–24 h following PCI.

Statistical analysis

Statistical analysis was performed using SPSS (SPSS Inc) statistical software package. Based upon previous data13 we hypothesised a mean IMR of 24.7 in patients with periprocedural myocardial injury, and 13.2 without, but with a SD of 10.2. With an α error of 0.05 and a power of 80%, we estimated a minimal sample size of 27 subjects in each arm would be required to demonstrate differences in IMR between patients with and without PPMI. Normality of data was assessed with the Kolmogorov–Smironov statistic. Non-parametric tests were used where appropriate. Logarithmic transformation of data was performed for non-normally distributed data. Continuous variables are summarised as mean±SD. Continuous variables were compared with the Student t test. Linear regression was used to assess predictors of troponin release following PCI. Multivariate logistic regression was performed to assess the impact of a set of variables on periprocedural MI. Univariate analysis was initially performed, and all the variables that exhibited a p<0.05 were entered in the multivariate model, along with other established risk factors for PPMI. Receiver-operator characteristic (ROC) analysis was used to determine the optimal cut-off value for IMR measured before and following PCI, total number of balloon inflations and rPIMR, for the identification of periprocedural myocardial infarction. The area under the ROC curve (AUC) was used as a measure of test accuracy. p<0.05 Was considered statistically significant.


Study population

Consecutive patients were screened to provide a study population that consisted of 54 patients undergoing elective PCI for stable angina. One patient with a major side branch occlusion with a clinically apparent PPMI was excluded from the analysis.

The mean troponin value following PCI was 0.37±0.8 ng/ml. Thirty-three (61%) patients fulfilled the criteria for PPMI in our study. Twenty-one patients (38%) had a troponin five times the upper reference range, and 16 (29.6%) patients had troponin elevation eight times the upper reference range. Mean CK following PCI was 106±99.

Baseline characteristics

Table 1 summarises the baseline, clinical and angiographic characteristics of the cohort divided into those that suffered from a PPMI and those that did not. All patients had successful PCI with TIMI III flow at the conclusion of the procedure. There were no major in-hospital complications. There was no clinically significant difference in age, sex, clinical comorbidities, haemodynamics or treatment between the two groups. Importantly, there was no difference in the proportion of patients treated with statins between the two groups or in the proportion of diabetic patients. No patient in the cohort received glycoprotein IIb/IIIa inhibition.

Table 1

Demographics and procedural characteristics of patients with and without PPMI

Coronary physiological data are presented in table 2. IMR pre-PCI was higher in patients who had a periprocedural myocardial infarction (21.2±2.1 PPMI vs 15.6±1.8 No PPMI, p=0.02, figure 1). IMR post-PCI was also significantly higher in patients who had a periprocedural myocardial infarction (25.9±3.8 PPMI vs 16.1±2.01 no PPMI, p=0.008). The rPIMR was higher in the PPMI group (rPIMR PPMI 1.48±0.19 vs no PPMI 0.89±0.13 p=0.03 figure 2). However, CFR (pre and post) was not significantly different between groups.

Table 2

Coronary physiological measures and procedural data of patients with and without PPMI

Figure 1

Mean IMR pre-PCI in patients with and without PPMI.

Figure 2

Mean rPIMR in patients with and without PPMI.

Patients with PPMI more frequently underwent pre- and post-dilatation (57.4% PPMI vs 42.6% no PPMI p=0.025, 76.4% PPMI vs 23.6 no PPMI <0.001, respectively). There were a higher number of balloon inflations (5.2±2.1 PPMI vs 3.8±2.0 no PPMI, p=0.02) in patients with PPMI. There was no difference in relation to stent size, length, number or FFR (pre or post) between the two groups.

The CIMRr was significantly lower in patients who developed PPMI (CIMRr PPMI 0.26±0.16 vs CIMRr No PPMI 1.71±0.33 p=0.03).

Table 3 provides information concerning different troponin thresholds and invasive coronary physiology. In essence, IMR pre-PCI, the rPIMR and the number of balloon inflations were higher in patients with higher troponin elevations.

Table 3

Mean differences in IMR and inflation number for troponin ×5 and ×8 ULN

Predictors of troponin post-PCI and periprocedural myocardial infarction

IMR pre-PCI and IMR post-PCI were significantly correlated with postprocedural troponin levels (r=0.43, p=0.001, r=0.34, p=0.01, respectively). The rPIMR was also positively correlated with troponin (r=0.33 0=0.017). There was also a significant correlation with the total number of balloon inflations and troponin release (r=0.3, p=0.02) and also with postprocedural CK levels (r=0.33, p=0.04). There was a negative correlation between the CIMRr and troponin (–0.35, p=0.014). However, there were no other statistically significant correlations between troponin and any other clinical and angiographic predictor inclusive.

Stepwise linear regression analysis performed to assess the determinants of troponin release following PCI included age, diabetes, eGFR lesion length and coronary physiological parameters. Of these, IMR pre-PCI appeared the only significant predictor of troponin (β 0.7, p=0.02) although the total number of balloon inflations approached predictive significance (β 0.24, p=0.06).

Tables 4 and 5 show the univariate predictors for PPMI using 3× and 5× ULN as the thresholds, respectively. Table 6 demonstrates results of multivariate regression using 3× ULN threshold IMR pre and post PCI, rPIMR and the number of balloon inflations were the major predictors of PPMI in our population.

Table 4

Univariate predictors of PPMI using troponin ×3 ULN

Table 5

Univariate predictors of PPMI using troponin ×5 ULN

Table 6

Multivariate predictors of PPMI using troponin ×3 ULN

At the highest troponin cut-off (×8) only the rPIMR and pre-PCI IMR were significantly predictive of PPMI (OR 1.07 (1 to 1.3) p=0.04, OR 17.3 (3 to 296.3) p=0.01, respectively). CFR was not significantly predictive of PPMI in our cohort.

By ROC curve analysis, the AUC was poor for IMR pre-PCI and inflation number to predict PPMI (AUC 0.64 and AUC 0.62, respectively). However, rPIMR performed better (AUC 0.68 (0.52 to 0.81) p=0.03). Using higher troponin values (×5), both IMR pre-PCI and the total number of balloon inflations performed better in detecting PPMI (AUC 0.72 (0.46 to 0.78) p=0.05, AUC 0.73 (0.54 to 0.87) p=0.02, respectively). However the rPIMR also performed better with an AUC 0.73 (0.58 to 0.85) p=0.005 (figure 3).

Figure 3

ROC curve of use of rPIMR in detection of PPMI (TnI ×5 ULN).

Using a troponin threshold of ×8 ULN the rPIMR performed well for detecting PPMI with an AUC of 0.75 ((0.51 to 0.86) p=0.02). Using the ROC curve the optimal cut-off value for rPIMR for the detection of PPMI (TnI ×5) was 0.88 and was associated with a sensitivity of 81%, a specificity of 70%, a positive predictive value (the number of true positives divided by the total number of true and false positive tests) of 67%, and a negative predictive value (the number of true negatives divided by the total number of true and false positive tests) of 80%.


We have shown that the rPIMR has the ability to identify patients who develop PPMI with good PPV and NPV. This was over a range of troponin thresholds and suggests that there are patients who can be prospectively identified and screened prior to PCI, with impaired baseline microvascular function, who are at greater risk of periprocedural injury. Furthermore, we have also shown that IMR pre-PCI and the total number of balloon inflations are important determinants of periprocedural troponin elevation.

Pathophysiology of PPMI

Although many factors are involved in the pathogenesis of PPMI, they all ultimately lead to the disruption of a portion of the coronary microcirculation producing myocardial necrosis. Contrast-enhanced cardiac MRI studies have shown that following coronary stenting, there are two major mechanisms responsible for PPMI.17 ,18 Side branch occlusions occurring peri-stent, that may not be clearly visible on angiography, are responsible for approximately 25% of cases.7 The other more common mechanism relates to distal embolisation. However, the relationship between baseline microvascular function and PPMI is not well described.

The clinical problem

The key issue concerning periprocedural myocardial infarction is that it remains clinically silent in the early phase post-PCI but portends a worse prognosis.19 The most common clinical scenario is that it is detected several hours following PCI on routine post-PCI biochemical evaluation. Therefore, strategies directed at the early detection of periprocedural myocardial infarction remain important on prognostic grounds.19

The prediction of periprocedural myocardial infarction

Large numbers of studies have identified risk factors for the development of PPMI. These include patient-related factors, such as advancing age, diabetes, renal impairment, multivessel disease and the presence of LV dysfunction, and also procedure-related factors, such as complex lesions, the use of rotational atherectomy and visible thrombus.8 However, despite knowledge of these risk factors and identification of the at-risk patient upfront, periprocedural myocardial infarction continues to occur in up to 30% of patients undergoing elective PCI.8 ,20 This has led to a number of investigators developing methods that allow for early identification of patients at risk for PPMI.

In 52 patients with stable angina and using Doppler derived indices, Bahrmann and colleagues demonstrated evidence of Doppler-detected microembolisation during PCI that was associated with periprocedural myocardial infarction.21 The group showed a count of 20 was an independent predictor of PPMI.22 However, other studies have shown a lack of correlation between Doppler signals and invasive coronary physiological measures.23

Cuisset et al demonstrated lower post-PCI IMR with a direct stenting strategy when compared with a conventional strategy that included predilation.13 Furthermore, patients with positive troponin levels post-PCI had a higher IMR values post-PCI. However, the pre-PCI IMR values were not disclosed in this study.

While our results are in accordance with prior studies, we have shown that an IMR pre-PCI is higher in patients who develop PPMI, and is independently predictive of periprocedural myocardial infarction. In addition, we have developed a novel ratio, the rPIMR, that is also predictive of PPMI over a range of troponin thresholds. Our optimal cut-off value of 0.88 was associated with a positive predictive value of 67% and a negative predictive value of 80%. However, unlike other studies, CFR was not predictive of periprocedural myocardial infarction in our population,

Prior studies utilising IMR as a technique have mainly been in the STEMI population, whereby a value of >3224 or >3325 is associated with a lack of left ventricular viability. Melikian and colleagues demonstrated that the mean IMR was 19 in patients with normal coronary arteries.26 We have shown a mean IMR of approximately 22 pre-PCI and 27 post-PCI in patients developing PPMI, which appears to be in line with previously published data.

When utilising the non-culprit vessel as a control, we assume homogeneity of microvascular resistance within the myocardium, an assumption that has underpinned other studies.16 ,26–28 Our results suggest that PPMI is more likely if IMR is higher in the culprit vessel compared with the non-culprit vessel, and strongly suggests that localised impairment of microvascular function predisposes to PPMI.

Previous data has highlighted the role of abnormal baseline microvascular function in determining biomarker release following PCI in patients with acute coronary syndromes.29 In accordance with this premise, our data demonstrates that the baseline IMR and rPIMR are different between patients with and without periprocedural myocardial infarction, and suggests that those patients who develop PPMI have a more diseased microvascular bed that is unable to cope with further ischaemic insults such as balloon inflations and stent deployment. This is also corroborated by the fact that the CIMRr is lower in patients who develop PPMI so that the change in status of the microcirculation is not as important in the development of PPMI than the resting status of the coronary microvasculature.

The fact that the total number of balloon inflations was predictive of troponin release following PCI is also an interesting concept and has not previously been reported. One hypothesis may be that increasing balloon/stent inflation may shower the distal microcirculation of the culprit vessel that if already impaired at baseline may be more susceptible to ischaemic insults resulting in periprocedural injury. In essence, the measurement of IMR pre-PCI and the rPIMR identify the ‘embolic potential’ of a lesion, and how balloon inflation/stent deployment may affect the distal microvasculature.

Clinical implications

We have shown that a high IMR pre-PCI and a high rPIMR are associated with the development of PPMI and independent predictors of PPMI. Clinically, the technique of measuring IMR may be used to identify patients at increased risk of developing PPMI. It can be easily performed during routine FFR assessment, so that once evaluation of the haemodynamic significance of a lesion has occurred, the risk of PPMI can also be evaluated. We have also shown that the total number of balloon inflations is also important in the development of PPMI. Prospective identification may allow escalation of therapy, such as the use of GPIIb/IIIa inhibitors or a direct stenting strategy, in an attempt to minimise the risk of periprocedural injury. Using IMR to identify the higher-risk patient and direct therapy requires validation in a larger cohort.

Study limitations

While we have shown important data concerning the prediction of PPMI, this is still a relatively small sample. However, given the complexity of the protocol, and in accordance with our power calculation, we feel that we have an adequate number of patients to ensure that our data are unlikely to reflect a type II statistical error.

While we have demonstrated the pre-PCI IMR is associated with troponin levels up to eight times the upper limit of normal, as there were limited numbers of patients with troponin elevations at the higher end of the spectrum, we have insufficient power to assess the ability of IMR to predict higher troponin levels. Recently, there has also been controversy within the literature as to the prognostic significance of a troponin three times the upper limit of normal, but a significant body of work linking a rise three times the upper limit of normal with worse outcomes.2 ,4


This is the first study to demonstrate that the pre-PCI IMR independently identifies periprocedural myocardial infarction in stable patients undergoing elective PCI, and that a novel ratio, the rPIMR, is also predictive of PPMI. Future studies in a larger cohort are required to establish the predictive ability of IMR in PPMI, and whether a targeted strategy aimed at reducing IMR pre-PCI leads to improved outcomes.


We would like to thank the nursing staff in the catheter laboratory and coronary care for their support


View Abstract


  • See Editorial, p 1471

  • Funding JJL is funded by an Australian National Health and Medical Research Council Postgraduate Scholarship.

  • Competing interests None.

  • Ethics approval Ethics approval was provided by Human Ethics Review Panel St Vincent's Hospital, Melbourne.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Linked Articles