Diagnostic relevance of angiography-derived coronary microcirculatory resistance in sudden cardiac death

Abstract

Background

Obstructive coronary disease remains a leading cause of sudden cardiac death (SCD). The management of SCD therefore involves coronary angiography. Our aim was to evaluate whether the non-hyperemic angiography-derived microcirculatory resistance index (NH-IMRangio) could be an easy-to-use tool for identifying patients with electrical heart disease (EHD) from patients with other causes of SCD.

Methods

A retrospective study was carried out on 30 patients who survived from SCD with no significant coronary lesions on coronary angiography. Etiological investigations enabled the classification of patients with myocardial disease (Group 1, n=20) and those with EHD without myocardial disease (Group 2, n=10). Myocardial disease was investigated by cardiac magnetic resonance imaging (CMR). NH-IMRangio was determined based on standard coronary angiographic views with 3-dimensional-modeling and computational analysis of the coronary flow.

Results

Patients were 46.4±15.9 years old and mostly male (73%). Group 1 included patients with dilated cardiomyopathy (n=7), non-dilated left ventricular cardiomyopathy (n=6), hypertrophic cardiomyopathy (n=2), arrhythmogenic right ventricular cardiomyopathy (n=1), myocardial infarction with non-obstructive coronary arteries (MINOCA) disease due to vasospastic angina (n=1), myocarditis (n=2), chemotherapy-induced cardiomyopathy (n=1). Group 1 presented a significantly higher NH-IMR angio compared to group 2 (46.5±13.1 vs. 34.1±10.8, P<0.02). An NH-IMR angio cut-off of 41.5 enabled an optimal classification of patients with or without myocardial disease.

Conclusions

A high NH-IMRangio could represent a useful tool for guiding the etiological diagnosis of SCD towards myocardial disease rather than EHD.

Introduction

Obstructive coronary artery disease remains a leading cause of sudden cardiac death (SCD), and immediate coronary angiography is therefore recommended in resuscitated cardiac arrest to exclude acute coronary occlusion (1). However, in a significant proportion of cases, angiography does not reveal obstructive lesions. In this context, other mechanisms such as epicardial coronary spasm, coronary microvascular dysfunction (CMD), or underlying structural myocardial disease must be considered (2).

In recent years, advances in computational fluid dynamics (CFD) have enabled the derivation of physiological indexes from angiography. Among them, the non-hyperemic angiography-derived index of microcirculatory resistance (NH-IMRangio) provides an estimate of microvascular function without the need for hyperemic agents (3-5). Importantly, CMD has been linked to abnormal findings on cardiac magnetic resonance (CMR), particularly in patients presenting with myocardial infarction with non-obstructive coronary arteries (MINOCA) (6).

Beyond ischemic mechanisms, structural cardiomyopathies—including dilated, hypertrophic, or infiltrative forms—represent another major cause of SCD in patients without obstructive coronary disease. Identifying these patients early after cardiac arrest is critical, as their prognosis, recurrence risk, and management strategies differ substantially from those with purely electrical heart disease (EHD).

On this basis, we hypothesized that the assessment of NH-IMRangio during post-cardiac arrest coronary angiography could help discriminate between patients with underlying myocardial disease and those with primary electrical disorders in the absence of myocardial involvement. We present this article in accordance with the STARD reporting checklist (available at https://qims.amegroups.com/article/view/10.21037/qims-2025-1601/rc).

Methods

Population

Patients were identified from the institutional cardiac arrest database of the University Hospital of Grenoble (CHU Grenoble Alpes) and cross-checked with electronic medical records. The patients had no obstructive coronary stenosis upon coronary angiography. An SCD workup was performed including a coronary pharmacological test (intravenous methylergometrine), a rhythmological work-up [ajmaline test, adrenaline test, electrophysiological test (EPS) and programmed ventricular stimulation (PVS)] and cardiac magnetic resonance imaging (CMR). Data were collected from an internal listing of patients with SCD and complemented with information extracted from the patients’ medical records. Patients were then classified into two groups:

• ❖ Group 1: SCD due to myocardial disease (cardiomyopathies, MINOCA, myocarditis);

• ❖ Group 2: EHD without myocardial disease (ventricular fibrillation in structurally normal heart, long QT syndrome, Brugada syndrome, catecholaminergic polymorphic ventricular tachycardia, the short QT syndrome and Severe conduction disturbance followed by torsade de pointes and subsequent ventricular fibrillation.

The study was registered at ClinicalTrials.gov (Identifier: NCT03479580). The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by West I Committee for the Protection of Persons (No. 2017T3-14) and informed consent was taken from all the patients.

NH-IMRangio

In cases where systolic blood pressure exceeded 90 mmHg, intracoronary nitroglycerin was administered prior to NH-IMRangio assessment. Coronary angiography was conducted using a high-speed contrast injection protocol spanning three cardiac cycles, with image acquisition at 15 frames per second, specifically targeting the left anterior descending artery (LAD). For each patient, the optimal end-diastolic frame—ensuring adequate contrast opacification and minimal vessel overlap—was selected from two angiographic projections separated by at least 25°. NH-IMRangio computation was carried out using QAngio XA 3D software (QFR^®, v2.1 Research Edition, Medis Medical Imaging, Leiden, the Netherlands), in accordance with established methodology (7-9). The analysis was performed blinded to clinical data. The protocol included: (I) identification of two suitable image sequences with sufficient angular separation; (II) selection of end-diastolic frames filled with contrast; (III) anatomical landmark registration for geometric correction; (IV) lumen segmentation in both views; (V) manual calculation of the TIMI frame count and input of mean resting aortic pressure (Pa); and (VI) three-dimensional (3D) reconstruction with automated computation of both QFR and NH-IMRangio.

The NH-IMRangio index was derived using the formula:

$$\text{NH-IMRangio}=\text{Pa}\left(\text{resting}\right)\times \text{QFR}\times \frac{\text{Nframes}\left(\text{resting}\right)}{\text{Fps}}$$ [1]

For each analysis, the portion of the artery considered for the QFR measurement was used to establish the vessel length. The number of frames (Nframes) corresponded to the time interval required for the contrast agent to progress along this segment, while the acquisition speed (fps) was standardized at 15 images per second.

All image interpretations were performed without knowledge of patient outcomes. Two experienced operators (B.S. and M.A) carried out the measurements independently, using a dedicated software package validated and provided by the manufacturer.

CMR

CMR examinations were performed using either 1.5T or 3T systems (Magnetom Aera or Skyra, Siemens Healthineers, Munich, Germany). The imaging protocol was adapted according to the clinical presentation and generally comprised cine steady-state free-precession sequences, T2-weighted STIR for edema detection, T2 mapping, perfusion imaging during the first contrast pass, and late gadolinium enhancement (LGE) for fibrosis assessment. Short-axis cine slices (10 to 12) were obtained to cover the entire left ventricle (LV) from base to apex, and long-axis views (2-, 3-, and 4-chamber) were acquired to evaluate myocardial morphology and function. For LGE analysis, a dose of 0.2 mmol/kg of a gadolinium-based contrast agent was administered intravenously. After a delay of approximately 8 to 10 minutes, the same imaging planes were repeated using a two- or three-dimensional inversion recovery sequence optimized for contrast. Quantitative post-processing was carried out with the QMass module (Medis, Leiden, Netherlands). LV and left atrial volumes, as well as myocardial mass, were calculated. Visual interpretation of LGE was performed by an expert observer (GB), blinded to the angiographic results and certified at level 3 (EACVI). LGE was considered present when focal hyperenhancement appeared consistently in short- and corresponding long-axis views.

Classification of myocardial involvement

Myocardial involvement was adjudicated by a multidisciplinary expert board (clinical cardiology, cardiac imaging, and electrophysiology), integrating cardiac MRI interpreted by experienced readers, clinical presentation, biomarkers, and ECG/echocardiographic findings. In the EHD group, cardiac MRI was performed early whenever possible to maximize sensitivity for transient myocardial edema. Disagreements were resolved by consensus.

Statistical analysis

Categorical variables were described using absolute counts and proportions. Comparisons between groups were performed using either the Chi-squared test or Fisher’s exact test, depending on expected frequencies. For continuous variables, data were summarized as mean ± standard deviation when normally distributed, or as median and interquartile range (25th–75th percentile) when distributional assumptions were not met. Group differences were evaluated using the Mann-Whitney U test or Kruskal-Wallis test, as appropriate. Receiver operating characteristic (ROC) analysis was conducted to evaluate the predictive performance of NH-IMRangio for the presence of LGE. The Youden index was applied to determine the optimal threshold value. To assess reproducibility, intra- and interobserver variability analyses were conducted using measurements from 10 patients randomly selected from the cohort. For intraobserver reproducibility, the same operator repeated the analysis; for interobserver agreement, two independent operators analyzed the same cases. Variability was quantified using intraclass correlation coefficients (ICC). Interobserver agreement was estimated with a two-way random-effects model (single-measure ICC), while intraobserver and intrasubject reliability were evaluated using a one-way random-effects model (average-measure ICC). All statistical analyses were performed using SPSS software (version 26.0, IBM Corp., Armonk, NY, USA), and graphical representations were generated using GraphPad Prism (version 8.0, GraphPad Software, San Diego, CA, USA). A two-tailed P value below 0.05 was considered statistically significant.

Results

Study population

The 30 patients with SCD (Group 1, n=20 and group 2, n=10) included in the study were on average 46.4±15.9 years old and in 73% of cases male. The pathologic mechanisms of SCD in groups are indicated in Table 1. CMR was performed on average 8 days (2–27 days) following SCD. Group 1 patients presented a significantly lower left ventricular ejection fraction compared with the group 2 (44.4%±13.6% vs. 55.5%±4.19%, P=0.03). Clinical data stratified by group are reported in Table 2.

Table 1. Cause of sudden cardiac death.

Group	Number
Group 1: myocardial disease	20
Dilated cardiomyopathy	7
Non-dilated left ventricular cardiomyopathy	6
Hypertrophic cardiomyopathy	2
Arrhythmogenic right ventricular cardiomyopathy	1
MINOCA due to vasospastic angina	1
Myocarditis	2
Chemotherapy-induced cardiomyopathy	1
Group 2: electrical heart disease	10
Ventricular fibrillation in structurally normal heart	8
Long QT syndrome	1
Atrioventricular block followed by torsades de pointes	1

MINOCA, myocardial infarction with non-obstructive coronary arteries.

Table 2. Baseline characteristics of patients.

Characteristics	All (N=30)	Myocardial disease (N=20)	Electrical heart disease (N=10)	P value
Age (years)	46.4±15.9	45.0±14.9	49.4±18.1	0.48
Male	22 [73]	15 [75]	7 [70]	0.77
Active smokers	5 [17]	3 [15]	2 [20]	0.55
Hypertension	9 [30]	5 [25]	4 [40]	0.43
Dyslipidemia	3 [10]	1 [5]	2 [20]	0.25
Diabetes	1 [3]	1 [5]	0 [0]	0.66
Familial history of CAD	1 [3]	1 [5]	0 [0]	0.66
CMR
LVEF	48.8±12.2	44.4±13.6	55.5±4.19	0.03
EDVi (mL/m2)	107.1±29.6	114.4±33.3	92.5±10.7	0.05
RVEF	52.3±9.28	50.9±10.6	55.0±5.09	0.26
LA volume	44.2±15.5	44.7±15.0	43.2±17.2	0.80
Presence of LGE	19 [63]	19 [95]	0 [0]	<0.0001
NH-IMRangio	42.3±13.6	46.5±13.1	34.1±10.8	0.02

Data are presented as mean ± standard deviation or number [%]. CAD, coronary artery disease; CMR, cardiac magnetic resonance imaging; LA, left atrial; LGE, late gadolinium enhancement; LVEF, left ventricle ejection fraction; NH-IMRangio, non-hyperemic angiography-derived index of microcirculatory resistance; RVEF, right ventricle ejection fraction.

Angiography derived IMR

The results are presented in Figure 1. Group 1 presented a significantly higher NH-IMR angio compared to group 2 (46.5±13.1 vs. 34.1±10.8, P<0.02). With regards to diagnostic performance, the ROC curve presented in Figure 2 indicated an area under the curve of 0.75 [95% confidence interval (CI): 0.58–0.92], corresponding to a parameter with relatively high discriminatory potential. The NH-IMRangio cut-off value for optimal classification of patients with or without myocardial disease was 41.5. Figure 3 displays a representative example of a patient with SCD with an elevated NH-IMRangio and a final diagnosis of non-dilated left ventricular cardiomyopathy (NDLVC) with CMR showing lateral fibrosis without acute inflammation.

Figure 1.

Box plot showing the significantly different distributions of NH-IMRangio in group 1 (myocardial disease) and group 2 patients (electrical heart disease). NH-IMRangio, non-hyperemic angiography-derived index of microcirculatory resistance.

Figure 2.

Receiver-operating characteristic curve of NH-IMRangio for the diagnosis of SCD due to myocardial disease. AUC, area under the curve; NH-IMRangio, non-hyperemic angiography-derived index of microcirculatory resistance; SCD, sudden cardiac death.

Figure 3.

Representative example of a patient with sudden cardiac death. Non-hyperemic angiography-derived microcirculatory resistance index was measured at 48 h by dedicated software analysis of coronary angiography images (A). The cardiac magnetic resonance short-axis view showed an abnormal image of late gadolinium enhancement in the lateral part of the myocardium (white arrow) with diagnosis final of non-dilated left ventricular cardiomyopathy (B). CRA, XXX; MLD, XXX; QFR, XXX; RAO, XXX.

Reproducibility

NH-IMRangio measurements demonstrated strong consistency, as reflected by intra-observer ICC of 0.91 (95% CI: 0.83–0.96) and inter-observer ICC of 0.90 (95% CI: 0.81–0.95).

Discussion

To the best of our knowledge, our study is the first to evaluate the potential of NH-IMRangio for guiding the etiological diagnosis of SCD towards myocardial disease rather than EHD. The main finding was that patients with SCD secondary to EHD had a significantly lower NH-IMRangio than patients in whom the cause of SCD involved myocardial disease. This suggests that angiography-derived microcirculatory indices may reflect distinct underlying mechanisms of SCD.

Under physiological conditions, the coronary microcirculation maintains stable myocardial perfusion through finely tuned autoregulatory mechanisms occurring at different levels: pre-arterioles respond mainly to variations in flow and pressure, while arterioles regulate vascular tone according to local metabolic demand. Capillaries and venules, with low resistance and high capacitance, play a critical role in oxygen delivery to myocardial tissue (10). Myocardial ischemia may thus result from structural alterations—such as lumen narrowing, microthrombi, capillary rarefaction, external compression by myocardial edema and interstitial fibrosis, or vessel infiltration—or from functional impairment, including impaired vasodilatation or excessive vasoconstriction such as microvascular spasm (11).

Four distinct forms of CMD have been described: primary dysfunction without epicardial disease, dysfunction associated with myocardial disease, dysfunction coexisting with obstructive coronary artery disease, and iatrogenic dysfunction (11).

In our cohort, and particularly in Group 1, CMD related to underlying myocardial involvement appeared the most relevant. Although this group included heterogeneous entities, all were characterized by microcirculatory impairment. This applies to MINOCA (6), as well as to several cardiomyopathies. CMD is observed in tako-tsubo syndrome following catecholaminergic surges and inflammation (12), in hypertrophic cardiomyopathy where it has prognostic value (13,14), and in dilated cardiomyopathy where capillary rarefaction plays a major role (15,16). Across these diseases, CMD has been consistently linked with adverse prognosis, reinforcing the hypothesis that it could contribute to mechanisms leading to SCD, in addition to the established roles of ischemia, inflammation, and fibrosis (17). CMD may also promote electrical instability by increasing ischemia-related heterogeneity of repolarization, as suggested by increased QT dispersion (18). By contrast, no role of CMD has been reported in EHD to date (11).

Wire-derived IMR is the standard tool for assessing CMD (19), especially in acute coronary syndromes (20). However, its use requires hyperemia and pressure wires, which add procedural complexity and costs. To overcome these limitations, angiography-derived IMR using computational flow has been developed as a reliable alternative, avoiding both hyperemia and additional equipment (3-5). Its diagnostic role has been demonstrated in MINOCA, particularly in subtypes of coronary origin, tako-tsubo, and infiltrative or inflammatory cardiomyopathy (6). Prognostic associations have also been reported in tako-tsubo syndrome, HCM, and DCM (21-24).

In the specific context of cardiac arrest, rapid neurological protection is a priority, and patients are sedated and often on vasopressors. This makes hyperemia-based indices such as thermodilution-derived IMR challenging. Angiography-derived indices extrapolating hyperemic conditions (such as angioIMR) have been validated in obstructive CAD (25,26), but not in conditions like MINOCA, myocarditis, or cardiomyopathies. In addition, vasopressor support further limits their reliability. For these reasons, we considered a resting index such as NH-IMRangio more appropriate in this clinical setting.

Perspectives

The present study suggests that coronary angiography performed in the setting of SCD management may provide etiological insights through NH-IMRangio determination, which may in turn suggest the need for CMR. In the case of low NH-IMRangio, a more rhythmologically oriented work-up may be preferable. Because CMR availability may be limited in some centers, a high NH-IMRangio could help prioritize further diagnostic investigations. Improvements in software for measuring angiography-derived IMR are likely to improve reproducibility and ease of use in the catheterization laboratory during the management of SCD. The advantage is that such indices are obtained from conventional images with no requirement for specific equipment.

While pathology-specific studies are still needed to define diagnostic and prognostic thresholds, our study provides initial evidence supporting the role of NH-IMRangio to differentiate entities marked by the presence or absence of myocardial disease.

However, due to the specific post-SCD context, direct comparison of NH-IMRangio values with thresholds validated in other clinical settings (such as stable CAD or STEMI) is difficult, as all patients were sedated and under catecholaminergic support, which can transiently increase coronary microvascular resistance. Consequently, the absolute NH-IMRangio values are globally higher, and the comparison between groups appears more relevant than adherence to fixed cut-offs. Future studies will be required to establish population-specific thresholds for post-SCD patients.

Limitations

Our study has several limitations. First, it is a retrospective analysis with a relatively small sample size. Although all patients underwent high-quality coronary angiography and cardiac MRI, the diagnostic work-up following sudden cardiac arrest was not fully standardized, and vasomotor or rhythmologic provocative testing was adapted to clinical context. Only patients with a confirmed final etiological classification were included, which may induce a selection bias.

Misclassification of myocardial involvement remains possible, particularly in myocarditis without fibrosis or in Takotsubo syndrome when MRI is performed after edema resolution. However, several elements mitigate this risk. Most patients in the EHD group (8/10) underwent MRI within the first 7 days, which substantially reduces the likelihood of missing transient edema. Furthermore, classification relied on multidisciplinary expert consensus. Importantly, any residual misclassification would bias the comparison toward the null, since patients with transient myocardial injury typically present higher NH-IMRangio values. Therefore, the between-group difference observed here is likely a conservative estimate rather than an overestimation.

Finally, NH-IMRangio measurements were performed off-line by experienced operators. Broader implementation would require training and real-time analysis workflows. Prospective studies are needed to confirm these findings and define their clinical integration.

Conclusions

In conclusion, our study suggests that NH-IMRangio could help orient the etiological diagnosis of SCD by distinguishing between patients with myocardial disease and those with EHD. This distinction is clinically important, as it may guide subsequent investigations: a high NH-IMRangio could prioritize CMR to explore structural disease, while a low NH-IMRangio may support a rhythmological work-up.

Because NH-IMRangio is derived from conventional angiographic images, it can be obtained quickly and without additional equipment, making it particularly suitable in the acute setting of cardiac arrest. Although our study is limited by its retrospective design, small sample size, and heterogeneity of causes in the myocardial group, it provides hypothesis-generating evidence supporting the role of microvascular indices in SCD evaluation.

Future prospective, multicenter studies are warranted to validate these findings, establish standardized diagnostic thresholds, and assess the prognostic implications of NH-IMRangio in the context of SCD.

Supplementary

The article’s supplementary files as

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. It was approved by West I Committee for the Protection of Persons (No. 2017T3-14) and informed consent was taken from all the patients.

Data Sharing Statement

Available at https://qims.amegroups.com/article/view/10.21037/qims-2025-1601/dss

qims-16-03-234-dss.pdf ^{(68.5KB, pdf)}

DOI: 10.21037/qims-2025-1601