The coronary sinus reducer implantation for ischemia reduction (CrossRoad) study (ClinicalTrials.gov identifier NCT04121845) was a randomized, single-center, double-blind, sham-controlled study that enrolled eligible patients who underwent treatment with CSR at UMC Ljubljana between 1st January 2019 and 31st December 2021. All patients had symptomatic angina for more than 3 months and were classified in class II–IV according to the Canadian Cardiovascular Society (CCS). Patients had to be treated with optimal medical therapy for at least one month and had reversible ischemia in the anterior, lateral, and inferolateral left ventricular walls confirmed by single photon emission tomography (SPECT). Patients with unstable angina within the last 30 days, acute myocardial infarction within the last 90 days, recent successful revascularization, decompensated heart failure, and severe valvular heart disease were excluded from the study. The study was approved by the national ethics committee. Before the inclusion, all patients signed the informed written consent to participate in the study.

Patients were randomized to either CSR implantation or a sham procedure. The CSR implantation technique is already described elsewhere [7, 8]. Following the right internal jugular vein puncture, the right atrial pressure was measured, followed by cannulation and venography of the coronary sinus. CSR was implanted in the distal coronary sinus with special care not to obstruct any greater tributary and with 10% oversizing to ensure stability. During CSR balloon inflation, occlusion pressure of the coronary sinus proximal to the CSR was measured. After final CSR positioning in the distal coronary sinus, the CSR balloon catheter was connected to the pressure transducer via incompressible plastic tubing. The pressure transducer was positioned at the level of the phlebostatic axis, and the reference point was set to the atmospheric pressure. During the inflation of the CSR balloon, the pressure waveform was recorded. The occlusion pressure was determined as a peak pressure recorded during systole (Fig. 1). Final venography confirmed the appropriate position of the device. A sham procedure was performed by the same experienced operator and in the same catheterization laboratory. Cannulation of the right internal jugular vein was followed by right atrial pressure measurement. The time of the procedure was similar to the CSR implantation. Both procedures were performed in hearing isolation. All medical personnel, apart from those performing the procedure, were blinded to the patient allocation.

Fig. 1.

Coronary sinus venography and occlusion pressure measurement. Left: Venography during coronary sinus reducer (CSR) implantation, showing the final position of the CSR in the distal coronary sinus. The CSR catheter balloon is still inflated. Right: Pressure tracing in the proximal coronary sinus during the balloon inflation. Systolic pressure was measured during this phase of the procedure to reduce heterogeneity associated with an extensive network of Thebesian veins.

Five-minute 12-lead high-resolution electrocardiogram (hrECG) (Cardiax computer ECG, IMED KFT., Budapest, Hungary) with a sampling rate of 1 kHz and 300 Hz low pass filter was recorded at baseline and after 6 months. The researcher who analyzed hrECG data was unaware of the patient allocation. Premature ventricular complexes and complexes with severe artifacts were removed from further analysis. The remaining QRS-T complexes were averaged to calculate median beats. Orthogonal leads were calculated using the Kors regression transformation method. T-peak and T-end interval (TpTe) was measured from vector signal and was defined as the time difference between the T wave’s peak amplitude and the T wave’s global end. The spatial angle between the QRS and T axis (QRS-T angle) was defined as the angle between the ventricular depolarization (QRS) and repolarization (T wave) vectors using maximal amplitudes of the QRS and T vectors, reported as QRSTP angle, and using mean amplitudes of the QRS and T vectors reported as QRSTM angle. QT interval for QT variability (QTV) calculations was normalized for heart rate ($\frac{dQT}{dRR}$) and measured from the constructed vector signal. QTV was reported as the QT variability index:

$QTVi=log\frac{QTVN}{HRVN}$, where $QTVN=\frac{SDQ{T}^2}{QTmea{n}^2}$ and $HRVN=\frac{HRVAR}{HRmea{n}^2}$ and as QT variability ratio:

$$VR=\frac{STVQT}{STVRR},whereSTVQT=\sum \frac{\left|QTn+1-QT\right|}{N\sqrt{2}}$$

[9].

Spatial ventricular gradient (SVG) amplitude was calculated as an integral of ECG voltages over the entire QRS-T complex:

$$SVG=[{\int }_{Qbeg}^{Tend}Vx\left(t\right)dt,{\int }_{Qbeg}^{Tend}Vy\left(t\right)dt,{\int }_{Qbeg}^{Tend}Vz\left(t\right)dt,]$$

obtained from all three axes of orthogonal ECG [10]. Heart rate variability parameters were calculated with power spectral density analysis using a Lomb-Scargle periodogram [11]. High, low, and total frequency powers are reported. As a measure of heart rate variability, we also included a time domain parameter reported as the standard deviation of the normal-to-normal intervals (SDNN).

Randomization was performed using block randomization with a block size of 4 generated by the online statistical software (http://www.jerrydallal.com/random/randomize.htm, visited on 6th May 2018). Categorical variables are presented as frequencies and percentages, and continuous variables as mean $\pm$ standard deviation (SD) or median and interquartile range. Categorical variables were compared using Chi-square or Fisher exact tests. The normality of distribution for continuous variables was evaluated by the Kolmogorov-Smirnov test. Intra- and intergroup comparisons were performed using paired or independent sample t-test, Wilcoxon signed-rank, and rank-sum test as appropriate. Statistical analysis was performed in SPSS Statistics, version 22.0 (IBM Corp., Armonk, NY, USA). A two-sided p-value of 0.05 was considered statistically significant.

A relatively small number of included patients due to the single-center design and a limited number of patients eligible for this treatment limits the strength of our findings. The study included eligible patients enrolled in the Crossroad study. Due to scarce data in the literature, prior calculation of the sample size was not possible. As it was a clinical study, the assessment of arrhythmic properties was limited to noninvasive analysis of hrECG parameters at rest, which may have underestimated the arrhythmic changes that might be evident with invasive testing or during exercise. While we measured the occlusion pressure during CSR implantation, we could not correlate ECG parameters to coronary sinus pressure during the recording.