PINOT NOIR: Pulmonic INsufficiency imprOvemenT with Nitric Oxide Inhalational Response

This was a single-center prospective study conducted at the Cleveland Clinic with approval from the Institutional Review Board and registered with the National Library of Medicine (NCT00543933). Subjects underwent appropriate consent with strict adherence to Health Insurance Portability and Accountability Act regulations. Consecutive individuals presenting for a clinically indicated CMR at the Cleveland Clinic with evidence of PI by echocardiography and a history suggestive of repaired TOF or PS were approached for enrollment. Patients who had a right ventricular-to-pulmonary artery conduit, underwent late pulmonary valve repair or replacement for PI, had residual shunt lesions, had residual pulmonary valvular/subvalvular (mean gradient >30 mmHg by echocardiography) or branch pulmonary stenosis, or evidence of pulmonary hypertension were excluded. Individuals were also excluded if contraindications to inhaled nitric oxide or CMR existed.

After consent was obtained, each individual had a brief assessment including collection of demographics, medical history and physical examination including vital signs and assessment of New York Heart Association (NYHA) functional class. Detailed medical history was taken directly from chart review. Elements of past surgical history including history of systemic to pulmonary shunts or other palliative shunts, type of repair performed, presence of right ventricular outflow tract (RVOT) aneurysm and presence of residual pulmonary stenosis were obtained.

CMR was performed at 1.5 Tesla (Magnetom Avanto; Siemens Medical Systems, Erlangen, Germany) with a multi-element phased-array surface coil. Studies were performed with patients in the supine position using a phased-array surface coil as a receiver and retrospective electrocardiographic gating. Images were obtained during end-expiratory breath holds preceded by brief hyperventilation. CMR images were downloaded to an off-line workstation (Leonardo; Siemens Medical Solutions) for analysis.

After obtaining standard localizer views and contiguous short-axis cine views covering both ventricles (encompassing the entire left and right ventricles from the base to the apex), two double oblique views oriented along the main axis of the pulmonary trunk and the aortic root were acquired with a standard steady-state free precession (SSFP) cine magnetic resonance sequence for the positioning of phase-contrast magnetic resonance images and to ensure that the imaging plane remained in the proper position throughout the whole cardiac cycle. Cine steady state free-precision image parameters were: TR (repetition time)/TE (echo time) 3.2 ms/1.6ms, flip angle (70°), an 8-mm section thickness with a gap of 2 mm (10–12 short axis images in total), matrix of 256 × 166 (typical), minimal field of view (typical), typical spatial resolution of 1.7 × 1.4 mm, acquired temporal resolution of 33–45 ms, 20–25 reconstructed cardiac phases (reconstructed by using view sharing when required), and a bandwidth of 900–1000 Hz/pixel. Endocardial contours in end systole and end diastole were manually traced to derive ventricular volumes using the disc-summation technique. The following measurements were then obtained for both the right and left ventricles: end diastolic volume, end systolic volume, stroke volume and ejection fraction.

Phase-contrast velocity-encoded CMR was acquired with a segmented fast gradient-echo sequence, with velocity encoding perpendicular to the imaging plane and a predefined upper velocity limit of 100 cm/s. If aliasing was noted, the velocity was progressively raised in 50-cm/s steps until aliasing disappeared. Imaging parameters included: TR/TE 7.5 ms/3.1 ms; flip angle, 15°; section thickness, 6 mm; field of view, 320–380 × 240–300 mm; matrix, 256 × 96 (typical in-plane resolution, 2.7 × 1.4 mm); number of signals acquired, one; temporal resolution, 35–45 ms; number of reconstructed cardiac phases, 25; bandwidth, 260 Hz/pixel. The typical breath-hold time for phase-contrast velocity-encoded sequences ranged from 15–20 seconds. If obvious breathing artifacts were noted, repeated acquisitions were attempted. Phase-contrast images were post-processed using commercially available software (Argus; Siemens Medical Solutions). The heart rate during CMR was recorded from the acquired images. The contours of the main pulmonary artery and aorta were automatically traced, with manual correction, simultaneously on magnitude and velocity-map images of all 25 reconstructed phases. Velocity in each of the voxels included within the contour was calculated and integrated over area and time to obtain the following parameters: average velocity, average flow, minimum area, maximum area, forward volume and reverse volume. Regurgitant fraction was calculated as reverse volume/forward volume. Tricuspid regurgitation was calculated as right ventricular stroke volume – pulmonary artery forward flow. Qp:Qs ratio was calculated as net forward pulmonary artery flow/net forward aortic flow.

The standard delivery method for iNO has been well described [12]. Medical grade nitric oxide gas (INOMAX®, IKARIA, Clinton, NJ) is typically delivered via the INOvent® delivery system (IKARIA, Clinton, NJ) from source tanks, and flushed through to the patient using supplemental O₂ flow at 5 L/min. The dose of iNO (ppm) is controlled by the INOvent regulator and confirmed by the electrochemical analyzer incorporated into the INOvent. Concentrations of NO (ppm), NO₂ (ppm) and O₂ (%) are continuously reported to the INOvent display from the analyzer. The magnetic field created by the CMR suite required modification to the standard delivery technique which has been previously described by our group [13].

Once baseline imaging was completed, iNO was continuously administered to the patient. After 5 minutes with the patient breathing inhaled nitric oxide to ensure steady state conditions, imaging was repeated. All volumetric measurements were indexed to body surface area using the DuBois method [14]. Two independent readers interpreted each study and the mean value of each measurement was used for analysis. The acquisition technique prevented blinding readers to the study sequence. A follow-up phone call was placed with each subject 1 day following the study to monitor for any adverse events.

Sample size was estimated from a CMR study measuring change in aortic regurgitant fraction in response to intravenous felodipine in 16 asymptomatic patients with aortic insufficiency [15]. Based on this study we anticipated that 22 patients would be required for a 90% power to detect a 6% decrease in regurgitation change with an estimated standard deviation of 6% using a significance of 0.05 accounting for a dropout rate of 40% in paired analysis. PASS 2005 (NCSS, 2005) was used for sample size determination.

Data are presented as mean ± standard deviation. Linear regression was used to test univariate associations. CMR volumetric and flow measurements were compared using paired t-tests to assess the effect of intervention between baseline measurements with repeated values obtained during the administration of iNO. Significance was assumed to be p<0.05 for all tests. All analyses were performed using JMP Pro 9 (SAS, 2010).

Twenty four patients were contacted regarding the study after initial screening (Figure 1). Two patients refused participation, 3 met exclusion criteria, 1 was unable to tolerate the CMR due to clostrophobia and 2 patients had incomplete CMR data available and were excluded from analysis. The final cohort included 16 patients (Table 1).

Half the cohort was female (n=8, 50%) and the median age [interquartile range] was 34.8 [23.4 – 40.8] years. The average body mass index of the cohort was 25.8 ± 5.4 kg/m² and most patients were in NYHA functional class I at the time of the study (n=12, 75%). Eleven patients reported intermittent symptoms (exercise intolerance, shortness of breath or palpitations) at the time of evaluation. Two patients had systemic hypertension and no patients were taking medications for or had symptoms of coronary artery disease.

Eleven patients (69%) had a history of TOF, while five (31%) had a history of isolated PS. The median age [interquartile range] for complete surgical repair among TOF and PS patients was 3.7 [1.8 – 4.6] years. Most patients had echocardiographic evidence of at least moderate PI (n=13, 87%) and all but one had mild tricuspid valve regurgitation or less (n=15, 93%). Mean estimated right ventricular systolic pressure by echocardiography was 35.8 ± 13.6 mmHg for the cohort. The mean QRS duration was 139.5 ± 28.2 ms on electrocardiography.

Volumetric results at baseline and with iNO are shown in Table 2. At baseline, right ventricular end diastolic volume index was 156.9 ± 32.7 mL/m², right ventricular end systolic volume index was 93.0 ± 19.6 mL/m², right ventricular stroke volume index was 63.9 ± 17.8 mL/m² and right ventricular ejection fraction was 40.5 ± 6.2%. All parameters remained unchanged during administration of iNO (p=0.95, p=0.39, p=0.21 and p=0.12 respectively).

At baseline, left ventricular end diastolic volume index was 83.0 ± 13.6 mL/m², left ventricular end systolic volume was 38.5 ± 8.7 mL/m², left ventricular stroke volume was 44.5 ± 7.7 mL/m² and ejection fraction was 53.8 ± 5.2%. All parameters remained unchanged during iNO administration (p=0.15, p=0.11, p=0.65 and p=0.63 respectively).

Pulmonary artery flow parameters are displayed in Table 3. Average velocity in the pulmonary artery was 15.3 ± 8.1 m/s at baseline and did not change with iNO (p=0.17). Average flow index was 2.7 ± 0.7 L/min/m² at baseline and did not change with iNO either (p=0.13). The forward volume across the pulmonary artery was similar before and after iNO administration (123.5 ± 26.8 vs. 122.7 ± 31.6 mL/beat/m², p=0.75) but regurgitant fraction (35.3 ± 16.4 to 33.3 ± 15.1%, p=0.02, Figure 2A) and reverse volume decreased with iNO (45.3 ± 25.0 vs. 42.4 ± 24.1 mL/beat/m², p=0.01, Figure 2B). Right ventricular end diastolic volume index appeared to be weakly correlated to the magnitude of PI (r²=0.22, p=0.07, Figure 3), but PI did not differ according to sex (p=0.34) or diagnosis (p=0.70).

No significant changes in aortic flow parameters were detected during administration of iNO (Table 3). Average velocity was 14.1 ± 6.2 m/s, average flow index was 2.9 ± 0.8 L/min/m², forward volume was 74.8 ± 17.2 mL/beat and reverse volume was 3.2 ± 4.4 mL/beat with no significant changes during iNO administration (p=0.10, p=0.71, p=0.052 and p=0.73, respectively). Aortic regurgitant fraction was minimal at baseline (4.5 ± 6.4) and did not change during iNO administration (p=0.51). Large % changes were observed in reverse volume and regurgitant volume with iNO administration but the absolute changes were minimal. Outliers in these fields were removed for improved data clarity.

The length of time from complete surgical repair to enrollment in the study did not appear to be associated with NYHA function class (r²=0.01, p=0.74), right ventricular end diastolic volume (r²=0.03, p=0.54), right ventricular ejection fraction (r²=0.18, p=0.10), QRS duration (r²=0.18, p=0.22) or the degree of pulmonary insufficiency (r²=0.13, p=0.17, Figure 4). Age at complete repair did not appear to be associated with these variables either. At baseline the mean Qp:Qs ratio was 1.13 and mean tricuspid regurgitation was 6.7%.