- Email: [email protected]
Evaluation of Right Ventricular Performance With a Right Ventricular Ejection Fraction Thermodilution Catheter and MRI in Patients With Pulmonary Hypertension
Evaluation of Right Ventricular Performance With a Right Ventricular Ejection Fraction Thermodilution Catheter and MRI in Patients With Pulmonary Hypertension* Marius M. Hoeper, MD; Joern Tongers; Andreas Leppert, MD; Stefan Baus, MD; Roman Maier; and Joachim Lotz, MD
Study objectives: We sought to compare catheter studies using a right ventricular ejection fraction (REF) catheter together with echocardiography and MRI in patients with pulmonary hypertension. Patients and methods: We compared hemodynamic findings, echocardiography, and MRI studies in 16 patients with pulmonary hypertension. Six healthy volunteers served as control subjects for the MRI studies. Results: MRI imaging provided accurate assessment of cardiac output in all but two patients. As compared with MRI, the REF catheter constantly underestimated the REF and overestimated right ventricular volumes in patients with pulmonary hypertension. REF, end-systolic and end-diastolic right ventricular volumes, and right ventricular muscle mass, as determined by MRI, were almost identical in patients with preserved cardiac function and those with low-output failure. The only factor that was different in both groups was the severity of tricuspid regurgitation. Conclusion: Right ventricular dimensions and muscle mass do not differ in patients with pulmonary hypertension who have low cardiac output and those who do not. According to our results, the major determinant of cardiac output in these patients appears to be the severity of tricuspid regurgitation. The REF catheter provides invalid data on right ventricular dimensions in patients with pulmonary hypertension. (CHEST 2001; 120:502–507) Key words: hypertension; MRI; pulmonary; right-heart failure; thermodilution Abbreviations: REF ⫽ right ventricular ejection fraction; RVEDV ⫽ right ventricular end-diastolic volume; RVESV ⫽ right ventricular end-systolic volume; SV ⫽ stroke volume
capacity of the right ventricle is a T hemajorfunctional prognostic determinant in pulmonary hypertension. It is unknown why some patients with markedly elevated pulmonary artery pressures maintain well-preserved cardiac function for several years, while others with equally or less severe pulmonary hypertension show rapidly progressive rightheart failure. One factor that has hindered the understanding of right ventricular performance in patients with pulmonary hypertension has been the *From the Departments of Pulmonary Medicine (Dr. Hoeper, Mr. Tongers, and Mr. Maier) and Diagnostic Radiology (Drs. Lotz, Baus, and Leppert), Hannover Medical School, Hannover, Germany. Manuscript received July 13, 2000; revision accepted March 2, 2001. Correspondence to: Marius M. Hoeper, MD, Department of Pulmonary Medicine, Hannover Medical School, 30623 Hannover, Germany; e-mail: [email protected] 502
lack of techniques that give a reliable picture of right ventricular remodeling. Recently, MRI has been shown to provide adequate information about right ventricular dimensions and muscle mass, and MRI may become the technique of choice to investigate right ventricular remodeling in patients with pulmonary hypertension.1,2 Information about right ventricular dimensions can also be derived from a right ventricular ejection fraction (REF) catheter, a modified thermodilution catheter. This catheter has been developed to estimate REF and right ventricular volumes based on a rapid beat-to-beat thermodilution technique, and has been found useful in patients with right heart dysfunction.3,4 We have shown5 that this catheter allows reliable assessment of cardiac output in patients with pulmonary hypertension. However, whether the REF catheter also provides valid data on Clinical Investigations
right ventricular dimensions and ejection fractions in this group of patients has not been investigated. To address this question, we compared hemodynamic data obtained with a REF catheter with the results from MRI imaging in patients with severe pulmonary hypertension. We also evaluated the relationship between hemodynamic findings and results from MRI studies and echocardiography to obtain information on right ventricular remodeling in patients with and without right ventricular failure. Materials and Methods Patients We studied 16 patients with primary (n ⫽ 14) or thromboembolic (n ⫽ 2) pulmonary hypertension (Table 1). The patients were in stable clinical condition, and medications were not changed between catheter studies and MRI examinations. In all patients, catheter examinations were done for diagnostic or therapeutic reasons unrelated to this study. The protocol was approved by the institutional review board, and all patients gave informed consent prior to the study. Control Population Magnetic resonance examinations were performed in six healthy volunteers (four women and two men; age range, 27 to 46 years) to assess right ventricular volumes, muscle mass, and flow in the pulmonary artery.
Table 1—Clinical and Hemodynamic Characteristics of the Patients* Characteristics Patients, No. Age, yr Range Male/female gender, No. Functional class, No. NYHA II NYHA III NYHA IV Hemodynamic variables (right-heart catheterization) PAP mean, mm Hg Range CO, L/min Range CI, L/min/m2 Range SV, mL Range PVR, dyne 䡠 s 䡠 cm⫺5 Range RA, mm Hg Range PCWP, mm Hg Range
Data 16 39 ⫾ 13 28–61 6/10 1 11 4
57 ⫾ 12 35–74 3.5 ⫾ 1.3 2.0–7.0 2.0 ⫾ 0.6 1.3–3.2 44 ⫾ 23 22–74 1,222 ⫾ 482 389–2,041 9.0 ⫾ 6.2 2–20 8.0 ⫾ 2.2 3–12
*Data are presented as mean ⫾ SD unless otherwise indicated. NYHA ⫽ New York Heart Association; PAP mean ⫽ mean pulmonary arterial pressure; CO ⫽ cardiac output; CI ⫽ cardiac index; PVR ⫽ pulmonary vascular resistance; RA ⫽ right atrial pressure; PCWP ⫽ pulmonary capillary wedge pressure.
Right-Heart Catheterization A 7.5F quadruple-lumen, balloon-tipped, flow-directed SwanGanz Catheter (No. 93A-434H7.5F; Baxter Edwards; Irvine, CA) was advanced through an 8F-introducer sheath via the right or left internal jugular vein. In order to ensure that the right atrial opening of the catheter was located directly before the tricuspid valve, the catheter was advanced into the pulmonary artery until the pressure recording through the proximal lumen revealed a ventricular pressure curve. Then, the catheter was slowly retracted until an atrial pressure recording was obtained. After insertion of the catheter, the patients were allowed to rest for at least 15 min before hemodynamic measures were recorded. The cardiac output was measured by the thermodilution technique with 10 mL of sterile, ice-cold, isotonic (0.9%) saline solution that was injected through the proximal (right atrial) lumen of the catheter, and the drop in temperature at the distal thermistor was recorded. Cardiac output was calculated using an analog computer system (REF-1; Baxter Edwards). Echocardiography Echocardiography was performed to exclude left-heart disease and intracardiac shunting and to assess the presence of tricuspid insufficiency. The echocardiographic examinations took place 1 to 3 days prior to the catheter studies. The severity of tricuspid regurgitation was graded as absent (grade 0), mild (grade I, jet area ⬍ 20% of the right atrial area), moderate (grade II, jet area between 20% and 33% of the right atrial area), and severe (grade III, jet area ⬎ 33% of the right atrial area), according to established grading systems.6 MRI Magnetic resonance studies were done within 24 h before (n ⫽ 5) or after (n ⫽ 11) the catheter studies. The radiologists were unaware of the catheter findings, and the pulmonologists were unaware of the MRI results. Volunteers and patients underwent ECG-gated MRI imaging (1.5 T MR Scanner; Signa Horizon EchoSpeed; General Electric; Milwaukee, WI). For determination of cardiac volumes, 8 to 10 slices along the short heart axis were obtained using a k space segmented, fast gradient-echo sequence (FastCard; GE-Medical Systems; Milwaukee, WI). Parameters included ECG gating, 8-mm slice thickness, 5-ms echo time, 11-ms repetition time, 20° flip angle, 32-cm field of view, and a matrix size of 256 ⫻ 160. Depending on the heart rate, an average of 23 phases per cardiac cycle was obtained. Flow measurements were obtained using a single double-oblique slice perpendicular to flow in the main pulmonary artery. In this setting, parameters used were ECG gating, 8-mm slice thickness, 6-ms echo time, 33-ms repetition time, 20° flip, 24-cm field of view, 200-cm/s velocity encoding, 32 phases per cardiac cycle, and a matrix size of 256 ⫻ 128. Data Processing Right ventricular volumes were assessed at a separate workstation (Sun SPARC 20; Advantage Windows 1.2; CAP Software; General Electric, WI). For each slice, right ventricular enddiastolic volume (RVEDV) and right ventricular end-systolic volume (RVESV) were calculated by semiautomatic outlining of the endocardium. Right ventricular volumes and ejection fractions were automatically calculated by the software. Ventricular muscle mass was measured by manual outlining of the epicardial and endocardial border of the right ventricle in each slice. The ventricular septum was defined as part of the left CHEST / 120 / 2 / AUGUST, 2001
503
ventricle. All volumes covering the right myocardium were added, and muscle mass was calculated by multiplying the muscle volume by the specific gravity of myocardium (1.05 g/cm). Flow measurements in the pulmonary artery were analyzed at the MRI console (image flow analysis software). For this purpose, the vessel was outlined in each of the 32 phases. From these areas, the software calculated minimum, maximum, and average flow rates, and, utilizing the vessel diameter, average flow rates per minute. Statistics All results are given as mean ⫾ SD. The Mann-Whitney U test was used to compare the MRI findings between the control population and patients with pulmonary hypertension, and to compare hemodynamic findings in patients with normal or low cardiac output. Variables obtained by MRI and with the REF catheter were compared by Wilcoxon’s signed-rank test. Fisher’s Exact Test was used to compare the prevalence of severe tricuspid regurgitation in patients with low or normal cardiac output. To address the question as to whether there were significant differences in right ventricular dimensions in patients with different degrees of tricuspid regurgitation, the KruskalWallis test was used. Linear regression analysis and Bland-Altman plots were used to compare catheter findings and MRI results.7 All p values ⬍ 0.05 were considered statistically significant.
mean difference between thermodilution and flow in the pulmonary artery as assessed by MRI was 0.1 ⫾ 1.0 L/min. In all but two cases, there was generally an acceptable overall agreement between these two techniques (Fig 1). In contrast to flow measurement in the pulmonary artery, volume measurements for determining cardiac output by MRI (that is, measuring stroke volume [SV] as the difference between RVEDV and RVESV, and calculating cardiac output as the product of SV and heart rate) constantly overestimated the cardiac output in patients with pulmonary hypertension. Flow measurements in the pulmonary artery and volume measurements of the right ventricle for calculation of cardiac output corresponded well in healthy volunteers, and were also acceptable in patients with mild tricuspid regurgitation. With increasing tricuspid insufficiency, the differences between volume and flow measurements for determining cardiac output by MRI became more pronounced (Fig 2). Right Ventricular Myocardial Mass and Hemodynamics
Results Cardiac Output by MRI and Thermodilution The results of catheter testing and magnetic resonance studies in healthy volunteers and patients with pulmonary hypertension are shown in Table 2. In patients with pulmonary hypertension, the cardiac output as determined by right-heart catheterization was 3.5 ⫾ 1.3 L/min. Assessment of volume flow in the pulmonary artery by MRI revealed a cardiac output of 3.4 ⫾ 1.8 L/min. The
The mean muscle mass of the right ventricle as determined by MRI was 44 ⫾ 7 g (range, 36 to 56 g) in the control group and 124 ⫾ 27 g (range, 74 to 165 g) in patients with pulmonary hypertension (p ⫽ 0.0004). There was no significant correlation between right ventricular muscle mass and mean pulmonary artery pressure (r ⫽ 0.37), cardiac output (r ⫽ ⫺ 0.07), right atrial pressure (r ⫽ ⫺0.24), or pulmonary vascular resistance (r ⫽ 0.17).
Table 2—Hemodynamic Findings and MRI Results in 16 Patients With Pulmonary Hypertension and 6 Healthy Volunteers Variables Cardiac output, L/min Thermodilution Flow in pulmonary artery (MRI) Volume measurement (MRI) RVEDV, mL REF catheter MRI RVESV, mL REF catheter MRI REF, % REF catheter MRI Myocardial mass (MRI), g Right ventricle
Control Subjects
Patients With Pulmonary Hypertension
p Value
Not done 6.0 ⫾ 0.71 5.8 ⫾ 0.61
3.5 ⫾ 1.3 3.4 ⫾ 1.8 5.0 ⫾ 1.5
0.0064* 0.34*
Not done 120 ⫾ 22
412 ⫾ 168 180 ⫾ 35
0.001† 0.0012*
Not done 47 ⫾ 13
372 ⫾ 178 120 ⫾ 32
0.001† 0.0004*
Not done 61 ⫾ 7
12 ⫾ 8 34 ⫾ 10
0.001† 0.0004*
124 ⫾ 27
0.0004*
44 ⫾ 7
*Control subjects vs patients with pulmonary hypertension (Mann-Whitney U test). †MRI vs REF catheter (Wilcoxon’s signed rank test). 504
Clinical Investigations
higher than the corresponding MRI values (range, 8 to 600 mL; p ⫽ 0.001), and the RVESV was 249 ⫾ 177 mL higher when assessed with the REF catheter compared to MRI (range, 7 to 634 mL; p ⫽ 0.001). The REF was 12 ⫾ 8% when measured with the REF catheter but 34 ⫾ 10% when assessed by MRI (p ⫽ 0.001). The differences between REF catheter and MRI became more pronounced with increasing tricuspid regurgitation. The discrepancies between REF catheter and MRI in determining the RVEDV and RVESV were 166 ⫾ 67 mL and 174 ⫾ 72 mL in patients with tricuspid regurgitation grade I and II, and 328 ⫾ 224 mL and 361 ⫾ 233 mL in patients with tricuspid regurgitation grade III, respectively. These differences, however, did not reach statistical significance. Determinants of Right-Heart Dysfunction Figure 1. Bland-Altman graph showing the differences of cardiac output measurements by thermodilution and pulmonary artery flow determination by MRI plotted against the mean of both techniques. NMR ⫽ nuclear magnetic resonance.
Right Ventricular Dimensions Assessed by MRI and by the REF Catheter When compared with MRI, the REF catheter constantly overestimated right ventricular volumes and underestimated REFs (Table 2). The RVEDV as assessed with the REF catheter was 231 ⫾ 165 mL
In order to address differences in patients with preserved cardiac function and those with low-output failure, we classified patients with pulmonary hypertension into two groups: those with low cardiac output, arbitrarily defined by a cardiac output ⬍ 3.0 L/min, and those with a near-normal or normal cardiac output of ⬎ 3.0 L/min. As shown in Table 3, both groups had similar pulmonary artery pressure levels, but the right atrial pressures were significantly higher in the lowoutput group (p ⫽ 0.02). Surprisingly, MRI revealed that both groups also had nearly identical right ventricular dimensions, eg, RVEDV and RVESV, as well as
Table 3—Hemodynamic Findings and MRI Results in 16 Patients With Primary Pulmonary Hypertension*
Figure 2. Differences between estimation of cardiac output by flow measurements in the pulmonary artery and right ventricular volume measurements (SV as determined by the difference between end-diastolic and end-systolic chamber volume times heart rate) in relation to the severity of tricuspid regurgitation. See text for the grading of tricuspid regurgitation.
Variables
“Normal” Output (n ⫽ 9)
Low Output (n ⫽ 7)
CO (thermodilution), L/min Range PAP mean, mm Hg Range PVR, dyne 䡠 s 䡠 cm⫺5 Range RA, mm Hg Range RVEDV (MRI), mL Range RVESV (MRI), mL Range SV (MRI), mL Range REF (MRI), % Range RV myocardial mass, g Range
4.3 ⫾ 1.1 3.3–7.0 55 ⫾ 13 41–74 942 ⫾ 374 389–1,476 6⫾4 2–13 178 ⫾ 18 143–205 119 ⫾ 19 79–144 59 ⫾ 21 21–87 33 ⫾ 11 15–52 125 ⫾ 28 74–165
2.3 ⫾ 0.21 2.0–2.6 54 ⫾ 9 39–71 1,583 ⫾ 295 1,138–2,041 12 ⫾ 7 2–20† 184 ⫾ 46 110–260 121 ⫾ 41 66–197 63 ⫾ 10 44–81 35 ⫾ 6 24–44 121 ⫾ 25 77–152
*Data are presented as mean ⫾ SD unless otherwise indicated; RA ⫽ right atrial pressure; RV ⫽ right ventricular; see Table 1 for expansion of other abbreviations. †p ⬍ 0.05 for comparison between groups. CHEST / 120 / 2 / AUGUST, 2001
505
similar right ventricular SVs, REFs, and right ventricular muscle masses (Table 3). The only remarkable difference between the two groups was the severity of tricuspid regurgitation, since severe (grade III) tricuspid regurgitation was present in six of seven patients in the low-output group but in none of nine patients in the “normal” output group (p ⫽ 0.0009). The RVEDVs as determined by MRI were 163 ⫾ 37 mL in patients with tricuspid regurgitation grade I (n ⫽ 4), 171 ⫾ 20 mL in patients with tricuspid regurgitation grade II (n ⫽ 6), and 201 ⫾ 38 mL in patients with tricuspid regurgitation grade III (n ⫽ 6). Thus, there was a trend toward larger right ventricular dimensions with increasing tricuspid regurgitation, but there was a wide overlap and the differences did not reach statistical significance. Discussion This study provides new insights in factors that influence right ventricular performance in patients with pulmonary hypertension. We expected the right ventricles of patients with low cardiac output to have larger dimensions and possibly lower muscle masses than those of patients with preserved cardiac output. This proved not to be true. In the present study, patients with low cardiac output (⬍ 3.0 L/min) and those with normal or near-normal cardiac output (⬎ 3.0 L/min) had similar right ventricular dimensions, right ventricular muscle masses, and REFs as determined by MRI. The only notable differences between both groups were the degree of right atrial pressure and the severity of tricuspid regurgitation. Patients with low cardiac output had significantly higher levels of right atrial pressure. Severe (grade III) tricuspid regurgitation was present in six of seven patients with low cardiac output but in none of nine patients with normal cardiac output. The right ventricular SV as assessed by MRI (RVEDV minus RVESV) was nearly identical in patients with low cardiac output and normal cardiac output, indicating that the reason for low output was not myocardial failure but ineffective blood flow due to regurgitation over the insufficient tricuspid valve. These findings underscore the pathophysiologic consequences of tricuspid regurgitation and right ventricular volume overload in patients with pulmonary hypertension and right-heart failure. Furthermore, the difference between the right ventricular output (calculated as difference between RVEDV and RVESV) and flow in the pulmonary artery seems to reflect regurgitating blood volume over the tricuspid valve, and may therefore allow indirect quantification of tricuspid regurgitation. With newer MRI software, it may be possible to directly quantify 506
regurgitating blood volume, but this technique was not available at the onset of this study. Our findings further suggest that the severity of tricuspid regurgitation is not simply determined by right ventricular dimensions. In our patients, the RVEDVs tended to be higher in patients with severe (grade III) tricuspid regurgitation than in patients with mild (grade I) or moderate (grade II) tricuspid regurgitation, but this difference did not reach statistical significance and there was wide overlap between these groups. This observation may indicate that factors other than ventricle size such as papillary muscle dysfunction may contribute to the severity of tricuspid insufficiency and thereby right ventricular failure in patients with pulmonary hypertension. Another finding of this study was that the REF catheter largely overestimated RVEDV and RVESV and underestimated REFs in patients with pulmonary hypertension. We have shown5 that thermodilution using the REF catheter allows accurate determination of cardiac output in these patients, and there was also good overall agreement between thermodilution and flow measurement in the pulmonary artery by MRI in the present study. It remains unclear why the REF catheter provides accurate measurements of cardiac output but invalid information on right ventricular volumes and ejection fractions in patients with pulmonary hypertension. Compared to conventional thermodilution catheters, the REF catheter has some modifications to allow fast determination of temperature changes: a rapid thermistor (95 ms) at the tip of the catheter, and two sensitive ECG electrodes for R-wave detection. As in other thermodilution computers, the REF-1 monitor uses a modified Steward-Hamilton equation to determine cardiac output from the integrated area under the thermodilution curve. In addition, the rapid detection of temperature changes combined with ECG analysis allows a beat-to-beat detection of
Figure 3. Thermodilution curve with an REF catheter. After injection of ice-cold, isotonic saline solution, the beat-to-beat analysis of the thermodilution curve shows plateaus (C1, C2, and C3). Two residual fractions (RF1 and RF2) are calculated from these plateaus are shown, and the mean residual fraction (RF) is used for calculation of the ejection fraction (EF). Clinical Investigations
blood flow that can be used for calculation of REF (Fig 3). Using this so-called plateau technique, the REF-1 computer calculates REF, RVEDV (SV divided by REF), and right ventricular SV (RVEDV minus RVESV). We assume that the presence of tricuspid regurgitation results in oscillations of the indicator solution and increased loss of temperature that cause a decrease of the beat-to-beat changes of the temperature. This in turn could lead to underestimation of REF and overestimation of RVEDV and RVESV. The shape of the thermodilution curve is flattened in the presence of tricuspid regurgitation, but the total area under the thermodilution curve that is used for calculation of cardiac output may not to be significantly affected. The interpretation of our results is based on the assumption that MRI provides reliable estimation of right ventricular dimensions and muscle mass. It is clearly a weakness of this study that no reference technique was used for determination of these variables. There is, however, no accepted “gold standard” for assessing right ventricular volume and muscle mass in the clinical setting, and others1,8 have validated the usefulness of MRI for right ventricular measurements. Furthermore, our MRI results compare well to published data on healthy volunteers and patients with pulmonary hypertension.1,9 There were several other limitations in the study. The relative small sample size requires confirmation of our results by larger studies. Furthermore, catheter studies and MRI examinations could not be performed simultaneously, which might have affected our findings. Nevertheless, all patients were in a stable clinical condition, their medications were not changed during this study, and the interval between both examinations was not ⬎ 24 h, which should make significant variations unlikely. In summary, our findings indicate that MRI may be helpful to understand right-heart performance in
patients with pulmonary hypertension. In these patients, the REF catheter seems to be invalid for determination of right ventricular dimensions. The major finding in this study was that the presence of right ventricular forward failure was not determined by ventricle size, muscle mass, or level of pulmonary hypertension, but primarily by the degree of tricuspid regurgitation. ACKNOWLEDGMENT: We thank Monica Jones, MD, for reviewing the article.
References 1 Katz J, Whang J, Boxt LM, et al. Estimation of right ventricular mass in normal subjects and in patients with primary pulmonary hypertension by nuclear magnetic resonance imaging. J Am Coll Cardiol 1993; 21:1475–1481 2 Tardivon AA, Mousseaux E, Brenot F, et al. Quantification of hemodynamics in primary pulmonary hypertension with magnetic resonance imaging. Am J Respir Crit Care Med 1994; 150:1075–1080 3 Jardin F, Gueret P, Dubourg O, et al. Right ventricular volumes by thermodilution in the adult respiratory distress syndrome: a comparative study using two-dimensional echocardiography as a reference method. Chest 1985; 88:34 –39 4 Dhainaut JF, Brunet F, Monsallier JF, et al. Bedside evaluation of right ventricular performance using a rapid computerized thermodilution method. Crit Care Med 1987; 15:148 –152 5 Hoeper MM, Maier R, Tongers J, et al. Determination of cardiac output by the Fick method, thermodilution, and acetylene rebreathing in pulmonary hypertension. Am J Respir Crit Care Med 1999; 160:535–541 6 Sagie A, Schwammenthal E, Padial LR, et al. Determinants of functional tricuspid regurgitation in incomplete valve closure: Doppler color flow study in 109 patients. J Am Coll Cardiol 1994; 24:446 – 453 7 Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986; 1:307–310 8 Longmore DB, Klipstein RH, Underwood SR, et al. Dimensional accuracy of magnetic resonance in studies of the heart. Lancet 1985; 1:1360 –1362 9 Lorenz CH, Walker ES, Morgan VL, et al. Normal human right and left ventricular mass, systolic function, and gender differences by cine magnetic resonance imaging. J Cardiovasc Magn Reson 1999; 1:7–21
CHEST / 120 / 2 / AUGUST, 2001
507
Study objectives: We sought to compare catheter studies using a right ventricular ejection fraction (REF) catheter together with echocardiography and MRI in patients with pulmonary hypertension. Patients and methods: We compared hemodynamic findings, echocardiography, and MRI studies in 16 patients with pulmonary hypertension. Six healthy volunteers served as control subjects for the MRI studies. Results: MRI imaging provided accurate assessment of cardiac output in all but two patients. As compared with MRI, the REF catheter constantly underestimated the REF and overestimated right ventricular volumes in patients with pulmonary hypertension. REF, end-systolic and end-diastolic right ventricular volumes, and right ventricular muscle mass, as determined by MRI, were almost identical in patients with preserved cardiac function and those with low-output failure. The only factor that was different in both groups was the severity of tricuspid regurgitation. Conclusion: Right ventricular dimensions and muscle mass do not differ in patients with pulmonary hypertension who have low cardiac output and those who do not. According to our results, the major determinant of cardiac output in these patients appears to be the severity of tricuspid regurgitation. The REF catheter provides invalid data on right ventricular dimensions in patients with pulmonary hypertension. (CHEST 2001; 120:502–507) Key words: hypertension; MRI; pulmonary; right-heart failure; thermodilution Abbreviations: REF ⫽ right ventricular ejection fraction; RVEDV ⫽ right ventricular end-diastolic volume; RVESV ⫽ right ventricular end-systolic volume; SV ⫽ stroke volume
capacity of the right ventricle is a T hemajorfunctional prognostic determinant in pulmonary hypertension. It is unknown why some patients with markedly elevated pulmonary artery pressures maintain well-preserved cardiac function for several years, while others with equally or less severe pulmonary hypertension show rapidly progressive rightheart failure. One factor that has hindered the understanding of right ventricular performance in patients with pulmonary hypertension has been the *From the Departments of Pulmonary Medicine (Dr. Hoeper, Mr. Tongers, and Mr. Maier) and Diagnostic Radiology (Drs. Lotz, Baus, and Leppert), Hannover Medical School, Hannover, Germany. Manuscript received July 13, 2000; revision accepted March 2, 2001. Correspondence to: Marius M. Hoeper, MD, Department of Pulmonary Medicine, Hannover Medical School, 30623 Hannover, Germany; e-mail: [email protected] 502
lack of techniques that give a reliable picture of right ventricular remodeling. Recently, MRI has been shown to provide adequate information about right ventricular dimensions and muscle mass, and MRI may become the technique of choice to investigate right ventricular remodeling in patients with pulmonary hypertension.1,2 Information about right ventricular dimensions can also be derived from a right ventricular ejection fraction (REF) catheter, a modified thermodilution catheter. This catheter has been developed to estimate REF and right ventricular volumes based on a rapid beat-to-beat thermodilution technique, and has been found useful in patients with right heart dysfunction.3,4 We have shown5 that this catheter allows reliable assessment of cardiac output in patients with pulmonary hypertension. However, whether the REF catheter also provides valid data on Clinical Investigations
right ventricular dimensions and ejection fractions in this group of patients has not been investigated. To address this question, we compared hemodynamic data obtained with a REF catheter with the results from MRI imaging in patients with severe pulmonary hypertension. We also evaluated the relationship between hemodynamic findings and results from MRI studies and echocardiography to obtain information on right ventricular remodeling in patients with and without right ventricular failure. Materials and Methods Patients We studied 16 patients with primary (n ⫽ 14) or thromboembolic (n ⫽ 2) pulmonary hypertension (Table 1). The patients were in stable clinical condition, and medications were not changed between catheter studies and MRI examinations. In all patients, catheter examinations were done for diagnostic or therapeutic reasons unrelated to this study. The protocol was approved by the institutional review board, and all patients gave informed consent prior to the study. Control Population Magnetic resonance examinations were performed in six healthy volunteers (four women and two men; age range, 27 to 46 years) to assess right ventricular volumes, muscle mass, and flow in the pulmonary artery.
Table 1—Clinical and Hemodynamic Characteristics of the Patients* Characteristics Patients, No. Age, yr Range Male/female gender, No. Functional class, No. NYHA II NYHA III NYHA IV Hemodynamic variables (right-heart catheterization) PAP mean, mm Hg Range CO, L/min Range CI, L/min/m2 Range SV, mL Range PVR, dyne 䡠 s 䡠 cm⫺5 Range RA, mm Hg Range PCWP, mm Hg Range
Data 16 39 ⫾ 13 28–61 6/10 1 11 4
57 ⫾ 12 35–74 3.5 ⫾ 1.3 2.0–7.0 2.0 ⫾ 0.6 1.3–3.2 44 ⫾ 23 22–74 1,222 ⫾ 482 389–2,041 9.0 ⫾ 6.2 2–20 8.0 ⫾ 2.2 3–12
*Data are presented as mean ⫾ SD unless otherwise indicated. NYHA ⫽ New York Heart Association; PAP mean ⫽ mean pulmonary arterial pressure; CO ⫽ cardiac output; CI ⫽ cardiac index; PVR ⫽ pulmonary vascular resistance; RA ⫽ right atrial pressure; PCWP ⫽ pulmonary capillary wedge pressure.
Right-Heart Catheterization A 7.5F quadruple-lumen, balloon-tipped, flow-directed SwanGanz Catheter (No. 93A-434H7.5F; Baxter Edwards; Irvine, CA) was advanced through an 8F-introducer sheath via the right or left internal jugular vein. In order to ensure that the right atrial opening of the catheter was located directly before the tricuspid valve, the catheter was advanced into the pulmonary artery until the pressure recording through the proximal lumen revealed a ventricular pressure curve. Then, the catheter was slowly retracted until an atrial pressure recording was obtained. After insertion of the catheter, the patients were allowed to rest for at least 15 min before hemodynamic measures were recorded. The cardiac output was measured by the thermodilution technique with 10 mL of sterile, ice-cold, isotonic (0.9%) saline solution that was injected through the proximal (right atrial) lumen of the catheter, and the drop in temperature at the distal thermistor was recorded. Cardiac output was calculated using an analog computer system (REF-1; Baxter Edwards). Echocardiography Echocardiography was performed to exclude left-heart disease and intracardiac shunting and to assess the presence of tricuspid insufficiency. The echocardiographic examinations took place 1 to 3 days prior to the catheter studies. The severity of tricuspid regurgitation was graded as absent (grade 0), mild (grade I, jet area ⬍ 20% of the right atrial area), moderate (grade II, jet area between 20% and 33% of the right atrial area), and severe (grade III, jet area ⬎ 33% of the right atrial area), according to established grading systems.6 MRI Magnetic resonance studies were done within 24 h before (n ⫽ 5) or after (n ⫽ 11) the catheter studies. The radiologists were unaware of the catheter findings, and the pulmonologists were unaware of the MRI results. Volunteers and patients underwent ECG-gated MRI imaging (1.5 T MR Scanner; Signa Horizon EchoSpeed; General Electric; Milwaukee, WI). For determination of cardiac volumes, 8 to 10 slices along the short heart axis were obtained using a k space segmented, fast gradient-echo sequence (FastCard; GE-Medical Systems; Milwaukee, WI). Parameters included ECG gating, 8-mm slice thickness, 5-ms echo time, 11-ms repetition time, 20° flip angle, 32-cm field of view, and a matrix size of 256 ⫻ 160. Depending on the heart rate, an average of 23 phases per cardiac cycle was obtained. Flow measurements were obtained using a single double-oblique slice perpendicular to flow in the main pulmonary artery. In this setting, parameters used were ECG gating, 8-mm slice thickness, 6-ms echo time, 33-ms repetition time, 20° flip, 24-cm field of view, 200-cm/s velocity encoding, 32 phases per cardiac cycle, and a matrix size of 256 ⫻ 128. Data Processing Right ventricular volumes were assessed at a separate workstation (Sun SPARC 20; Advantage Windows 1.2; CAP Software; General Electric, WI). For each slice, right ventricular enddiastolic volume (RVEDV) and right ventricular end-systolic volume (RVESV) were calculated by semiautomatic outlining of the endocardium. Right ventricular volumes and ejection fractions were automatically calculated by the software. Ventricular muscle mass was measured by manual outlining of the epicardial and endocardial border of the right ventricle in each slice. The ventricular septum was defined as part of the left CHEST / 120 / 2 / AUGUST, 2001
503
ventricle. All volumes covering the right myocardium were added, and muscle mass was calculated by multiplying the muscle volume by the specific gravity of myocardium (1.05 g/cm). Flow measurements in the pulmonary artery were analyzed at the MRI console (image flow analysis software). For this purpose, the vessel was outlined in each of the 32 phases. From these areas, the software calculated minimum, maximum, and average flow rates, and, utilizing the vessel diameter, average flow rates per minute. Statistics All results are given as mean ⫾ SD. The Mann-Whitney U test was used to compare the MRI findings between the control population and patients with pulmonary hypertension, and to compare hemodynamic findings in patients with normal or low cardiac output. Variables obtained by MRI and with the REF catheter were compared by Wilcoxon’s signed-rank test. Fisher’s Exact Test was used to compare the prevalence of severe tricuspid regurgitation in patients with low or normal cardiac output. To address the question as to whether there were significant differences in right ventricular dimensions in patients with different degrees of tricuspid regurgitation, the KruskalWallis test was used. Linear regression analysis and Bland-Altman plots were used to compare catheter findings and MRI results.7 All p values ⬍ 0.05 were considered statistically significant.
mean difference between thermodilution and flow in the pulmonary artery as assessed by MRI was 0.1 ⫾ 1.0 L/min. In all but two cases, there was generally an acceptable overall agreement between these two techniques (Fig 1). In contrast to flow measurement in the pulmonary artery, volume measurements for determining cardiac output by MRI (that is, measuring stroke volume [SV] as the difference between RVEDV and RVESV, and calculating cardiac output as the product of SV and heart rate) constantly overestimated the cardiac output in patients with pulmonary hypertension. Flow measurements in the pulmonary artery and volume measurements of the right ventricle for calculation of cardiac output corresponded well in healthy volunteers, and were also acceptable in patients with mild tricuspid regurgitation. With increasing tricuspid insufficiency, the differences between volume and flow measurements for determining cardiac output by MRI became more pronounced (Fig 2). Right Ventricular Myocardial Mass and Hemodynamics
Results Cardiac Output by MRI and Thermodilution The results of catheter testing and magnetic resonance studies in healthy volunteers and patients with pulmonary hypertension are shown in Table 2. In patients with pulmonary hypertension, the cardiac output as determined by right-heart catheterization was 3.5 ⫾ 1.3 L/min. Assessment of volume flow in the pulmonary artery by MRI revealed a cardiac output of 3.4 ⫾ 1.8 L/min. The
The mean muscle mass of the right ventricle as determined by MRI was 44 ⫾ 7 g (range, 36 to 56 g) in the control group and 124 ⫾ 27 g (range, 74 to 165 g) in patients with pulmonary hypertension (p ⫽ 0.0004). There was no significant correlation between right ventricular muscle mass and mean pulmonary artery pressure (r ⫽ 0.37), cardiac output (r ⫽ ⫺ 0.07), right atrial pressure (r ⫽ ⫺0.24), or pulmonary vascular resistance (r ⫽ 0.17).
Table 2—Hemodynamic Findings and MRI Results in 16 Patients With Pulmonary Hypertension and 6 Healthy Volunteers Variables Cardiac output, L/min Thermodilution Flow in pulmonary artery (MRI) Volume measurement (MRI) RVEDV, mL REF catheter MRI RVESV, mL REF catheter MRI REF, % REF catheter MRI Myocardial mass (MRI), g Right ventricle
Control Subjects
Patients With Pulmonary Hypertension
p Value
Not done 6.0 ⫾ 0.71 5.8 ⫾ 0.61
3.5 ⫾ 1.3 3.4 ⫾ 1.8 5.0 ⫾ 1.5
0.0064* 0.34*
Not done 120 ⫾ 22
412 ⫾ 168 180 ⫾ 35
0.001† 0.0012*
Not done 47 ⫾ 13
372 ⫾ 178 120 ⫾ 32
0.001† 0.0004*
Not done 61 ⫾ 7
12 ⫾ 8 34 ⫾ 10
0.001† 0.0004*
124 ⫾ 27
0.0004*
44 ⫾ 7
*Control subjects vs patients with pulmonary hypertension (Mann-Whitney U test). †MRI vs REF catheter (Wilcoxon’s signed rank test). 504
Clinical Investigations
higher than the corresponding MRI values (range, 8 to 600 mL; p ⫽ 0.001), and the RVESV was 249 ⫾ 177 mL higher when assessed with the REF catheter compared to MRI (range, 7 to 634 mL; p ⫽ 0.001). The REF was 12 ⫾ 8% when measured with the REF catheter but 34 ⫾ 10% when assessed by MRI (p ⫽ 0.001). The differences between REF catheter and MRI became more pronounced with increasing tricuspid regurgitation. The discrepancies between REF catheter and MRI in determining the RVEDV and RVESV were 166 ⫾ 67 mL and 174 ⫾ 72 mL in patients with tricuspid regurgitation grade I and II, and 328 ⫾ 224 mL and 361 ⫾ 233 mL in patients with tricuspid regurgitation grade III, respectively. These differences, however, did not reach statistical significance. Determinants of Right-Heart Dysfunction Figure 1. Bland-Altman graph showing the differences of cardiac output measurements by thermodilution and pulmonary artery flow determination by MRI plotted against the mean of both techniques. NMR ⫽ nuclear magnetic resonance.
Right Ventricular Dimensions Assessed by MRI and by the REF Catheter When compared with MRI, the REF catheter constantly overestimated right ventricular volumes and underestimated REFs (Table 2). The RVEDV as assessed with the REF catheter was 231 ⫾ 165 mL
In order to address differences in patients with preserved cardiac function and those with low-output failure, we classified patients with pulmonary hypertension into two groups: those with low cardiac output, arbitrarily defined by a cardiac output ⬍ 3.0 L/min, and those with a near-normal or normal cardiac output of ⬎ 3.0 L/min. As shown in Table 3, both groups had similar pulmonary artery pressure levels, but the right atrial pressures were significantly higher in the lowoutput group (p ⫽ 0.02). Surprisingly, MRI revealed that both groups also had nearly identical right ventricular dimensions, eg, RVEDV and RVESV, as well as
Table 3—Hemodynamic Findings and MRI Results in 16 Patients With Primary Pulmonary Hypertension*
Figure 2. Differences between estimation of cardiac output by flow measurements in the pulmonary artery and right ventricular volume measurements (SV as determined by the difference between end-diastolic and end-systolic chamber volume times heart rate) in relation to the severity of tricuspid regurgitation. See text for the grading of tricuspid regurgitation.
Variables
“Normal” Output (n ⫽ 9)
Low Output (n ⫽ 7)
CO (thermodilution), L/min Range PAP mean, mm Hg Range PVR, dyne 䡠 s 䡠 cm⫺5 Range RA, mm Hg Range RVEDV (MRI), mL Range RVESV (MRI), mL Range SV (MRI), mL Range REF (MRI), % Range RV myocardial mass, g Range
4.3 ⫾ 1.1 3.3–7.0 55 ⫾ 13 41–74 942 ⫾ 374 389–1,476 6⫾4 2–13 178 ⫾ 18 143–205 119 ⫾ 19 79–144 59 ⫾ 21 21–87 33 ⫾ 11 15–52 125 ⫾ 28 74–165
2.3 ⫾ 0.21 2.0–2.6 54 ⫾ 9 39–71 1,583 ⫾ 295 1,138–2,041 12 ⫾ 7 2–20† 184 ⫾ 46 110–260 121 ⫾ 41 66–197 63 ⫾ 10 44–81 35 ⫾ 6 24–44 121 ⫾ 25 77–152
*Data are presented as mean ⫾ SD unless otherwise indicated; RA ⫽ right atrial pressure; RV ⫽ right ventricular; see Table 1 for expansion of other abbreviations. †p ⬍ 0.05 for comparison between groups. CHEST / 120 / 2 / AUGUST, 2001
505
similar right ventricular SVs, REFs, and right ventricular muscle masses (Table 3). The only remarkable difference between the two groups was the severity of tricuspid regurgitation, since severe (grade III) tricuspid regurgitation was present in six of seven patients in the low-output group but in none of nine patients in the “normal” output group (p ⫽ 0.0009). The RVEDVs as determined by MRI were 163 ⫾ 37 mL in patients with tricuspid regurgitation grade I (n ⫽ 4), 171 ⫾ 20 mL in patients with tricuspid regurgitation grade II (n ⫽ 6), and 201 ⫾ 38 mL in patients with tricuspid regurgitation grade III (n ⫽ 6). Thus, there was a trend toward larger right ventricular dimensions with increasing tricuspid regurgitation, but there was a wide overlap and the differences did not reach statistical significance. Discussion This study provides new insights in factors that influence right ventricular performance in patients with pulmonary hypertension. We expected the right ventricles of patients with low cardiac output to have larger dimensions and possibly lower muscle masses than those of patients with preserved cardiac output. This proved not to be true. In the present study, patients with low cardiac output (⬍ 3.0 L/min) and those with normal or near-normal cardiac output (⬎ 3.0 L/min) had similar right ventricular dimensions, right ventricular muscle masses, and REFs as determined by MRI. The only notable differences between both groups were the degree of right atrial pressure and the severity of tricuspid regurgitation. Patients with low cardiac output had significantly higher levels of right atrial pressure. Severe (grade III) tricuspid regurgitation was present in six of seven patients with low cardiac output but in none of nine patients with normal cardiac output. The right ventricular SV as assessed by MRI (RVEDV minus RVESV) was nearly identical in patients with low cardiac output and normal cardiac output, indicating that the reason for low output was not myocardial failure but ineffective blood flow due to regurgitation over the insufficient tricuspid valve. These findings underscore the pathophysiologic consequences of tricuspid regurgitation and right ventricular volume overload in patients with pulmonary hypertension and right-heart failure. Furthermore, the difference between the right ventricular output (calculated as difference between RVEDV and RVESV) and flow in the pulmonary artery seems to reflect regurgitating blood volume over the tricuspid valve, and may therefore allow indirect quantification of tricuspid regurgitation. With newer MRI software, it may be possible to directly quantify 506
regurgitating blood volume, but this technique was not available at the onset of this study. Our findings further suggest that the severity of tricuspid regurgitation is not simply determined by right ventricular dimensions. In our patients, the RVEDVs tended to be higher in patients with severe (grade III) tricuspid regurgitation than in patients with mild (grade I) or moderate (grade II) tricuspid regurgitation, but this difference did not reach statistical significance and there was wide overlap between these groups. This observation may indicate that factors other than ventricle size such as papillary muscle dysfunction may contribute to the severity of tricuspid insufficiency and thereby right ventricular failure in patients with pulmonary hypertension. Another finding of this study was that the REF catheter largely overestimated RVEDV and RVESV and underestimated REFs in patients with pulmonary hypertension. We have shown5 that thermodilution using the REF catheter allows accurate determination of cardiac output in these patients, and there was also good overall agreement between thermodilution and flow measurement in the pulmonary artery by MRI in the present study. It remains unclear why the REF catheter provides accurate measurements of cardiac output but invalid information on right ventricular volumes and ejection fractions in patients with pulmonary hypertension. Compared to conventional thermodilution catheters, the REF catheter has some modifications to allow fast determination of temperature changes: a rapid thermistor (95 ms) at the tip of the catheter, and two sensitive ECG electrodes for R-wave detection. As in other thermodilution computers, the REF-1 monitor uses a modified Steward-Hamilton equation to determine cardiac output from the integrated area under the thermodilution curve. In addition, the rapid detection of temperature changes combined with ECG analysis allows a beat-to-beat detection of
Figure 3. Thermodilution curve with an REF catheter. After injection of ice-cold, isotonic saline solution, the beat-to-beat analysis of the thermodilution curve shows plateaus (C1, C2, and C3). Two residual fractions (RF1 and RF2) are calculated from these plateaus are shown, and the mean residual fraction (RF) is used for calculation of the ejection fraction (EF). Clinical Investigations
blood flow that can be used for calculation of REF (Fig 3). Using this so-called plateau technique, the REF-1 computer calculates REF, RVEDV (SV divided by REF), and right ventricular SV (RVEDV minus RVESV). We assume that the presence of tricuspid regurgitation results in oscillations of the indicator solution and increased loss of temperature that cause a decrease of the beat-to-beat changes of the temperature. This in turn could lead to underestimation of REF and overestimation of RVEDV and RVESV. The shape of the thermodilution curve is flattened in the presence of tricuspid regurgitation, but the total area under the thermodilution curve that is used for calculation of cardiac output may not to be significantly affected. The interpretation of our results is based on the assumption that MRI provides reliable estimation of right ventricular dimensions and muscle mass. It is clearly a weakness of this study that no reference technique was used for determination of these variables. There is, however, no accepted “gold standard” for assessing right ventricular volume and muscle mass in the clinical setting, and others1,8 have validated the usefulness of MRI for right ventricular measurements. Furthermore, our MRI results compare well to published data on healthy volunteers and patients with pulmonary hypertension.1,9 There were several other limitations in the study. The relative small sample size requires confirmation of our results by larger studies. Furthermore, catheter studies and MRI examinations could not be performed simultaneously, which might have affected our findings. Nevertheless, all patients were in a stable clinical condition, their medications were not changed during this study, and the interval between both examinations was not ⬎ 24 h, which should make significant variations unlikely. In summary, our findings indicate that MRI may be helpful to understand right-heart performance in
patients with pulmonary hypertension. In these patients, the REF catheter seems to be invalid for determination of right ventricular dimensions. The major finding in this study was that the presence of right ventricular forward failure was not determined by ventricle size, muscle mass, or level of pulmonary hypertension, but primarily by the degree of tricuspid regurgitation. ACKNOWLEDGMENT: We thank Monica Jones, MD, for reviewing the article.
References 1 Katz J, Whang J, Boxt LM, et al. Estimation of right ventricular mass in normal subjects and in patients with primary pulmonary hypertension by nuclear magnetic resonance imaging. J Am Coll Cardiol 1993; 21:1475–1481 2 Tardivon AA, Mousseaux E, Brenot F, et al. Quantification of hemodynamics in primary pulmonary hypertension with magnetic resonance imaging. Am J Respir Crit Care Med 1994; 150:1075–1080 3 Jardin F, Gueret P, Dubourg O, et al. Right ventricular volumes by thermodilution in the adult respiratory distress syndrome: a comparative study using two-dimensional echocardiography as a reference method. Chest 1985; 88:34 –39 4 Dhainaut JF, Brunet F, Monsallier JF, et al. Bedside evaluation of right ventricular performance using a rapid computerized thermodilution method. Crit Care Med 1987; 15:148 –152 5 Hoeper MM, Maier R, Tongers J, et al. Determination of cardiac output by the Fick method, thermodilution, and acetylene rebreathing in pulmonary hypertension. Am J Respir Crit Care Med 1999; 160:535–541 6 Sagie A, Schwammenthal E, Padial LR, et al. Determinants of functional tricuspid regurgitation in incomplete valve closure: Doppler color flow study in 109 patients. J Am Coll Cardiol 1994; 24:446 – 453 7 Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986; 1:307–310 8 Longmore DB, Klipstein RH, Underwood SR, et al. Dimensional accuracy of magnetic resonance in studies of the heart. Lancet 1985; 1:1360 –1362 9 Lorenz CH, Walker ES, Morgan VL, et al. Normal human right and left ventricular mass, systolic function, and gender differences by cine magnetic resonance imaging. J Cardiovasc Magn Reson 1999; 1:7–21
CHEST / 120 / 2 / AUGUST, 2001
507