ESTIMATED RIGHT VENTRICULAR SYSTOLIC TIME INTERVALS FOR THE ASSESSMENT OF RIGHT VENTRICULAR FUNCTION IN ACUTE RESPIRATORY DISTRESS SYNDROME

Abstract

Right ventricular (RV) systolic time intervals (STIs) have been shown to accurately reflect RV function in patients with acute respiratory distress syndrome (ARDS). The measurement of RVSTIs requires phonocardiography to define the time of RV end systole. If RV end-systolic pressure (RVESP) can be derived from peak pulmonary artery (PA) systolic pressure, then the time of RV end systole, and hence, RVSTIs can be deduced without phonocardiography. We tested this possibility. In 34 patients with ARDS, RVESP was determined on the PAP curve at RV end systole, which was defined by phonocardiography. The ratios of RVESP/peak PA systolic pressure were obtained in each patient, the mean of which was 0.90 ± 0.006. With an application of this value, the estimated RVSTIs were determined in other groups of patients. Right ventricular end-systolic pressure was estimated from the peak PA systolic pressure by multiplying 0.9. Then the point of RV end systole was determined on the electrocardiographic tracing that coincided with the point of RVESP on the PAP curve by simultaneous display of electrocardiograph and PAP curve. Total electromechanical systole was measured from the onset of the QRS complex to the point of RV end systole on the electrocardiograph. The onset of RV ejection was defined by PAP curve. The validity of this estimated RVSTIs was tested by comparing with the measured RVSTIs. By Bland-Altman analysis, the mean difference in RVSTIs between the two methods was 0.007, and bias was 0.0036, suggesting close agreement. The estimated RVSTIs can be used to accurately assess RV function.

ABBREVIATIONS-RVSTIs-right ventricular systolic time intervals; ARDS-acute respiratory distress syndrome; ESP-end-systolic pressure; RVESP-right ventricular end-systolic pressure; PAP-pulmonary artery pressure; PA-pulmonary artery; QS₂-interval of electromechanical systole; PPASP-peak pulmonary artery systolic pressure; PEP-pre-ejection period; RVET-right ventricular ejection time; Pao-partial pressure of oxygen in arterial blood; Fio-inspired oxygen concentration in fraction

INTRODUCTION

In patients with acute respiratory distress syndrome (ARDS) complicated by pulmonary hypertension, the elevated ratio of physiological dead space to tidal volume has been shown to be associated with high mortality (^{1, 2}). Because right ventricular function has been shown to be a major determinant of the ratio of physiological dead space to tidal volume in these patients (^{1, 3}), right ventricular dysfunction would be a prognostic indicator of poor outcome. Thus, it is important to detect any such dysfunction. Right ventricular ejection fraction has been used previously to assess right ventricular function (⁴), but it may not be used as an independent index of right ventricular function in patients with ARDS because of high load dependency (³). Also, right ventricular ejection fraction is affected by a tricuspid regurgitation, which is often present in pulmonary hypertension (^{5, 6}). Right ventricular end-systolic pressure-volume relation has been accepted as a gold standard in assessing right ventricular contractile function (⁷), but the definition of the end-systolic pressure-volume relationship requires alteration of the loading condition of the right ventricle (RV) and the measurement of right ventricular volume.

Right ventricular systolic time intervals (RVSTIs) have been used to assess right ventricular function (^8-10). Although RVSTIs have not been studied as thoroughly as left ventricular systolic time intervals, previous studies indicate that, in the right ventricle, systolic time intervals offer a measure of overall right ventricular systolic performance (^{11, 12}). Right ventricular systolic time intervals are expressed as a ratio of pre-ejection period (PEP) to right ventricular ejection time (RVET). Without using echocardiography, however, noninvasive measurement of PEP/RVET ratio is not possible because of the inability to define accurately the onset of right ventricular ejection. In patients with an indwelling pulmonary artery (PA) catheter in place, the onset of right ventricular ejection can be easily defined as the beginning of the rapid upstroke of the pulmonary artery pressure (PAP). Nevertheless, phonocardiography is required to define the end of right ventricular systole, with which the interval of electromechanical systole (QS₂) can be determined. With respect to the left ventricle, end-systolic pressure (ESP) has been accurately predicted from the peak arterial systolic pressure (¹³). Likewise, if right ventricular ESP (RVESP) can be estimated from the peak PA systolic pressure (PPASP), then the end of right ventricular systole can be defined as the time at which PAP reaches the RVESP. If so, RVSTIs can be measured by means of PAP and electrocardiograph without requiring phonocardiography or echocardiography.

In the present study, we examined this possibility by testing whether the estimation of RVESP from the PPASP is consistent among patients with ARDS, and determining if the estimated time of right ventricular end systole could be applied to yield reliable estimates of RVSTIs. The reliability of this estimated RVSTIs was tested in other groups of patients from previous data.

MATERIALS AND METHODS

Patients with ARDS in whom an indwelling PA catheter had been placed were included in the study. The institutional review board approved the study with a waiver of the informed consent. Thirty-four patients were included. The patient group consisted of 21 men and 13 women, ranging in age from 20 to 69 y (mean age, 49 y). Acute respiratory distress syndrome followed sepsis in 13 patients, trauma in 10, major surgery in 7, acute necrotizing pancreatitis in 2, and others. All patients had the acute hypoxic type of respiratory failure (ratio of partial pressure of oxygen in arterial blood to inspired oxygen concentration in fraction [Pao/Fio ratio] <200 and acute onset of bilateral infiltrates on chest radiography) necessitating mechanical ventilatory support with positive end-expiratory pressure (Table 1). No patient was considered for study if there was an evidence of left ventricular failure (pulmonary capillary wedge pressure >18 mmHg) or a history of chronic obstructive pulmonary disease. No patient was receiving vasodilators or vasopressors during the time of study.

TABLE 1: Demographic data of patient groups

All patients received mechanical ventilation, using a volume-controlled ventilator. The partial pressure of carbon dioxide in arterial blood was maintained between 40 and 45 mmHg. An inspired oxygen concentration and positive end-expiratory pressure were adjusted to maintain the Pao higher than 70 mmHg. No patient was receiving a positive end-expiratory pressure of higher than 10 cm H₂O during the study. Ventilator settings were chosen for optimum patient care without regard to the study.

Pulmonary artery pressures were measured with transducers (Baxter Healthcare, Santa Ana, Calif) at the end of expiration (¹⁴). Curve tracings of PAP, lead II of electrocardiography, and phonocardiography were recorded simultaneously on magnetic tape. The recorded PAP was corrected for the frequency and phase response. A 7.5F VIP thermodilution PA catheter with a natural frequency of 33 Hz was used. The effect of frequency and phase response on the amplitude response and phase lag of the manometer-catheter system, which we used in this study, has been described previously (¹⁵). Modulus and phase angle derived from Fourier coefficients were corrected in accordance with the measured amplitude response and phase response of the pressure-measuring system before they were applied. With an ASYST-based (ASYST, Asyst Software Technologies Inc, Rochester, NY) fast Fourier transform program, the analysis automatically included up to the 64th harmonic.

Estimation of RVESP

By simultaneous graphic display of the recorded electrocardiograph, phonocardiograph and PAP curves, RVESP was determined on the PAP curve at right ventricular end systole, which was defined as the closure of the pulmonic valve, as reflected in the pulmonic component of the second heart sound (Fig. 1). An argument whether the right ventricular pressure at the closure of the pulmonic valve is ESP or end-ejection pressure is meritless in the present study (³), because the difference between the two pressures is not significant owing to elevated PAP and pulmonary vascular resistance in patients with ARDS. The RVESP/PPASP ratios were measured in each patient. The consistency of this ratio was examined. By multiplying the RVESP/PPASP ratio, RVESP can be estimated from PPASP. The point of right ventricular end systole can be determined on the electrocardiographic tracing that coincides with the point of RVESP on the PAP curve tracing by simultaneous graphic display of electrocardiogram and PAP curves. And then QS₂ can be estimated by measuring the intervals from the onset of the QRS complex to the point of right ventricular end systole on the electrocardiogram without requiring phonocardiography.

Fig. 1: Simultaneous display of the electrocardiograph (ECG), phonocardiograph (PCG), and PAP for the determination of RVESP on the PAP curve and hence the ratio of RVESP/PPASP. With an application of this ratio, the point of right ventricular end systole can be determined on the ECG tracing that coincides with the point of RVESP on the PAP curve by simultaneous display of ECG and PAP curve, because RVESP can be estimated from the PPASP by multiplying the ratio (0.9). And then QS₂ (interval of electromechanical systole) can be estimated by measuring the interval from the onset of QRS complex to the point of right ventricular end systole on the ECG without requiring PCG. Right ventricular ejection time was measured as the phase of systole from the beginning of the rapid upstroke of the PAP curve to the dicrotic notch. Pre-ejection period was derived by subtraction of the RVET from the interval of QS₂. Adapted with permission from Her et al. (¹²) and publisher.

Reliability of estimated QS2

To determine if the estimated QS₂ can be applied to yield the reliable estimates of RVSTIs, which were expressed as a PEP/RVET ratio, the values by two methods, the estimated PEP/RVET ratio using estimated QS₂ and the measured PEP/RVET ratio using measured QS₂ with phonocardiography were compared in other groups of patients from previous data (Table 1), 20 patients with ARDS (pressure overload) (¹¹) and 25 patients with morbid obesity (volume overload) (¹²). In those previous data, the PEP/RVET ratio was measured by simultaneous graphic display of the electrocardiograph, phonocardiograph, and PAP curves. The QS₂ was determined as the interval from the onset of the QRS complex on the electrocardiograph to the closure of the pulmonic valve, as reflected in the pulmonic component of the second heart sound. Right ventricular ejection time was the same in both methods, which was measured as the phase of systole from the beginning of the rapid upstroke of the PAP curve to the dicrotic notch. Pre-ejection period, the interval from the onset of ventricular depolarization to the beginning of ejection, was derived by subtraction of the RVET from the interval of QS₂ (Fig. 1). Fifteen measurements of the phases of systole were determined, and the average value of PEP/RVET ratio was taken.

In addition, we extended the application of estimated RVESP to use as a reliable estimate of RVESP. In the previous studies (^{11, 12}), the assessment of right ventricular function by RVSTIs was compared with that by right ventricular end-systolic pressure-volume relation, for which PA diastolic notch pressure was used as an estimate of RVESP, and there was a good correlation between the PEP/RVET ratio and the slope of right ventricular end-systolic pressure-volume relationship line. In the present study, we reassessed the slope of right ventricular end-systolic pressure-volume relationship line from previous data, using estimated RVESP, instead of PA diastolic notch pressure, and re-examined the correlation with the estimated PEP/RVET ratio.

Statistical analysis

Bland-Altman analysis was used to assess the agreement between two methods, the measured and the estimated (¹⁶). Also, regression analysis was used to compare the data by two different methods. The method of least squares was used for regression. Analysis of covariance was used to compare the slopes of multiple regression lines. To determine whether the fitted model of regression was correct, we examined the residuals against the fitted values (ŷ) (¹⁷). To detect a certain type of serial correlation, the correlation between observational errors, we used the Durbin-Watson test at the level 2α = 0.02 (¹⁸). All data including those in Table 1 were expressed as a mean ± 1 SD. A two-tailed P < 0.05 was considered significant.

RESULTS

The number of measurements of RVESP/PPASP ratio in each patient varied, depending on the number of clear PAP curves available in the record, and the mean was 8 ± 0.7. The mean RVESP/PPASP ratio with SD was 0.90 ± 0.008. The mean SD of each patient's measurements was 0.014 ± 0.006, which was very low. Data indicate that there was relatively little variation in each patient's measurement and among the patients.

Because a previous study (¹²) has shown that the correlation between PEP/RVET ratio and the slope of right ventricular end-systolic pressure-volume relationship line in patients with ARDS is not different from that in patients with morbid obesity, the previous two data (^{11, 12}) were combined. Bland-Altman analysis showed that there was a good agreement between the measured PEP/RVET ratio and the estimated as shown in Figure 2. The mean difference was 0.007, and the standard error of the bias (difference) was 0.0036. The standard error of the 95% limits of agreement was 0.0062. The 95% confidence interval for the upper limit of agreement was 0.0358 to 0.0545. The 95% confidence interval for the lower limit of agreement was −0.0531 to −0.0407. These intervals are reasonably narrow, suggesting a very close agreement of two methods. Also, there was a strong correlation between the measured PEP/RVET ratio and the estimated (R = 0.965, P < 0.0001). The correlation slope with standard error of estimate was 0.94 ± 0.023. When the residuals from the regression were plotted against the fitted value (ŷ), the residuals were randomly scattered, indicating that the errors were independent, and there was no serial correlation (Durbin-Watson d statistic was 1.997).

Fig. 2: Bland-Altman analysis for assessing agreement between the values by two methods, the estimated pre-ejection period (PEP)/RVET ratio using estimated electromechanical systole (QS₂) interval and the measured PEP/RVET ratio using measured QS₂ interval with phonocardiogram in patients with ARDS (open circles) and with morbid obesity (closed circles).

There was good agreement between the slope of right ventricular end-systolic pressure-volume relationship line determined with PA diastolic notch pressure and the slope determined with estimated RVESP as shown in Figure 3. The mean difference was 0.0061, and the standard error of the bias was 0.00438. The standard error of the 95% limits of agreement was 0.0075. The 95% confidence interval for the upper limit of agreement was 0.0562 to 0.0712. The 95% confidence interval for the lower limit of agreement was −0.059 to −0.044. These intervals are reasonably narrow, suggesting a very close agreement of two methods. Also, there was a strong correlation between two slopes with different estimates for RVESP (R = 0.97, P < 0.0001). The correlation slope with standard error of estimate was 1.042 ± 0.03.

Fig. 3: Bland-Altman analysis for assessing agreement between the values by two methods, the slope of right ventricular end-systolic pressure-volume relationship line using the estimated RVESP and the slope using PA dicrotic notch pressure in patients with ARDS (open circles) and with morbid obesity (closed circles).

There was a good inverse correlation between the estimated PEP/RVET ratio and the estimated slope of right ventricular end-systolic pressure-volume relationship line (with estimated RVESP) as shown in Figure 4. The correlation slope with as standard error of estimate was −0.592 ± 0.0723 (R = 0.664, P < 0.0001). When the residuals from the regression were plotted against the fitted value (ŷ), the residuals were randomly scattered, indicating that the errors were independent, and there was no serial correlation (Durbin-Watson d statistic was 2.132). When this correlation line between the estimated two values was compared with the correlation line between the measured PEP/RVET ratio and the measured slope of right ventricular end-systolic relationship line from the previous data (¹²), there was no difference between two correlation lines.

Fig. 4: A linear relationship between the estimated PEP/RVET ratio and the estimated slope of right ventricular end-systolic pressure-volume relationship (RV ESPVR) line ( R ² = 0.66, P < 0.0001) in patients with ARDS (open circles) and with morbid obesity (closed circles). PEP/RVET ratio = 0.72 - 0.59 × slope of RV ESPVR line.

DISCUSSION

Previous studies have shown that RVSTIs offer a measure of overall right ventricular systolic performance in patients with ARDS (¹¹), in patients with primary pulmonary hypertension (⁹), and in patients with morbid obesity (¹²). For the measurement of RVSTIs, the onset of right ventricular ejection and the end of right ventricular systole should be defined. In patients with an indwelling PA catheter, the onset of right ventricular ejection can be easily defined without using echocardiography. To define the end of right ventricular systole, however, phonocardiography is needed. When considering the left ventricle, ESP has been accurately predicted from the peak arterial systolic pressure (¹³). Likewise, if RVESP can be estimated from the PA systolic pressure, then the end of right ventricular systole can be defined as the time at which PAP reaches the RVESP. And then, RVSTIs can be measured without requiring phonocardiography. In this study, we tested this possibility by determining if the estimation of RVESP from the peak PA systolic pressure is consistent among patients with ARDS, and if the estimated time of right ventricular end systole could be applied to yield reliable estimates of RVSTIs in other groups of patients. The study documented that there was relatively little variation of RVESP/PPASP ratio among the patients and that there was a good agreement between two methods for the measurements of PEP/RVET ratio, the estimated and the measured. The data suggest that RVSTIs can be determined reliably using the estimated RVESP in patients with an indwelling PA catheter, without using phonocardiography or echocardiography.

Right ventricular systolic time intervals have been used to assess right ventricular systolic function in patients with primary pulmonary hypertension and in patients with congenital heart diseases. A previous study has shown that in primary pulmonary hypertension that there is a strong correlation between PEP/RVET ratio and PA diastolic pressure, suggesting that the higher PA resistance is associated with the worse right ventricular systolic function (⁸). Another study has demonstrated significant correlation between the PEP/RVET ratio and clinical severity of right ventricular dysfunction in patients with primary pulmonary hypertension (⁹). In patients with congenital transposition of great arteries, in which left ventricle empties into the lower resistance pulmonary circuit while the right ventricle ejects into the high resistance systemic circuit, PEP/RVET ratio has been found elevated as a result of right ventricular myocardial dysfunction, indicating that the right ventricle is poorly adapted to assume the function of systemic ventricle (¹⁰). As such, the elevated PEP/RVET ratio should be a reliable index of right ventricular dysfunction.

Some patients with ARDS develop right ventricular dysfunction (^{1, 3}). In right ventricular dysfunction, right ventricular dilatation and pressure overload lead a leftward shift of the interventricular septum, changing left ventricular geometry. Right ventricular dilatation also increases the constraining effect of the pericardium. These changes contribute to the low cardiac output state by decreasing left ventricular distensibility and preload. Right ventricular dysfunction impairs right ventricular filling and increases right ventricular pressures, which can lead to fluid retention and congestive hepatopathy (¹⁹). Right ventricular dysfunction also impairs distribution of pulmonary blood flow, leading to poor gas exchange (^{1, 3}). Therefore, it is imperative to detect those patients with right ventricular dysfunction, so that therapeutic measures such as inhaled nitric oxide (for right ventricular afterload reduction and for improvement of oxygenation), inotropics (²⁰), optimization of preload (²¹), and protective ventilation can be applied. A previous report has suggested that protective mode of ventilatory support may reduce the incidence of right ventricular dysfunction (²²). The measure to assess right ventricular function should be simple, easy, and accurate. Such an example is the use of RVSTIs.

If the estimated RVESP can be used to define right ventricular end-systolic pressure-volume relation instead of using PA diastolic notch pressure, one can avoid the problem with obtaining clear diastolic notch pressure reading. The study documented that there was a good agreement between two methods for the slope of right ventricular end-systolic pressure-volume relationship line, one with an estimated RVESP, the other with a pulmonary artery diastolic notch pressure, suggesting that the estimated RVESP in the present study is a reliable estimate of RVESP. Actually, PA diastolic notch pressure is a rough estimate of RVESP.

Previous studies have shown that thermodilution technique for the measurement of right ventricular ejection fraction has a tendency to underestimate ejection fraction and overestimate right ventricular volumes, when compared with three-dimensional echocardiography and magnetic resonance imaging technique (^{6, 23}). Because the validity of RVSTIs as a measure of right ventricular systolic function was based on comparison with right ventricular end-systolic pressure-volume relation for which right ventricular end-systolic volume was derived by thermodilution technique in previous studies (^{11, 12}), one may argue that the validity of RVSTIs is questionable. However, the accuracy of some of the data in those previous studies is questionable. In a previous study (²³), in which thermodilution technique was compared with three-dimensional echocardiography, right ventricular ejection fraction by thermodilution technique in patients without right ventricular dysfunction was very low (0.33) compared with that in other studies (0.45 and 0.47) (^{24, 25}). In a previous study (²⁴), right ventricular ejection fraction by thermodilution technique was 0.45 in patients without right ventricular dysfunction in whom right ventricular dysfunction was ruled out by right ventricular end-systolic pressure-volume relation despite the presence of pulmonary hypertension. This value of 0.45 in the presence of pulmonary hypertension is very close to the value (0.48) of normal right ventricular ejection fraction measured by three-dimensional echocardiography (²⁶). In the absence of right ventricular dysfunction, right ventricular ejection fraction measured by radionuclide technique in patients with pulmonary hypertension was 0.39 (⁴). Thus, in the presence of pulmonary hypertension, the value of right ventricular ejection fraction, 0.39 measured by thermodilution technique in the previous studies (^{12, 27}) appears to be accurate and not underestimated. Interestingly, there are considerable variations in normal values of right ventricular ejection fraction measured by three-dimensional echocardiography or magnetic resonance imaging techniques among previous studies. Right ventricular ejection fraction by three-dimensional echocardiography was 0.48 in a study (²⁶) and 0.55 in another study (²³) and, by magnetic resonance imaging technique, was 0.61 (⁶). These higher values of right ventricular ejection fraction may indicate that three-dimensional echocardiography or magnetic resonance imaging technique overestimates right ventricular ejection fraction. Nevertheless, there is a close agreement of the measured values of right ventricular ejection fraction in patients with right ventricular dysfunction between thermodilution technique (0.34) (¹²) and magnetic resonance imaging technique (0.32, 0.34) (^{6, 28}). Taken together, it is conceivable that RVSTIs offer a valid measure of right ventricular systolic function.

Transthoracic echocardiography with tissue Doppler imaging technique such as tricuspid annular systolic velocity has been used to assess right ventricular systolic function. However, the assessment of right ventricular function with tricuspid annular systolic velocity has limitations (^29-31). A particular limitation can be encountered in the cases of elevated right ventricular pressure as in patients with ARDS or with morbid obesity. A negative correlation between the mean PAP and tricuspid annular systolic velocity in a previous study (³²) suggests that a significantly elevated right ventricular pressure can mimic a depressed right ventricular function (²⁹). An animal study has shown that right ventricular myocardial acceleration during isovolumic contraction (right ventricular peak systolic velocity divided by the time interval from the onset of the wave during isovolumic contraction to the time at peak velocity of the wave) is a measurement of right ventricular contractile function that is unaffected by loading conditions, when compared with right ventricular end-systolic pressure-volume relation, but tricuspid annular systolic velocity during isovolumic contraction is load-dependent (³³). It is interesting to note that isovolumic myocardial acceleration in that study is heart rate-dependent. Thus, the assessment of right ventricular function by echocardiography with tissue Doppler imaging technique still has limitations.

For the patients with an indwelling PA catheter in place in critical care settings, the measurement of RVSTIs with PAP curve and electrocardiogram can be considered noninvasive. Other advantage of using RVSTIs includes a very low measurement error. For an accurate measurement of RVSTIs, it is important to define the time of right ventricular end systole accurately. In a previous study (³⁴), in which QS₂ was overestimated by measuring the intervals from QRS complex to the dicrotic notch of PAP curve, the PEP was erroneously prolonged. With the estimated RVESP, QS₂ can be determined easily and accurately.

In addition, it is interesting to note that the RVESP/PPASP ratio for the right ventricle is the same as the ratio of left ventricular ESP to peak arterial systolic pressure for the left ventricle (0.9) (¹³). This may be because, in patients with pulmonary hypertension complicating ARDS, pulmonary arterial system may have become a high-pressure system. If so, the RVESP/PPASP ratio may not be 0.9 in patients without pulmonary hypertension. Further study would be needed.

In conclusion, our data indicate that RVESP can be estimated from PPASP accurately and that the estimated time of right ventricular end systole, which is derived from the point of RVESP on the electrocardiographic curve without using phonocardiography, can be applied to yield reliable estimates of RVSTIs in patients with ARDS. Also, the estimated RVESP can be used to define right ventricular end-systolic pressure-volume relation instead of using PA diastolic notch pressure, avoiding the problem of obtaining clear diastolic notch pressure reading.