Abstract
Background The frequency and clinical correlates of global right ventricular (RV) dysfunction in patients treated with primary percutaneous coronary intervention for a first acute ST-elevation myocardial infarction (STEMI) without a coexisting RV infarction is not well known.
Materials and Methods One hundred seven consecutive patients underwent conventional echocardiography and pulsed-wave tissue Doppler imaging (TDI) within 72 hours after a successful primary percutaneous coronary intervention to assess their RV function. Global RV function was quantified with the RV myocardial performance index (MPI) by pulsed-wave TDI. An abnormal TDI-derived RV MPI was defined as greater than the upper reference limit of 0.55.
Results Global RV dysfunction was present in 18 (17%) of the 107 patients enrolled. The patients with global RV dysfunction had significantly higher glucose levels on admission (216 ± 102 vs 163 ± 86 mg/dL; P = 0.027), higher peak creatine kinase (4027 ± 2171 vs 2660 ± 1980 IU/L; P = 0.014), and more frequently had anterior infarcts (89% vs 58%; P = 0.016) than those without RV dysfunction. Patients with global RV dysfunction also had a significantly lower left ventricular (LV) ejection fraction (45.1 ± 10.8% vs 51.1 ± 9.7%; P = 0.021), a higher global wall motion score index (1.9 ± 0.3 vs 1.7 ± 0.4; P = 0.007), and greater LV MPI (0.65 ± 0.19 vs 0.47 ± 0.11; P = 0.001) than patients without. With the use of multivariate regression analysis, TDI-derived LV MPI (odds ratio [OR], 3.40; 95% confidence interval [CI], 1.20–9.67; P = 0.022), the ratio of transmitral peak early (E) to late diastolic filling (A) velocities (E/A ratio) (OR, 0.41; 95% CI, 0.18–0.92; P = 0.031), and admission plasma glucose level (OR, 1.01; 95% CI, 1.0–1.02; P = 0.039) were independently associated with the presence of global RV dysfunction.
Conclusions In patients with a first acute STEMI without an associated RV infarction, depressed global LV function reflected by increased TDI-derived LV MPI, a lower mitral E/A ratio, and a higher glucose level on admission are independent correlates of early global RV dysfunction. Routine assessment of global RV function should be implemented in patients with STEMI with these characteristics.
Right ventricular (RV) involvement after an acute left ventricular (LV) myocardial infarction (MI) has been shown to be associated with higher morbidity and mortality.1,2The prevalence of RV involvement in acute LVMI has been reported to range from 50% to 80% in postmortem and animal studies but is frequently underestimated in the clinical setting owing to the diagnostic limitations of the electrocardiogram (ECG) and echocardiography.3–5In fact, extensive work has been done to evaluate RV dysfunction in patients with ECG- and/or echocardiography-documented RV infarction.1,2,6,7However, still remaining is the question of how the RV function is altered in patients with acute LVMI but without the accompanying classic diagnostic criteria of RV infarction, especially in the era of primary percutaneous coronary intervention (PCI) as the criterion standard treatment for these patients. Because recent studies have shown that RV function may provide incremental prognostic information in patients with acute ST segment elevation MI (STEMI) treated with primary PCI,8,9it might be useful in clinical practice to identify the factors associated with RV dysfunction during the events.
Quantitative assessment of RV function is often difficult using the various noninvasive imaging modalities owing to the inherently complex geometry of the right ventricle. Cardiac magnetic resonance imaging has been proposed to be a superior imaging tool to detect RV involvement in a STEMI,8,10,11but it is time consuming and relatively expensive. Echocardiography remains the most commonly used technique for RV function measurement in clinical practice because of its widespread availability. Initially presented by Tei in 1996, the myocardial performance index (MPI) of RV based on conventional Doppler echocardiography has proven useful in the evaluation of RV function and in prognostication for patients with primary pulmonary hypertension, congenital heart disease, and systolic heart failure.12–16It has been recommended as one of the initial quantitative measurements of RV function in daily practice according to the guidelines of the American Society of Echocardiography.17An RV MPI provides a combined RV systolic and diastolic function assessment and may be a sensitive tool for detecting “occult” RV dysfunction in acute LVMI, as the time intervals of RV contraction, ejection, and relaxation are all evaluated. Nevertheless, the conventional Doppler-derived MPI has one important limitation, namely, that the interval between the end and onset of tricuspid inflow and the ejection time are measured sequentially and not on the same cardiac cycle. In contrast, tissue Doppler imaging (TDI) can simultaneously measure these important time intervals on the same cardiac cycle, making the TDI-derived MPI superior in the estimation of global RV function.18Accordingly, our aim in this study was to use the TDI-derived RV MPI as a measurement of global RV function and investigate the frequency and clinical correlates of global RV dysfunction in patients with a first acute STEMI, without a coexistent RV infarction, who have been successfully treated by primary PCI.
MATERIALS AND METHODS
Patients and Protocols
Between December 2007 and June 2011, we studied 214 patients selected from 229 consecutive patients with acute STEMI who had received emergency coronary angiography in the catheterization laboratory of Buddhist Tzu Chi General Hospital, Taipei Branch, Taiwan. The study inclusion criteria were as follows: (1) confirmed first acute STEMI, based on the presence of characteristic chest pain and a new ST-segment elevation at the J point in 2 or more contiguous leads with the cutoff points of 0.2 mV or greater in leads V1, V2, or V3 and 0.1 mV or greater in the other leads; (2) successful primary coronary angioplasty (defined as achievement of thrombolysis in myocardial infarction flow grade 3 and residual stenosis of the infarct-related artery (IRA) <30%) performed within 12 hours of the onset of symptoms; and (3) the procedure of angioplasty was confined to the IRA only. There was no age limit. All patients who fulfilled the inclusion criteria (n = 214) received 2D echocardiography and pulsed-wave TDI studies within 72 hours after undergoing primary PCI.
The exclusion criteria were the following: (1) the presence of RV infarction, which was defined by an ST-segment elevation of 0.1 mV or greater in leads V4R on electrocardiograms at presentation (before PCI), and/or any RV wall motion abnormality detected by echocardiography with visual assessment (n = 23); (2) shock from any cause (n = 18); (3) presence of atrial fibrillation or bundle branch block (n = 10); (4) previous history of MI or coronary revascularization (n = 17); (5) moderate or severe valvular heart disease (n = 4); (6) presence of a coexisting clinical condition that might affect RV function, ]including pericardial disease, chronic lung disease, pulmonary hypertension, or connective tissue disorder (n = 9); and (7) consent refusal (n = 11), yielding a total of 122 patients selected for the study. Loading doses of dual antiplatelet agents (aspirin and clopidogrel) and unfractionated heparin were administered to all the patients before undergoing PCI.
Clinical history and baseline laboratory data were collected for all patients. Systemic hypertension was defined as a history of mean systolic blood pressure of 140 mm Hg or higher or diastolic blood pressure of 90 mm Hg or higher measured on 3 separate occasions with or without treatment before admission. Patients were defined as having diabetes mellitus if they had a previous history or current diagnosis of diabetes. Dyslipidemia was defined as a serum total cholesterol level greater than 200 mg/dL or a triglyceride level greater than 150 mg/dL, or current treatment with lipid-lowering medications. Current smoking was defined as having active smoking over the past 3 months. Significantly obstructive non–infarct-related coronary arteries were defined as 50% or greater reduction of the internal diameter of the non–infarct-related vessels. The institutional review board approved the protocol, and informed consent was obtained from each patient.
Echocardiography and Pulsed-Wave TDI analysis
Using a Philips SONOS 7500 (Agilent Technologies, Andover, MA) phased-array system equipped with TDI technology, we conducted a comprehensive 2D Doppler echocardiography and pulsed-wave TDI while the patients were lying in the partial left lateral decubitus position during either shallow respiration or a breath hold. All the image acquirement and analyses were performed by the same experienced echocardiologist who was blind to the patients’ clinical data. The LV dimensional measurements were obtained from an M-mode recording. The LV ejection fraction (LVEF) and left atrial volume were estimated using the modified Simpson method according to the recommendations of the American Society of Echocardiography.19The volume indexes were derived by dividing the volumes by the body surface area at each time point. A 16-segment model for LV segmentation was used to evaluate regional wall motion abnormalities.19The wall motion of each segment was scored from 1 (normal) to 4 (dyskinetic), and the global wall motion score index (WMSI) was calculated as a sum of all the scores divided by the number of segments visualized. In the apical 4-chamber view, the RV end-diastolic dimension was assessed at the mid-cavity of the right ventricle, the RV fractional area change (RVFAC) was calculated as the RV end-diastolic area—the end-systolic area / the RV end-diastolic area, and the right atrial area was estimated by planimetry at the end of ventricular systole as recommended.17The tricuspid annular plane systolic excursion (TAPSE) was measured from the apical 4-chamber view at the RV free wall level as previously described.17The transmitral and transtricuspid Doppler flow velocities were recorded from an apical 4-chamber view with the sample volume placed between the tips of the mitral and tricuspid valves, respectively, and the peak early filling velocity (E), peak atrial velocity (A), E/A ratio, and E-wave deceleration time (DT) corrected for the heart rate were also measured.
Pulsed-wave TDI images were acquired using transducer frequencies of 2.5 to 4.0 MHz, adjusting the Nyquist limit and gain setting to avoid aliasing and minimize noise. From the standard apical 4-chamber view, a 5.2- mm sample volume was placed at the lateral tricuspid annulus, mitral septal, and mitral lateral annular sites to obtain the spectral pulsed tissue Doppler data. Care was taken to make the Doppler beam as parallel as possible to the direction of the tricuspid and mitral annular motion. The parameters measured from the septal and lateral sides of the mitral annulus were averaged for further analysis. Three cardiac cycles were averaged for each TDI measurement. A diagram of the tricuspid annular TDI measurement can be seen in Figure 1. Peak systolic annular velocity (S m), early diastolic annular velocity (E m), and late diastolic annular velocity (A m) of both ventricles were measured offline; and the derived indices, the E m/A m and E/E m ratios, were computed. Myocardial relaxation time (RT; measured at the time interval from the end of S m to the onset of E m), isovolumic contraction time (measured at the time interval from the end of the A m to the beginning of the S m), and ejection time (measured from the onset to the end of S m) of both ventricles were calculated to obtain the TDI-derived LV and RV MPI with the formula: RT + isovolumic contraction time / ejection time.13The upper reference limit of 0.55 was used as the cut-off value for an abnormal RV MPI based on the recommendations of the American Society of Echocardiography.17
Statistical Analysis
The statistical analyses were performed using STATA software version 10.0 (Stata Corp, College Park, TX). Continuous data are presented as mean ± SD and were compared using the Student t test if they were normally distributed (as verified by the Skewness-Kurtosis test). Wilcoxon rank sum test was used to compare the continuous variables that were not normally distributed. Categorical variables are expressed as percentages and numbers and were analyzed using the χ2 or Fisher exact test as appropriate. Multivariate logistic regression analyses were applied to identify the baseline clinical and echocardiographic correlates of global RV dysfunction. The potential parameters on the basis of known clinical relevance and the variables showing P ⩽ 0.05 between the patients with and without global RV dysfunction were selected to enter into the multivariate regression models. The correlations between the TDI-derived RV MPI and other RV functional parameters (TAPSE, tricuspid S m velocity, and RVFAC) were also explored using the Pearson or Spearman test whenever appropriate. P < 0.05 was considered significant for all statistical tests.
Reproducibility of the TDI-derived MPI was analyzed with repeated measurements in 20 randomly selected individuals by the original echocardiologist in a blind fashion and by a second experienced physician on different occasions. Interobserver and intraobserver variability for the TDI-derived LV and RV MPI were expressed as correlation coefficients between the measurements of the 2 investigators and 2 sets of observations using linear regression analysis.
RESULTS
Patients’ Characteristics
Of the 122 patients initially recruited for the study, 15 patients were further excluded because of inadequate echocardiographic image quality, so that a total of 107 patients represented the final study population. The presence of an abnormal RV MPI was used to dichotomize the patients: group 1 (n = 18 [17% of the study population]) was composed of patients with impaired RV global function (RV MPI > 0.55) and group 2 (n = 89; 83% of the study population) included patients with normal RV global function (RV MPI ⩽ 0.55). The median time delay from PCI to echocardiographic evaluation was 30 hours (25th–75th percentiles, 18–46 hours), which did not differ between the 2 groups (35 hours in group 1 vs 29 hours in group 2; P = 0.42). The baseline clinical characteristics of the study population are included in Table 1. There were no statistically significant differences between the 2 groups in age, sex, body mass index, smoking status, proportions of hypertension and dyslipidemia, systolic and diastolic blood pressure, Killip class, door-to-balloon time or medication. The patients in group 1 had a slightly higher proportion of diabetes mellitus and significantly higher glucose levels on admission than those in group 2. Heart rates were significantly higher in the patients in group 1 than those in group 2. The patients in group 1 were more likely to have anterior wall infarction (89% vs 58%; P = 0.016) and higher peak creatine kinase levels (4027 ± 2171 vs 2660 ± 1980; P = 0.014). An abnormal TDI-derived RV MPI was present in 24% of the patients with anterior wall infarction and only 5% in the patients with inferior/posterior infarcts. There were 55 patients with significant right coronary artery disease; 9 patients (50%) in group 1, and 46 patients (52%) in group 2. No significant differences in the percentages of stenotic coronary arteries and right coronary artery dominance were observed between the 2 groups.
Conventional and Tissue Doppler Echocardiographic Characteristics
Overall, the patients in group 1 had a significantly lower LVEF, higher global WMSI, and higher LV MPI than those in group 2, suggesting that the myocardium was more extensively jeopardized during the index event in the patients in group 1 (Table 2). The peak mitral E/A ratio, mitral annular E m velocity, and mitral annular E m/A m ratio were significantly lower and the mitral annular RT was significantly longer in group 1 than in group 2, reflecting that some of parameters of LV diastolic function were different between the 2 groups. Only 4 (4%) of the 107 patients had the mitral E/A ratios greater than 1.5, 6 patients (6%) had a mitral DT of less than 160 milliseconds, and 4 patients (4%) had left atrial volume index greater than 34 mL/m2. As for the RV functional parameters other than the RV MPI (Table 3), the peak tricuspid E/A ratio was significantly lower and the tricuspid annular RT was significantly longer in the patients in group 1 as compared to those in group 2. The peak tricuspid annular systolic velocity was significantly lower in group 2 but still within the reference range.17There were no significant correlations between the TDI-derived RV MPI with TAPSE, tricuspid S m velocity, and RVFAC (r = −0.185, P = 0.115; r = 0.053, P = 0.590; r = −0.184, P = 0.053, respectively).
Correlates of Early Global RV Dysfunction
Multivariate logistic regression analyses were conducted with the following variables: age, hypertension, diabetes mellitus, peak creatine kinase level, admission glucose level, heart rate, door-to-balloon time, anterior wall infarction, LVEF, WMSI, TDI-derived LV MPI, mitral E/A ratio, mitral annular E m velocity, mitral annular E m/A m ratio, and mitral RT. Diabetes mellitus was not retained in the model owing to potential interaction with plasma glucose levels. Patients with increased TDI-derived LV MPI as well as higher admission glucose levels were more likely to present with global RV dysfunction, and a higher mitral E/A ratio was significantly less likely to present with global RV dysfunction (Table 4).
Reproducibility of Measurement
The correlation coefficient (r) of the intraobserver variability was 1.04 for the TDI-derived LV MPI (standard error [SE], 0.05; P < 0.001) and 0.89 for the TDI-derived RV MPI (SE = 0.06; P < 0.001). For the interobserver variability analysis, r was 0.92 for the TDI-derived LV MPI (SE = 0.05; P < 0.001) and 0.79 for the TDI-derived RV MPI (SE = 0.08, P < 0.001).
DISCUSSION
In this observational study of 107 patients treated with primary PCI for a first acute STEMI without coexistent RV infarction, we demonstrated that global RV dysfunction, as reflected by an increased TDI-derived RV MPI, was present in 17% of these patients. The individuals with early global RV dysfunction usually had a larger infarct size, lower LVEF, higher TDI-derived LV MPI, more depressed LV diastolic function, greater proportions of diabetes mellitus, and higher plasma glucose levels on admission; they also presented more frequently with anterior wall infarctions. Moreover, TDI-derived LV MPI, mitral E/A ratio, and admission plasma glucose levels are the independent factors associated with global RV dysfunction in this prespecified population. These data suggest that global RV dysfunction may be present even if after successful reperfusion of an acute STEMI episode and does not exclusively occur in patients with inferior-posterior LV infarcts associated with clinically documented RV infarction. Because RV involvement may provide prognostic information in patients treated with primary PCI,8,9routine RV function assessment should be implemented in these patients, particularly in those with depressed global LV function (reflected by increased TDI-derived LV MPI), a lower mitral E/A ratio, and a higher glucose level on admission.
Right Ventricular Involvement After Acute Myocardial Infarction
Because robust data have shown that the prognosis of patients with acute MI is mostly related to LV function, the primary focus for clinicians is generally the left ventricle. The relevance of RV function, on the other hand, is often overlooked and is less defined in post-MI patients. Most of the earlier investigations of RV involvement after an acute MI included patients with documented RV infarction and heterogeneous reperfusion status.2,20,21In the echocardiographic substudy of the Survival And Ventricular Enlargement (SAVE) trial,2RV dysfunction, as assessed by RV fractional area change, was found in approximately one fifth of the post-MI patients with LV dysfunction (LVEF ⩽40%). Popescu et al.20evaluated the patterns of RV functional recovery using TAPSE in 500 patients with acute MI (GISSI-3 Echo substudy) and also found that early (24–48 hours) RV systolic function was significantly lower in patients with LVEF less than 45% than in patients with LVEF of 45% or greater. In the contemporary era of primary PCI as the criterion standard for STEMI management, the degree of LV dysfunction is generally mild, and data are scarce on the prevalence of RV involvement in this currently growing population of post-MI patients.8,11Masci et al.11characterized the patterns of early post-infarction RV ischemic injury and its impact on RV function in patients with acute STEMI using comprehensive cardiovascular magnetic resonance (CMR) images. They found that RV involvement is frequent (approximately one-third to one-half of the patients) in patients without clinical evidence of hemodynamic RV compromise. In a study of 50 patients with a first STEMI who underwent CMR immediately after successful PCI, Jensen et al.8found that RV involvement was detected in 27 patients (54%). However, as mentioned previously, these studies did not exclude patients with documented RV infarction. The actual occurrence rate of global RV dysfunction in patients with acute STEMI without coexistent RV infarction is not well known. Our study exclusively enrolled patients with acute STEMI without concomitant RV infarction and showed that impaired global RV function is present in 17% of this prespecified cohort, even if they underwent a successful primary PCI and had relatively preserved LV systolic function. Because strict exclusion criteria and a different methodology for evaluating RV function were used, both may account for the lower incidence of RV involvement than was reported in earlier studies.
Notably, global RV dysfunction occurred in a considerable portion (24%) of the patients with anterior LV infarcts, and more than 80% of our patients with global RV dysfunction presented with anterior wall infarction. In a recent CMR imaging study of Masci et al.,11they found that RV involvement was present in up to 33% of those patients with reperfused anterior wall STEMI. Although RV dysfunction has been demonstrated in patients with anterior wall STEMI when compared with healthy controls,21,22whether anterior infarcts independently correlated with global RV dysfunction is not well investigated in the previous works. In the univariate analysis of our study (data not shown), anterior wall infarction was a significant factor correlated with global RV dysfunction (odds ratio, 5.69; 95% confidence interval, 1.23–26.3; P = 0.026). However, the infarct location (anterior wall or not) is no longer an independent correlate when adjusted for the other confounding variables including the parameters of LV global and diastolic function in the multivariate analyses. Therefore, we speculate that the effect of an acute anterior wall LVMI on RV functional changes may be still dependent on the LV function.
Correlates of Postinfarction RV Dysfunction
In the multivariate analyses, we found that global RV dysfunction is significantly related to the echocardiographic parameters of LV function. As we have known, the 2 ventricles of the heart are not only interconnected anatomically but are also functionally dependent.23Alteration of the LV function may impair RV function on the systole as well as on the diastole, and vice versa. It is not surprising, therefore, that a more depressed LV function is likely to result in a secondary impairment of RV performance via ventricular interaction.24In the GISSI-3 Echo substudy, it was shown that RV function deteriorated in patients with acute MI; as the LVEF recovered over time, RV function improvement was also observed.20Moller et al.25investigated the serial changes in RV function after acute MI with a conventional Doppler-derived RV MPI. They found that the RV MPI correlated significantly with the LV MPI and was significantly higher in patients with shortening of mitral E-wave DT that was presumed to be related to ventricular interaction. In our study, global RV dysfunction was also significantly associated with the TDI-derived LV MPI, which is in agreement with the aforementioned published studies that included patients with RV infarction and with heterogeneous reperfusion status. The result further supports that although in the patients with reperfused STEMI without concomitant RV infarction, ventricular interdependence may still, at least, partially contribute to the occurrence of global RV dysfunction.
The inverse relation of the mitral E/A ratio with the TDI-derived RV MPI is somewhat intriguing and cannot be fully elucidated in the present study. As we have known, the LV filling pattern usually has a U-shape relationship with LV diastolic function, which may affect the RV function via the ventricular interaction. However, the mitral E/A ratio should not be used universally to infer the presence and the grade of LV diastolic function owing to variable mean left atrial pressures in some patients.26Because the mean mitral E/A ratio of our patients with global RV dysfunction was 0.73 (range, 0.43–1.17), the relationship between the mitral E/A ratio and global RV function may be interpreted by ventricular diastolic interaction only in this range.
In addition to the echocardiographic parameters, the plasma glucose level on admission was shown to be an independent, albeit weak, predictor of global RV dysfunction in the present study, even after adjusting for the LV functional parameters and the infarct size in the multivariate analyses. Elevated plasma glucose is a common feature during the early hours after an acute MI in patients with or without diabetes.27,28Admission hyperglycemia, regardless of diabetic status, has been shown to be associated with larger infarct size and adverse outcomes both in the prethrombolytic therapy era and in patients treated with PCI.29–32To our knowledge, this is the first study that has demonstrated the relationship between the admission plasma glucose level and global RV function after an acute LVMI in the era of primary PCI. However, the mechanism(s) of the effect of admission glucose level on RV function cannot be fully addressed in this observational study. Acute hyperglycemia may attenuate endothelial function, impair coronary microvascular function, activate inflammatory and blood coagulation responses, and increase myocyte apoptosis that exaggerates LV remodeling, all of which may increase the infarct size and cause deterioration of the LV function.32Additional well-designed studies need to explore (1) whether similar effects of stress hyperglycemia on the LV may simultaneously affect the RV, and (2) whether the association of admission hyperglycemia with poor outcomes could be, in part, causally related to global RV dysfunction.
Limitations
Our study has several potential limitations. First, we used TDI-derived RV MPI as the quantitative measurement of RV function rather than used a more novel technique such as RV strain with speckle-tracking analysis, as there is no consensus about the best echocardiographic quantification of RV function at the time of the study, there are no normal limits for the RV strain, and facilities for strain study were not available at our institution. The results of the study, therefore, might have changed if a different modality of RV function measurement had been chosen. Second, because RV functional recovery may occur over time after an acute MI episode,20we could not exclude the possibility that different timing of the assessment would have resulted in different findings. However, the timing of the examinations did not indicate significant differences between the 2 study groups. Third, most of the patients were in Killip class 1 or 2, and the patients with cardiogenic shock were excluded. Our results, therefore, cannot be simply extrapolated to the full spectrum of patients with acute STEMI. Finally, the lack of clinical follow-up data makes it difficult to address the clinical implications of subtle global RV dysfunction specifically.
CONCLUSIONS
In patients with a first acute STEMI without associated RV infarction who have undergone a successful primary PCI, global RV dysfunction is not uncommon and did not exclusively occur in patients with inferior-posterior wall infarction. Depressed global LV function, a lower mitral E/A ratio, and a higher plasma glucose level on admission are independent correlates of early global RV dysfunction. Although ventricular interdependence may partly explain the global RV dysfunction after an acute reperfused MI, further studies are necessary to explore other possible mechanisms that may contribute to these findings and evaluate the impact of early global RV dysfunction on prognosis, all of which might refine medical management beyond the current standards for this group of patients.