Low-Dose Dobutamine Stress Echocardiography to Assess Left Ventricular Contractile Reserve for Cardiac Resynchronization Therapy: Data From the Low-Dose Dobutamine Stress Echocardiography to Predict Cardiac Resynchronization Therapy Response (LODO-CRT) Trial
Saverio Iacopino, MD, Sant’Anna Hospital, Viale Pio X, 111, 88100 Catanzaro, Italy
Cardiac resynchronization therapy (CRT) is an effective methodology indicated in selected heart failure patients. Identifying responders to the therapy is still challenging. Most studies report that at least 30% of the patients are nonresponders. Baseline characteristics of the Low-Dose Dobutamine Stress Echocardiography to Predict Cardiac Resynchronization Therapy Response (LODO-CRT) trial population are presented. The study investigates dobutamine stress echocardiography’s role in predicting CRT response. Two hundred seventy-one CRT candidates were studied. Mean age was 67±10 years, 69% were male, 96% had New York Heart Association class III disease, and 39% had heart failure of ischemic etiology. Mean QRS and left ventricular ejection fraction were 146±24 ms and 26%±6%, respectively. Seventy-seven percent of participants showed contractile reserve. Left ventricular end-diastolic volume was shown to be independently associated with contractile reserve presence. In particular, more dilated ventricles are associated with a lower chance of having contractile reserve. The LODO-CRT trial enrolled a cohort of patients fulfilling criteria for CRT. Dobutamine stress echocardiography was highly feasible and safe in this population. Contractile reserve was associated with healthier ventricles. Congest Heart Fail. 2010;16:104–110. © 2010 Wiley Periodicals, Inc.
Cardiac resynchronization therapy (CRT) is an effective therapeutic option for heart failure (HF), since it improves New York Heart Association (NYHA) functional class, left ventricular ejection fraction (LVEF), exercise tolerance, and quality of life and reduces hospitalizations, mortality, and costs.1–5 Currently, American College of Cardiology/American Heart Association/European Society of Cardiology guidelines rate CRT in class Ia for HF patients with NYHA functional class III or IV, QRS ≥120 ms, and LVEF ≤35%.6,7
Despite these favorable results, the identification of potential responders to the therapy is still a challenging task. Responders’ proportion appears to be strictly dependent on the definition of response. Studies report that a “clinical” response is observed in 60% to 75% of patients8,9 with current CRT indications, but according to several authors, this rate drops to 50% if objective parameters such as left ventricular (LV) reverse remodeling or oxygen consumption are used to define the response.10
The presence of LV dyssynchrony before implantation has often been associated with positive response to CRT. Over the last 10 years, many different methods have been proposed to estimate LV dyssynchrony by means of echocardiographic indexes, measured both with traditional echocardiography and with tissue Doppler imaging.11–13 Results from single-center studies have been promising, but the Predictors of Response to CRT (PROSPECT) study14 raised substantial concerns about the reproducibility of these methods and concluded that no single echocardiographic measure of dyssynchrony may be recommended to improve patient selection for CRT beyond current guidelines. Up to now, we could not rely on dyssynchrony indexes alone in selecting candidates for CRT in an attempt to increase the overall proportion of responders.
Other factors must be taken into account when selecting a patient for implantation, such as etiology15 or the extent and position of the scar.16,17 Moreover, lead placement may be a key factor in determining the response.18
In search for further factors associated with response to CRT, we hypothesized that LV contractile reserve (LVCR) presence can be of great importance to increase the chances of positive response to the therapy. Early findings claim a strong relationship between LVCR and response to CRT,19–21 bringing preliminary evidence that a necessary condition to gain improvement in LV function is the availability of a relevant amount of viable myocardial cells. In particular, from these preliminary works it seems that in advanced stages of HF, the presence of fibrotic or scar tissue may inhibit effective biventricular stimulation. LVCR presence can be easily detected in patients with HF of any etiology using dobutamine stress echocardiography (DSE), which was shown to be feasible and safe.22–25
The Low-Dose Dobutamine Stress Echocardiography to Predict Cardiac Resynchronization Therapy Response (LODO-CRT) trial is a prospective, multicenter, observational study aimed at evaluating the role of DSE in selecting suitable candidates for CRT. We hypothesize that patients showing presence of acute LVCR on DSE are more likely to positively respond to CRT at follow-up.26 In this article, the baseline clinical and echocardiographic parameters of the population enrolled in the study are presented.
LODO-CRT Trial Design. A detailed description of the rationale and methods of the LODO-CRT trial has been previously published.26 Briefly, this study was designed to assess the predictive value of LVCR, determined by means of DSE (defined as an LVEF increase >5 units), in predicting CRT response.
The study is designed in 2 sequential phases in which instrumental and clinical response to CRT is evaluated. The primary end point of phase 1 is a reduction of the LV end-systolic volume of at least 10% at 6-month follow-up with respect to the value at baseline, while in phase 2 the primary combined end point is LV end-systolic volume reduction and improvement in clinical status, assessed via a clinical composite score at 12-month follow-up. Predictive value of LVCR will also be compared to other measures, such as LV dyssynchrony measures, through adjusted multivariable analysis.
Inclusion and Exclusion Criteria. LODO-CRT26 included patients for whom CRT is indicated according to current guidelines, which prescribe resynchronization for those with NYHA functional class III or IV disease, despite optimized treatment for HF and diuretics administration, who have an LVEF ≤35% with evidence of LV dilatation (LV end-diastolic dimension ≥55 mm) and a QRS interval ≥120 ms in at least 2 electrocardiographic leads. All patients provided written informed consent before enrollment. Patients with permanent or persistent atrial fibrillation or flutter were excluded.
Patients with myocardial infarction, revascularization, or other major cardiovascular events in the 6 weeks prior to randomization were excluded. Patients who had previously received an implantable pulse generator and showed a documented pacing percentage >10% were excluded.
Patient Subgroups. Data are presented separately for the subgroups of patients with and without LVCR according to DSE and, with regard to DSE test outcome, for ischemic and nonischemic patients.
Study Procedures. At baseline, complete assessment of patient status was performed, including inclusion/exclusion criteria verification, clinical and anamnestic evaluation, echocardiographic examination with evaluation of LV volumes and LVEF, mitral regurgitation grade, DSE, and left interventricular and intraventricular dyssynchrony estimation. Afterward, a CRT device was implanted.
LVCR Assessment. The DSE test was conducted as follows. Inotrope infusion was initiated at a rate of 5 μg/kg/min. The infusion rate was then increased by 5 μg/kg/min in 5-minute steps. Maximum dosage allowed was 20 μg/kg/min.27 The test was interrupted in case of arrhythmias or occurrence of any complications or adverse effects. Measurements taken at each infusion level (ie, end-diastolic and end-systolic volumes) were assessed offline by means of the 2-dimensional biplane disk method. An absolute increase in LVEF >5 points with respect to baseline value demonstrated the presence of LVCR and was considered a positive response to the test.23
Echocardiographic Evaluation of Dyssynchrony and Cardiac Function. A widely accepted measure28 of interventricular dyssynchrony was taken by measuring the interventricular delay (ie, the difference between aortic and pulmonary pre-ejection time). Moreover, the presence of restrictive filling pattern and nonrestrictive filling pattern were distinguished on the basis of the E/A ratio and E-wave deceleration time values.29,30
Comparison With Other Trials. Baseline data of the enrolled population were compared with those from the Cardiac Resynchronization in Heart Failure (CARE-HF)3 Study and the PROSPECT14 study, which both enrolled HF patients with similar conditions and assessed the predictive value of several baseline clinical and echocardiographic characteristics.
Statistics. Descriptive statistics were applied as appropriate according to the distribution of variables. Categorical data were reported as incidence (percentage), and noncategorical data were reported as mean/median and standard deviation. Differences between mean data were compared by a t-test for Gaussian variables and by Mann–Whitney or Wilcoxon nonparametric test for non-Gaussian variables, for independent or paired samples, respectively. Differences in proportions were compared by a chi-square analysis or Fisher’s exact test as appropriate. Univariable Cox regression analysis was used for the analysis of the association between other baseline measures and LVCR presence. All variables associated to a P value <.1 at univariable analysis were entered into the multivariable Cox regression analysis. All analyses were conducted in SPSS version 12.0 (SPSS, Inc. Chicago, IL).
From July 2006 to April 2008, 297 patients were enrolled in the study. Twenty-six patients were excluded due to implant failure (7 patients) or due to inadequate or missing echocardiographic information (19 patients). Baseline characteristics of the remaining 271 patients are presented hereinafter (Table I).
Table I. Baseline Characteristics of the Overall Population
|Male, No. (%)||188 (69)|
|NYHA class III, No. (%)||259 (96)|
|Months from HF symptoms’ first show||23±44|
|Ischemic etiology, No. (%)||106 (39)|
|Hypertension, No. (%)||168 (62)|
|Heart rate, bpm||67±13|
|Systolic blood pressure, mm Hg||71±10|
|Diastolic blood pressure, mm Hg||110±15|
|Previous MI, No. (%)||92 (34)|
|Valvular disease, No. (%)||50 (16)|
|QRS duration, ms||146±24|
|Left bundle branch block, No. (%)||206 (76)|
|Previous atrial arrhythmias, No. (%)||41 (15)|
|β-Blockers, %||76 |
|ACE inhibitors, %||77 |
|Spironolactone, %||51 |
|Diuretics, %||96 |
|LV end-diastolic diameter, mm||67±9|
|LV end-systolic diameter, mm||56±10|
|LV end-diastolic volume, mL||207±78|
|LV end-systolic volume, mL||153±64|
|LV ejection fraction, %||26±6|
|Interventricular delay, No. (%)||126 (46)|
|E-A duration, ms||387±138|
|E-wave deceleration time, ms||166±91|
|RFP presence, No. (%)||79 (29)|
|Severe mitral regurgitation (grade 3–4)||67 (25)|
Demographics, Clinical Assessment, and Medications. The mean age of the patients was 67±10 years. One hundred eighty-eight (69%) patients were men and 83 (31%) were women. Investigators reported ischemic HF etiology in 106 (39%) patients. HF symptoms in the study population were noticed on the average from 23 months.
Among these patients, a history of previous myocardial infarction was documented in 92 (34%); 45 (17%) of those were treated with a stent and 31 (11%) with a coronary artery bypass graft. Underlying valvular disease was reported in 50 (16%) patients.
Hypertension was the most commonly reported concomitant condition (62%), while a past history of atrial arrhythmias was detected in 15% of the enrolled patients. The vast majority of patients were reported to be in NYHA functional class III (96%).
Ninety-six percent of patients were receiving diuretics, 51% were taking spironolactone, 77% an angiotensin-converting enzyme inhibitor or angiotensin receptor blocker, and 76% a β-blocker. Nitrates, statins, anticoagulants, and aspirin were also commonly prescribed.
Electrocardiographic and Echocardiographic Assessment. Enrolled patients had a mean QRS duration of 146±24 ms (left bundle branch block was present in 76% of patients) and a depressed LVEF of 26%±6%. A complete echocardiographic assessment was performed according to the study protocol. Mean aortic pre-ejection time was 143±40 ms. Of interest, 126 (46%) patients showed presence of interventricular delay. Moreover, a presence of restrictive pattern due to a diastolic dysfunction was reported in 79 (29%) patients. The estimated mean mitral regurgitation degree was 1.8±1.0, and 67 (25%) patients presented moderate or severe mitral regurgitation.
Dobutamine Stress Echocardiography. DSE was feasible without complications in 268 of 271 patients (99%). The test was interrupted in 3 patients due to ventricular arrhythmia onset, all of which self-terminated.
The measured mean LVEF at rest was 26%±6%, while at peak stress LVEF increased to 36%±9%. According to the predefined cutoff, 206 (77%) patients showed presence of LVCR. Reported maximal heart rate during test was 97±23 bpm.
Differences Between LVCR and Non-LVCR Patients. LVEF was not significantly different in the 2 groups of patients with and without LVCR (26%±5% and 27%±6%, respectively; P=.433). Baseline characteristics of the 2 groups are reported in Table II and Table III. No significant differences in clinical baseline characteristics could be detected between the groups.
Table II. Baseline Characteristics of Patients Without Contractile Reserve (No LVCR) and Patients With Contractile Reserve (LVCR)
|Male, No. (%)|| 40 (66)||144 (70)||.488|
|NYHA class III, No. (%)||57 (92)||199 (97)||.119|
|Ischemic etiology, No. (%)||25 (40)||80 (39)||.833|
|QRS duration, ms||150±23||145±24||.071|
|Hypertension, No. (%)||37 (59)||131 (63)||.571|
|Previous MI, No. (%)||25 (41)||66 (32)||.226|
|Left bundle branch block, No. (%)||47 (77)||155 (76)||.805|
|Previous atrial arrhythmias, No. (%)||15 (25)||26 (13)||.026|
Table III. Echocardiographic Parameters for Patients Without Contractile Reserve (No LVCR) and Patients With Contractile Reserve (LVCR)
|LV end-diastolic diameter, mm||71±9||66±8||.001|
|LV end-systolic diameter, mm||59±10||55±9||.005|
|LV end-diastolic volume, mL||237±91||197±72||.001|
|LV end-systolic volume, mL||178±74||145±59||.001|
|LV ejection fraction, %||26±5||27±6||.433|
|Interventricular delay presence, No. (%)||36 (58)||89 (43)||.040|
|E-A duration, ms||390±161||386±133||.871|
|E-wave deceleration time, ms||133±65||171±82||.024|
|RFP presence, No. (%)||26 (42)||52 (25)||.011|
|Severe mitral regurgitation (grade 3–4)||23 (38)||44 (22)||.012|
The echocardiographic assessment showed that patients without LVCR presented significantly larger ventricles and a higher prevalence of moderate or severe mitral regurgitation (38% vs 22%, respectively; P=.012). Finally, there were more patients with interventricular delay (58% vs 43%, respectively; P=.040) and restrictive filling pattern (42% vs 25%, respectively; P=.011) in the group without LVCR than in the group with LVCR.
Differences Between Ischemic and Nonischemic Patients in DSE Response. Considering the 2 groups of ischemic and nonischemic patients, no significant differences could be observed in LVEF at rest (27%±5% vs 26%±6%, respectively; P=.572) and at peak stress (36%±9% vs 35%±9%, respectively; P=.376) or in the proportion of patients with LVCR (76% vs 77%, respectively; P=.833). Maximal heart rates during testing were 94±23 and 101±223, for ischemic and nonischemic patients, respectively (P=.032)
Independent Predictors of Contractile Reserve Presence. A univariable logistic regression was performed for all relevant baseline parameters to assess the association between LVCR presence and other clinical and echocardiographic baseline measures.
Only ventricular volumes, previous atrial arrhythmia history, and severe mitral regurgitation were significantly associated to the DSE outcome (Table IV).
Table IV. Univariable Logistic Regression
|Bundle branch block||0.92||0.46–1.81||.805|
|Previous atrial arrhythmia history||0.44||0.21–0.93||.030|
|LV end-diastolic volume (10-mL variation)||0.94||0.91–0.98||.001|
|LV end-systolic volume (10-mL variation)||0.93||0.89–0.97||.001|
|Interventricular dyssynchrony (%)||0.55||0.30–1.02||.057|
|Severe mitral regurgitation (grade 3–4) ||0.45||0.14–0.87||.018|
All the variables with a P value <.1 were then included in a multivariable analysis to detect the independent predictors of LVCR presence. Results show that only LV end-diastolic volume independently predicts LVCR presence (Table V).
Table V. Multivariable Logistic Regression
|Previous atrial arrhythmia history||0.44||0.21–0.93||.030|| || || |
|LV end-diastolic volume (10-mL variation)||0.94||0.91–0.98||.001||0.93||0.99–1.00||.004|
|LV end-systolic volume (10-mL variation)||0.93||0.89–0.97||.001|| || || |
|Interventricular dyssynchrony, No. (%)||0.55||0.30–1.02||.057|| || || |
|RFP presence||0.46||0.20–1.10||.076|| || || |
|Severe mitral regurgitation (grade 3–4) ||0.45||0.14–0.87||.018|| || || |
Reproducibility of DSE Measurements. Although a systematic review of all echocardiographic measures has not been carried out, DSE data of about one-third of the enrolled population were stored and sent to a core lab in which LVEF measurements were repeated to assess the agreement of test results in terms of detection of LVCR. For all reassessed DSE findings, the result in the core lab was the same as the one obtained during execution.
Consistency With Other Studies. Some of the main baseline characteristics of the LODO-CRT patients are comparable with those from CARE-HF and the PROSPECT trial, as depicted in Table VI. In particular, mean age, male sex, and proportion of patients in NYHA class III are practically identical in the 3 populations, while mean LVEF, QRS duration, and volumes are significantly different.
Table VI. Comparison With Other Studies
|Patients, No.||409||426||271|| || |
|Mean age, y||67||68||67||NS||NS|
|Ischemic heart disease, %||40||54||39||NS||<.001|
|NYHA class III, %||94||96||96||NS||NS|
|LV ejection fraction, %||25||24||26||NS||<.001|
|LV end-systolic volume, (mL)||121 mL/m2 (index)||168||153||N/A||.02|
|LV end-diastolic volume, (mL)||N/A||230||207||N/A||.01|
The LODO-CRT trial successfully enrolled a cohort of patients fulfilling current criteria for CRT device implantation. Patients showed major left ventricular systolic dysfunction and generally persistent symptoms of HF, despite conventional pharmacologic therapy.
The reported feasibility of the low-dose DSE protocol in the LODO-CRT framework was remarkably high. Dobutamine side effects were negligible in 99% of patients. The test had to be interrupted before reaching a clear indication of presence or absence of LVCR only in 3 cases in which a ventricular tachycardia was induced but eventually self-terminated. These data are in line or slightly better than those from Cigarroa and coworkers,27 who reported an overall test feasibility of 93% in 49 patients, and confirm the results from the Beta-Blocker Evaluation in Survival Trial (BEST) on safety and feasibility of the low-dose test protocol in 79 HF patients.31 However, to our knowledge this is the first large study confirming that DSE is feasible and safe in selected patients with symptomatic chronic HF.
Overall, in the LODO-CRT population, 77% of patients showed presence of LVCR. In the preliminary experience on the role of DSE in predicting response to CRT, carried out by Tuccillo and coworkers21 in 42 patients, the percentage of responders to DSE was 64%. However, the definition of response that was used was different (25% LVEF relative increase) from the one adopted in the present study.
The proportion of patients showing LVCR in the LODO-CRT trial is remarkably higher than those reported by Da Costa and coworkers19 and by Ypenburg and coworkers,20 who both identified approximately 50% of DSE responders in a population of candidates to CRT.
The main reason for this discrepancy is the different test protocol they used. In fact, in both experiences dobutamine infusion did not exceed 10 μg/kg/min, while in LODO-CRT investigators were allowed to increase the drug dosage to 20 μg/kg/min if needed. Moreover, the cutoff for positive response to the test was set at 25% LVEF relative increase and 7.5 points LVEF absolute increase, respectively, while in LODO-CRT an LVEF increase >5 absolute points was used. Furthermore, in Da Costa’s experience, a higher percentage of patients in NYHA class IV were present; this most probably influenced the average results of DSE.
In the LODO-CRT trial framework, the predefined cutoff for DSE outcome was identified according to Agricola and coworkers,23 who determined in their review on DSE that an increase in LVEF of at least 5% in absolute terms is the critical value to determine the presence of LVCR.
The evaluation of the study’s principal outcomes for specific subgroups of patients is of interest, since the characteristics of the subgroups may explain differences in outcome or baseline received therapy. Subgroup analysis of baseline characteristics identified few clinically relevant differences between patients with or without LVCR. As expected, absence of LVCR identified “sicker” patients. In particular, these patients showed more impaired ventricular function (larger volumes with worse diastolic patterns) with a higher grade of mitral regurgitation. Furthermore, in this group there were more patients showing interventricular delay. Finally, multivariable analysis showed that a more dilated left ventricle is the only parameter associated with a lower probability of LVCR presence.
To our knowledge, such a comparison is reported only by Ramahi and coworkers,32 who evaluated the association between the DSE test response and other clinical parameters in a nonischemic HF population similar to the LODO-CRT sample. In their experience, patients with low response to dobutamine presented with a greater mean NYHA functional class, larger diastolic volumes, and a higher mortality rate after 3 years from the test.
Our data confirm in an HF population of any etiology that a higher response to dobutamine identifies healthier patients, characterized by better functional capacity and smaller ventricles. These data are indeed promising, since the co-presence of factors like ischemic etiology of HF or larger ventricles has already been associated with lack of response to resynchronization.33
The resulting association between LVCR presence and other baseline conditions is not detrimental for DSE relevance. In fact, only DSE is able to detect the residual amount of viable tissue that would respond to the electrical stimulus provided by resynchronization therapy. Thus, dobutamine echocardiography provides unique information in selecting potential responders to CRT, since it takes into account the resynchronization’s own action mechanism, that is, the recruiting of a certain amount of cardiac cells, which would not actively contract if not electrically stimulated, due to the modifications induced by LV dilation.
Response to DSE seems to be independent of HF etiology. The 2 groups of ischemic and nonischemic patients did not show any significant differences in LVEF and proportion of patients with LVCR. Apparently, these data conflict with those reported in BEST in which the ischemic etiology in an unselected HF population was highly associated with absence of LVCR.31
Finally, the LODO-CRT population presented similar baseline characteristics to those of the CARE-HF and PROSPECT studies, although higher mean LVEF, shorter QRS duration, and smaller volumes in LODO-CRT may identify slightly less impaired patients. However, as the enrolled population adheres to current guidelines for CRT device implantation, the expected DSE results in terms of ability to predict the response to CRT can be compared with other predictors of response evaluated in other trials.
Study Limitations. This study presents data about the prevalence of LVCR in a population of CRT candidates and its association with other baseline clinical conditions, but it provides no information about the outcome of CRT with respect to LVCR presence. Moreover, the association between intraventricular dyssynchrony and LVCR has not been assessed yet. These two main points are now under evaluation; results will be published as soon as they become available.
Preliminary single-center experiences suggest a strong relationship between the presence of LVCR and a positive response after CRT.22–24 These data need confirmation in multicenter framework in a larger population of CRT candidates. Should the correlation between LVCR presence and response to CRT be confirmed, DSE could represent a highly feasible and easy procedure for the selection of patients to be treated with resynchronization.
Acknowledgments and disclosures: The authors thank all the investigators of the LODO-CRT trials and Lorenza Mangoni and Tiziana De Santo from Medtronic Italy for the statistical support. Mario Davinelli and Vanessa Novelli are Medtronic Italia employees. No other conflicts of interest exist. No extra funding has been provided.
Appendix 1. Participating Centers
LODO-CRT Trial Participating Investigators.
Dr Saverio Iacopino, Dr Rossella Alemanni, Sant’Anna Hospital – Catanzaro
Dr Maurizio Gasparini, Dr Renato Bragato, Istituto Clinico Humanitas – Rozzano (MI)
Dr Carmine Muto, Dr Bernardino Tuccillo, Ospedale Santa Maria di Loreto Mare – Napoli
Dr Francesco Zanon, Dr Silvio Aggio, Ospedale Civile – Rovigo
Dr Cosimo Dicandia, Dr Ennio Pellegrino, Clinica Città di Lecce – Lecce
Dr Giuseppe Distefano, Dr Maria Barillà, Centro Cuore Morgagni – Pedara (CT)
Dr Antonio Curnis, Dr Ermanna Chiari, Spedali Civili – Brescia
Dr Paolo Diotallevi, Ospedale Santi Antonio e Biagio e Cesare Arrigo - Alessandria
Dr Roberto Donati, Dr Antonio Cesario, Ospedale G. B. Grassi – Ostia (Roma)
Dr Carlo Peraldo Neja, Ospedale Fatebenefratelli – San Giovanni Calibita – Roma
Dr Valeria Calvi, Dr Salvatore Scandura, Ospedale Ferrarotto – Catania
Prof Stefano Favale, Dr Vito Marangelli, Policlinico di Bari
Dr Gabriele Giannola, Dr Gioachino Giarratana, Istituto San Raffaele G. Giglio – Cefalù (PA)
Dr Cosimo Dicandia, Dr Mariarosa Liccese, Clinica Anthea – Bari
Dr Biagio Sassone, Ospedale Consorziale – Bentivoglio (BO)
Dr Rosario Foti, Ospedale Civico San Vincenzo – Taormina (ME)
Dr Serafino Orazi, Ospedale San Camillo de’ Lellis – Rieti
Dr Franco Badessa, Presidio Ospedaliero – Milazzo (ME)
Dr Diego Vaccari, Presidio Ospedaliero – Montebelluna (TV)
Dr Roberto Verlato, Ospedale Civile – Camposampiero
Dr Massimo Sassara, Ospedale Belcolle – Viterbo