| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Guest Access | Sign In via User Name/Password
|
|
|
|||||||
Guest Access | Sign In via User Name/Password |
|||||||||
* From the Department of Medicine, Los Angeles Biomedical Research Institute at Harbor-UCLA Medical Center, Torrance, CA.
Source Investigators and Investigational Centers for the study are listed in the Appendix.
Corresponding author: Karlman Wasserman, MD, PhD, Department of Medicine, Los Angeles Biomedical Research Institute at Harbor-UCLA Medical Center, 1124 W Carson St, Torrance, CA 90502; e-mail: kwasserman{at}labiomed.org
Abstract
Background: Biventricular pacing (BVP) is used for cardiac resynchronization therapy in wide-QRS-complex heart failure. We sought to quantify the effect of BVP on the exercise pathophysiology of heart failure patients.
Methods: Using cardiopulmonary exercise testing, we analyzed exercise data for a multicenter study sponsored by St. Jude Medical. Patients had pacemaker electrodes implanted in both ventricles in the standard manner and were randomized by St. Jude before exercise testing. Exercise measurements included peak oxygen uptake (
O2), peak O2 pulse, anaerobic threshold (AT), and ventilatory equivalent for CO2 (reflecting change in peak exercise cardiac output, stroke volume, maximal sustainable exercise capacity, and ventilation-perfusion mismatching, respectively) at baseline and at a 6-month follow-up. The studies included progressively and uniformly increasing work rate to maximum tolerance. The investigators were blinded both to sponsor-controlled randomization and pacemaker status. There were 239 paired 6-month studies, as follows: 47 studies served as the control with the pacemaker off (ie, the BVP-OFF group); and 192 patients received pacing (ie, the BVP-ON group).
Results: The BVP-ON group significantly improved in all exercise parameters in contrast to the control group (p < 0.0001). When baseline measurements for the BVP-ON group were ranked in quintiles, only patients in the three functionally worst quintiles improved significantly at 6 months (peak O2 < 11.6 mL/min/kg, AT < 7.6 mL/min/kg, peak O2 pulse < 12.0 mL/beat, and minute ventilation/CO2 ratio at AT > 38.1) [p < 0.01 to < 0.0001].
Conclusion: BVP benefited aerobic function and ventilation-perfusion mismatching most in those patients with the greatest physiologic impairment.
Key Words: anaerobic threshold • biventricular pacing • cardiopulmonary exercise testing • cardiac resynchronization therapy • heart failure • oxygen transport • peak O2 pulse • peak O2 • ventilatory efficiency
Patients with left ventricular failure (LVF) have major physiologic defects that limit their ability to exercise. Exercise fatigue can be partially attributed to the failure of the circulation to deliver sufficient O2 to the muscles for the aerobic regeneration of high-energy phosphate needed to sustain muscular work. This results in a low maximal exercise capacity (low peak oxygen uptake [O2]),12 low anaerobic threshold (AT),3 and a reduced O2 pulse for the rate of work performed.2 The AT, signaling the start of the buffering of exercise-induced lactic acidosis, occurs in both healthy subjects4 and patients with LVF56 when exercise cannot be sustained solely by aerobic metabolism. Inappropriately high (wasted) ventilation in patients with LVF (increased physiologic dead space/tidal volume ratio)7 often causes exertional dyspnea. Biventricular pacing (BVP) is being used with increasing frequency to treat LVF patients with wide-QRS-complex heart failure.8910 The concept behind BVP therapy is that cardiac contraction is resynchronized, thereby improving stroke volume (SV).
Aerobic function and ventilation-perfusion mismatching are abnormal in patients with LVF, and the magnitudes of these abnormalities are useful in prognosticating survival.11121314 As the core exercise laboratory in a multicenter study (Resynchronization for Hemodynamic Treatment for Heart failure Management [or RHYTHM] study at St. Jude Medical), we had the opportunity to measure indexes of exercise pathophysiology in LVF patients with or without BVP therapy who had wide-QRS-complex (ie, > 120 ms) disease with an average New York Heart Association (NYHA) symptom class of 2.9. We compared the 6-month to baseline exercise data on all patients to determine whether BVP improved peak O2, AT, peak O2 pulse, and the minute ventilation (E)-carbon dioxide output (CO2) relationship, reflecting peak exercise cardiac output, maximal sustainable aerobic exercise capacity, SV, and ventilation-perfusion mismatching, respectively.
Materials and Methods
Study Population
The study was initiated and financed by St. Jude Medical, Inc. Informed consent was obtained at each site of the study. Our study had Harbor-UCLA Institutional Review Board approval. The cardiologists at the patient care centers (see "Appendix") and St. Jude Medical, selected NYHA class II-IV LVF patients with wide-QRS-complexes (> 120 ms) for the implantation of BVP electrodes. Before cardiopulmonary exercise tests (CPETs) were performed, St. Jude Medical randomized the patients into a control group that did not use BVP (ie, the BVP-OFF group) or a group that used BVP (ie, the BVP-ON group) [Table 1
]. Initially, there were two BVP-ON groups, one labeled "synchronized" (ie, simultaneous stimulation) and one labeled "optimized" (ie, stimulation of the right and left ventricles slightly differently to achieve optimized contraction). After the study, because our analysis revealed no difference between the two BVP-ON groups, we combined them to simplify the presentation of the results.
|
During the testing of a given patient, a St. Jude field clinical engineer was sometimes present. However, the engineers were not involved in performing the CPET or adjusting the CPET equipment and could not modify the electronic data that were sent directly from the site to the core laboratory.
Exercise Protocol
The protocol required measurements during 3 min of rest, 3 min of unloaded cycling or comparable warm-up on the treadmill, followed by a progressively increasing work rate in ramp pattern or 1-min intervals to maximum tolerance. The schedule for increasing the work rate was designed to have the subjects smoothly reach their maximum tolerated work rate in 6 to 15 min of exercise whether on the cycle or treadmill.15 A total of 45% of the sites used the cycle, and the remainder used the treadmill (Table 1). Whether a cycle or treadmill was used, the exercise protocol was the same. Work rate was increased at 5 to 15 W/min depending on the patient’s NYHA class and gender, using a schedule provided by us to each site. For each paired study, the baseline and 6-month protocols were identical.
Prior to patient testing, each site had to demonstrate that they could follow the exercise protocol satisfactorily for two consecutive tests on a healthy subject with good reproducibility in peak O2, AT, peak O2 pulse, ventilation (ie, E), and CO2 relationships that fit normal values.16 The mean ± SD between replicate tests for the 60 qualified sites were 5.1 ± 4.1%, 5.2 ± 5.3%, 5.0 ± 5.8%, and 4.0 ± 3.0% for the four parameters, respectively. Sites that failed to qualify were disqualified from study participation. Our core laboratory reviewed the tabular and graphical data on each study for evident breaches in protocol and calibration.
Measurements
Gas exchange measurements were made with multiple commercial systems (eg, Medical Graphics [St. Paul, MN], SensorMedics [Anaheim, CA], Cosmed [Rome, Italy], Jaeger [Wurzburg, Germany], and ParvoMedics [Sandy, UT]). We converted the electronic files of all studies into a single uniform tabular format and a nine-panel graphical display of 15 exercise variables with scaling optimized to view the physiologic responses, as previously described.17 Breath-by-breath data were interpolated to second-by-second values, sequentially averaged in 10-s bins, and plotted as illustrated in Figure 1
. Uniform plotting and scaling allowed us to superimpose the baseline and 6-month studies and to evaluate changes in all variables, graphically and numerically. All studies were reviewed by at least one of the Core Laboratory investigators who was not involved in the primary data analysis.
|
O2 (in milliliters per minute) was measured as the highest average O2 over 30 s, as determined from three sequential 10-s periods at peak exercise (Fig 1, top right, 3). Peak O2 (in milliliters per minute per kilogram) was calculated as the dividend of peak O2 and the patient’s baseline body weight.
The AT (in milliliters per minute) was measured by the V-slope method, in which CO2 is plotted against O2 on X-Y coordinates of equal scale (ie, enlarged Fig 1, center, 5).16 The AT is the O2 where CO2 increases faster than O2 (Fig 1, center, 5, vertical arrow), demonstrating that HCO3– is in the process of buffering lactic acid. Hyperventilation does not confound this measurement because arterial and end-tidal PCO2 are constant or increasing rather than decreasing at this time. AT (in milliliters per minute per kilogram) was calculated as the dividend of O2 at AT and the patient’s baseline body weight. Peak O2 pulse (in milliliters per beat) was calculated as the peak O2 (in milliliters per minute) divided by the simultaneous heart rate (HR; in beats per minute) [Fig 1, top center, 2].
Ventilatory efficiency (ie, ventilation-perfusion mismatching) was measured in two ways. The slope of E vs CO2 was determined from the start of exercise to the ventilatory compensation point (VCP)1819 (Fig 1, center left, 4). The E/CO2 ratio at the AT was determined as the average ratio from the AT to 1 min after the AT1920 (Fig 1, center right, 6). If the AT was indeterminate, theE/CO2 ratio at the VCP was used.
The significance of each graph of the nine-panel plot is described in the legend of Figure 1, including how they become abnormal in patients with lung and heart diseases.17 Based on the reproducibility of prior patient studies,20 which showed a 6% improvement to be significant, we considered a 
10% improvement in a given parameter for a given patient as significant. Predicted values16 take into account age, size, gender, and form of ergometry (ie, cycle or treadmill [10% higher for the latter]).
Statistical Analysis
For the analysis, both the BVP-ON and BVP-OFF groups were divided into quintiles of equal size for O2, AT, peak O2 pulse, and the E-CO2 relationship. Thus, group 1 had the 20% most abnormal values and group 5 had the 20% least abnormal values. There were 38 to 39 patients in each BVP-ON quintile and 9 to 10 patients in each BVP-OFF quintile. Because no significant differences were found in the percentage of change in any of the measured parameters in the BVP-OFF quintiles, the entire BVP-OFF group was recombined to create a group with a size similar to that of the BVP-ON quintiles. The data are expressed as the mean ± SD, and were statistically analyzed by analysis of variance, paired t test, unpaired t test, and 
2 test. A p value of < 0.05 was considered to be significant.
Results
The patients’ demographics, NYHA class, resting and peak exercise HR and systolic BP, and baseline and 6-month CPET results are shown in Table 1. Despite randomization prior to baseline CPET studies, the baseline peak O2 and AT values in the BVP-ON group were significantly lower than those of the BVP-OFF group. However, the baseline peak O2 pulse values, ventilation-perfusion mismatching indexes (E/CO2 ratio at AT), and peak respiratory exchange ratios (RERs) did not significantly differ.
Effect of BVP on Aerobic Function
Compared to the BVP-OFF group, the BVP-ON group had highly significant increases in the percentage of patients who improved each parameter of aerobic function and ventilation-perfusion mismatching 10%, whether calculated in absolute units (Table 2
) or as percentages of predicted values (Table 3
). By analysis of variance, there were no differences at 6 months in quintile responses in the BVP-OFF group.
|
|
O2:
In contrast to the BVP-OFF group, the three lowest peak O2 quintiles of the BVP-ON group, with average baseline values of < 10.5 mL/min/kg (< 50% predicted), showed significant improvement at 6 months (Fig 2
), as did the number of patients who benefited (Tables 2, 3). The lowest baseline peak O2 quintile showed the greatest improvement and the most patients (66 to 71%) whose conditions had improved 10% at 6 months. When related to a peak O2 of 
10 mL/min/kg or 50% predicted (Table 4
), the BVP-ON groups increased 20% and 19%, respectively (p < 0.0001). Neither the BVP-ON group with a peak O2 of > 10 mL/min/kg (or > 50%) nor both BVP-OFF groups improved, significantly.
|
|
10%.
|
10%.
|
E/CO2 at AT and slope of the E-CO2 linear relationship1819 as related to peak O2 (in milliliters per minute per kilogram) for all baseline studies are shown in Figure 5
. In all LVF subgroups, both ratios and slopes were abnormally elevated, with patients having the lowest peak O2 having the highest values. The E/CO2 at AT ratios had less variability than the slope of the E-CO2 linear relationship (SD, 9.1 vs 11.4, respectively). Because of this lesser variability,20 we present changes in ventilatory efficiency in Figure 6
as the E/CO2 ratio at AT.
|
|
E/CO2 ratio at AT were unchanged in the BVP-OFF group (Fig 6). In contrast, the ratios significantly decreased from baseline (ie, improved matching of ventilation to perfusion) for the three most abnormal quintiles in the BVP-ON group, expressed as percent predicted or absolute values, with most patients (61%) increasing 10% in the most abnormal quintile (Tables 2, 3).
Correlation of Baseline Measures of Aerobic Function
Historically, peak O2 has been the primary measure of exercise impairment in patients with LVF. The high and positive correlation between AT and peak O2 for the LVF population (Fig 7
) suggests that this submaximal measurement (AT) could be used to classify impairment when patients do not achieve peak exercise effort.
|
10% at 6 months (Table 2). When baseline values were expressed as a percentage of predicted values, the values of 61 to 74% of patients in the most abnormal quintile had improved by 10% by 6 months (Table 3). These values contrast with 15 to 28% of the BVP-OFF patient group. The values of patients in the two least abnormal quintiles were not significantly improved, compared to those of patients in the BVP-OFF group for any of the four parameters studied. Discussion
In this study, the effect of BVP on pathophysiology was determined using measurements of peak O2, AT, peak O2 pulse, and E/CO2 ratio that are typically abnormal in patients with LVF. We compared the effect of treatment vs a control group (implanted but unstimulated electrodes) on these measurements, and investigated the importance of severity of physiologic impairment on the ability to improve function. With the BVP-ON group, patients in baseline peak O2 quintiles of < 11.6 mL/min/kg (or < 62% predicted) showed statistically significant improvement in that parameter (Fig 2). For AT, patients in quintiles of < 7.6 mL/min/kg (or < 66% predicted) showed significant improvement in that parameter (Fig 3). For peak O2 pulse, the patients in quintiles of < 68% predicted had significant improvement in that parameter (Fig 4). For E/CO2 at AT, the patients in quintiles with a ratio of > 38 or > 130% predicted showed significant improvement in that parameter (Fig 6). For all four parameters, the improvement over the BVP-OFF group (Figs 234, 6) and the percentage of patients whose conditions had improved 10% for each parameter (Tables 2, 3) were greater for patients in the more abnormal quintiles at baseline.
For BVP to be beneficial to the patient, the physiologic abnormalities that reduce exercise tolerance should be improved. An improvement in symptoms (usually exertional fatigue and/or dyspnea) should be matched by improvements in the major physiologic deficiencies in LVF (ie, the reduced cardiac output and SV responses to exercise and the abnormal ventilation-perfusion relationships that accompany severe LVF).71821 These deficiencies are reflected in the measurements chosen for this study.
Since there is a direct relationship between peak cardiac output and peak O2 (O2 = HR x SV x C(a-v)O2 [where C(a-v)O2 is the arterial-mixed venous oxygen concentration difference]), an increase in peak cardiac output at 6 months might be expected to be reflected in an increase in peak O2 and peak O2 pulse, assuming that peak C(a-v)O2 had not changed. If the interval hemoglobin concentration or arterial oxyhemoglobin saturation (usually normal in stable patients with LVF)19 does not change, the increase in peak O2 pulse (ie, SV x C(a-v)O2) signifies an increase in SV.
Since O2 pulse is influenced by the size, age, and gender of the patient, we expressed baseline peak O2 pulse in absolute terms and as the percentage of predicted values. Both were significantly improved after 6 months of treatment with BVP in the three lowest quintiles compared to the control group (Fig 4) and by a significantly higher fraction of patients than the control group (Tables 2, 3).
The predicted peak O2 pulse values were derived from healthy subjects16 who were not receiving treatment with ß-adrenergic blockade that might slow HR and elevate O2 pulse values. Despite the fact that 81% of the BVP-ON patients and 85% of the BVP-OFF patients were receiving ß-adrenergic blockade therapy, the reduced peak O2 pulse values (Table 1) confirm that SVs were likely reduced in these patients.
Because patients may not always exercise to their peak capacity, submaximal parameters, such as AT and E/CO2 ratio at AT, have been found to be useful measures of LVF severity1118 and to correlate with peak O2 (Figs 5, 7). Improvements in these parameters reflect an improved ability to sustain exercise and a greater uniformity in lung perfusion relative to ventilation.
For the purpose of measuring ventilatory efficiency in patients with heart failure, the greater variability of the slope of the E-CO2 linear relationship compared to the E/CO2 ratio at AT deserves special comment. If the data points establishing the slope went through the origin of the E-CO2 plot, the ratio and the slope would be equal (Fig 1, center left, 4, and center right, 6). However, because the slope usually has a positive intercept, its value is usually 1 to 3 points less than the ratio.19 The positive intercept is variable and attributable to a reduction in the dead space/tidal volume ratio after the start of exercise and/or variable degrees of early exercise hyperventilation.
Three aspects of the analysis might have affected the study outcome and should be addressed. (1) The small size of the BVP-OFF group may have resulted in a random nonchanging exercise capacity at 6 months compared to the much larger BVP-ON group. However, when dividing the BVP-ON group into quintiles to obtain groups similar in size to the control group, the 6-month values were systematically increased over baseline values for all four parameters used to assess function. The pattern of improvement showed that only the three worst quintiles improved, with the order depending on the severity of impairment. The two best functional quintiles did not improve at 6 months, which is similar to the situation in the control group. (2) The lower baseline values in peak O2, AT, and peak O2 pulse might have led to a better response in the BVP-ON group compared to the BVP-OFF group. We, of course had no control over the randomization, which was done prior to testing. However, when the BVP-ON group was divided into equal quintiles with a size similar to that of the BVP-OFF group, with overlapping values, the analysis shows that only patients in the three most impaired BVP-ON quintiles improved. (3) The possibility is suggested that the changes at 6 months are random and that the patients with the lowest values would improve to the natural mean of the population and those with the highest values would decrease toward the mean. However, the lowest quintile remains the lowest quintile and the highest quintile remains the highest quintile at 6 months. Thus, the argument that the improvement seen is due to random changes, depending on the baseline values, does not fit well with the observations.
Because these studies are from 60 cardiology sites with differing experiences in cardiopulmonary exercise testing, technical variability might have obscured the physiologic changes. Despite this, the results are orderly and interpretable, with each parameter measured showing that within-group improvement was dependent on the extent of physiologic impairment. Thus, our results suggest that the degree of exercise impairment, assessed by cardiopulmonary exercise testing, helps to select those LVF patients who are most likely to benefit from BVP.
In summary, four exercise parameters (reflecting peak exercise cardiac output, maximal sustainable exercise capacity, SV, and ventilation-perfusion mismatching) were used to grade the severity of physiologic impairment, and to measure change following treatment with BVP in LVF patients with wide-QRS-complex systolic dysfunction. Those patients who had the most abnormal values benefited the most from BVP, as measured by four exercise parameters that characteristically have abnormal values in patients with LVF.
Appendix
Source Investigators and Investigational Centers
Thomas Arne, Jr., Sarasota Memorial Hospital, Sarasota, FL; Charles Athill, Sharp Hospital, San Diego, CA; Steven Bailin, Iowa Heart Center, Des Moines, IA; James Baker II, St. Thomas Hospital, Nashville, TN; Scott Beau, Arkansas Heart Hospital, Little Rock, AR; David Benditt, St. Cloud Hospital, St. Cloud, MN; Sheldon Brownstein, University of Cincinnati Medical Center, Cincinnati, OH; Albert Camacho, Legacy Good Samaritan Hospital and Medical Center, Portland, OR; James Cook, Baystate Medical Center, Springfield, MA; Rafael Corbisiero, Deborah Heart and Lung Center, Browns Mills, NJ; Wynne Crawford, Baptist Health Medical Center, Montgomery, AL; Dan Dan, Piedmont Hospital, Atlanta, GA; Prakash Desai, Northwest Texas Healthcare System, Amarillo, TX; Rahul Doshi, Sunrise Hospital and Medical Center, Las Vegas, NV; Thomas Edel, EMH Regional Medical Center, Elyria, OH; Andrew Epstein, University of Alabama at Birmingham, Birmingham, AL; Kenneth Ellenbogen, Medical College of Virginia Hospitals, Richmond, VA; Richard Fogel, St. Vincent Hospital, Indianapolis, IN; G. Joseph Gallinghouse, Texas Cardiac Arrhythmia Institute, Austin, TX; Michael Gold, Medical University of South Carolina, Charleston, SC; G. Stephen Greer, Baptist Health Medical Center, Little Rock, AR; John Herre, Sentara Norfolk General Hospital, Norfolk, VA; Evelyn Horn, New York Presbyterian Hospital, New York, NY; Steve Hsu, Shands Jacksonville Medical Center, Jacksonville, FL; David Huang, Strong Memorial Hospital, Rochester, NY; John Ip, Ingham Regional Medical Center, Lansing, MI; Roy John, Lahey Clinic Medical Center, Burlington, MA; Steven Kalbfleisch, Riverside Methodist Hospital, Columbus, OH; Erol Kosar, Centinela Hospital, Inglewood, CA; Robert Kowal, University of Texas Southwestern Medical Center, Dallas, TX; Jeffrey Kushner, St. Mary’s Hospital Medical Center, Madison, WI; Steven Kutalek, Drexel University College of Medicine, Philadelphia, PA; Todd Langager, Mercy Medical Center, Cedar Rapids, IA; Charles Love, Ohio State University Medical Center, Columbus, OH; Richard Luceri, Holy Cross Hospital, Ft. Lauderdale, FL; Ali Massumi, St. Luke’s Episcopal Hospital, Houston, TX; Thomas Mattioni, Arizona Arrhythmia Consultants, Phoenix, AZ; John McKenzie, Glendale Memorial Hospital and Health Center, Glendale, CA; Davendra Mehta, Mount Sinai Medical Center, New York, NY; Theofanie Mela, Massachusetts General Hospital, Boston, MA; Magdy Migeed, Crouse Hospital, Syracuse, NY; Sanjiv Narayan, Veteran Administration Medical Center (UCSD), San Diego, CA; Mark Niebauer, University of Nebraska Medical Center, Omaha, NE; Jess Oren, Geisinger Medical Center, Danville, PA; Antonio Pacifico, Methodist Hospital, Houston, TX; Luis Pires, St. John Hospital and Medical Center, Detroit, MI; James Porterfield, Methodist University Hospital, Memphis, TN; Arun Rao, Wellmont Holston Valley Medical Center, Kingsport, TN; Derek Rodrigues, Overlake Hospital Medical Center, Kirkland, WA; Lawrence Rosenthal, University of Massachusetts Medical Center, Worcester, MA; Steven Rothman, Temple University Hospital, Philadelphia, PA; Andrea Russo, Hospital of the University of Pennsylvania and Presbyterian Medical Center, Philadelphia, PA; William Sanders, Jr., University of North Carolina at Chapel Hill, Chapel Hill, NC; Arjun Sharma, Mercy General Hospital, Sacramento, CA; Kalyanam Shivkumar, University of California, Los Angeles Medical Center, Los Angeles, CA; Richard Soucier, St. Francis Hospital, Hartford, CT; Bruce Stambler, University Hospitals of Cleveland, Cleveland, OH; Kyong Turk, Saint Elizabeth Regional Medical Center, Lincoln, NE; Jesus E. Val-Mejias, Galichia Heart Hospital, Wichita, KS; and Bruce Wilkoff, Cleveland Clinic Foundation, Cleveland, OH.
Footnotes
Abbreviations: AT = anaerobic threshold; BVP = biventricular pacing; C(a-v)O2 = arterial-mixed venous oxygen concentration difference; CPET = cardiopulmonary exercise test; HR = heart rate; LVF = left ventricular failure; NYHA = New York Heart Association; RER = respiratory exchange ratio; SV = stroke volume; CO2 = carbon dioxide output; VCP = Ventilatory compensation point; E = minute ventilation; O2 = oxygen uptake
This study was started before clinical trial numbers were issued.
This study and the partial salaries of the authors were supported by a contract with the Los Angeles Biomedical Research Institute at Harbor-UCLA Medical Center by St. Jude Medical, Sunnyvale, CA.
The authors have reported to the ACCP that no significant conflicts of interest exist with any companies/organizations whose products or services may be discussed in this article.
Received for publication December 4, 2006. Accepted for publication March 9, 2007.
References
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
