Use of intravenous EDTA to treat atherosclerotic diseases. Original Article

Transcription

1 Original Article Quality-of-Life Outcomes With a Disodium EDTA Chelation Regimen for Coronary Disease Results From the Trial to Assess Chelation Therapy Randomized Trial Daniel B. Mark, MD, MPH; Kevin J. Anstrom, PhD; Nancy E. Clapp-Channing, MPH; J. David Knight, MS; Robin Boineau, MD, MA; Christine Goertz, DC, PhD; Theodore C. Rozema, MD; Diane M. Liu, BA; Richard L. Nahin, PhD, MPH; Yves Rosenberg, MD, MPH; Jeanne Drisko, MD; Kerry L. Lee, PhD; Gervasio A. Lamas, MD; on behalf of the TACT Investigators* Background The National Institutes of Health funded Trial to Assess Chelation Therapy (TACT) randomized 1708 stable coronary disease patients aged 50 years who were 6 months post myocardial infarction ( ) to 40 infusions of a multicomponent EDTA chelation solution or placebo. Chelation reduced the primary composite end point of mortality, recurrent myocardial infarction, stroke, coronary revascularization, or hospitalization for angina (hazard ratio, 0.82; 95% confidence interval, ; P=0.035). Methods and Results In a randomly selected subset of 911 patients, we prospectively collected a battery of quality-of-life (QOL) instruments at baseline and at 6, 12, and 24 months after randomization. The prespecified primary QOL measures were the Duke Activity Status Index (Table I in the Data Supplement) and the Medical Outcomes Study Short-Form 36 Mental Health Inventory-5. All comparisons were by intention to treat. clinical and QOL variables were well balanced in the 451 patients randomized to chelation and in the 460 patients randomized to placebo. The Duke Activity Status Index improved in both groups during the first 6 months of therapy, but we found no evidence for a treatment-related difference (mean difference [chelation placebo] during follow-up, 0.9 [95% confidence interval, 0.7 to 2.6; P=0.27]). There was no statistically significant evidence of a treatment-related difference in the Mental Health Inventory-5 during follow-up (mean difference, 1.0; 95% confidence interval, 0.1 to 2.0; P=0.08). None of the secondary QOL measures showed a consistent treatment-related difference. Conclusions In stable, predominantly asymptomatic coronary disease patients with a history of myocardial infarction, EDTA chelation therapy did not have a detectable effect on QOL during 2 years of follow-up. Clinical Trial Registration URL: Unique identifier: NCT (Circ Cardiovasc Qual Outcomes. 2014;7: ) Key Words: atherosclerosis coronary disease quality of life Use of intravenous EDTA to treat atherosclerotic diseases dates back to the mid-1950s. Clinicians using the drug to treat lead toxicity noted a concomitant unexpected improvement in patient-reported angina symptoms. Because the drug s only known mechanism of action was to bind heavy metal cations (eg, calcium, magnesium, and lead) and facilitate their excretion, early users of this therapy speculated that EDTA provided benefit by chelating the calcium in atherosclerotic plaques, thereby facilitating a debulking of symptom-causing plaques. As the understanding of the pathobiology of atherosclerosis advanced, the implausibility of this mechanism, together with the absence of supporting clinical outcome evidence beyond the level of case reports, led mainstream cardiovascular practitioners to reject EDTA chelation therapy. Some practitioners, however, continued to use chelation for atherosclerotic disease based on anecdotal impressions of clinical improvement. A 2007 survey suggested that > patients in the United States receive this treatment each year. 1 Received February 27, 2014; accepted June 11, From the Outcomes Research Group (D.B.M., K.J.A., J.D.K., N.E.C.-C., D.M.L.) and Duke Clinical Research Institute (K.L.L.), Duke University Medical Center, Duke University, Durham, NC; National Heart, Lung, and Blood Institute, Bethesda, MD (R.B., R.L.N., Y.R.); Palmer Center for Chiropractic Research, Davenport, IA (C.G.); Biogenesis Medical Center, Landrum, SC (T.C.R.); Integrative Medicine, University of Kansas Medical Center, Kansas City (J.D.); and Columbia University Division of Cardiology at Mount Sinai Medical Center, Miami, FL (G.A.L.). *A list of the TACT Investigators is provided in the Data Supplement. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Heart, Lung, and Blood Institute, the National Center for Complementary and Alternative Medicine, or the National Institutes of Health. The Data Supplement is available at Correspondence to Daniel B. Mark, MD, MPH, Duke Clinical Research Institute, 2400 Pratt St, Room 0311, Durham, NC daniel.mark@duke.edu 2014 American Heart Association, Inc. Circ Cardiovasc Qual Outcomes is available at DOI: /CIRCOUTCOMES

2 Mark et al QOL Outcomes With Disodium EDTA Chelation for CAD 509 WHAT IS KNOWN Small case series of EDTA chelation therapy in coronary disease patients from the 1950s and early 1960s reported unusual symptomatic relief of angina. No large randomized trial has examined the effects of EDTA chelation therapy on quality of life in patients with coronary artery disease. WHAT THE STUDY ADDS The Trial to Assess Chelation Therapy trial enrolled a stable, post myocardial infarction population; 21% of the patients had angina at baseline, and 20% had heart failure symptoms. There was no difference in cardiac-related functional status (Duke Activity Status Index) or psychological distress (Short Form-36 Mental Health Inventory-5) between the EDTA chelation arm and placebo arm. EDTA chelation therapy did not improve quality of life relative to placebo during 2 years of follow-up The National Institutes of Health released a request for applications in 2001 and then funded the Trial to Assess Chelation Therapy (TACT) in 2002 to evaluate, in a rigorous randomized clinical trial, the benefits and harms of this therapy in coronary disease patients treated with modern evidence-based therapies. Assessment of patient-reported quality-of-life (QOL) outcomes was a major secondary aim of the TACT research effort, and this article reports the QOL results in the chelation arms of the trial. Methods Patient Population and Primary Clinical Results The TACT is a randomized, double-blind, placebo-controlled, 2 2 factorial trial comparing 40 infusions of a multicomponent EDTA chelation solution with placebo and oral high-dose multivitamin and mineral supplement with placebo in stable post myocardial infarction (MI) patients. The primary end point was a composite of death from any cause, reinfarction, stroke, coronary revascularization, or hospitalization for angina. Complete inclusion and exclusion criteria have been previously published. 2 All patients provided informed consent, and study protocol approval was obtained from each site s institutional review board or ethics committee. A total of 1708 patients were enrolled (839 to chelation and 869 to placebo) at 134 clinical sites in the United States and Canada between September 10, 2003, and October 4, Median follow-up was 55 months. The median age was 65 years, 18% were women, and 94% were white. At baseline, the use of evidence-based medicines was high and did not differ between the EDTA and placebo groups: aspirin, warfarin, or clopidogrel, 91%; β-blockers, 72%; statins, 73%; angiotensin-converting enzyme inhibitor or angiotensin receptor blocker, 63%. Follow-up medication use remained at high levels through the last follow-up and did not differ by treatment group: aspirin, warfarin, or clopidogrel, 87%; β-blocker, 72%; statins, 68%; angiotensin-converting enzyme inhibitor or angiotensin receptor blocker, 61%. Coronary revascularization during follow-up occurred in 15% of the EDTA chelation patients and 18% of the placebo patients. 3 As previously reported, the primary intention-to-treat comparison showed that the primary end point occurred in 26% of the patients assigned to chelation and in 30% assigned to placebo (hazard ratio, 0.82; 95% confidence interval [CI], ; P=0.035). 3 The magnitude of treatment effect of chelation on the nonfatal components of the primary end point was consistent with its overall effect (no single component dominant). No evidence for a mortality effect was seen, but power for this end point was low. Two prespecified subgroups, patients with diabetes mellitus and patients with a prior anterior MI, received particular benefit from the chelation strategy. 4 Patient Selection and Health-Related QOL Data Collection As a result of budgetary considerations, a decision was made at the outset of TACT to enroll 50% of the total study cohort into the QOL substudy. Patients were selected for QOL data collection by simple random selection at the time of treatment assignment. Health-related QOL data were collected using interviews at baseline and at 6, 12, and 24 months after randomization. Site coordinators were specially trained to administer the QOL baseline interviews. Follow-up interviews were conducted via telephone by trained interviewers from the Duke Clinical Research Institute s Outcomes Research Group. Proxy forms were collected for patients who were unable to participate in the full interview because of illness or incapacity ( 3% 5% of surviving patients at each follow-up assessment; Figure 1). New York Heart Association class and Canadian Cardiovascular Society angina 1708 stable patients with coronary disease were randomized into the Trial to Assess Chelation Therapy : 911 patients were randomly assigned to the quality of life substudy 911 responses (100%) from baseline to 6 months EDTA Infusion N=451 eligible 451 (100%) patient-reported surveys 2 Died 12 Withdrawn 6-month interval: 871 patients eligible for questionnaire 824 responses (94.6%) from 6 months to 12 months EDTA Infusion N=437 eligible 398 (91.1%) patient-reported surveys 15 ( 3.4%) proxy-reported surveys 24 ( 5.5%) not done 2 unknown 9 refused 11 unable to locate 2 other 12-month interval: 835 patients eligible for questionnaire 796 responses (95.3%) from 12 months to 24 months EDTA Infusion N=414 eligible 8 Died 15 Withdrawn 378 (91.3%) patient-reported surveys 15 ( 3.6%) proxy-reported surveys 21 ( 5.1%) not done 5 unknown 6 refused 8 unable to locate 2 other 10 Died 9 Withdrawn 24-month interval: 789 patients eligible for questionnaire 757 responses (95.9%) EDTA Infusion N=395 eligible 368 (93.2%) patient-reported surveys 12 ( 3.0%) proxy-reported surveys 15 ( 3.8%) not done 9 unknown 4 refused 2 unable to locate 0 other Placebo Infusion N=460 eligible 460 (100%) patient-reported surveys Placebo Infusion N=434 eligible 4 Died 22 Withdrawn 389 (89.6%) patient-reported surveys 22 ( 5.1%) proxy-reported surveys 23 ( 5.3%) not done 3 unknown 11 refused 6 unable to locate 3 other Placebo Infusion N=421 eligible 4 Died 9 Withdrawn 384 (91.2%) patient-reported surveys 19 ( 4.5%) proxy-reported surveys 18 ( 4.3%) not done 8 unknown 7 refused 2 unable to locate 1 other Placebo Infusion N=394 eligible 6 Died 21 Withdrawn 359 (91.1%) patient-reported surveys 18 ( 4.6%) proxy-reported surveys 17 ( 4.3%) not done 10 unknown 4 refused 3 unable to locate 0 other Figure 1. Reasons for missing quality of life data in the Trial to Assess Chelation Therapy.

3 510 Circ Cardiovasc Qual Outcomes July 2014 class were collected on the clinical case report form at baseline and each follow-up interval. QOL Measures A battery of validated measures was collected to provide a comprehensive but efficient assessment of QOL in TACT. Two principal QOL measures were prespecified during the design phase of the study as being most likely to show a treatment effect: the Duke Activity Status Index (DASI) and the Short Form-36 (SF-36) Mental Health Inventory-5 (MHI-5). The DASI is a 12-item instrument used to assess cardiac-related functional status. 5 Scores range from 0 (worst) to 58.2 (best) and reflect the ability of patients to do physical activities without difficulty or assistance in 12 domains (see Appendix in the Data Supplement). For an individual patient, a clinically significant change is considered to be 4 points. 6 Item selection and weighting in DASI were performed using a cardiopulmonary exercise testing reference standard. 5 The MHI-5 assesses psychological well-being, including depression and anxiety. 7 It is scored from 0 (worst) to 100 (best). A clinically significant change for an individual patient is approximated by a one-fourth SD difference, 5 points. We also collected 3 scales from the Seattle Angina Questionnaire (SAQ) to assess the impact of angina symptoms on QOL outcomes. 8 The SAQ anginal frequency scale assesses the frequency of angina symptoms in the previous 4 weeks. Higher scores reflect fewer angina symptoms. The SAQ anginal stability scale measures changes in angina frequency, with a score of 50 representing no change and higher scores indicating less frequent angina compared with the previous month. The SAQ QOL scale measures the effect of angina symptoms on patients perceptions of their QOL, with higher scores being more favorable. For SAQ scales, 5 points represent a benchmark for a clinically significant change for an individual patient. 8 To provide an overview of functioning and well-being from a generic perspective, we also collected individual scales from the Medical Outcomes Study SF-36: general health, physical function, role functioning physical, role functioning emotional, bodily pain, social function, and vitality (energy/fatigue). 7 These instruments asked patients to recall health status during the past 4 weeks. Higher scores reflect more favorable outcomes. These SF-36 scales can also be used to score 2 additional summary scales: the SF-36 Physical Component and Mental Component. 9 These scores are standardized to a population norm-based average value obtained from the 1998 US general population, with a mean of 50 and an SD of 10. A clinically significant change for a patient using this scoring system has not been established but can be approximated by a one-fourth SD or 2.5 points. We collected the 5-dimensional generic health status index EQ-5D to measure patient utilities (health states weighted by the population preferences for that state), which indicate the relative desirability of each patient s health state on a scale from 0 (dead) to 1 (best possible health). 10 Statistical Analyses Primary Analyses All primary comparisons were performed according to the principle of intention to treat. Descriptive statistics included percentages for discrete variables and medians with 25th to 75th percentiles plus means with SDs for continuous variables. The χ 2 test was used for discrete variable comparisons. Treatment comparisons of continuous variables were performed using a regression model to account for repeated measures within a patient, with treatment group, time point, and the respective QOL measure from the baseline questionnaire as predictor variables. Using PROC MIXED in SAS version 9.2, the 6-, 12-, and 24-month measurements within a patient were fitted using an unstructured covariance matrix. 11 At each time point, estimated treatment differences, 95% confidence intervals, and P values were obtained using estimate statements. These models assumed that missing data were missing at random. Time and treatment groups were coded using a cell mean model, with 8 binary variables representing the 4 QOL assessment points and 2 treatment groups. All reported P values were 2-sided, and the conventional Neyman Pearson decision rule for rejecting the null hypothesis was used (P<0.05). No adjustment was made in significance levels for multiple comparisons. As an aid to interpretation, we considered evidence supporting a QOL effect to be present if we saw either a consistent treatment difference over multiple consecutive follow-up contacts or a concordant treatment difference in multiple-related QOL measures. Subgroups We compared QOL outcomes according to the prespecified subgroups set forth for the overall TACT analysis: sex, minorities, elderly (>70 years), anterior MI, diabetes mellitus, metabolic syndrome, statin therapy, interaction with high-dose vitamin/mineral arm, and type of enrolling site (chelation practice or not). 3 Two additional subgroups were examined that were not prespecified: angina at baseline (Canadian Cardiovascular Society class >0) and heart failure at baseline (New York Heart Association class >0). Results Characteristics and QOL Data Collection Rates We randomly selected 911 (53.3%) of the 1708 patients enrolled in the TACT trial for the QOL substudy. In this QOL cohort, 451 patients were randomized to EDTA and 460 to placebo. The baseline characteristics of the study population were representative of the parent study population and were well balanced between the 2 treatment arms (Table 1). Overall, the median age was 65 years, 19% were women, and 93% were non-hispanic whites. From a total of 3403 expected follow-up patient QOL contacts out to 24 months, 3288 (96.6%) QOL questionnaires were collected. Expected contacts exclude deaths and withdrawals. The rate of missing QOL assessments did not differ by treatment group at baseline or any follow-up interval (Figure 1). QOL Outcomes Prespecified Principal Patient-Reported QOL End Points Both treatment groups showed an improvement in patientreported cardiac-related physical functioning, as assessed using the DASI, in follow-up compared with baseline assessments (Table 2; Figure 2). At baseline, the chelation group had a mean DASI score of 24.6, which improved to 29.1 at 6 months and 29.4 at 12 months but then fell to 27.1 at 24 months. The placebo group had a mean DASI score of 23.5 at baseline, 27.0 at 6 months, 26.3 at 12 months, and 25.1 at 24 months. After adjustment, between-group differences averaged 0.9 DASI points during the 24 months of QOL follow-up (95% CI, 0.7 to 2.6) and were not statistically different at any time point. Both treatment groups had baseline MHI-5 scores that were at the population-level benchmark (score of 50), indicating that there was no initial deficit in this QOL domain. Comparison of adjusted follow-up values showed that there were no clinically or statistically significant differences at any time point (mean overall adjusted difference between treatment groups, 1.0 [95% CI, 0.1 to 2.0]; Table 2; Figure 3). Additional Patient-Reported QOL Measures The SAQ anginal frequency scale exhibited a ceiling effect at baseline (median value was 100, which is the highest

4 Mark et al QOL Outcomes With Disodium EDTA Chelation for CAD 511 Table 1. Characteristics of Overall Cohort and Quality-of-Life Subset in the Trial to Assess Chelation Therapy QOL Subset by Treatment Group Overall TACT (n=1708) QOL TACT (n=911) Chelation Therapy (n=451) Placebo (n=460) Age, y, median (25th 75th) 65 (59 72) 65 (58 71) 64 (58 71) 65 (58 72) Female, % Nonwhite BMI, median (25th 75th) 30 (27 34) 30 (27 34) 30 (27 34) 30 (27 35) Used, % Married, % High school graduate or higher, % CCS angina, % None I II III IV <1 <1 <1 <1 NYHA heart failure, % None I II III Medical history, % Hypercholesterolemia Hypertension Angina Diabetes mellitus Heart failure Stroke Any cardiac revascularization medications, % β-blocker Statin ACE or ARB Aspirin, warfarin, or clopidogrel Multivitamin Other vitamins/minerals Herbal products Laboratory, mg/dl Total cholesterol 165 ( ) 164 ( ) 164 ( ) 165 ( ) High-density lipoprotein 43 (36 51) 43 (37 51) 43 (36 52) 43 (37 51) Low-density lipoprotein 89 (67 115) 87 (66 114) 87 (65 111) 88 (66 117) Triglycerides 139 (96 204) 141 (98 204) 134 (97 210) 147 ( ) ACE indicates angiotensin-converting enzyme; ARB, angiotensin receptor blocker; BMI, body mass index; CCS, Canadian Cardiovascular Society; NYHA, New York Heart Association; QQL, quality of life; and TACT, Trial to Assess Chelation Therapy. value of the scale, indicating no angina), corresponding to a baseline prevalence of Canadian Cardiovascular Society class >0 of 21%. Median observed values remained at 100 for both treatment groups during follow-up, and no difference in the distribution of adjusted values was evident by treatment (Table 3). The SAQ QOL scale, reflecting patient satisfaction with aspects of their life affected by angina, showed improvement in both treatment groups from baseline to initial follow-up, with little change thereafter and with no differential pattern according to treatment group (Table 3). Other secondary QOL measures, including the SF-36 generic health status scales and summary measures, did not show any clinically or statistically significant treatment differences between groups at any follow-up interval (Tables II and III in the Data Supplement). Neither the New York Heart

5 512 Circ Cardiovasc Qual Outcomes July 2014 Table 2. Prespecified Primary Quality-of-Life Measures by Intention to Treat in the Trial to Assess Chelation Therapy EDTA Chelation* (n=451) Placebo* (n=460) Adjusted Difference Between Chelation and Placebo (95% CI) Adjusted P Value Duke Activity Status Index Higher scores=better cardiac-related physical functioning; scale of 0 (worst) to 58.2 (best) Number of measurements Median (interquartile range) 21 (10 39) 19 (10 37) Mean±SD 24.6± ± mo Number of measurements Median (interquartile range) 27 (15 45) 26 (11 43) Mean±SD 29.1± ± ( 1.2 to 2.5) mo Number of measurements Median (interquartile range) 29 (14 45) 24 (10 42) Mean±SD 29.4± ± ( 0.4 to 3.6) mo Number of measurements Median (interquartile range) 25 (12 43) 23 (10 40) Mean±SD 27.1± ± ( 1.6 to 2.6) 0.63 All follow-up Number of measurements Median (interquartile range) 27 (13 44) 24 (11 42) Mean±SD 28.6± ± ( 0.7 to 2.6) 0.27 SF-36 Mental Health Inventory-5 0 to 100 scale, with higher scores representing better psychological well-being Number of measurements Median (interquartile range) 53 (44 57) 53 (44 60) Mean±SD 50.0± ± mo Number of measurements Median (interquartile range) 53 (48 60) 53 (46 60) Mean±SD 52.1± ± ( 0.8 to 1.7) mo Number of measurements Median (interquartile range) 55 (48 60) 53 (46 60) Mean±SD 52.1± ± ( 0.2 to 2.5) mo Number of measurements Median (interquartile range) 55 (48 60) 53 (46 60) Mean±SD 52.5± ± ( 0.1 to 2.6) 0.07 All follow-up Number of measurements Median (interquartile range) 55 (48 60) 53 (46 60) Mean±SD 52.3± ± ( 0.1 to 2.0) 0.08 Continuous variables shown as medians (25th to 75th percentiles) and means with SDs. CI indicates confidence interval; and SF-36, Short Form-36. *Observed data. Adjusted for repeated measures within a patient, with treatment group, time point, and the respective quality-of-life measure from the baseline questionnaire as predictor variables.

6 Mark et al QOL Outcomes With Disodium EDTA Chelation for CAD 513 Figure 2. Duke Activity Status Index. Scores range from 0 (worst) to 58.2 (best) and reflect the ability of patients to do physical activities without difficulty or assistance in 12 domains. Graph represents observed data shown by intention to treat. Association class nor the Canadian Cardiovascular Society angina class showed any evidence of a treatment-related difference (Table IV in the Data Supplement). Prespecified Subgroup Analysis None of the prespecified subgroups showed a treatment effect on QOL (data not shown). In particular, there were no treatment-specific QOL differences among patients with diabetes mellitus or anterior MI. Of the 2 post hoc subgroups we examined, no treatment effect was seen in the patients with heart failure symptoms at baseline. The 21% of TACT patients with angina symptoms at baseline (Canadian Cardiovascular Society class >0) had a 2.44 difference in DASI points at 6 months (P=0.20), a 4.99-point difference at 1 year (P=0.019), and a 1.21-point difference at 24 months (P=0.57), with an overall average difference of 2.86 DASI points (P=0.097) favoring EDTA chelation. No differences were seen in the SF-36 MHI-5 or the secondary QOL measures. Discussion The National Institutes of Health funded TACT in 2002 to assess whether EDTA chelation had therapeutically useful biological activity in coronary artery disease. TACT was also charged with defining the safety of this therapy. The primary clinical results of TACT have now been reported and show that during a 55-month median follow-up period, EDTA chelation did produce a clinically important and statistically significant reduction in the composite primary end point of the trial. 3 Of particular note, the pattern of benefit was not confined to a single component of the composite end point but was consistent across all the nonfatal components. No treatment effect was seen for mortality, but the study lacked power to make this evaluation. A major prespecified secondary end point of TACT was health-related QOL. Because chelation therapy was used in TACT in a stable post-mi population in anticipation of future events more than as a treatment for ongoing cardiac symptoms, its likely effects on QOL during the duration of the study were uncertain at the time the substudy was designed. In addition, we had no prior data involving the use of chelation therapy for coronary artery disease that would have allowed us to anticipate the specific QOL benefits of the chelation strategy being tested in TACT. Therefore, we made the general assumption that most likely beneficial QOL effects of chelation in coronary disease patients, if there were any to be found, would manifest as less angina and consequent improved functional status. We assembled a comprehensive but efficient QOL battery that was expected to provide reasonable ability to detect both positive and negative effects of chelation therapy. Recognizing that comparison of multiple QOL scales at multiple different time points might lead to differences because of chance rather than treatment and without any relevant prior chelation therapy research to guide us, we prespecified 2 QOL measures, DASI and MHI-5, as our primary measures of interest and assigned all the remaining measures to a secondary status. Our primary finding in the present study is that chelation therapy did not have a detectable effect on health-related QOL in TACT. Averaged during 24 months of follow-up, the EDTA chelation arm differed from the placebo arm by 0.9 DASI units (95% CI, 0.7 to 2.6) and 1 unit on the MHI-5 0 to 100 scale (95% CI, 0.1 to 2.0). Based on power projections made Figure 3. Short Form-36 (SF-36) Mental Health Inventory-5 (MHI-5). The MHI-5 assesses psychological well-being, including depression and anxiety with scores ranging from 0 (worst) to 100 (best).

7 514 Circ Cardiovasc Qual Outcomes July 2014 Table 3. Seattle Angina Questionnaire by Intention to Treat in the Trial to Assess Chelation Therapy EDTA Chelation* (n=451) Placebo* (n=460) Adjusted Difference Between Chelation and Placebo (95% CI) Adjusted P Value Seattle Angina Questionnaire Anginal frequency Higher scores=less frequent angina Number of measurements Median (interquartile range) 100 (80 100) 100 (80 100) Mean±SD 86.3± ± mo Number of measurements Median (interquartile range) 100 (80 100) 100 (90 100) Mean±SD 89.8± ± ( 4.1 to 1.3) mo Number of measurements Median (interquartile range) 100 ( ) 100 (90 100) Mean±SD 92.6± ± ( 2.7 to 2.7) mo Number of measurements Median (interquartile range) 100 (90 100) 100 (90 100) Mean±SD 92.9± ± ( 1.2 to 4.0) 0.29 All follow-up Number of measurements Median (interquartile range) 100 (90 100) 100 (90 100) Mean±SD 91.7± ± ( 2.2 to 2.2) 0.99 Anginal stability Score of 50=no change Number of measurements Median (interquartile range) 50 (50 50) 50 (50 50) Mean±SD 50.8± ±16.6 Quality of life 0 to 100 scale, with higher scores representing better quality of life Number of measurements Median (interquartile range) 75 (50 92) 75 (50 92) Mean±SD 69.8± ± mo Number of measurements Median (interquartile range) 83 (67 100) 87 (67 100) Mean±SD 79.4± ± ( 4.4 to 2.2) mo Number of measurements Median (interquartile range) 83 (75 100) 92 (67 100) Mean±SD 80.5± ± ( 4.3 to 2.6) mo Number of measurements Median (interquartile range) 83 (67 100) 83 (67 100) Mean±SD 79.4± ± ( 3.3 to 3.6) 0.92 (Continued)

8 Mark et al QOL Outcomes With Disodium EDTA Chelation for CAD 515 Table 3. Continued EDTA Chelation* (n=451) Placebo* (n=460) Adjusted Difference Between Chelation and Placebo (95% CI) Adjusted P Value All follow-up Number of measurements Median (interquartile range) 83 (67 100) 83 (67 100) Mean±SD 79.8± ± ( 3.5 to 2.3) 0.70 Continuous variables shown as medians (25th to 75th percentiles) and means with SDs. CI indicates confidence interval. *Observed data. Adjusted for repeated measures within a patient, with treatment group, time point, and the respective quality-of-life measure from the baseline questionnaire as predictor variables. during the planning phase of the study, 900 patients should have provided us with >90% power to detect the minimal clinically important difference in DASI (4 DASI units or 1/4 of an SD). Because the mechanism of the therapeutic action of EDTA chelation demonstrated in the overall TACT results remains unclear, the absence of a QOL effect is consistent with several different explanations. One possibility is that EDTA chelation alters aspects of coronary artery disease progression that do not have symptomatic correlates. The previously reported clinical results showed a treatment effect on coronary revascularization and hospitalization for angina. Although both of those end points would be expected to have a symptomatic element, it is possible that the duration of symptoms was truncated by application of effective therapies, and no long-term decrements in QOL were produced. Another possible explanation for our negative findings is that TACT selected a population in which there was little opportunity for a QOL benefit to be produced because of the low level of cardiac symptoms present at the time of enrollment. For example, if only some patients with coronary artery disease have the potential to experience a QOL benefit from EDTA chelation, such as those with baseline anginal symptoms (comprising 21% of our sample), our power would have been much lower than the estimate calculated above under the assumption that all patients could experience benefit. This possibility suggests that the effectiveness of this form of therapy to improve symptoms and functioning in patients with chronic cardiac symptoms will require additional study. Little modern clinical research on EDTA chelation therapy has been published. Several small trials of a 20-week regimen of chelation have been performed in intermittent claudication patients. 12,13 These did not show any evidence of a therapeutic effect. No prior trials have examined the effects of this therapy on either clinical events or QOL in a population similar to TACT. Examination of subgroup effects in a clinical trial is routine but must be approached with caution because of both the reduced precision of estimated treatment effect sizes and the possibility of chance variations in the data being misinterpreted as a treatment-related effect. None of the prespecified subgroups we examined showed any treatment signal on QOL outcomes. Of the 2 post hoc subgroups we evaluated, only the patients with angina at baseline showed a suggestion of a treatment effect evident on cardiac-related functional status (DASI) at 6 months, peaking at 1 year and then greatly attenuated at 2 years. These data, therefore, are not clear enough to be confident that a chelation-related treatment effect is responsible for the observed differences. Rates of use of evidence-based medicine were generally high in this trial, and no treatment-related differences were seen during follow-up. Coronary revascularization was reduced by 3 per 100 by EDTA chelation at the end of followup, 3 but no difference was seen by the 2-year follow-up in our QOL sample. Thus, the likelihood that differential use of medical therapies or revascularization affected our results so as to have masked a treatment-related difference seems small. Several caveats should be considered in the interpretation of our results. As mentioned earlier, the selection of a trial population that was largely asymptomatic at baseline may have reduced our chances for observing an effect of treatment on QOL. Additional study in patients with chronic angina will be required to assess the possibility that EDTA does have an effect on ischemic symptoms. Because of a limitation in funding for this part of the TACT program, we were not able to obtain follow-up QOL assessments beyond 2 years. Although the anecdotal reports of symptomatic benefit of chelation suggest an effect that should be detectable within the first 2 years, the pattern of prognostic benefits evident in the overall TACT trial raises the possibility that the effects of chelation are biologically persistent long after the infusion phase of therapy has concluded. Another potential limitation of our study relates to the amount of missing QOL data we had (Figure 1). Most analytic approaches to QOL data make the useful, but unprovable, assumption that missing data are missing at random. However, when data are missing because of either death or illness severity, the missing at random assumption is unlikely to be true. In the present study, treatment had no effect on mortality, and relatively equal numbers of patients were missing QOL data but alive at each follow-up point. These 2 factors make it unlikely that the missing data caused a clinically important bias in measured QOL effects, resulting in the masking of a true QOL benefit of chelation. In conclusion, our study shows that in a population of stable, largely asymptomatic coronary artery disease patients with prior MI, use of EDTA chelation therapy did not produce a measurable change in health-related QOL during 2 years of follow-up. Acknowledgments We are particularly indebted to the coordinators at the Trial to Assess Chelation Therapy (TACT) sites who collected data and to the patients who agreed to provide their quality-of-life data for the trial.

9 516 Circ Cardiovasc Qual Outcomes July 2014 Sources of Funding This work was supported by grants from the National Heart, Lung, and Blood Institute (U01 H092607) and from the National Center for Complementary and Alternative Medicine, National Institutes of Health, Bethesda, MD (U01 AT001156). Disclosures Dr Mark has received grant funding from Eli Lilly & Company (significant), AstraZeneca (significant), Medtronic (significant), and Gilead (significant) and has served as a consultant for Janssen. Dr Anstrom has received research support from AstraZeneca (significant), Eli Lilly & Company (significant), and Medtronic (significant); has served as a consultant for Abbott Vascular (modest), AstraZeneca (modest), Bristol Meyer Squibb (modest), Pfizer (modest), GSK (modest), and Ikaria (modest); and has served on Data Monitoring Committees for Pfizer (modest) and Vertex (modest). The other authors report no conflicts. References 1. Barnes PM, Bloom B, Nahin RL. Complementary and alternative medicine use among adults and children: United States, Natl Health Stat Report. 2008;12: Lamas GA, Goertz C, Boineau R, Mark DB, Rozema T, Nahin RL, Drisko JA, Lee KL. Design of the Trial to Assess Chelation Therapy (TACT). Am Heart J. 2012;163: Lamas GA, Goertz C, Boineau R, Mark DB, Rozema T, Nahin RL, Lindblad L, Lewis EF, Drisko J, Lee KL; TACT Investigators. Effect of disodium EDTA chelation regimen on cardiovascular events in patients with previous myocardial infarction: the TACT randomized trial. JAMA. 2013;309: Escolar E, Lamas GA, Mark DB, Boineau R, Goertz C, Rosenberg Y, Nahin RL, Ouyang P, Rozema T, Magaziner A, Nahas R, Lewis EF, Lindblad L, Lee KL. The effect of an EDTA-based chelation regimen on patients with diabetes mellitus and prior myocardial infarction in the Trial to Assess Chelation Therapy (TACT). Circ Cardiovasc Qual Outcomes. 2014;7: Hlatky MA, Boineau RE, Higginbotham MB, Lee KL, Mark DB, Califf RM, Cobb FR, Pryor DB. A brief self-administered questionnaire to determine functional capacity (the Duke Activity Status Index). Am J Cardiol. 1989;64: Hlatky MA, Rogers WJ, Johnstone I, Boothroyd D, Brooks MM, Pitt B, Reeder G, Ryan T, Smith H, Whitlow P, Wiens R, Mark DB. Medical care costs and quality of life after randomization to coronary angioplasty or coronary bypass surgery. Bypass Angioplasty Revascularization Investigation (BARI) Investigators. N Engl J Med. 1997;336: Ware J, Jr, Snow KK, Kosinski M, Gandek B. SF-36 Health Survey: Manual and Interpretation Guide. Boston, MA: The Health Institute, New England Medical Center; Spertus JA, Winder JA, Dewhurst TA, Deyo RA, Prodzinski J, McDonell M, Fihn SD. Development and evaluation of the Seattle Angina Questionnaire: a new functional status measure for coronary artery disease. J Am Coll Cardiol. 1995;25: Ware J, Jr, Kosinski M, Keller SD. A 12-Item Short-Form Health Survey: construction of scales and preliminary tests of reliability and validity. Med Care. 1996;34: EuroQol--a new facility for the measurement of health-related quality of life. The EuroQol Group. Health Policy. 1990;16: Moser EB. Repeated measures modeling with PROC MIXED. 29th SAS Users Group International Conference. Montreal, Canada van Rij AM, Solomon C, Packer SG, Hopkins WG. Chelation therapy for intermittent claudication. A double-blind, randomized, controlled trial. Circulation. 1994;90: Guldager B, Jelnes R, Jørgensen SJ, Nielsen JS, Klaerke A, Mogensen K, Larsen KE, Reimer E, Holm J, Ottesen S. EDTA treatment of intermittent claudication a double-blind, placebo-controlled study. J Intern Med. 1992;231: