Why the NIH Trial to Assess Chelation Therapy (TACT) Should Be Abandoned

Kimball C. Atwood IV, MD; Elizabeth Woeckner, AB, MA; Robert S. Baratz, MD, DDS, PhD; Wallace I. Sampson, MD


Medscape J Med. 2008;10(5):115 

In This Article

III. State of the Evidence

In the American Heart Journal in 2000, Dr. Gervasio Lamas argued that regardless of disconfirming RCTs, chelation for CAD ought to be studied in a large human trial: "...case reports and case series encompass thousands of patients with successful outcomes...; a small or moderate benefit of EDTA chelation cannot be excluded...; we believe that a state of clinical equipoise exists and support a definitive trial....[29]" In TACT protocols submitted in 2001 and 2003, he reiterated this assertion, citing "biologically plausible mechanisms," "ample suggestions of benefit" from "the very large number of published case reports," and the small sample sizes of controlled trials to date.[4,5] According to the 2003 protocol, "the totality of the evidence on chelation therapy includes basic research, clinical investigation, descriptive and observational epidemiologic studies and 3 small, randomized trials.[5]"

By 2003, the reports of the Canadian PATCH study of Chelation for Ischemic Heart Disease -- which, in a sample size of 84, had found no effect on "exercise time to ischemia, exercise capacity, quality of life measurements" or, in a substudy of 47 subjects, endothelial function -- had been published.[27,28] This brought to 275 the number of subjects acknowledged by Dr. Lamas to have been studied in adequately designed, controlled trials of IV Na2EDTA for either CAD or PVD. In none of those trials was a treatment effect demonstrated.[21,22,25,27,28]

Another small RCT of chelation for CAD, not cited by Dr. Lamas, had found no difference between active and placebo groups of 8 subjects each.[26] Dr. Lamas himself has conducted a "pilot" RCT, "PACT," in which 30 of 40 subjects were to have received Na2EDTA.[5] It should have been completed several years ago, but its results have not been made public. We suspect that it also did not support chelation, because otherwise it would have been a newsworthy event that the NCCAM and its ACAM collaborators could hardly have resisted publicizing.

Antioxidant vitamins, also part of the TACT regimen, have been extensively studied for the prevention or treatment of atherosclerosis and have failed to demonstrate a favorable effect.[151]

The relevant basic research, previously cited, suggests the opposite of what the TACT investigators have hypothesized: EDTA is more likely to act as a pro-oxidant, particularly in the presence of iron and vitamin C, than as an antioxidant.[147,148,149] More recent work has supported this finding.[236] Others have noted pro-oxidant effects of vitamin C and its tendency, when taken in large doses, to enhance iron absorption from the gut.[150,237,238,239,240] The importance of potential pro-oxidant effects of vitamin C in vivo is uncertain.[241]

In scarce recent studies of EDTA in animal models of atherosclerosis, the results have been conflicting.[242,243] The 2001 TACT protocol does not cite animal studies; the 2003 protocol cites only one,[244] published in a journal devoted to chelation advocacy[82] and co-authored by Edward McDonagh, who had previously admitted under oath that he had "kept slipshod records, recording only positives about his experiments, to avoid possible liability lawsuits.[44]"

Thus Dr. Lamas' "totality of the evidence" in support of chelation for CAD appears limited to case series reporting "thousands of patients with successful outcomes.[29]" We therefore consider the case series in some detail. Next we discuss the RCTs that had been completed by the time the TACT began; there is no need to do so at length because the TACT protocols do not dispute their validity, only their "power." We do, however, comment on the import of those trials when contrasted with the case series. We also compare the existing evidence for chelation with evidence for other health claims that were long ago dismissed.

Very few case series of chelation for atherosclerosis, including none of the large ones, are credible. Only the early series (from the 1950s and 1960s) are worthy of consideration, but the TACT protocols misrepresent their findings by omitting pertinent information, reporting erroneous data, and citing "positive" data pools more than once. A naive reader could become aware of such errors and duplications only by consulting the primary sources. The TACT protocols also fail to observe that enthusiastic reports among the early series are best explained by nonspecific effects that were known, even at that time, to be common to uncontrolled trials of treatments for angina pectoris. The most plausible of the early reports concluded that there was no evidence for an effect of chelation on the course of CAD[54]; the 2003 TACT protocol summarizes that report as having suggested almost the opposite.[5]

The later "positive" series cited by the 2003 TACT protocol are even less credible than the early ones: The largest was unpublished, and its author denied well-documented complications and deaths. Each series reported outcomes that are unlikely in the extreme. All were written by a few fringe practitioners -- "pseudoscientific zealots," in the words of an expert who studied both chelation and the social phenomenon of chelationists[9,10,11] -- with significant financial interests in chelation, Laetrile, IV hydrogen peroxide, "ozone therapy," "electrodermal screening," and other dubious methods. Most were published in a journal created to make chelation and other dubious methods appear to have scientific backing. The journal is the mouthpiece of the ACAM, the major chelation advocacy organization.[82] Neither the TACT protocols nor Dr. Lamas' editorial in the American Heart Journal suggests such distinctions.[4,5,29] On the contrary, the TACT protocols call the pseudoscientific zealots "prominent experts.[4,5]"

On page 8 of the 2003 version of the TACT protocol is a table purporting to summarize the results of case series of chelation for CAD. We reproduce it here ( Table 2 ), save the original reference numbers.

Perusing Table 2 , it can be seen that there is a temporal gap between the early studies, ending with Lamar (1966), and the later ones, beginning with Evers (1979). As suggested above, the early articles were part of the legitimate medical literature of the time, whereas the later ones were either unpublished or were published in non-refereed, advocacy journals. The early investigations abruptly ended when it became clear that chelation did not change the natural history of CAD, and that whatever subjective and ECG effects it seemed to have were transient and similar to those reported in early trials of other treatments later shown to be ineffective. Neither Table 2 nor the text of the TACT protocols gives any suggestion of that history.

Table 2 is also misleading in a more direct sense: several of the cited case series are redundant; several are erroneously reported; and several did not address the relevant question. We therefore discuss the early series in some detail, contrasting them with their portrayals in Table 2 .

Clarke (1956)[245]: In 1955 Clarke and colleagues[245] reported the results of IV Na2EDTA given to a single subject with "extensive nephrocalcinosis." The subject also had impaired hearing and a history of an ulcer with "3 attacks of intestinal bleeding." According to the authors, "50-65% of the metastatic calcium [was] dissolved out of the kidney"; there was a "striking improvement" in hearing; and "ulcer symptoms [were] completely relieved." The authors mentioned having treated 22 patients with "EDTA for various related diseases and the general results will be the subject of a future paper." They did not report those results in the 1955 article. Thus the entry in Table 2 -- "Sample Size 22, some improvements" -- is erroneous.

Clarke (1956)[32]: In 1956 Clarke and colleagues[32] reported the results of IV Na2EDTA given to 20 subjects with angina pectoris. Their rationale was that the drug was a potent chelator of Ca++, and therefore might "disintegrate the organic atheromatous matrix and allay the symptoms of angina pectoris." The report was of an uncontrolled series lacking eligibility or inclusion criteria other than "angina pectoris," the definition of which was not stated. Diagnoses were not standardized, but appear to have been made by 1 or more of the following criteria: frequency of chest pain, exertion necessary to bring on chest pain (usually measured by the number of city blocks walked), number of nitroglycerin tablets used per day, ability to work, history of MI, and resting ECG. The "disease was confirmed independently by two or more physicians," but there was no mention of potential or actual diagnostic disagreements.

Most subjects were followed from 1 to 12 months; 5 were followed for up to 22 months. The study authors reported dramatic, subjective improvements in frequency of angina and in exercise tolerance for 19 subjects, and improved ECG findings in 6 subjects. Table 3 shows the differences in self-reported horizontal walking distances before the onset of angina for the 8 subjects in which the study authors reported both pre- and posttreatment figures.

The only relevant data from a controlled trial, with which to compare these walking distances, comes from the Canadian PATCH study.[27] It did not report distances per se, but "time to ischemia," defined as "time to reach at least of 1 mm ST-segment depression" on a treadmill. Among 84 subjects with stable angina, after 27 weeks and 33 Na2EDTA or placebo infusions, the mean time to ischemia increased from approximately 9.5 minutes to approximately 10.5 minutes. There was no difference between the active and placebo groups, and the PATCH study authors observed that the improvement was "consistent with a combination of placebo and training effects commonly seen in studies of angina patients.[27]"

In the Clarke study, 1 subject died; the authors "suspected that he died from a calcium embolus that had been freed from a large arterial plaque.[32]" At autopsy, "his aorta showed extensive atheromatosis....The coronary artery sections showed intense medial sclerosis with calcification....[32]" The autopsy report did not suggest "disintegration" or decalcification of atherosclerotic plaques.

As they had in their previous report, Clarke and colleagues also reported dramatic effects for conditions other than angina, including complete resolution of intermittent claudication, musculoskeletal pain, and ventricular ectopy. They reported, in 3 subjects who may or may not have been among those treated for angina, the following: healing of a necrotic neck ulcer containing "extensive deposits of metastatic calcium" that had been present in a man for 7 years after thyroidectomy and radiation therapy for cancer; continued dissolution of kidney stones, seemingly in the subject described in their previous report; and removal of calcium from the mitral valve of a woman with rheumatic heart disease, whose "immediate symptomatic relief resemble[d] a successful surgical commissurotomy."

The man with the neck ulcer "extruded through a sinus tract many small to large calcified plaques" for at least 10 months after chelation treatments had ended. The man with nephrocalcinosis "continued to pass large amounts of sand and stones in his urine" for at least a year after the last treatment.

Boyle (1957)[246]: Boyle and colleagues,[246] including Clarke, published 7 detailed case reports among a group of "20 patients with angina." These appear to be identical to the 20 subjects with angina reported in "Clarke (1956).[32]" Thus, Table 2 is in error in reporting them twice. Boyle 1957, moreover, refers to "the numerous patients with angina treated by the Providence Hospital group in Detroit," suggesting that considerably more than 20 had been treated. This raises the possibility of selection bias.

Clarke (1960)[50]: In 1960 Clarke and colleagues[50] published another uncontrolled series of Na2EDTA treatments in 76 subjects with angina pectoris, 31 subjects with intermittent claudication, and 25 subjects with cerebrovascular disease. Unstated numbers of subjects in each of the latter groups appear to have also been included in the angina group.

The authors raised the possibility of a controlled trial, but dismissed it because "the double blind test for appraising new drugs has been found to be far from infallible...." They did not identify any subjects as having been involved in their previous series. Eligibility criteria were vague. The diagnostic criteria for angina pectoris appear to be similar to those in their previous report, but in the newer report there were no individual case descriptions. The diagnostic criteria for intermittent claudication seem to have included some combination of a history of rest pain or pain with walking, physical exam suggestive of circulatory insufficiency, and "greatly decreased oscillometric readings."

The authors reported that they had followed all subjects for 2 years. They rated 87% of subjects with angina "much improved," defined as "90% to 100% symptom relief," and 9.2% as "improved," defined as "75% symptom relief." Paradoxically, 13% died. There were no reports of autopsies, ECGs, or other objective outcomes. The TACT protocol's "Results" entry for the 1960 Clarke series -- "58 improved" -- does not follow from the report ( Table 2 ).

Of the 31 subjects with intermittent claudication, 23 were rated "90% to 100% relieved" and 4 were rated "75% relieved." Three died: 2 of myocardial infarction" and 1 of "C.V.A." Among the 25 subjects with cerebrovascular disease, "all who had cerebellar vertigo or dizziness had complete relief and the manifestations of senescence were significantly alleviated...; the paralysis caused by recent cerebrovascular accidents seemed to improve faster and more completely than usually expected and tinnitus, even of long standing, has been relieved." Six of these subjects died, 2 of "C.V.A." and 4 of "congestive heart failure." No other quantitative outcome data were reported.

As they had done previously, the authors also reported beneficial effects on other conditions: Several subjects, including 2 with "familial hypercholesterolemia," experienced dramatic reductions in serum cholesterol lasting up to 2 years after their final chelation treatments.

In an editorial, also published in 1960, Clarke referred to "an accumulated experience of several hundred patients...[with]...occlusive vascular disease...," including "a large [number unstated] series of patients with angina pectoris.[247]" The 1960 series, however, appears to be the final report from the Providence Hospital group of their own subjects treated with EDTA for atherosclerosis. This constitutes more evidence of selection bias.

Meltzer (1960)[51]: In 1960 Meltzer and colleagues[51] reported an uncontrolled series of Na2EDTA treatments in 10 male subjects with "severe, persistent angina pectoris which had not been controlled with previous therapies." Six had previous MIs; 9 had abnormal ECGs "compatible with coronary artery disease"; 3 had "radiographic evidence of cardiomegaly"; and 3 had intermittent claudication. After 2-3 months the treatments were stopped because of an apparent lack of efficacy, but after 3 more months the subjects were noticeably better: "Nine out of ten men had a significant reduction in the number and severity of anginal attacks, five out of nine electrocardiograms showed improvement, and all three patients with cardiomegaly showed a reduction in heart size."

The study authors also reported x-ray evidence of reduction in calcification of an aortic valve in 1 subject, "rather dramatic improvement" of "long-standing extensive keratotic skin lesions" in another, and "definite decreases in severity of...symptoms" in 2 subjects with arthritis. The intermittent claudication in 3 subjects, however, did not improve.

Kitchell (1961)[248] and Boyle (1961)[249]: The 10 subjects reported by "Meltzer (1960)" appear identical to the 10 subjects attributed in Table 2 to "Kitchell (1961)" and to "Boyle (1961)." These 10 are also among those reported in "Kitchell (1963),[54]" in which they and 28 new subjects were treated and observed for a substantially longer period. Thus, "Kitchell (1963)" is the only report from which conclusions about the original 10 can legitimately be drawn, as will be further discussed. Table 2 has erred in listing the original series of 10 subjects 3 times, thus implying that there were 3 distinct series, and in failing to acknowledge the final report of that series.

Meltzer (1961)[250]: This study of chelation treatments in 81 subjects did not report therapeutic outcomes at all; it reported, as suggested by its title, side effects and toxicity of Na2EDTA. Table 2 quotes the article as having reported that the treatment was "effective" at producing an outcome that was not stated. This appears to have been inferred from a phrase in the article's final paragraph: "...our plan of alternate day infusions has been therapeutically effective...." Two sentences later, however, Meltzer and colleagues made it clear that the intended meaning of "effective" in this context was not what the author of Table 2 must have imagined: "We are currently studying the maximum effectiveness (calcium excretion) with varying periods."

There is language in the article suggesting that Meltzer and Kitchell's group may also have been guilty of selection bias: "...infusions of disodium EDTA were given to 81 subjects in a study of the effectiveness of this therapy in coronary artery disease during a two-year period...; 62 patients have now been followed from 8 to 30 months." Yet these investigators eventually reported outcomes in only 38 subjects.

Kitchell (1963)[54]: In 1963 Kitchell and Meltzer's group issued a "reappraisal" report. They had found that after observing and treating their original 10 subjects for up to 2 more years, and having added and followed another 28 subjects for 18 months, the apparent early improvements had proved illusory. MIs and deaths occurred at rates that would have been expected without chelation treatments. Five of the original 10 were dead, all by MI. Among the survivors, initial symptomatic and ECG improvements had largely disappeared; only 2 of the original 10 were thought still "improved."

Among the 28 new subjects, 16 had considered themselves "improved" and 2 "markedly improved" after 6 months; those numbers could be the source of the "18 improved" entry in Table 2 . After the full 18 months of the study, however, 7 of the 28 were dead (6 of MI, 1 in congestive heart failure); 2 were worse, 6 were deemed "no change"; and 13 "remained improved."

Of the 13 subjects who died, 6 had autopsies: none showed evidence of decalcification or attenuation of coronary plaques.

The study authors concluded that chelation "does not significantly alter the natural history of coronary artery disease, nor does it offer any protection against repeated infarctions and death." They granted that it "may slightly influence the anginal syndrome," but cautioned that "similar percentages of improvement...have been reported after varied medical and surgical therapies.[54]"

Thus the 2003 TACT protocol has misrepresented the findings of Kitchell and colleagues by citing the original positive sample 3 times, by failing to acknowledge the subsequent unfavorable outcome of that sample, and by reporting erroneous results of the additional sample of 28 ( Table 2 ). The table's entries for "Kitchell (1963)" are wrong in every category: Sample Size, Outcome Measures, and Result.

The discussion in the text of the protocol is only slightly less misleading: "In 1963, Kitchell and co-workers reported on 28 patients...; within 3 months of therapy, about 60% of patients reported improvement based both on patients' impression and their documented exercise tolerance. Nonetheless, the benefit was not felt to be long-lasting.[5]" There is no mention of the original 10 subjects, the deaths, the MIs, the autopsies, or the study authors' conclusions, including their telling remark about "similar percentages of improvement." A reader of the 2003 TACT protocol would likely suppose that the outcomes of "Kitchell (1963)" were nearly the opposite of what Kitchell and colleagues reported.

Lamar (1964)[52]: In 1964 Lamar reported the results of Na2EDTA treatments in a group of diabetics:

Of 15 diabetic patients suffering from severe vascular complications, all were relieved of their various degrees of peripheral vascular insufficiency: 8 were restored to normal after suffering from strokes or advanced brain syndrome; 4 were improved of their cardiac failure and 3 patients with different degrees of diabetic retinopathy obtained dramatic subjective and objective benefits....Seven of these 15 diabetics have shown persistent improvement with reductions of their insulin needs after EDTA.[52]

There were no eligibility criteria other than "diabetic with severe occlusive vascular disease," which seems to have included PVD affecting the lower extremities in all subjects, and cerebrovascular disease as evidenced by stroke or "chronic brain syndrome" in 8 subjects. None of the subjects were reported to have had angina or CAD prior to treatment. The diagnostic criteria for PVD were not standardized or specified for each subject, but were reported to be "some degree of peripheral vascular insufficiency manifested by atrophic pedal changes, diminished pulses and oscillometric readings, vascular calcifications observed in x-ray films, and by intermittent claudication or rest pain of the lower extremities."[52]

Lamar provided detailed case reports of 8 subjects, mentioning evidence of PVD in only two. Three of the 8 died, but only after each had demonstrated, according to Lamar, dramatic clinical improvements. ("After one of these active, fun-filled days, she was found dead one morning....") Autopsies were performed in 2 subjects: In the first, "Severe generalized arteriosclerosis involved primarily the abdominal aorta and its branches and as well as the coronary arteries...[there was] fresh myocardial necrosis of the anterior wall and interventricular septum..."; in the second there was "...unusually severe generalized arteriosclerosis....Death had occurred from recent thrombosis of the basilar artery. The walls of the cerebral artery system were markedly calcified [italics in the original]." Lamar's curious interpretation of 3 dead subjects having improved suggests that there may have been other deaths, left unmentioned.

Lamar also reported a case of a nondiabetic 54-year-old man, "disabled for 5 years by hypertrophic osteoarthritis of the hands...and a painful calcific tenosynovitis of the left shoulder with a large calcified area visible on the x-ray film and easily felt by palpation." The man had requested chelation treatments

...after the satisfactory results observed in his own severe diabetic 76-year-old and senile mother...for whom EDTA totally relieved severe rest leg pains and intermittent claudication...[and who] became a more stable diabetic....Above all, the most appreciated benefit she enjoyed was disappearance of senile irritability and intolerance, and conversion to a pleasant and cheerful attitude.[52]

Lamar described the results of EDTA given to the son:

After receiving 45 gm of EDTA in 3 weeks, he reported excruciating pain at the shoulder. Palpation gave the feeling of many crepitant, small, gravel-like pieces, apparently fragments of the calcified area previously felt as a solid hard mass. X-ray films confirmed this impression and a local injection of a steroid and local anesthetic combination resulted in immediate and lasting relief. At the completion of the series of 30 injections, there was no pain at all; no calcification was detectable in the x-ray film or by palpation; and full mobility was restored to the shoulder and to the hands and fingers. A year later, this very grateful patient was found still entirely normal on follow-up examination.[52]

Lamar, like the other authors, revealed that his report might have suffered from selection bias: "The cases described above belong to the group of the first 15 diabetics among some 40 patients affected with severe occlusive vascular disease....[52]"

Thus Table 2 's designation of "15 improved" for the Lamar report is in error, even discounting the absence of angina or CAD as the target of the intervention. Atherosclerosis per se is a reasonable basis for including an article in a literature search for evidence that chelation might be effective for CAD, but this article is not accurately represented by Table 2 : At least 3 of the 15 subjects died, and among the rest the diagnostic criteria were mostly unstated and not rigorous enough to draw conclusions. The naive enthusiasm of the author and the evidence of selection bias further suggest that the stated results are doubtful, at best. Lamar's subsequent article, discussed next, offered no grounds for granting a benefit of that doubt.

Lamar (1966)[53]: The table's entry for this report does not appear to support a role for Na2EDTA in treating CAD or atherosclerosis, nor does it make arithmetic sense: "Sample size: 3. Result: 1 improved, 1 died." We summarize it only to further characterize the lack of rigor both in Lamar's work and in the table's entries. In 1966 Lamar reported:

A carefully selected group of 50 patients with various forms of occlusive atherosclerosis, have been under study for the last six years [sic]....The consistently good results may be reasonably interpreted as due to the breakdown of the atherosclerotic plaques as EDTA chelated and removed metastatic calcium from the lesions, thus widening the contracted arterial lumen [etc.]....In a private practice it is not possible to carry on double-blind or cross-over control studies, so we must rely upon the experience and clinical judgment acquired in four decades of intensive medical practice and observation.[53]

Unfortunately, criteria for "careful selection" of patients are almost entirely absent from the report, and outcomes are described for only 4 subjects. One died; his autopsy, like the two in the previous report, showed that "the amount of severe atherosclerosis in most...areas of the body was quite extensive...." Of the other 3 subjects, one was described as having CAD: "...at age 70 years [he] had had severe congestive left ventricular failure for five years following a myocardial infarction." His ECG "show[ed] first-degree heart block, complete RBBB, old posterior wall infarction, left ventricular hypertrophy and strain...toxic effects of digitalis." He had:

...labored breathing, ascites, marked edema of the feet and legs, cardio-hepatomegaly....He had been intensely digitalized and was also receiving 2 to 3 intramuscular injections of mercurial diuretics per week and daily oral doses of thiazide diuretics....His diet was rigidly restricted in sodium. Nausea, anorexia and severe mental depression had developed.[53]

After he began chelation treatments, he was said to have improved markedly over a period of about 2 months: He lost 12 lb and his dyspnea, peripheral edema, hepatomegaly, ascites, and cardiomegaly were reported to have completely resolved. The ECG now showed "elimination of the complete RBBB, isoelectric S-T segments in all leads, improvement in myocardial depolarization, and disappearance of the signs of left ventricular strain and of digitalis toxicity." During that time, however, his digitoxin dose had been reduced from 0.25 mg daily to 0.1 mg daily, and his mercurial diuretics had been discontinued. Lamar did not provide information about serum electrolytes, nor did he comment on whether the man was hospitalized or was otherwise in bed.

This man's marked improvement is most simply explained by relief from digitoxin toxicity, effective diuresis, and gradual replenishment of whole-body potassium and sodium after the cessation of the mercurial diuretic. At most, Na2EDTA treatments may have tweaked the process by transiently lowering serum calcium and increasing intravascular volume during and shortly after infusions -- thus possibly reducing the toxic effects of digitoxin[55] while increasing the glomerular filtration rate -- but there is no need to invoke other, specifically antiatherosclerotic effects.

In this article Lamar also reported an incidental, complete reversal of "prostatism" in a man who was being treated for PVD. According to Dr. Lamar, the softening and shrinking of the prostate were heralded, after 6 months of chelation treatments, by "the passing of large amounts of urinary 'sand'...." Within 2 months the man's prostate was normal.

By 1966 Lamar should have been aware of the 1963 Kitchell reappraisal report. He did not cite it, although he cited all of the Clarke articles and even "Kitchell (1961)." Such selective citations, together with further evidence of selective reporting and a barely qualified enthusiasm for chelation as a near-panacea for all manner of "calcinoses," render Lamar's 2 reports unworthy of being taken seriously. Table 2 's entry for the report, in any case, bears no relation to the report itself. The apparent reason for that will become clear.

Soffer (1964)[55]: This report is notable for not being cited by the TACT protocols. Alfred Soffer, later to become the editor of both Chest and the Archives of Internal Medicine, gave 2 courses of 20 treatments of Na2EDTA to each of 5 patients with "signs and symptoms of peripheral atherosclerosis":

With one exception, intermittent claudication, standardized walking tolerance, oscillometric data, dependent rubor or blanching upon elevation of the extremity, and strength of pulses were relatively unaffected by therapy.[55]

Soffer also found the treatment ineffective in 3 patients with scleroderma heart disease and in 1 patient with aortic stenosis. Together with findings of no improvement in the only 3 subjects with intermittent claudication among Meltzer and Kitchell's original group of 10,[51] Soffer's small series raised a red flag.

Others had noticed, even by the 1950s, that several proposed treatments for angina pectoris had at first been reported to have dramatic effects, only to be found ineffective when subjected to more rigorous study -- usually in the form of controlled trials. In 1961 Beecher illustrated this by reviewing the history of bilateral ligation of the internal mammary arteries, a method that had enjoyed a brief popularity at about the same time as the early chelation reports.[251] Beecher contrasted the reports of "enthusiasts," who reported outcomes comparable to the early chelation series, with those of "skeptics" who were wary enough to perform controlled trials. In the case of internal mammary artery ligation, 2-3 negative controlled trials involving a total of about 40 subjects convinced experts in the field to dismiss prior positive case series involving thousands. The practice was promptly stopped.

Later, Benson and McCallie expanded the discussion. They reviewed the history of 5 ineffective treatments for angina pectoris -- 3 medical and 2 surgical -- ranging from the 1920s to the 1960s, and found a striking pattern:

Data from enthusiasts' studies reveal that subjective improvement was seen in 82.4±9.7 per cent (mean ± S.D.). This small variance in 13 studies of 1187 patients, evaluating five different inactive treatments over the course of four decades, strongly attests to the constancy and predictability of the placebo effect in the relief of angina pectoris.[252]

The authors found that objective improvements, including increased exercise tolerance, reduced use of nitroglycerin, and improved ECGs, were also common to early studies by "enthusiasts," although not so common as subjective improvements. Finally, they demonstrated a consistent finding of treatments that were eventually studied in controlled trials: About 35% of subjects in each group, "active" or placebo, reported improvement.

We summarize the early "case series" cited by the 2003 TACT protocol as follows. They were small, uncontrolled, vaguely described series involving only 3 groups of investigators. Two of the groups (Clarke and colleagues and Lamar) were unrepentant enthusiasts who touted chelation for multiple conditions, and the third group (Kitchell and colleagues) was, at first, only slightly less enthusiastic. Many of the individual case reports are quite fanciful. There is evidence of selective reporting by all 3 groups. Only the Kitchell "reappraisal" report demonstrated the objective outcomes, attention to follow-up, and skepticism that can make an uncontrolled case series useful, but it was misrepresented by the TACT protocol. The other notable disconfirming report of the time, by Soffer, is not cited by the TACT protocols.

In a review of chelation for occlusive vascular disorders, published in the Federal Register in 1970, the National Academy of Sciences--National Research Council's "Panel on Cardiovascular Drugs" summarized only the Kitchell and Soffer reports. The Panel called "investigations by other authors," citing Clarke and Lamar, "little more than testimonials.[253]"

Even prior to the Kitchell "reappraisal," the previous reports' deaths by MI and autopsies showing extensive atherosclerosis had suggested the truth of the matter: The early chelation reports had demonstrated the same pattern of error that had occurred in reports of other ineffective treatments for angina pectoris.[243] Only Kitchell and colleagues eventually acknowledged that history, possibly because of what they had learned from having also been among the early enthusiasts for mammary artery ligation.[251,254] Their reappraisal and Soffer's 1964 report brought chelation back to reality: There was no credible evidence for an effect of IV Na2EDTA on CAD or on atherosclerosis in general.

Later case series. The later case reports provide even less support for chelation. Not long after the 1963 reappraisal article,[54] reports of trials of Na2EDTA as a treatment for atherosclerosis largely disappeared from the mainstream medical literature for nearly 3 decades. When they returned, they did so in the form of reasonably well-designed, controlled trials, all of which were negative.[21,22,23,24,25,26,27,28] The remaining positive case series cited by Dr. Lamas have thus emanated from a tiny group of practitioners who share an enthusiasm for chelation, Laetrile, IV hydrogen peroxide, and other fringe practices, and who fail to appreciate the pitfalls of uncontrolled case series performed by enthusiasts. Their reports are not credible, and not merely because they were uncontrolled.

What follows are summaries of the 3 largest case series listed in Table 2 , which provide the bulk of Dr. Lamas' "thousands of patients with successful outcomes,[29]" and a summary of "Casdorph (1981)[255]" because it is discussed in the text of the 2003 TACT protocol. There are also brief discussions of 3 other pertinent references. The first is a "meta-analysis,[95]" the first author of which is L. Terry Chappell. That article, reporting 19 studies involving more than 22,000 subjects, is the source for much of the "evidence" cited by the second reference, a book written by the TACT Trial Chelation Consultant, which the 2001 TACT protocol cites in support of "published case series suggesting clinical benefits.[4,57]" The third reference is a case report of a single subject followed by academic investigators at the same time that he was being treated at a "chelation clinic.[24]" The discrepancies in reported outcomes between the 2 settings are revealing.

Evers (1979)[256]: This is an unpublished "review of a six-year report of chelation therapy given to about 3,000 patients..., which has primarily been directed at the treatment of vascular insufficiency due to atheromatous vascular involvement." Inclusion criteria appear to be only what can be gleaned from that statement; Evers did not report how many subjects were considered to be in each of the several possible subcategories. Diagnostic criteria, when mentioned, and outcomes were described only retrospectively:

In arteriosclerotic obliterans cases, specific changes include: definite improvement of pedal artery pulsation, gain in color, return in normal temperature and improvement in tissue quality of the feet. We find that ninety percent of these problems in the lower extremities make significant gains including regaining the ability to walk long distances comfortably, freedom from claudication and evidence of improved distal circulation. Those whose cerebral vascular system is severely damaged by arteriosclerosis and/or micro-circulation thrombosis, suffering from amnesia, confusion, aphasias, and motor coordination have improved [sic]. There has been a notable improvement in coronary circulation in all cases of angina, characterized by the patient having no need for vaso-dilators after about the fifth infusion. An interesting, but not predictable dividend in some cases consist [sic] of improved renal functions, [sic] reduction of prostatic obstruction by calculi, decrease in the degree of insulin required by the diabetic, almost normal breathing in emphysematous patients, great improvement in arthritic patients and even in Parkinson's Disease sufferers.[256]

Evers offered no other evidence for his claims of clinical benefits. Thus Dr. Lamas' implication, that Evers' 3000 cases constitute a significant portion of those that challenge the negative findings of controlled trials, must depend not on rigorous data, because Dr. Evers did not supply any. Rather, it must depend on Dr. Evers' credibility, which we now examine.

In this case series Dr. Evers denied complications more serious than "simple drug sensitivity," and in a late 1970s brochure for patients he stated: "We have never had a case yet where there were any adverse reactions and there have been no deaths from it.[257]" However, in a 1976 trial in federal court, expert witnesses testified that at least 13 patients, and possibly as many as 21, had died as a direct result of Na2EDTA administered by Evers at his own hospital (Meadowbrook Hospital in Belle Chasse, Louisiana) over a 2-year period from 1973 to 1975.[18] He had neglected to seek consultations for patients with renal failure and other life-threatening complications, treating them instead, for example, with "Myoflex": a tiny electric current applied to the skin, which, according to Evers, "increases blood flow to the kidneys.[65]"

One of the government's expert witnesses was Dr. J. David Spence, now Professor of Neurology and Clinical Pharmacology at The University of Western Ontario, London, Ontario, Canada, and Director of the Stroke Prevention and Atherosclerosis Research Centre at the Robarts Research Institute. He concluded his report to the Bureau of Health Insurance of the Department of Health, Education, and Welfare with these words:

The total picture is so appalling that I strongly urge the Bureau of Health Insurance not only to withhold certification for Medicare, but to take all steps possible within the law to expedite the closing of this hospital and to protect future patients from being subjected to these dangerous practices.[65]

In 1975 Evers had been found "medically incompetent" by the Louisiana State Board of Medical Examiners.[258] In 1986 the Alabama State Board of Medical Examiners revoked his license after he had

...represented to...a patient under his care that a proposed treatment for malignant carcinoma of the breast using Ointment A and Ointment Z would cause the malignant tumors to come out through the skin when in fact the Respondent knew or reasonably should have known that such representation was false, untrue, or deceptive.[259,260]

The Alabama Board also found that Evers had

...caused to be performed or ordered numerous x-rays, laboratory and diagnostic tests [and] medical services including hyperbaric oxygen treatments, hydrotherapy, coffee enemas, chelation, accupath 1000 & iridology, electrical muscle stimulants, massages, mineral sit baths, ozone therapy, and spinal checks which the Respondent knew or reasonably should have known were not medically necessary in the diagnosis and treatment of the medical condition from which the patient suffered.[259,260]

The Alabama Board also asserted that Dr. Evers had "wrongfully conducted medical experiments upon his patient...without adequately informing the patient of the risk of harm to her health by the use of untested and unproven drugs, medicines and therapies.[259]" Dr. Evers' "Agreement Between Physician and Patient" included language such as this:

For, and in consideration of, medical care rendered to me by Dr. Ray Evers (herein referred to as Physician)...

Should the government's insurance carriers or intermediaries or third parties find reason to deny me insurance benefits because of supposed "lack of medical necessity," I shall not hold my Physician responsible for any of my expenses for clinic treatment...

I specifically request that you, as my attending Physician, prescribe for my diagnosis and treatment based upon your individualized judgment and discretion, and not upon numerical averages in rulebooks or so-called "norms", "criteria", or "standards", or by whatever name, as established by any individual, group, agency, insurance company, or "service corporation", government agency, program, union, fraternal body, or any other organization or person...

I understand that the type of therapy given at RA-MAR CLINIC may not be in perfect agreement with the so called orthodox methods of treatment as approved by the AMA, FDA, or HEW. I understand that the type or therapy given here is the type that the Physician and I both agree is the correct future of medicine. (By the use of nutrition, enzymes, physical therapy, magnetic medicine, use of pyramids, etc. or any other modalities that may be used to benefit mankind)....[261]

Dr. Spence, after interviewing several of Evers' patients at Meadowbrook Hospital, wrote: "These interviews indicate clearly that the patients, in signing the so-called informed consent form, are not giving informed consent, and that they are not properly informed of the adverse effects of chelation therapy.[65]"

Evers was not pleased with those who disagreed with him or who inconvenienced him. He brought $1 million libel suits against Dr. Spence and at least one other expert witness in the 1976 case (Spence JD. Personal communication. 2006).[56] In the same year he was convicted of "intimidating and impeding two officers of the Internal Revenue Service" (IRS) after he had threatened to "get my gun and blow your brains out," had driven his car straight at one of the officers, and had called that officer a "yellow son-of-a-bitch" and the other officer a "black bastard.[62]"

The TACT protocols mention none of these facts with regard to Evers. Otherwise, it would have been evident to naive readers, such as members of the NIH Special Emphasis Panel[39] and human studies review boards, that Evers' "3000 cases" provide evidence only of the dangers of chelation and chelationists.

Casdorph (1981)[255]: This article, from the Journal of Holistic Medicine, reported an increase in ejection fractions (EFs) in 17 of 18 subjects who underwent chelation. The study was uncontrolled. There were no rigorous inclusion criteria; the author asserted, without documentation other than a history of coronary artery bypass graft in 4 subjects, that all subjects "had documented evidence of arteriosclerotic heart disease." Resting EFs were measured by radionuclide ventriculography before and after 20 Na2EDTA infusions. The average increase was 5.8%, called "highly significant" by Dr. Lamas.[5]

It is not clear what this series has to do with evidence for chelation as a treatment for CAD, or even as a treatment for left ventricular dysfunction. (Congestive heart failure is an exclusion criterion for the TACT.[5]) Casdorph did not specify the timing of the post-treatment ventriculograms; if they occurred immediately after the final chelation infusion, increases in EFs likely resulted from increases in preload due to the sodium content of the chelation treatment (600 mg in 3 g of Na2EDTA[65]), combined with afterload reductions that accompanied transient hypocalcemia. Thirteen of the 18 subjects had normal EFs to begin with, and only 1 subject had an EF below 46%; thus, the Frank-Starling relation predicts that an increase in preload alone would have enhanced systolic function. The mean increase of 5.8%, although perhaps statistically significant, can hardly be considered clinically significant in such a sample. That increase has also been reported in 20% to 30% of subjects in congestive heart failure who received placebo.[262,263]

Other aspects of the article cast doubt on the objectivity of the author. He cited the early Clarke studies as authoritative evidence for the ability of Na2EDTA to remove "metastatic calcium deposits" and to provide "remarkable benefits and improvement in patients suffering from angina pectoris.[255]" He neglected to cite the Kitchell reappraisal or other references to misinterpretations of uncontrolled trials of treatments for angina pectoris. He begged the chelation question: "...it is not too surprising that the administration of agents which remove calcium from the arterial wall do bring about dramatic relief of signs of arterial insufficiency and angina pectoris.[255]" He blurred the distinction between that assertion and the action of calcium channel blockers: "This presumably takes place because of removal of metastatic calcification in the arterial wall as well as interference with slow calcium currents which were recently described.[255]" His final sentence reveals a grandiose opinion of his study and an ignorance of cardiac physiology: "Thus EDTA must be added to the other new drugs popularly referred to as calcium blocking agents which have been shown to be potent coronary vasodilators.[255]"

Dr. Casdorph has since changed his emphasis about chelation, from one stressing its effects on calcium and atherosclerosis to one claiming that it is the cure for "toxic metal syndrome," the major cause of dementia: "...no other persons with our combination of medical knowledge and writing skill are privy to this proposed book's information about the dramatically successful therapy to antidote Alzheimer's disease and allied forms of dementia [sic].[264]"

Casdorph, who was the ACAM's president from 1987 to 1989, has also authored a book proffering "indisputable evidence" for faith healing.[265] He was listed as a TACT co-investigator by ClinicalTrials.gov as recently as March 2007,[38] but was not so listed by the NCCAM as of July 26, 2007.[117]

Olszewer (1988)[90,91]: This is a retrospective report of 2870 subjects treated with Na2EDTA between 1983 and 1985 and followed for 28 months in a clinic in Sao Paulo, Brazil. They were diagnosed mainly with CAD (844 subjects), PVD (1130 subjects), or cerebrovascular disease (504 subjects). Subjects with CAD were diagnosed by a combination of history, ECG, and the "stress test." Outcome criteria seem to have been assigned only retrospectively. A subject with "marked improvement" was defined as "a patient whose stress test was previously positive and became negative, and who was previously symptomatic, and became asymptomatic, while off all drugs, after a course of 20-30 treatments." A subject with "good improvement" was defined as "a patient who had an improved stress test, with a normal ST segment, and who became asymptomatic, off all drugs, but who still had evidence of arrhythmia or hypertension."

Among CAD subjects, after treatment 649 (77%) were judged to have made a "marked improvement" and 140 (17%) were judged to have made a "good improvement." Only 1 subject was judged worse. In the PVD group, the results were even more dramatic: 91% of subjects were judged to have made a "complete recovery."

The authors reported that 120 subjects were "lost to follow-up and...not included in this report.[90,91]"

Like the early series, the report lacks rigorous inclusion, diagnostic, or outcome criteria, but there are more serious problems. The number of reported deaths over a period of 28 months was 7; 2 were in the CAD group. This is not credible for 2870 subjects with life-threatening cardiovascular diseases, followed for 28 months in the early 1980s, most of whom were eventually "off all drugs." It remains not credible even if all 120 subjects "lost to follow-up" are assumed to have died.[266] It suggests a departure from the rigor that is required of any study, but especially a human trial of a purportedly lifesaving treatment. For example, many subjects may have been misdiagnosed, or the reporting of outcomes may have been biased in the extreme. The authors may argue that the results demonstrated that chelation is a miracle cure for atherosclerosis, but if so this should have been evident in the more recent RCTs.

Bias or naiveté is evident in other parts of the report. The authors cited the early Clarke studies with approval, but ignored the Kitchell reappraisal and Soffer's negative report. They argued that their own "clinical response rates...are much too high...to be attributed to placebo-effect alone...." They added that their results "are remarkably similar to those obtained by Clarke...," implying that this enhanced the validity of that assertion. They were apparently unaware that Clarke's and their own "clinical response rates" were remarkably similar to the "constant and predictable placebo effects" found in other reports by "enthusiasts.[251,252]" The glaring difference in death rates between their own study and Clarke's, however, renders the Olszewer study invalid unless it can be satisfactorily explained.

In this article, Olszewer and Carter also reported the "preliminary results" of "a randomized, double-blind, cross-over study...on a small group of 10 patients with peripheral vascular disease." The subsequent report of that study, published in 1990,[92] offered more specific inclusion and outcomes criteria than the authors' previous case series, including objective measures of walking distance (whether "pain-free" or maximal is not clear); the Master's 2-step exercise test; and a bicycle stress test, although it seems unlikely that "the Bicycle Stress Test included patients with claudication before 3 minutes at 50 kilometers per hour.[92]" The methods of randomization, allocation concealment (if any), and blinding were inadequately described, as was a comparison of baseline characteristics of the treatment and control groups. The dose of EDTA was 1.5 g per infusion; the frequency was not reported. In the authors' retrospective series, the dose had been 3.0 g per infusion, given 2-3 times per week.

The reported results were striking. After "active" and placebo groups of 5 subjects each were given the first 10 of 20 planned treatments, 5 subjects had improved so much -- had, on average, doubled their walking distances -- that the authors "therefore decided to break the code." All of those who had improved had received EDTA; those who had received placebo had not improved at all. The authors proceeded to give all subjects EDTA for the final 10 treatments, and by the end the original placebo group had experienced improvements comparable to those of the active group after its first 10 treatments. The original active group also continued to improve: The mean walking distance for those who received all 20 EDTA infusions increased from 160 to 602 m, or almost 400%. By comparison, in 2 subsequent controlled trials of chelation for PVD, involving a total of 181 subjects and using twice the EDTA dose, those in the placebo and active groups each increased pain-free and maximal walking distances by 20% to 50%.[21,25]

Because the sample size was so small, Olszewer, Sabbag and Carter's "double-blind" results have reasonably been dismissed as evidence for chelation being an effective treatment for cardiovascular disease. In the first TACT protocol, Dr. Lamas appears to do this without citing the article: "...two additional very small trials of 9 and 10 patients are not interpretable and not reviewed here.[4]" However, there is a reason for examining the trial despite its small sample size: to help determine whether Olszewer and Carter's previous case series, which reported the results of 844 subjects with CAD and 1130 with PVD, warrants inclusion among the references cited by Dr. Lamas as evidence for "thousands of patients with successful outcomes."

We have already argued that "Olszewer (1988)" ought not to be included, because of its inadequate methods, its bias, and its implausible death rate. Dr. Lamas did not acknowledge those problems in the text of page 7 of the 2003 TACT protocol, where he summarized the report.[5] The 1990 double-blind report gives more reasons to question any report by Olszewer and Carter. The small sample size notwithstanding, the results were highly improbable, as others have also concluded.[267] Bias continued to be evident in the authors citing the early Clarke articles but failing to cite the disconfirming reports of Kitchell and Soffer. Most telling, however, was their assertion that EDTA's

...safety, when administered according to the American College for Advancement in Medicine protocol, is no longer a major concern, in view of the fact that more than 500,000 patients have been treated in the US alone, without a single reported incident of renal failure or death since 1960.[92]

That statement was false. Co-author Dr. James Carter, who lives in Louisiana, was fully aware of Dr. Evers' cases of renal failure and of his deaths. In 1992 Carter published a book alleging a conspiracy of the American Medical Association, the FDA, the US Postal Service, the FTC, the IRS, the NIH, Medicare, Blue Cross Blue Shield, other private health insurers, drug companies, the American Hospital Association, the American Cancer Society, the AHA, medical journals, medical schools, state licensing boards, and other entities.[94] These institutions, according to Carter, were in collusion to "suppress" Laetrile, krebiozen, chelation, chiropractic, homeopathy, antineoplastons, "body detoxification," "immune system stimulation," "herbal medicine," coffee enemas, numerous other "alternatives," and even the Church of Scientology. In his book, Carter argued that there is no question that "chelation works," citing "published medical evidence" (eg, the articles reviewed here) and using phrases, such as "beyond a shadow of a doubt" and "the case against EDTA chelation therapy, on the basis of lack of efficacy and scientific medicine, is closed!"[94]

Carter discussed "pioneering chelation doctor Ray Evers" at some length, arguing that the deaths reported in the 1976 federal case had been due not to EDTA, but to "underlying progressive illness.[94]" He also suggested that Dr. Evers may

not have been strictly following the [ACAM] chelation protocol developed in the 1970s....Dr. Evers was aware, however, that if there was any suggestion of impaired or reduced kidney function, the patient was not a good candidate for chelation therapy; if it were administered under those circumstances, the patient should be monitored closely.[94]

Carter attributed Evers' legal troubles, in part, to the chairman of the Louisiana Medical Board who, according to Carter, may have held a grudge against Evers because the two had interned together and had, at the time, "dated a mutual woman friend." Finally, "Dr. Evers, a devout Christian, never turned away any patients who came to him for help."

Dr. Evers, however, was a member of ACAM, then called the AAMP. His "3000 cases" article had described Na2EDTA dosing essentially similar to that of the current ACAM "protocol.[33,256]" Dr. Spence, the expert witness, reported: "These patients were receiving chelation therapy in the manner advocated by the AAMP, under the direction of a member of the AAMP, who had on his office wall a certificate of merit from the AAMP.[66,268]"

In evaluating Evers' deaths, Dr. Spence was careful to distinguish the effects of EDTA treatments from "underlying progressive illness." Commenting on Dr. Evers' having neglected to seek expert consultations for complications of EDTA, Dr. Spence wrote:

If the patients were all very advanced in age, and had died of hopeless diseases for which they were receiving no therapy, a consultation rate of one in twenty-one deaths might be considered within the realm of acceptable medical practice...however, patients as young as 47 years of age are permitted to die in congestive heart failure, and patients as young as 24 years of age are allowed to die of renal failure... [emphasis in the original].[65]

In a letter to the California State Department of Health, Dr. Spence wrote: "The causes of death included...renal failure, which, in a number of cases, was not present before administration of EDTA.[66]"

Elsewhere in his book, Carter called Evers "the eminent Dr. Ray Evers[94]" and revealed that he, Carter, had his own grudge against the Louisiana Medical Board, which he referred to as a "kangaroo court.[94]" The Board, curiously, dropped an apparently strong case against Carter after it had discussed the matter with the Tulane University School of Public Health and Tropical Medicine, where Carter was Chairman and Professor of the Department of Nutrition.[258] In regard to Dr. Evers being "a devout Christian" who "never turned away any patients," one need only review his threats to the 2 IRS officers and his "Agreement Between Physician and Patient" to wonder about the integrity of such assertions.[62,261]

It is reasonable to conclude that the cases reported by Olszewer and Carter, like those of Evers, have no place in a body of evidence purporting to justify an NIH-sponsored, phase 3 trial exposing 2000 subjects to a dangerous drug that certainly doesn't work as claimed by its advocates. Carter, who was the Chairman of the GLACM IRB that approved Henry Heimlich's "induced malaria therapy for HIV" trial,[133,134] is now a TACT co-investigator.

Hancke (1993)[269]: This is a report from Denmark of a retrospective series of 470 subjects with "claudication and/or angina pectoris" who "usually required 30 [chelation] treatments over a period of 3-4 months." They were assessed "on completion of treatment and again 2 months later." The authors might have followed the subjects for a longer period, because they also reported that "our experience covers a period of 6 years and we saw no severe side-effects or casualties arising from the treatment."

Inclusion criteria are absent. Diagnostic criteria are more specific than in most reports, but the reader can't tell whether they were assigned prospectively or retrospectively:

Patients with claudication had their ankle-arm index, walking distance, foot temperature, pain at rest, skin color of feet and healing of wounds assessed and registered. Subjective judgement [sic] of resting pain was rated on a scale of 1-3. Patients with angina pectoris had their working capacity measured on a treadmill, and ST depression by electrocardiogram....The subjective judgement [sic] of results was rated on a scale from 1 to 3 with regard to the number of attacks of angina pectoris and consumption of nitroglycerin (1:worse, 2: unchanged ± 10%, and 3: improved).[269]

The results were similar to those of other reports by enthusiasts: Of 265 subjects with "myocardial ischemia," more than 90% "improved" and only 2 worsened. Eighty-five percent had improved "working capacity"; more than 90% "reduced their consumption" of nitroglycerin; 91% with angina pectoris "improved," as did 69% with "ST-depression" and 62% with "arrhythmia." Further, "of 65 patients referred for bypass surgery, 58 did not require it after their course of chelation.[269]"

In 262 subjects with claudication (some subjects were in both groups), the authors reported "improvement in 82%," including 31 of 44 who had "problems with wound healing" and 110 of 137 "who complained of cold feet." The ankle/arm blood pressure ratios improved in 82% and walking distance improved in 87%. Unfortunately, the report lacks raw data to compare these results with those of other, plausible trials. One statement, however, described individual results more impressive than even those reported by Olszewer, Sabbag, and Carter[92]: "Some patients with claudication who were unable to walk more than 100 feet could walk painlessly for 2 miles or ride several miles on a bicycle after their treatment.[269]" "Of the patients referred for amputation, 24 of 27 legs were spared following their course of chelation.[269]"

The authors reported no deaths. They concluded that:

...the beneficial results that we observed were far in excess of the 10-15% improvement that is usually seen in the placebo group of a controlled study....Our results are identical with those of other similar studies....On the basis of such well published data, it seems to us that it is unethical to wait for a randomized, double-blind, cross-over study to approve chelation for the treatment of arteriosclerosis....[269]

Ironically, and in rebuttal to Dr. Lamas' assertion that a properly controlled trial "leads to changes in clinical practice,[29]" such a randomized study had recently been done, right in Denmark.[21] It had failed to find evidence of a treatment effect for chelation. Hancke and Flytlie cited it only to dismiss it as "seriously defective" and as a manifestation of "the harsh attitude of the Danish vascular surgeons." Instead of changing their own clinical practices, Hancke and Flytlie filed a complaint against the trial authors to the Danish Committees on Scientific Dishonesty (DCSD).[153] The DCSD subsequently found "no evidence of scientific dishonesty on the part of the accused.[270]"

Other statements in the article by Hancke and Flytlie cast doubt upon their own integrity as investigators. In asserting that chelation "has been proved effective in a number of clinical trials," they cited 6 references. Two are the nearly identical reports of the same retrospective study of Olszewer and Carter[90,91]; one is the "double-blind" study of Olszewer, Sabbag, and Carter[92]; two are by Edward McDonagh, the chelationist who later admitted under oath that he had falsified his data.[44,89]

Hancke and Flytlie also wrote that "evidence for effectiveness of EDTA chelation therapy is cumulative over many years," citing 6 more references. One of these (and probably two others) was concerned solely with the common use of EDTA as an antioxidant for in vitro biochemical preparations[271]; it is barely relevant to the proposition of chelation therapy, and not at all relevant to the practice. Another reference was to an editorial in which Alfred Soffer unequivocally denied the existence of evidence for effectiveness of Na2EDTA in treating cardiovascular disease.[9] It seems doubtful that Hancke and Flytlie could have read that article, which stated, in part: "There are several sites...where this therapeutic fad currently is in vogue and where the zealot peddles these wares to the nave afflicted. Symposia and miniconventions have been organized to extol its virtues....[9]"

Among those "zealots," of course, were Hancke and Flytlie.

Jorgensen, Guldager, and Jelnes (1992)[24]: This is another report notable for not being cited by the TACT protocols. It is by the authors of the Danish RCT cited above, working at the Hillerod Central Hospital, Hillerod, Denmark. It is a case report of a single subject, but provides at least as much useful information as the other reports in the later series. The full report is in Danish, but the abstract suffices to make the point:

A 62-year-old man suffering intermittent claudication had 20 infusions of EDTA in a double-blind, placebo-controlled trial. No effect on symptoms or systolic ankle/brachial blood pressure index was found. Following the trial, he received 30 further infusions of EDTA in a private clinic. The systolic ankle/brachial index was unchanged throughout the total period as measured in Hillerod Central Hospital. However, the private clinic found a significant increase in the index following EDTA treatment. The reason for this discrepancy could be poor technique in the clinic or it could be due to bias or manipulation. The discrepancy explains the difference between the positive results claimed by the private EDTA clinics and the results of the double-blind, placebo-controlled Danish trial.[24]

Chappell (1993)[95]: This "meta-analysis" is not cited by the TACT protocols, but is pertinent because it is the article most frequently referenced by an article that is so cited: "Dayton 1995,[57]" which is summarized next. It is also pertinent because the first author, L. Terry Chappell, led the ACAM's mission at the hearing of Rep. Dan Burton's House Committee on Government Reform, discussed previously.[41] Chappell also sat on the NIH Special Emphasis Panel that passed judgment on Dr. Lamas' application for the TACT award[39] and was named, in that application, as one of the "prominent experts in chelation therapy" assigned to the "TACT Liaison Committee to the American College for Advancement in Medicine.[4]" Chappell is a TACT co-investigator.[7]

The article purported to demonstrate that in 19 studies, 87% of 22,765 subjects treated with EDTA "improved," with "a correlation coefficient of 0.88, which indicates a high positive correlation between EDTA therapy and improved cardiovascular function." There are numerous objections to this article, beginning with its spurious claim to being a meta-analysis and including its reliance on a single unpublished "study" for the vast majority of its subjects (19,147). That study appears to have used only one, atypical outcome measure -- thermography -- and was apparently funded by the company that made the thermography device.[272]

Most troubling is how Chappell and his co-author dealt with the lack of control groups in the reports they reviewed:

Even though placebo control has, for the most part, been absent in existing EDTA studies, this does not mean that the data collected are not valid or useful...

For the purpose of this meta-analysis...simply consider the existing study data to be the data for the treatment group and compare the improvement in cardiovascular function of the treatment group to a control group defined to have no improvement in cardiovascular capability.[95]

As unlikely as it may seem, the authors simply added imaginary controls and assumed that had they existed they would have experienced "no improvement." One might imagine that this betrayed an ignorance of the natural history of disease and the design and interpretation of clinical trials. Not quite: "It needs to be shown that the assumption of a no-treatment group with no improvement is reasonable. This meta-analysis will use the blinded study by Olszewer, Sabbag and Carter to show this.[95]"

We agree with other reviewers that "the invention of a null-effect group in this way is clearly invalid and renders the meta-analysis meaningless.[267]"

Dayton (1995; revised in 1999 and 2006)[57]: On page 261 of the 2001 TACT protocol, Dr. Lamas cites this short, online book as support for "published case series suggesting clinical benefits.[4]" It is not a case series, but a sales pitch to potential consumers. It lists some 70 "conditions or combination of conditions which have been reported to improve following intravenous chelation therapy." It dismisses "criticisms of chelation therapy by doctors" with pithy, facile rebuttals, eg, "does a person need to have the education of a sky rocket scientist to admit rockets work?" It has a "frequently asked questions" section that ensures readers that chelation is "safer than aspirin," and "unlike surgical approaches, no strokes, deaths nor heart attacks have been reported to be due to intravenous chelation therapy, and fewer side effects are reported than with many pharmaceutical medical treatments.[57]"

Elsewhere it asserts:

Although published scientific studies have proven chelation to be effective in regard to hardening of the arteries, the existence of such studies is still unknown to many physicians.

In addition to being known as a treatment for metal toxicity, such as found in lead poisoning, chelation has been successfully used to overcome various conditions associated with aging. Impaired circulation due to hardening of the arteries and discomfort due to arthritis are among the most notable. Numerous scientific articles reflecting effectiveness and safety have been published. In one study evaluating over 22,000 patients, 87% demonstrated objective improvement. Millions of chelation administrations have been performed over 40 years, world wide. According to the American College for Advancement in Medicine not one fatality proven to be caused by chelation therapy has been reported when appropriate protocol is followed.[57]

Chelation is often used as a safer method to replace much costlier conventional surgical and related medical procedures. Hardening of the arteries is a lucrative, multi- billion dollar industry. Generally those who profit from an industry, whether it be a surgical/medical industry or any other industry, tend to be antagonistic to changes that reduce profits.

Chelation therapy is used in conjunction with programs, exercise, and stress reduction. It also complements standard medical and surgical therapy. With the improvement of health, the medical necessity for harmful and potentially harmful pharmaceutical drugs may be reduced. Although better known as a substitute for cardiovascular surgery, chelation also has been used as a strategy to prevent reclogging of arteries after surgical bypass and balloon angioplasty.[57]

The reader will by now recognize the usual gambits common to treatises by chelation advocates: "metal toxicity"; the conflation of Na2EDTA and the safer CaEDTA (only the latter is approved for the treatment of lead poisoning, but it is overwhelmingly the former that is used for nonstandard treatments, including the TACT); invoking "published scientific studies" that are more accurately characterized as pseudoscientific studies (especially the Chappell "meta-analysis," presumably chosen because "22,000" is impressive to the scientifically naive); the false assertion of "not one fatality proven to be caused by chelation therapy"; the suggestion that chelation's lack of acceptance is due not to its lack of scientific support, but to the conflicting financial interests of cardiologists, surgeons, and drug companies; and finally, that chelation is a safer and less expensive alternative to cardiovascular procedures, surgery, and "potentially harmful pharmaceutical drugs."

In addition to the Chappell "meta-analysis," to which Dayton alludes several times, his book cites several of the other studies reviewed here and more than 10 articles by Edward McDonagh, the chelationist who admitted in a court of law that he had falsified data.[44] Dayton, like other advocates,[152,153,154,155,156,157,158,159,160,161,162] also claims that all negative studies of chelation, including all controlled trials, were actually positive.

The 2001 TACT protocol named Dayton the "Trial Chelation Consultant" and 1 of 9 "prominent experts in chelation therapy" who were to be members of the TACT Liaison Committee to the ACAM (Chappell is another).[4] According to the 2001 TACT protocol, Dr. Dayton is "the former Director of the Scientific Research Committee of the ACAM [and] has clinical experience with over 75,000 chelation infusions.[4]" At $100 per infusion,[58] "nutritional supplements in the range of $20 to $200 per month [and] diagnostic study costs and professional fees...ranging from a few hundred to several thousand dollars,[57]" such experience may have made Dr. Dayton rich.

But has it made him an expert? In a 2000 article extolling "detoxification," Dr. Dayton was quoted as saying that "dramatic increases in life span are found with chelation" and "chelation favorably impacts all four major causes of death in the United States (heart disease, cancer, cerebrovascular disease, and lung disease)." He recommended EDTA chelation for anyone over the age of 30 because "modern people are overwhelmed by pollutants." He claimed that the American College of Physicians "recommend[s] chelation therapy as a preferred treatment" for "unclogging carotid blockage.[234]"

Dayton's practitioner profile for the Florida Department of Health lists 2 federal criminal convictions: 18USC 371, "Conspiracy to commit offense or to defraud United States," and 18USC 1341, "Mail Fraud: Frauds and Swindles.[61]" Dayton is a TACT co-investigator.[36]

The later case series, which provide the vast majority of Dr. Lamas' "thousands of patients with successful outcomes," contribute nothing -- not even the "suggestive" evidence that might be granted to legitimate case reports -- to the assertion that chelation for CAD may have "a small or moderate benefit.[29]" The TACT protocols fail to acknowledge the complications and deaths of Evers[18] and fail to mention abundant, additional evidence of incompetence and/or bias, eg, the miniscule death rate reported in "Olszewer 1988[90,91]" or the inaccurate ankle/brachial index measurements performed by chelationists in Denmark, as reported by Jorgensen and colleagues.[24] The 2003 TACT protocol's characterization of such case series as "observational epidemiologic studies[5]" is unworthy of a document approved by the NIH.

The errors and misrepresentations in the TACT protocols are a serious problem. They are the basis for Dr. Lamas' argument that chelation for CAD is worthy of study in a large, human trial. Human studies review boards must have accepted that argument in order to approve a phase 3 trial of a drug that had failed in several phase 2-like trials, contrary to the usual policies of the NIH and to the US Code of Federal Regulations.[177,178]

Dr. Lamas does not appear to be the original author of Table 2 , even if his cause has benefited from its errors. Table 2 is almost an exact duplicate, albeit printed without attribution, of a table that appeared in the 2000 article by Edzard Ernst, in the American Heart Journal, that Dr. Lamas' editorial rebutted.[29,166] We reproduce that as Table 4 .

All of the erroneous entries in Dr. Lamas' table are also present in Dr. Ernst's. There are a few differences; one is worth mentioning, if only for its absurdity: "Clarke (1960)" in Ernst's table, in which 87% of 700 subjects are said to have improved, is not present in Dr. Lamas' table. However, that article is not the report of a study at all; it is an editorial that we have previously cited.[247] ("Clarke 1960[50]" in Lamas' table is designated "Clarke et al 1960" in Ernst's). Nowhere in Clarke's editorial does the number "700" appear. We can only imagine, referring to the sentence that we quoted from the article, that someone changed "several hundred" to "seven hundred." A potential source for such a change is an abbreviation in a table on page 143 of Chappell's 1993 "meta-analysis," in which the Clarke editorial is listed as having presented the results of "sev hundred" subjects.[95] Chappell's table also misspells "Winebaugh" as "Wirebaugh," as does Lamas' table but not Ernst's.

That suggests that Dr. Ernst was also not the sole author of the erroneous table. Ernst, moreover, used it to argue against the efficacy of Na2EDTA for CAD, so he would have had no reason to exaggerate positive data. In addition to the table in Chappell's article, another table may be pertinent. It is found in a 1993 article by Grier and Meyers in the Annals of Pharmacotherapy.[273] We reproduce it as Table 5 .

Several of the errors in Drs. Lamas' and Ernst's tables ( Table 2 and Table 4 , respectively) are present in the table of Grier and Meyers ( Table 5 ), even if the 3 tables are not identical. Examples of common errors are the "58 improved" entry for Clarke et al (1960),[50] redundant entries for the original Meltzer/Kitchell cohort of 10 subjects,[51,54,248,249] and failure to include those 10 subjects in the entry for Kitchell (1963).[54] Nonsensical numerical entries in Ernst's table (Meltzer 1960,[51] Kitchell 1963,[54] and Lamar 1966[53] in Table 4 ) and in Lamas' table (Kitchell 1963[54] and Lamar 1966[53] in Table 2 ) can be explained as incomplete copies of the corresponding entries in the table of Grier and Meyers ( Table 5 ). Drs. Lamas and Ernst each cited the article by Grier and Meyers, although neither cited it as a source for his own table. Ernst cited Chappell; Lamas did not.

By the time the 2003 TACT protocol was issued, a total of nearly 300 subjects had been studied in 4, adequately designed RCTs and several useful substudies of Na2EDTA for either CAD or PVD.[21,22,23,24,25,26,27,28] In no trial was chelation found to be superior to placebo.

Dr. Lamas reports having studied an additional 40 subjects, of whom 30 were to have been randomized to receive Na2EDTA, in a "pilot study," PACT. According to the 2003 TACT protocol, the PACT should long ago have been completed.[5] We have asked the investigators and the NIH for its results, even if only in qualitative form, but they have denied us most of this information. We are concerned that there is no published report of the PACT. As previously stated, we suspect that this means that it did not demonstrate a favorable treatment effect.

Are the RCTs to date, even excluding the PACT, adequately powered to detect or exclude a clinically important benefit of chelation for atherosclerosis, contrary to Dr. Lamas' opinion? We think so. The Canadian PATCH study alone was designed to

...provide 90% power to detect a 60-second difference in mean change in exercise time from baseline to the 27-week follow-up, assuming an SD of 80 seconds within each group. The 60-second difference was based on a minimally important difference determined by a consensus of Calgary cardiologists.[27]

Independent reviewers from the United Kingdom agreed that the PATCH study was so powered, and called it "well designed and conducted.[267]" Those reviewers, who did not consider the case series at all, observed that it

...would require a very large number of participants to test the hypothesis that chelation is more effective than placebo. A study with these numbers is unlikely to be undertaken because it would require considerable funding; it is questionable whether the small effect sizes that may exist have any clinical significance.[267]

We agree that the hypothesis that chelation is merely more effective than placebo is not the correct question to be asking. The correct questions are two, and they are closely related. First, is there support for the standard claims made by chelation advocates over the last 50 years? If not, is there support for a lesser effect that is likely to be useful enough, ie, through a favorable combination of safety and efficacy, to warrant "add[ing] EDTA chelation to the armamentarium for clinicians who treat coronary disease?[2]" In hypothesizing the second question, Dr. Lamas ignored its dependence on the first.

Recall that the standard claims of advocates are that chelation dramatically improves symptoms, objective outcomes, and life expectancy in 80% to 90% of subjects with CAD (and 70 other conditions).[57,152,234] Dr. Lamas has conceded that those claims, at least in regard to cardiovascular disease, have been disproved by the existing RCTs.[29] However, what about the possibility that some lesser but still useful effect is the explanation for chelationists' perceived outcomes? If so we would expect that, among the many series, at least some would have reported such an effect. However, there have been no studies whatsoever -- no case series and no RCTs -- reporting the "small or moderate benefit" proposed by Dr. Lamas.[29] All studies by advocates and "enthusiasts" have been strikingly, implausibly positive; all studies by sober academics, including all credible controlled trials, have been unequivocally negative.

The authors of the PATCH report cautioned that their results "can be applied to fit only a similar population to that studied," and that there "was no difference [between the active and placebo groups] in the number of clinical events, but our study was not powered to detect any such differences." The authors also stated, "Larger trials with a broader range of patients will be needed to assess the safety and impact of EDTA therapy on clinical event rates.[27]"

Such caveats accompany every honest report of a clinical trial. They do not mean, unless the greater scientific and clinical question warrants it and it can be ethically justified, that there must be numerous, additional RCTs to cover every contingency. The PATCH authors, moreover, were referring to those caveats when they wrote that "larger trials...will be needed." If they had considered the low prior probability that chelation works as claimed, the low likelihood that a large trial would yield a clinically significant, as distinguished from a statistically significant, result, and if they had considered the case series and external evidence for toxicity of Na2EDTA, they might have been more circumspect.

If, as the TACT investigators argue, the existing chelation data truly support pursuing a phase 3 human trial, what about numerous other health claims that have come and gone over the years? Certainly many of those have been disqualified by the same sort of "meager clinical trial data" that, according to Dr. Lamas, "Ernst and others over-interpreted...to come to [the] conclusion" that "chelation for CAD should be discarded.[29]" In fact, many have been so discarded, and rightly so. Two pertinent examples follow.

We have previously mentioned another failed treatment for CAD, bilateral internal mammary artery ligation, which enjoyed brief popularity in the United States at the same time as the early chelation series. It was a simple surgical procedure that could be done with small incisions and local anesthesia.[274] There was an anatomic rationale: Small vessels connect the internal mammary arteries to the coronary arteries. Flow from the former to the latter could, in theory, be enhanced by the increased pressure gradient resulting from downstream mammary artery ligation.[275] Uncontrolled case series involving hundreds of subjects with angina (thousands in the world literature) reported dramatic improvements.[251] Eventually, 3 controlled trials, totaling 47 subjects, reported no differences in outcomes between mammary artery ligation and sham operations, and the practice abruptly ceased.[251,274,276,277]

Thus the evidence against internal mammary artery ligation was similar in kind, but considerably less powered, than is the evidence against chelation. That is, responsible positive case series of mammary artery ligation involved thousands of subjects, whereas negative controlled trials involved only 47 subjects. In the case of chelation, responsible positive case series have involved a few tens of subjects at best. Negative RCTs of chelation have involved almost 300. Yet no competent physician or scientist would consider studying mammary artery ligation now, even though it is simpler, safer, and cheaper than coronary artery bypass grafting, and at least simpler and cheaper than coronary angioplasty -- characteristics frequently cited to justify chelation.[57,106]

Laetrile is the most notorious quack product of modern times.[70,71,72,73,74,75,76,78] It is ineffective and toxic. Its purveyors, many of whom were and are chelationists, prey on desperate people. In justifying the existence of the NCCAM shortly after his appointment as director, the late Stephen Straus cited Laetrile as an example of a treatment whose public demand had "abated markedly only after competent studies showed it to be ineffective.[278]" By "competent studies," Straus was referring almost exclusively to a single, uncontrolled case series involving 178 subjects.[279] Laetrile apologists criticized the study for being uncontrolled,[280] but its design was similar to contemporary, preliminary studies of numerous chemotherapeutic agents that were consequently discarded because of no clear evidence of effect.[281]

Thus the trial evidence against Laetrile was considerably less rigorous than that against chelation. However, no competent physician or scientist would now call for more trials of Laetrile, although a Cochrane Review has done exactly that.[282] That is because its advocates made sensational claims, but as is the case for chelation, Laetrile lacked a scientific basis and the first credible case series showed no favorable effects -- only toxicity. It was obvious to ethical physicians at the time that Laetrile was fraudulent.[283,284,285]

Dr. Straus, moreover, was wrong about the demise of Laetrile. Its sales had abated markedly 2 years prior to the publication of the case series cited above, for the simple reason that it had become illegal -- against the formal objection of the ACAM.[71,73,76,77,286] Demand, however, persisted, as Straus ought to have known.[287,288] At least 2 TACT co-investigators, Michael Schachter and Jack Slinguff, have prescribed Laetrile within the past few years, and another, Stuart Freedenfeld, commends it on his Web site.[191,289,290] Slinguff's Laetrile business led to the suspension of his Ohio osteopathic license in 2004, but ClinicalTrials.gov still listed his Lake Cable Medical Center as a chelation study site as recently as March 2007.[38]


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.