But "sweeping" calcium away is a double-edged sword. While it might help unclog coronary arteries, it might also lead to deadly low calcium levels in the blood, as happened to the three people whose deaths were described in the CDC report.
That's one reason why the National Institutes of Health required that the infusions be done at a slow rate -- over a period of three hours or more -- so problems like low calcium levels could be caught easily.
In a series of letters to the study's authors, the Department of Health and Human Service's Office for Human Research Protection cited several concerns about the study, including that the researchers hadn't followed the rules about doing the infusions at a slow rate.
In one letter, the government overseers expressed concern that infusions were performed in "shorter than recommended" times in 440 instances, involving 251 subjects.
The study was stopped from September 2008 until June 2009 to respond to the government's concerns.
A question of significance
In addition to possible safety problems, there are concerns that the study was poorly done and doesn't actually show that chelation works.
In the study, researchers divided the patients into two groups. One group of 839 patients received 40 infusions of a chemical drug solution called disodium EDTA.
Another group of 869 patients received infusions with a placebo solution of salt and sugar water. Neither the patients nor the doctors knew who was receiving which treatment.
Among the patients receiving chelation, 26% went on to have a cardiovascular event, such as death, heart attack or stroke. Among the placebo group, 30% went on to have a cardiovascular event.
According to the authors' analysis, this four-percentage-point difference was barely statistically significant, leading Lamas to note that the difference might have been by chance, not because chelation actually worked.
Further watering down the report is the fact that a relatively large number of patients -- 17% -- dropped out of the study.
Experts say patients usually drop out of studies when things aren't going well -- when they're having bad side effects, for example, or when the therapy isn't working. But since they dropped out, their results can't be included in the study.
"This study was so badly done, and the results are so marginal from a statistical perspective that this therapy can't be recommended," said Nissen, the Cleveland Clinic cardiologist.
"Making a treatment decision about a life-threatening disease based on poor-quality scientific data can be dangerous," he added. "History has taught us that when this happens, considerable harm can come to patients."
The future of chelation
Kalidas, the chelation practitioner in Florida, said he hopes these new study results will help persuade insurance companies to start paying for chelation.
But insurance companies usually only pay for treatments that are FDA approved -- and there's little chance the FDA would approve chelation for heart disease based on this study, experts say.
"This would never pass muster at the FDA," said Nissen, the former chairman of the FDA advisory panel that examines applications to put new heart drugs on the market.