Chelation Therapy

Elmer M. Cranton, M.D.


ABSTRACT: Trace and toxic element testing is important in the practice of chelation therapy. No single clinical test for adequacy or toxicity of metallic elements exists that does not also have serious weaknesses. Each element has its own unique distribution in body fluids, tissues and organs. Elemental blood concentrations fluctuate widely throughout the day while levels in solid tissues vary more slowly. Factors that alter elemental concentrations in body fluids and tissues include diet, nutritional supplements, atmospheric and environmental contamination, industrial exposure, and physical exertion. Concentrations in fluids and tissues in different parts of the body accumulate and fluctuate in unique and different ways for each element. Acute exposures have different effects than gradual accumulation over longer periods of time. Extremely low levels of trace elements in biological specimens can make contamination of test specimens a serious source of error—beginning with initial collection, during transport and storage, and during laboratory handling and measurement. Extremely low concentrations can make accurate laboratory testing difficult and expensive. Provocative urine testing and chelation of mercury is covered in detail because of a recent upsurge in interest for that particular toxic element.


Each type of testing for adequacy or toxicity of metallic elements has its own unique strengths, weaknesses, and potential sources of error, which also differ for each element tested. Medical, dietary, and occupational histories (combined with careful physical examination) are helpful, but not always adequate to make a reliable diagnosis.

The most accurate assessment for trace and toxic elements would require multiple measurements of a variety of specimens--including whole blood, erythrocytes, leukocytes, serum, plasma, urine, hair, nails, muscle, liver, and other tissues. High cost and impracticalities of specimen collection place limits on what can be done in a clinical setting.  Provocative urine testing following a chelator seems a good compromise.

When testing urine for metallic elements, it has been customary to compensate for intraday dilutional fluctuations by collecting 24-hour, timed specimens.  A serious limitation of 24-hour urine collection is lack of patient compliance. Accurate collection of 24-hour specimens has proven unreliable in an ambulatory setting.(1) Specimens become contaminated by skin, feces, vaginal secretions, menstrual blood, perspiration, and fingers in transfer containers. Some urine may be lost during defecation, and voidings are missed entirely. Errors from contamination are compounded by collection of multiple voidings over a 24-hour period. Female patients have difficulty voiding into narrow-neck containers, and transfer from wide-mouth cups adds to the risk of contamination. Even with careful and detailed instructions prior to collection, followed by questioning of patients when specimens are returned, it is not possible to consistently obtain accurate 24-hour collections.(1)  A six-hour urine collection obtained in the doctor’s office under supervision is much more likely to be accurate.

By measuring elements in micrograms or nanograms per gram of urinary creatinine, clinically meaningful data can be obtained without the need to collect 24-hour urine specimens.(2) By collecting a six-hour timed urine specimen under supervision, a provocative chelation test can thus be accurately performed.  When testing urine for nutritional trace elements and minerals, it is necessary to avoid intake of nutritional supplements containing those elements for at least 36 hours prior to urine collection. Otherwise test results will reflect recent intake and not tissue levels.  

Testing urine is less expensive and more convenient than testing blood. Whole blood, erythrocytes, plasma, and serum are all too viscous to nebulize in the test instrument without first digesting with acid, which dilutes those specimens before measurement, reducing sensitivity and accuracy at the very low levels of some trace and toxic elements. Like blood and serum, hair and other biopsy specimens must also be digested in acid prior to testing. Because addition of acid dilutes the elements being measured, sensitivity of hair testing may be less than that for urine.  Urine is not viscous and can be tested directly without digestion or dilution. An urine concentrations can be greatly increased by provocative chelation.

Using multi-channel test instruments, accurate measurements can be made of a wide spectrum of low-level elements in urine with a single pass through the instrument.  This multi-element procedure results in a relatively low cost per element. Because of the need for digestion and dilution, measurements of equal sensitivity for the very lowest-level elements in blood may require individual tests for each element using more sophisticated and more expensive instruments. The cost for each element might then equal or exceed the cost to measure a spectrum of 20 or more elements in urine.

Minerals and trace elements in systemic circulation are continuously filtered through renal glomeruli into the urine. Absorption and excretion by renal tubules maintains equilibrium between blood and urine. Metallic elements in blood maintain equilibrium with body cells and tissues. Trace elements and minerals in urine thus reflect concentrations of those elements throughout the body.

Regardless of whether blood, urine, hair, saliva or another type of tissue biopsy is used, certain principles must be considered relative to the test protocol and interpretation.  Mercury testing is described in detail below as just one example of the principles, pitfalls and precautions of trace element testing.


People frequently contact me with questions about mercury test results performed elsewhere after being told that their mercury level is high.  Often they have been told that they require a course of mercury chelation therapy, based on those test results.  Upon reviewing the laboratory reports, I frequently find that urine was tested immediately after administering DMPS or a similar chelating agent. The resulting test, however, is then reported incorrectly on a form using reference ranges derived from urine collected without a prior provocative chelation--without using DMPS or any other chelator.  If that is done, the result is always much higher than references ranges on the report form and an attempt at interpretation is not possible. That type of incorrect reporting sometimes frightens patients into believing they are poisoned and need mercury chelation, when in fact no interpretation can be made without proper reference ranges.

Every person has mercury and other toxic elements in their body.  That was true since time began, long before present-day environmental and technological pollution. If DMPS (dimercapto propane sulfonic acid), DMSA (dimercapto succinic acid) or any other chelator of mercury is administered before urine collection, mercury excretion will increase greatly for everyone, even for people with relatively low and nontoxic levels of mercury.  The question to answer is whether or not the amount of mercury measured is enough to cause symptoms of ill health.  Is it a toxic level?  Or is it below a non-toxic and well-tolerated threshold. The only way to know that is to use accurate reference ranges for interpretation and comparison. Physicians and patients are advised to contact the laboratory and insure that the reference ranges on the report form are correct for the provocative method used to collect the urine, as described below.

Following a single dose of the mercury chelators DMPS or DMSA (but not EDTA), urine mercury will increase enormously in healthy subjects without toxicity, up to 4,000 percent even if the person tested has relatively low and nontoxic amounts of mercury in the body. If a mercury chelator is given before collection, test results should therefore be interpreted using reference ranges derived in that same laboratory, using the same instruments, from a large number of people, and following the same protocol of provocative chelation for urine collection.  Mercury chelators will always cause a great increase of mercury in the urine, regardless of the amount in the body.  The safe, nontoxic threshold will always be much higher after a chelator.  If proper procedure is followed, a large majority of healthy Americans tested will be in the nontoxic range.  The way mercury testing has been reported by many laboratories, provoked urine measurements are printed on forms intended for urine collected without a chelator, and every patient mistakenly reads very high, regardless of the amount in the body.

It is true that there is no "good" level for mercury.  It is a toxin.  It is also true that every toxin has a safe level below which it is well tolerated. But if tens of millions of people can live in good health and without toxic symptoms, at a tolerably low level of mercury, it is highly misleading to tell patients they have mercury toxicity and need to undergo a course of mercury chelation when they are at or below that same level.

To correctly interpret a laboratory test report and to know if a patient really has a mercury problem, it is necessary to compare results with mercury levels measured by that same laboratory in urines from a large population of typical Americans, most of whom have no symptoms of toxicity. Reference ranges must be derived using the same method of provocative testing and specimen collection. Laboratories should also print the means and standard deviations on the report forms, computed from test results of a large series (200 consecutive healthy test subjects or more) using the same collection protocol.  At the mean level, approximately 50-percent of all people tested will be higher.  At the mean plus one standard deviation (SD), approximately 16-percent will be higher.  At mean plus two SD, 2½-percent will be higher.  At mean plus three SD, less than 1-percent will be higher.

Reference ranges reported by clinical laboratories are customarily set to fall between the 2.5th percentile and the 97.5th percentile (mean +/- 2 SD). That is the method recommended by the International Federation of Clinical Chemistry (IFCC).(3) For toxic elements with no known nutritional value, such as lead, the reference range might better be established below 95th percentile (approximately the mean plus 1.8 SD). Those percentiles will vary somewhat if the distribution of values is skewed, but this method allows a reasonable approximation. Clinicians are often not aware that ranges printed on laboratory forms are rarely represented as “normal” by the lab. It is properly left to the doctor to interpret what is normal and what is abnormal.  Laboratory reports generally have reference ranges based the mean +/- 2 SD, derived a large number of similar tests at that same laboratory.

It is not justified to diagnose metal toxicity at levels lower than those found in the majority of the population with no symptoms of toxicity. In my practice the cut-off point for toxicity is set at 1.8 SD above the mean for most toxic metals (approximately the 95th percentile), and sometimes lower for mercury, depending on clinical findings. In the absence of scientific evidence to indicate otherwise, that is the accepted scientific procedure laboratories follow when establishing reference ranges printed on report forms.  If that level does not cause symptoms in two hundred million Americans who are below the 95th percentile, then that same level is not likely to be causing symptoms in the person being tested.

Hair analysis can also be used a screening test, but hair is subject to external contamination.  If hair analysis is used, results are best confirmed by provocative urine testing before a course of therapy is undertaken.  The sensitivity of measurements of hair elements is generally lower because of the need to first dilute and digest specimens in solution.  On the other hand, the concentration of elements in hair is usually higher.

It is not advisable to make a diagnosis using hair analysis alone.  If the laboratory report on hair is accurate (not always the case), and if hair mercury is less than 1.5 to 1.8 standard deviations above the mean for that laboratory, then it is unlikely that mercury is a significant cause of symptoms.  If hair mercury is greater than 1.5 standard deviations above the mean, a confirmatory DMSA provocative urine test will determine the extent to which mercury toxicity might be a problem and will rule out hair contamination. In my experience, reference ranges listed on laboratory report forms are sometimes quite arbitrary. I recommended using means and standard deviations, as described above.

It is rarely advisable to make any type of final diagnosis or to begin a lengthy course of therapy based on a single laboratory test.  Laboratory error is not uncommon..  For example, it is not acceptable to begin life-long treatment for diabetes with only one blood sugar reading, or to treat gout after a single uric acid measurement.

Accurate mercury testing is very difficult and not all laboratories have proper quality control.  In the past, I have submitted multiple samples of the same specimens to several different licensed laboratories to compare results.  The reports varied widely.  I have also submitted homogeneous portions of the same specimen to the same laboratory at different times, again with widely different results.  Some of that type of error will be cancelled out, if the laboratory is consistent and if the means and standard deviations for that laboratory are used to interpret results as described above. Reputable laboratories will usually refund charges for duplicate test specimens submitted only to check their accuracy and reproducibility. It helps the laboratory and adds to their internal quality control.

I do not recommend blindly accepting reference ranges listed on laboratory report for toxic elements.  I have personally computed the means and SD for each metal using a large number of test results from my own patients, using the method described below. With the computed means and SDs. I have established my own reference ranges. I do this because laboratories often do not provide me with that kind of data and ranges on the forms often seem so arbitrary.

Provocative Test for Mercury

To test for mercury and a spectrum of toxic elements I administer a single 500-mg dose of DMSA by mouth on an empty stomach with a glass of water, preferably in the morning before breakfast. All urine is then carefully collected for six hours, while under supervision in my office. Care is taken not to contaminate the urine or the container. Using this method, a wide range of toxic metals (mercury, lead, arsenic, cadmium, nickel, and others) are be measured relative to urine creatinine.  I no longer use 24-hour urines, as explained earlier, because they are subject to much greater collection errors.(1) Urine is more concentrated with fasting and has higher concentrations of trace metals. Concentrated urine facilitates testing and improves accuracy for metals with the lowest levels. Creatinine is excreted in urine at a relatively constant rate throughout the day. On average, 1.2 grams of creatinine is excreted daily—varying with weight and decreasing with age. If urine is more dilute, creatinine is more dilute. By measuring mercury and other metals relative to urine creatinine, a partial correction is made for variations in dilution. All chelation, including DMSA, should be stopped at least four weeks before initial or follow-up provocative urine testing to allow equilibration throughout in the body. I periodically update reference ranges I personally use in my practice for this test protocol and post them on the website below:


In my opinion, intravenous DMPS is obsolete and no longer has a place in medicine.  I have heard from many people who received DMPS elsewhere and were sicker by it—one was permanently disabled.  Even if DMPS is used, it can effectively be given by mouth without intravenous infusions.  EDTA is a very weak chelator of mercury and is not an effective treatment for mercury toxicity. For mercury removal, I recommend DMSA, by mouth as a the safest, most effective, and least expensive treatment for mercury removal.(4)

The form of mercury with greatest toxicity is methyl mercury and related compounds.  Those have a high affinity for the brain and nervous system. DMSA crosses the blood-brain barrier and removes mercury from the brain and spinal cord. DMPS is much less effective in that regard. DMPS is also three times more toxic than DMSA, based on LD-50.  Animal studies show DMSA to be approximately three times more effective than DMPS in removing mercury from the brain.(4) DMSA has the great advantage that it is taken by mouth in capsule form.  As reported at the fifth Nordic Symposium on Trace Elements in Human Health and Disease, Norway, in 1994, "DMSA may now be considered as the treatment of first choice in cases of acute or subacute lead poisoning and in mercury poisoning.  All experimental and clinical experiences show a low toxicity for this drug. Some chemically sensitive patients react adversely to any type of medicine, but DMSA is so safe that the FDA approves its use in small children.

It has been mistakenly said that EDTA and DMSA can cause mercury to enter the brain.  That is false.  Those chelators bind tightly and inertly to metals and remove them from the body in urine and feces. According to some studies, much more is removed in feces than in urine.

The general principles applied above to mercury can also be applied to testing and interpretation of other metals, with appropriate modifications.


1. Cockcroft DW, Gault MH. Prediction of creatinine clearance from serum creatinine. Nephron. 1976;16:31-41.

2. Skerfving S. Toxicology of inorganic lead. In: Prasad AS, ed.: Essential and Toxic Elements in Human Health and Disease. New York, Alan R Liss Inc; 1987:619.

3. IFCC Expert Panel on Theory of Reference Values. Part 5, Statistical treatment of collected reference values, determination of reference limits. J Clin Chem Clin Biochem. 1983;21:749-760.

4. Aaseth J, Jacobsen D, Andersen O, Wickstrom E. Treatment of mercury and lead poisonings with dimercaptosuccinic acid (DMSA) and sodium dimercaptopropanesulfonate (DMPS). Analyst 1995 Mar;120:853ff. (Presented at the fifth Nordic Symposium on Trace Elements in Human Health and Disease, Loen, Norway, June 19-20, 1994).

[Return to menu]