In a recent study of amalgam in children, the authors said in their conclusions, "…although it is possible that certain especially sensitive children could be affected by low-dose mercury exposure from amalgam, the factors which might produce enhanced sensitivities are unknown." (Bellinger et al, Neuropsychological and Renal Effects of Dental Amalgam in Children; A Randomized Clinical Trial: JAMA, April 19, 2006; Vol 295, No 15, 1775-83).
Many metals are poisonous at medium to high levels, but are essential to our well being at very low levels. Mercury is one of the few exceptions. There is no intake level at which mercury is good for us. Nevertheless we have ways of dealing with it so that at very low levels it may be a nuisance but is not harmful. Between the level where mercury is toxic to everybody and the much lower level where it affects nobody, there is a fairly large range where most people are unaffected but where some of us, the so-called mercury sensitive, are harmed (often in ways that remain invisible for years).
As to the factors which cause the sensitivity, they may be unknown, but it is not too difficult to identify a few suspects. Of these, genetic weaknesses are probably most important but there are some dietary and environmental factors as well.
1. Dietary deficiencies: (a) Most of the specialized chemicals which comprise our body's defenses are manufactured in the cells of our body using simpler chemicals supplied by the food we eat. Many food chemicals are used in this manufacturing process, but some are required in greater quantities than others. In particular, selenium and sulphur are especially important for making the chemicals which deal with heavy metals so a low level of selenium or sulphur in our diet would reduce our natural defenses against mercury. (b) In addition to those food chemicals which are used as building blocks, some food components, like the vitamins, can be used without further modification to make life easier for the first line of defense. For example, glutathione (GSH), which combats mercury, also has antioxidant duties – antioxidants like vitamin C and vitamin E can shoulder part of the load, leaving GSH free to deal with heavy metals. Conversely, a lack of these antioxidants in the diet puts more pressure on GSH.
2. Free radicals: all foods contain free radicals and more are created as food is digested. Most foods also contain enough antioxidants to neutralize most of the free radicals. But there is a variety of ways in which the balance might be shifted: commercial agricultural methods, which go heavy on the macronutrients (NPK) and don't bother with the micronutrients, produce foods lower in antioxidants; refining and processing have a tendency to reduce the antioxidant content of food; cooking in hot oil creates free radicals. An excess of free radicals causes oxidative stress which in turn makes us more susceptible to harm from heavy metals.
3. Environmental pollutants: heavy metals in our food or in the air use up part of our defensive capacity and hence make us more sensitive to mercury from amalgam. Also, some industrial and agricultural chemicals use up the resources of our immune system which are needed to deal with heavy metals.
4. Physical insults: serious illnesses like encephalitis, certain injuries like concussion, routine medical procedures like a general anaesthetic, and physical stress like heat stroke can cause permanent damage to the brain and central nervous system. These things are seldom serious enough by themselves to make us aware that we have been wounded, but they reduce our capacity to withstand oxidative stress making us more vulnerable to heavy metals.
Any one of the above is capable of making a genetically robust person mercury sensitive. But for most of us, the above factors will not be severe enough to do the job by themselves — there will have to be an additional contribution from one of more of the following factors, most of which have a genetic component.
5. Glutathione (GSH): present in every active cell in our body. It is a bodyguard. It makes no contribution to the business of the cell until an unwelcome intruder enters. If the intruder is a molecule of mercury, two GSH molecules wrestle it to the ground and all three go out with the trash. These self-sacrificing GSH molecules have to be replaced. Some of us have a defective enzyme in charge of manufacturing GSH which reduces the amount we can make. A reduced supply of GSH (or a defect anywhere in the GSH system) would be a factor in making us mercury sensitive.
6. Metallothionein (MT): not a specific chemical; rather it is a general name for a large group of closely related chemicals, first identified 50 years ago. What they have in common is that they are specialized proteins, on the small side for proteins, and they have a very high metal and sulphur content. They have constructive roles in seeing that zinc and copper get to where they are needed. And they have a defensive role in protecting against poisonous metals like cadmium and mercury.
Our body is capable of making many different MTs, depending on what is required. There is a lot of room here for insufficient or substandard MTs, whether from inadequate raw materials or sloppy workmanship (enzymes not up to scratch) or flaws in the genetic blueprints.
7. Apolipoproteins: involved in the transport of fat and cholesterol. There are a number of distinct apolipoproteins (identified by letters from A through to E) with slight differences depending on their home and destination and also on some secondary functions they are able to perform. ApoE comes in three forms, apoE2, apoE3, and apoE4, but nobody has all three forms. Our genes are coded to manufacture apoE in pairs and the pair can be a matched pair of the same type, or an unmatched pair in any combination. The pair we are born producing is the pair we are stuck with for life. A matched pair of two apoE3's is the most common — 60%. 21% of us have an apoE3/apoE4 combo. 11% have apoE3/apoE2. The remaining 8% of the population has one of 2/4, 4/4 or 2/2.
All three forms do the transport job, though apoE3 and apoE4 are best at that. The apoE2 combinations are just barely adequate under the best of circumstances; and the few people (less than 1%) who have the apoE2/apoE2 combination are most likely to have health problems connected with fat or cholesterol transport. But while apoE2 is not great at its main job, it is very good at protecting against mercury. Conversely, apoE4 is a first rate fat transporter but pretty much useless in protecting against mercury. So those of us with a 3/4 combination, or worse yet, a 4/4 combination, are more susceptible to harm from small amounts of mercury.
8. Neurotrophic repair proteins: in the past 10 years there has been a lot of scientific interest in the identification and study of proteins which not only assist in the growth of new cells (notably nerves and neurons) but also repair and rejuvenate sick and tired cells. The first such neurotrophin was discovered in the 50's but intensive study did not begin until 1993 when the high powered GDNF was identified. At first it was thought that GDNF was specific to dopamine producing neurons and might be the key to understanding Parkinson's Disease. It probably is important to PD, but not just to PD. GDNF has been found in cells all over the body. And GDNF is accompanied by a whole family of neurotrophic proteins (including neurturin which is just as powerful as GDNF) which work together as a maintenance crew.
Any problem, genetic or otherwise, which reduces the activity of neurotrophic proteins will reduce the ability of injured cells to bounce back from an encounter with a toxin like mercury and hence increase sensitivity.
9. Melatonin: when we are about to fall asleep, the pineal gland starts releasing melatonin. We tend to think of melatonin as producing sleep, but from the point of view of the pineal gland, it is approaching sleep which triggers melatonin production. And melatonin, in turn, acts as an alarm clock for the night crew — there are some chores that are best done at night when the systems connected with physical and mental activity are largely shut down. One of these is cell repair.
Melatonin signals to GDNF and the rest of the neurotrophic crew that it is time to get to work. Without melatonin, much less gets done. So anyone with a defective pineal gland will have reduced GDNF activity and therefore an increased sensitivity to mercury. In addition to genetic flaws, the pineal gland is subject to environmental effects. Most recently it has been shown that fluoride reduces the ability of the pineal gland to make melatonin.
10. CPOX4: this is a gene designation referring to the section of our genetic map that has the blueprint for coproporphyrinogen oxidase (an enzyme for making a component of blood that is a first cousin to hemoglobin). CPOX4 is polymorphic which means that one or more alternate versions of the gene is fairly common (the gene for hair colour, for example, is polymorphic). Some people have a version of CPOX4 which makes them more susceptible to the toxic effects of mercury.
Our defenses against mercury are many and varied with backup systems and backups for the backups. For most individuals who are mercury sensitive, there are probably a variety of factors, working in combination, rather than some single deficiency.
We generally think of mercury as a neuropoison but it affects all organs, including the kidney. Whether the kidney of a mercury sensitive person will be affected depends on the amount of mercury and the nature of the deficiencies which are causing the sensitivity. Not everyone who suffers harm from low levels of mercury will have kidney damage. But the kidney has the virtue that a loss in efficiency is easily measured — it may be the easiest organ to check for damage. And there is growing evidence that the amount of mercury that is absorbed from amalgam fillings is sufficient to cause harm to the kidneys of a significant portion of those people who are mercury sensitive.
The conclusion of the amalgam study, cited above (Bellinger et al, Neuropsychological and Renal Effects of Dental Amalgam in Children; A Randomized Clinical Trial: JAMA, April 19, 2006; Vol 295, No 15) is that for the average child amalgam is safe. And the safety of amalgam is what was reported in the news without, for the most part, making the point that there may be mercury sensitive children for whom this is not true. Also not mentioned in the news was an editorial published in the same issue of JAMA where H.L. Needleman wrote, "It is predictable that some outside interests will expand the modest conclusions of these studies to assert that use of mercury amalgam in dentistry is risk free. This conclusion would be unfortunate and unscientific." (JAMA, April 19, 2006; Vol 295, No 15, p.1836). Both Needleman and the study authors point out that some of the children suffered kidney damage during the five years of the study and while the data from the study does not prove that mercury from amalgam caused kidney damage, that possibility cannot be ruled out either.
While this experiment was not designed to address the question of mercury sensitivity or whether there might be occasional bad reactions to amalgam, I find the data more suggestive in this regard than do the authors. Furthermore, I think there is a hint as to what percent of the population is mercury sensitive.
The main focus of the study was on possible effects on IQ. The ACR (albumin/creatinine ratio) kidney data took a back seat with results reported in a single sentence: "Among the 180 participants in the amalgam group, the unadjusted mean (SE) albumin level at year 5 was 32.8 (6.9) mg/g of creatinine (median, 7.5) and among the 183 in the composite group, it was 23.7 (5.0) mg/g of creatinine (median, 7.4) …"
On the subject of kidney damage: "There were 77 children with microalbuminuria (albumin >30 mg/g of creatinine) during the trial with no significant difference between treatment groups."
Looking at microalbuminuria first, an ACR of 30 mg/g is considered the threshold for mild kidney damage (macroalbuminuria starts at an ACR of 300 mg/g and that is kidney transplant territory). The figure of 77 children with microalbuminuria is a total for the two groups. That works out to 19% of the 409 participants for which renal data was obtained.
When extensive testing of children was done in the USA, 9.5% had microalbuminuria (Jones et all, Microalbuminuria in th US Population: Third National Health and Nutrition Examination Survey, Am J Kidney Dis, Vol 39, No 3, 2002, 445-459). So 19% is an unusual amount of kidney damage and highlights a weakness of the amalgam study. They assumed at the outset that composite fillings could be used as a control to compare with amalgam fillings. But while composites have been used for some time, negative health effects are still coming to light and it is not impossible that something in composites is hard on the kidney. This is conceded (obliquely) by the study authors in their concluding remarks: "Finally, the choice of composite for comparison was based on widespread use and availability. The safety of the composite used is itself not established nor could it be assessed in this trial."
The authors don't report separate microalbuminuria numbers for the amalgam and composite groups but, since they didn't find a significant difference between the groups, it is fair to assume that kidney damage in both groups was unusually high. It turns out that it would have been a good idea to include a third group, in the study, which didn't need any fillings (or who needed fillings but were asked to wait a while).
There are lots of things that are hard on kidneys including some environmental factors. To give just one example, the municipal water supply in Walkerton, Ontario was contaminated by Escherichia coli in 2000, which caused diarrhea associated hemolytic uremic syndrome (HUS). 19 children who recovered from HUS we matched with 38 children who showed no symptoms and both groups were tested, three years later, for ACR levels. (Garg et al, Microalbuminuria three years after recovery from Escherichia coli O157 hemolytic uremic syndrome due to municipal water contamination, Kidney International, 2006, Vol 67, 1476-82). After three years, 32% of the HUS children developed microalbuminuria versus 5% of the controls.
Returning to the amalgam study, if a no-fillings group from the same geographical area as the other two groups also had high ACR values, it would let composites and amalgam off the hook (but would suggest there was an environmental factor at work in New England which ought to be investigated and eliminated before it did serious damage). On the other hand, if a no-fillings groups (taken from the same geographical and economic area) had less than 10% microalbuminuria, it would be evidence that the fillings were responsible for some of the cases of microalbuminuria.
The median values of ACR for the amalgam and composite groups were very close: 7.5 mg/g for the amalgam group and 7.4 mg/g for the composite group. The median is obtained by plotting all the measurements in order of increasing magnitude — the measurement in the middle is the median. For the two medians to be so close, the measured values of ACR were probably increasing in lock step, indicating that for well over half the participants there was no difference between the two groups (and the ACR levels in both groups were quite acceptable). Without looking further it would be reasonable to conclude that for most people, amalgam has little or no effect on the kidney.
If the increase in ACR values had been linear (same rate of increase after the midpoint as before), then the mean (arithmetical average) value for ACR would have been the same as the median. The fact that the means were much higher than the medians (32.8 versus 7.5 for amalgam and 23.7 versus 7.4 for composites) suggests that somewhere after the midpoint, the ACR values for both groups started to increase at an accelerating rate (producing an exponential curve). And for these higher values to have such a marked effect on the overall average, there must have been quite a few children with high levels of ACR.
When statisticians analyse a set of measurements that increase exponentially, they make adjustments; in this case they applied a log transformation to the ACR measurements to prevent a few large values from having an undue influence on the whole group (before comparing with another group). If the data had been analysed without being adjusted, the conclusion probably would have been that there was a statistically significant difference between the amalgam and composite groups. After the adjustment they concluded there was no significant difference. This is the sort of thing that makes outsiders suspicious of statisticians. They appear to be manipulating the numbers until they get the result they want. There are times, however, when some sort of manipulation is fully justified.
Suppose you are comparing household incomes in small towns of 1000 households. In one such town, 999 have incomes from $15,000 to $90,000; plus there is one billionaire who lives in a modest house on the edge of town. A simple average would show incomes of over one million for each household in this town. This would produce a grossly misleading picture of the typical household. Better to call the billionaire an outlier and exclude him.
Now, suppose that, rather than a single billionaire there is a classy subdivision of 20 estates down on Lakeview Avenue. These households have incomes in the one to five million range. The other 980 households have the same income distribution as the last town. What do you do? If the 20 estates are included, the average income in the town is over $100,000 which still presents a distorted view of the typical town household. But there are too many estate households for the analyst just to ignore them. Besides, those estates are going to affect the town in many ways and the data should be treated in a way that reflects this.
If you are trying to determine the purchasing power of a typical household you may want to emphasise the median rather than the mean. But if you are looking at tax base, those estates are going to be making a large and significant contribution which you would not want the analysis to ignore. And if you are in the luxury vehicle business, then estate subdivisions will be your primary interest
There are several questions that need to be answered about amalgam. First, is whether the small amount of mercury released from amalgam fillings is sufficient to harm the average person. This study presents compelling evidence that the average person is not affected in any discernable way by mercury from amalgam. And, contrary to the news reports, that is all this study claimed to show.
A second matter of considerable interest is whether some fraction of the population is sufficiently sensitive to mercury to be harmed by amalgam fillings. The high mean of the amalgam group suggests that this is the case but does not prove it. Nevertheless, the children in the amalgam group who had high ACR scores are of particular interest because it is likely that there are more mercury sensitives among them than among the rest of the cohort. Some of those high ACR measurements could have come from other causes and could have gotten into the amalgam group by chance. Nevertheless, that high mean for the amalgam group should make amalgam a prime suspect.
One minor curiosity of the two groups is that the amalgam group was half and half, boys and girls, whereas, the composite group had more girls than boys (58% girls, 42% boys at the start of the study). And girls tend to have higher ACR levels than boys. So if both amalgam and composites had zero effect on the kidney, we would have expected the composite group (with its larger proportion of girls) to have a higher mean ACR, rather than the other way around. The study authors would have taken this into account when determining whether the higher mean of the amalgam group was significant.
When two groups are being compared, a statistical analysis is used to come up with a P-value which gives the odds that any difference between the two groups occurred by chance. If the P-value is 1, the odds are 100% that observed differences (if there were any) occurred purely by chance. If the P-value is 0.05, there is a 5% possibility that the differences occurred by chance (or a 95% possibility that the differences between the two groups are due to the factor being tested). In medical studies, a P-value of 0.05 or less is considered highly significant and the proposition is considered to be proven (subject to corroborating tests).
What about a P-value of 0.35? Not significant. It means there is a 35% possibility that observed differences occurred by chance. Or just a 65% possibility that the proposition is true (not even close to the required 95%). This is reasonable when establishing medical facts. But if the thing being tested is something to do with my health, am I going to be reassured because it is not proven to be harmful — that there is only a 65% chance it will do me harm?
It seems to me there is something backwards here. It would appear that amalgam is going to continue to be considered safe until it is demonstrated with 95% certainty that it is harmful. Should we not be taking the opposite tack? If there is more than a 5% possibility of harm then we should be concerned. In other words, in a test such as this if the P-value is less than 0.95 it should be cause for concern.
The higher mean of the amalgam group must have been analysed for significance but the P-value is not reported. Just that the difference between the two groups was not significant, so we know that the P-value was greater than 0.05 but we don't know how much greater. My guess is that, even after a log-transformation of the data, the P-value for the ACR measurements was somewhat less than 0.95 and that we should be concerned.
In a list of seventeen adverse health conditions (Table 4 in the amalgam study), some health conditions had more cases in the amalgam group, some conditions had more in the composite group. But there were four conditions with noticeably more cases reported for the amalgam group (P-values ranging from 0.42 to 0.25). In any group of 17 there are certain to be a few striking differences due to random occurrences. If in a succeeding study with the same parameters there were a different four conditions with the lowest P-values, it would show that the results in the first study were meaningless. If, on the other hand, one or more of these conditions did a repeat performance, it should be taken seriously (even if the combined P-value was still greater than 0.05).
'Not proven to be harmful' is quite different from 'proven not to be harmful.' Tobacco was not proven to be harmful for most of the last century but that didn't keep it from doing harm.
Returning to the 77 children who developed microalbuminuria, what are we to make of the fact that 19% is much more than the 9.5% that would be expected in a random group of children? The authors of the amalgam study did not break down the microalbuminuria numbers for the two groups and while it would be mathematically possible for the composite group to have had more than half of the microalbuminuria cases, it is likely that the amalgam group with its much higher ACR mean had a few more cases of microalbuminuria than the composite group. This would make the microalbuminuria rate in the amalgam group a little over 20%
The interesting thing here is that if amalgam were responsible for the extra microalbuminuria cases in the amalgam group (over and above 9.5%) then the children who got microalbuminuria because of the amalgam fillings must be mercury sensitive. This gives a base level for mercury sensitivity of approximately 10% of the population.
Of course, there are many things that can cause kidney damage so even if sensitivity to amalgam is responsible for some of these cases, it is unlikely that mercury sensitivity would account for the whole 10%. Nevertheless, amalgam is implicated by this study. It is the prime suspect. And even though there are many possible reasons for there being 10% more cases in the amalgam group (than would be expected from comparing with the US national average), we should take a second look at the suspect in custody before we go chasing phantoms.