Measuring mercury exposure (and why provoked urine testing is the wrong way to do it)

A commenter by the name of “Robin” asked for information on mercury toxicity a while back due to her husband having reported high mercury levels. My workload is finally getting close enough to being under control that I can tackle this. However, it’s a complicated topic with a few different aspects. Before I actually get into some of the effects of mercury, I think a bit of context is important. So for this first post I’m going to talk about how mercury exposure is measured and how to know if mercury levels really are elevated. I’ll follow that up with some future posts (hopefully within the next week or so) about where this mercury exposure is coming from, and what the effects can be [Update: part 2 on some causes of high mercury levels is here, and part 3 on the different types of mercury is here]. Some of the concepts in this post build on an earlier entry on measuring chemicals in blood and urine.

In order to understand any sort of testing for mercury levels in humans, it is important to understand what happens to mercury in the body. In general, mercury is distributed throughout the body after exposure; generally the highest levels end up in the kidneys and brain. Generally inorganic mercury has a half-life on the order of days in most tissues, but can stay longer in the kidneys and brain. Organic mercury, particularly methylmercury (the mercury compound of concern from eating fish), tends to stay in the body longer. Eventually both inorganic and organic mercury are excreted in urine and feces.

The most common way to measure mercury exposure is through urine samples, since it’s fairly simple to collect these samples. While it is total mercury (organic plus inorganic) that is usually measured, generally the results of urine testing are a better indicator of inorganic mercury than organic mercury. Both Health Canada and the US CDC have conducted population monitoring of mercury concentrations in urine. The average level in the US adult population was about 0.5 μg/L (just over 0.5 μg/g creatinine), but there was considerable variation; 95% of the population had levels less than 3.11 μg/L (2.31 μg/g creatinine) in 2005-2006. The Canadian data only reported inorganic mercury in urine; the concentrations were a bit lower, with 50% of the population aged 20-29 having levels less than 0.23 μg/L (0.2 μg/g creatinine) and 95% of this population less than 2.31 μg/L (1.93 μg/g creatinine); concentrations were similar in 60-79 year olds, but slightly higher in 40-59 year-olds. Other studies (summarized by ATSDR) have shown average levels in the range of 4-5 μg/L, however.

Some health practitioners (I use that term loosely here) have been using what is called a “provoked” urine test. This involves having administering a chemical to the person being tested that basically forces the mercury (and other metals) that are circulating in the blood, kidneys, and other organs to be rapidly excreted. The result is that the levels in urine are temporarily highly elevated (potentially around 10 times what they would be in an unprovoked test). The practitioners and labs using this test then tend to compare the results to the normal (unprovoked) range, making it look like the person has an unusually high mercury level, which is then often followed by recommendations for therapies or treatments to reduce the mercury level. This test should, for all intents and purposes, be considered a scam; even a normal person with an average level of mercury exposure will show up as having a high level of mercury. The only way you could meaningfully interpret the results would be to compare the results to reference ranges for provoked mercury in the general population, but there aren’t really good reference ranges for this, and there isn’t really a good reason to develop one since the test results aren’t particularly meaningful to start with. But if you want to persuade someone that they have high levels of mercury or other metals so you can sell them expensive and unnecessary treatments, this is the way to go.

Measurement of mercury levels in blood is also relatively common, and is also included in the Health Canada and CDC studies. Blood tests pick up more of the organic mercury, and in particular methylmercury; as a result people who consume a lot of fish tend to have higher levels of mercury in blood. In the US, average levels in blood are around 1 μg/L, and 95% of the tested population had a concentration less than 5.31 μg/L in 2005-2006. The Canadian levels were fairly similar in this case. ATSDR suggests average levels are around 1 to 8 μg/L, with a mean concentration of 2 μg/L for people who don’t eat fish, but this is based on older data; concentrations as high as 200 μg/L have been reported in people who eat a lot of fish.

Mercury exposure can sometimes be measured in exhaled air, but this is really only an indicator of very recent exposure to mercury vapours and only of use in some occupational settings.

Both blood and urine concentrations can fluctuate a fair amount and can be affected by recent high exposures. For measurement of longer-term (on the order of about a month) average exposures, hair samples are sometimes used. Mercury circulating into the body is incorporated into hair as it grows, and once there the concentration doesn’t change. Concentrations in hair have been found to correlate well with concentrations in organs where mercury may accumulate, such as the brain and kidneys, particularly for methylmercury. If you really want to evaluate whether someone has long-term high mercury exposure, particularly methylmercury, hair samples are probably a lot more reliable than blood or urine. However they can still be affected by airborne mercury vapours or some hair treatment products, and concentrations can also be affected by the rate of hair growth, and so these tests aren’t perfect either. The data on typical levels are also a lot more limited; the limited data suggest a concentration of about 0.25 μg/g, ranging up to about 2.5 μg/g, is typical in European populations, while in Japan levels tend to be more in the 2 to 4 μg/g range (presumably due to higher fish consumption). CDC measured mercury in hair of women and children in one of their surveys, and found generally consistent results, with median and geometric mean concentrations of about 0.2 μg/g in women and 0.38 μg/g in women who ate fish frequently.

The other approach to measuring mercury exposure is to measure how much mercury there is in various media that a person is exposed to. For example, mercury concentrations can be measured in the air someone breathes, the food someone eats, and the water someone drinks. This type of monitoring is more useful for occupational scenarios where a large number of people may be exposed to mercury-contaminated air in a single location, or for determining the average and range of mercury concentrations in a particular species of fish or fish from a specific lake (for example); testing all potential sources of exposure for a single person would be prohibitively expensive and impractical.

While there are several ways of measuring mercury concentrations in biological samples, it is important to stress that all of these methods can have highly variable results. Furthermore, and perhaps most important, while these tests (aside from the provoked urine test, which is completely unreliable), if done properly using standardized methods, may provide evidence about whether a person’s mercury exposure is higher than normal, a higher than normal mercury exposure does NOT automatically mean that the exposure level is harmful, particularly since most common tests don’t differentiate between the different mercury compounds (more on this in an upcoming entry).

Advertisement