HORMESIS REVISITED ONCE MORE

Milk, honey, clean water and outdoor exercise have a deservedly good reputation, yet in sufficiently high amounts each of them is lethal. (No? Try 5 gallons of milk for breakfast.) This is so obvious that the effect does not even have a name.

But there is also the opposite phenomenon: deadly poisons are beneficial in very small quantities. And that does have a name: hor-mesis. For chemical and biological poisons, it has long been known to physicians, but it is a comparatively newly discovered pheno-menon for ionizing radiation: at low levels, radiation is beneficial to the health of living organisms as measured by resistance to disease, fertility, longevity, and other indicators. The evidence was there in experiments since the late 19^th century, but it was not consistently collected before Prof. Luckey's book Hormesis by Ionizing Radia-tion (CRC Publ., 1980). Today international conferences are being held on the subject, and radon measurements are putting to rest the "no threshold, linear" hypothesis, according to which the risk of radiation is directly proportional to the dose from zero up. In reality hormesis causes the risk of cancer to be smaller for a little radiation than for none at all, though of course, the curve soon turns round and then rises until the level becomes lethal.

There is a number of seemingly curious observations that by themselves would be quite worthless, and even in larger amounts they still do not amount to much, but they are consistent with hor-mesis, and at the very least they are no longer surprising. One of these surfaced last month when the EPA found that among US states, Iowa had the highest average levels of radon in the home, with 70% to 75% of its homes (as estimated from a representative sample of 1,500 homes) exceeding the EPA's 4 pCi/l borderline. When the finding was announced on 10/15/89, Gov. Terry Branstad said that "Iowans should not panic." Statistics can be misleading, opined a state official. "Despite" the high radon levels, he said, Iowa's cancer rate is lower than the national average.

But all 8 of the mountain states, with roughly double the radia-tion background of the US average, have a cancer incidence way below that of Iowa's, and the correlation between background and cancer incidence for the 50 states is strongly negative (the more radiation, the less cancer).

There is a "Free Enterprise Mine" in Boulder, Montana, which is used by arthritis and other patients because of its high radon concentration (it is high also in the miraculous waters of Lourdes, Bath, Bad Gastein, and other European spas). I interviewed the owner last month, and he claimed the radon concentration was 1500 pCi/liter. That's 375 times above EPA standard (set for lifetime exposure, of course), and people pay good money to lie inside for a few hours.

All of this proves nothing, but fits the general pattern. For lower organisms hormesis is well established, and for lower mammals such as mice the evidence is highly persuasive, too. For human ex-posures, the evidence is not yet fully accepted, but some of it is very persuasive, and there seems to be a near total lack of counter-evidence.

One of the interesting, though not conclusive, examples is given by John Fremlin, professor emeritus of radiology at the University of Birmingham (England), in an article that I listed in [AtE Jul 89]. It is the case of leukemia and other cancers in Hiroshima and Nagasaki among 9,000 children under 10 at the time of the bomb-ing. At levels below 50 rem the disease incidence is actually lower than among the control group of 5,800 children who received no exposure at all. I have previously explained [AtE Sep 89] why for a rare disease such as child leukemia these seemingly large numbers are quite insufficient for drawing firm conclusions; but clearly the data lend no support to the linear hypothesis.