There’s no junk science in pesticide regulation
In a June editorial, Terence Corcoran dismissed my arguments questioning pesticide regulation in this country as “junk science.” As a medical professional, I take science very seriously and have little patience with risk assessments and political agendas that take scientific findings out of context. When this happens, one truly encounters “junk science.” I will explain why I, the Ontario College of Family Physicians, the provincial government, Public Health and the Canadian Cancer Society got it right.
We cannot assume that simply because a product is available in Canada that it is “safe.” The recent re-registration of the common herbicide 2,4-D, the most widely used herbicide in the world, is a good case in point.
It’s not that the final decision to allow the use of 2,4-D was right or wrong — it may be legitimate to conclude that the industrial and commercial benefits of products containing 2,4-D could possibly outweigh adverse health and environmental impacts. But the evaluation process is so severely flawed, it is impossible to tell. Although the scientists working for Health Canada or in industry are usually conscientious professionals trying to do the best job possible, the constraints under which they operate make it difficult to arrive at scientific conclusions about safety.
In 2005, I sat on Health Canada’s Pest Management Advisory Council. We advised the Pest Management Regulatory Agency to cease calling registered pesticides “safe” when label directions were followed. To its credit, the leadership of PMRA followed this recommendation, switching to “acceptable for use” or “poses an acceptable risk,” acknowledging that all pesticides have inherent hazards to human and ecosystem health. Two years later the PMRA has reverted back to using the word “safe,” not because the science has suddenly changed, but in response to public polling demonstrating the public wish for reassurance.
Qualifying “safe,” with “when used according to label directions” is critical to Mr. Corcoran, but this is not based on science. To deem any pesticide safe today would be no different than calling DDT and organochlorines as safe in the 1960s before their carcinogenicity was established. Even our drug regulator is not so arrogant as to consider licensed pharmaceuticals safe. To its credit, Health Canada regularly reconsiders drug registrations and, in the U.S., the odds that products licensed by the FDA will be taken off the market or have a blackbox warning placed within 25 years of licensure is at least 20%. Sweden and Norway, believing themselves no less scientific, have de-registered 2,4-D, and Denmark has severely restricted its use. And sadly, those who may be the most biologically vulnerable — children, the elderly and pets — may be the least able to read or follow label directions.
But let’s stick to the core issue — does the Canadian registration process mean that registered pesticides will be safe for typical use?
The PMRA relies heavily on toxicological risk assessment, employing “safety factors,” better termed uncertainty factors. Safe doses are arbitrarily scaled from the dose known to cause no observed effects in a target species studied, multiplying by factors of up to 10 to account for each of the following unknowns: differences between animals tested and humans; variations among humans; data gaps; extra sensitivity of vulnerable populations such as the young, ill, pregnant or elderly; database deficiencies, such as a lack of a multi-generational or developmental neurotoxicity study; and extrapolations from subchronic to chronic exposures.
The scientific basis for giving each factor a value up to 10 is not clear and the process to arrive at the number is not transparent. The highest known factor employed by the PMRA is 3,000.
While in medical school, I had a summer job developing models to determine human effects based on experiments with rats which inhaled photocopying toner. This proved next to impossible, as interspecies differences between rats and humans were large, from detoxifying enzymes to fluid characteristics, to cellular makeup, to the degree of branching of airways. Such differences allow rats to live in environments that are inhospitable to humans but make modeling extremely complex.
Human-to-human variation, such as compromised excretory pathways including renal and hepatic insufficiency, cause vast differences in the human ability to detoxify pesticides. Only in the last decade have we discovered genetic polymorphism of the liver’s Cytochrome P 450 enzyme system. Coffee is normally not a risk factor for heart disease, but the genetic variation you have on Cytochrome P450 1A2 can triple your risk of a heart attack when you take coffee. Those with the 1A variant, the rapid metabolizers, may have lower risk with coffee ingestion but those with the 1F variant may be slow metabolizers whose risk increases substantially. Such variations alone could account for order of magnitude differences among humans detoxifying pesticides. Extrapolating short-term to long-term exposure, or to all life stages, including the developing foetal brain, is fraught with further uncertainty.
Many would say that a fundamental flaw of industry-run studies is that there is an inherent and inevitable bias in favour of any product being examined. The EPA in the U.S. found evidence of falsification of data, for example, on glyphosate till the early 1990s. However I am a bit less concerned about this with pesticides than drug trials, for unlike our drug regulator, PMRA is able to ask for the specific studies it wants. Funding may even allow industry to produce high quality studies and analysis.
What is of more major concern is what the PMRA chooses to ask or not ask. Industry critic Meg Sears, who has a chemical engineering degree (as do I), notes that the PMRA does not require testing for contaminants on real batch conditions, the equivalent of random drug testing. Complex breakdown products may be even more toxic than the original product. Sears asserts that PMRA has ignored data on 2,7-DCDD, a potentially carcinogenic contaminant, primarily because it is produced in such high amounts that they would not know how to evaluate it.
Epidemiological studies of people who have been exposed accidentally, in jobs where they are frequent users (golf courses, agriculture workers, pesticide applicators), and cross-sectional surveys of large populations, point to an association of 2,4-D with Non-Hodgkin lymphoma. Definitive proof of harm is difficult to establish because of confounders such as multiple exposures and the faulty recall of people. A high level of expertise is required to interpret relevant data. Unfortunately, until recently, the PMRA did not have a single epidemiologist on staff and made several fundamental errors in the 2,4-D decision.
For example, the statement of noted researcher Renu Gandhi that “one half of the studies indicated higher rates of Non-Hodgkin’s lymphoma among populations exposed to 2,4-D, the other half did not” was interpreted by PMRA as meaning there was no reason for concern. Mills’ study on California agricultural workers purported to show lower incidence of NHL in those using phenoxy herbicides, an astonishing conclusion derived from comparing those using phenoxy (2,4-D) herbicides with a very small group of 31 individuals using other pesticides.Compared to the much larger group that did not use pesticides at all, the phenoxy herbicide group was actually about three times more likely to develop the cancer.
Unlike Europe’s requirement to use the least toxic product for a particular purpose, Canadian regulations default to needing to demonstrate definitive harm before a product is disallowed. Whether harm is found depends on what you look for and with what tools. PMRA was willing to re-register 2,4-D with incomplete data, only requiring developmental neurotoxicity studies a year and a half from now. Canada is also lacking a systematic process to look at the synergistic effects of pesticides of the same class (e.g., organophosphates, operating on the same organ, or having the same effects on neurons).
Risk assessors recognizes that risk tolerance by a population is normative — that is, even if the science were solid and the risk knowable, legislators would be entitled and indeed obliged to make their own judgment of societal tolerance for risk.
Public health authorities must prudently weigh risks and benefits and differentiate caution from scaremongering. Family physicians help people make what may be life and death decisions each day, on products licensed by the government from medications for blood pressure, cholesterol or depression. In the Ontario debate on cosmetic use of pesticides we have each arrived at the conclusion that the benefits do not outweigh their risk of harm.
Neil Arya
Dr. Neil Arya is a member of the environmental health committee of the Ontario College of Family Physicians and an adjunct professor of environment and resource studies at the University of Waterloo. He represented the OCFP on the pest management advisory council until 2007.
