Tuesday, February 9, 2016

Cadmium, Arsenic, and glass manufacturing in Portland: What do the test results mean?


Tonight at Cleveland High School (in Portland, OR) the Multnomah County government convened a public meeting to discuss the recent pinpointing of elevated cadmium (Cd) and arsenic (As) levels in Portland to the Bullseye Glass manufacturing facility at 3722 SE 21st. Avenue. I was frankly hoping that the representatives of the Department of Environmental Quality (DEQ) and Oregon Health Authority (OHA) would put the levels DEQ measured of those metals in October into a health context. Unfortunately, they did not, and the meeting degenerated into an expression of public anger that didn't answer many questions.

Fortunately, we have well-established models of cancer risk from the Environmental Protection Agency (EPA) and reference exposures from the California EPA for the threshold of likely health risks in general. Those models and exposures are based on data from industrial use of Cd and As, as well as documented environmental exposures from published studies. In other words, we have an evidence-based context for the numbers DEQ measured in the air outside Bullseye's production facility. In hopes of helping people understand the big question voiced at the meeting tonight ("What do the emissions mean for my family?"),  I've compiled the results from air quality monitoring with those reference levels.

To make things easier to compare, I've put all the exposure data and relevant reference levels in a table with everything in the same units (micrograms per cubic meter).


Let's break these numbers down a bit. First off, the DEQ numbers come from the agency's monitoring station in the parking lot of Fred Meyer corporate headquaters, 120 meters from the Bullseye Glass factory. You can get those numbers here. Note: DEQ reported nanograms per cubic meter, so you have to divide those numbers by 1,000 to get micrograms per cubic meter.

The Oregon DEQ benchmarks are reported here. These numbers are set based on the EPA lung-cancer model, where the benchmark is the level at which exposure would increase cancer risk by 1 in one million for a person who breathed that air for their entire lifetime (chronic exposure).

The EPA cancer risk and CalEPA reference exposures come from this page (arsenic) and this page (cadmium). I used the 1 in 10,000 lung cancer risk estimate from EPA because it was closest to the exposures reported near Bullseye. It's a linear model, so it shouldn't be surprising that you can just multiply the DEQ benchmark by 100 to get that number as well. Note: The CalEPA reference exposures are given in milligrams per cubic meter, so you have to multiply those numbers by 1,000 to get to micrograms per cubic meter.

From a first look, you may notice that the average exposure DEQ measured falls near the 1 in 10,000 risk for lung cancer from the EPA model (it's actually about 1 in 6,300 for As, and 1 in 20,400 for Cd). The averages are also 2-3 times the CalEPA threshold below which no health effects at all are expected.

Hopefully you've noticed something else from the table above: the DEQ benchmarks are quite stringent. They are far below the CalEPA thresholds for any likely health effects at all from inhaling either cadmium or arsenic, and they correspond to a 1 in one million increase in lung cancer risk (based on the EPA models). For some perspective, men have a 1 in 13 chance of developing lung cancer in their lifetime (and a 1 in 2 chance of any cancer), while women have a 1 in 16 lifetime chance for lung cancer and a 1 in 3 lifetime chance for all cancers (check out the data from the American Cancer Society). Taken another way, if one million people lived at the DEQ benchmark, we would expect one lung cancer due to cadmium exposure, and about 70,000 lung cancers from other causes. Even going up to the 1 in 10,000 risk in the table above, you would expect around 50 cadmium-related lung cancers in a population of one million who all lived at the monitoring station, as compared to the 70,000 cases from other causes. Did I say stringent?

It's important to repeat that the data above come from right next to Bullseye. If you go even a short distance away, mixing in the atmosphere will lower the concentration of arsenic and cadmium. DEQ has even published a preliminary map of expected cadmium exposures here, based on relative concentrations of the elements in moss samples collected by the US Forest Service. Here's a screenshot:
These numbers are in nanograms per cubic meter again, so you have to divide by 1,000 to compare with the table. It may help to think of the boundary between the highest two concentrations (around 30 ng/cubic meter) as the average from monitoring (29.4 ng/cubic meter). By the time you get to the Fred Meyer day care, the estimated levels have dropped to about 0.01 micrograms/cubic meter - or right at the CalEPA threshold below which any health effects are unlikely. Farther out, at Cleveland HS or Winterhaven Elementary, levels are more like 0.005 micrograms/cubic meter, half the CalEPA health threshold and in the 1 in 100,000 probability of cancer in an individual due to the cadmium exposure. You may also notice a smaller point-source in north Portland, which looks like it's at Ostrom Glass and Metal Works.

So what's the conclusion?

Even for students at the Fred Meyer day care facility, inhaling the air is unlikely to cause any ill effects at all based on the CalEPA threshold and EPA cancer models. The cancer risk estimate is already down to 1 in 60,000. In other words, if there were 60,000 students in that day care, we would expect one of them to get a lung cancer from Cd exposure. There are in fact approximately 100 children who attend the facility, and they will not breathe this air 24/7 for 70 years (the assumptions of the EPA model). As for other health effects, the exposure looks like it would be right at the CalEPA threshold outside (and probably lower inside).

Further away (for example, in the residential neighborhoods east of 26th Ave. and south of Gladstone St., exposures should be lower. While there are reports of a detectable level of Cd and As in the soil at the day care facility, soil and air tests at Cleveland HS and Winterhaven Elementary ordered by Portland Public Schools did not detect any Cd or As. (I hope to link to those results if and when they are released).

In short, the levels DEQ detected are at or below where one would expect much of an elevation in lung cancer risk or other health consequences. Personally, I don't see grounds to get particularly worried about human health given the numbers that have been reported.

Let me be clear: It would be great to get cadmium and arsenic exposures down to the DEQ benchmarks, especially since they are elements and do not "break down" in the environment the way molecular toxins can. Given the possibility of soil accumulation, DEQ's efforts make sense. I hope that the employees at Bullseye are being evaluated for As and Cd exposure, since their exposures are potentially much higher inside the factory. And re-evaluating the pollution control measures at Bullseye is a no-brainer, since we now know that it is a significant point source for As and Cd. There is lots of work to be done.

However, the community anger expressed at tonight's meeting is, in my opinion, totally unjustified. It was clear from DEQ and Forest Service presentations that sleuthing out the source of Cd and As was an involved process, from DEQ identifying the two elements as more concentrated in monitoring than predicted from modeling known sources, to the Forest Service using a moss and lichen model to establish relative exposures throughout the City, to the DEQ doing targeted monitoring near Bullseye to confirm the source and document the exposure levels. Put simply, government agencies discovered the problem and tracked it to its source before anybody noticed any consequences. To me, that looks like effective detective work on the part of DEQ and USFS. The first commenter, reading a prepared statement from Neighbors for Clean Air, characterized the process as "an abysmal failure of government." While I understand that many of my neighbors are worried and angry, I think we can do better than that.


UPDATES:
As the conversation has broadened to include emissions of hexavalent chromium, I've written another post dealing with that metal.

Also, soil testing results from the nearby daycare have been released, and I've worked through those in another post.

The OR state cancer registry (OSCaR) has released the results of an analysis of lung and bladder cancer in the area around Bullseye glass; here are those results with my analysis.

35 comments:

  1. A few minor quibbles aside (the measurements at PPS were done after Cd emissions stopped, so say nothing about potential exposure of students), I think this analysis is very consistent with others I have talked to, and I think it is a good way to understand what little exposure information we currently have. Thank you for posting it.

    One area of disagreement is not technical -- it's the evaluation of the anger that some (but far from all) of the commenters last night expressed. While some of the comments were clearly over the top, I do think people have had the feeling that DEQ was acting a watchdog, and were upset to learn this was not the case. Betrayal is one of the bitterest of human emotions.

    Though it doesn't feel it, Portland's air quality is pretty bad, and I hope people will channel their anger toward the productive goal of changing the rules, tightening standards, and giving the DEQ the teeth it needs to be our watchdog.

    ReplyDelete
  2. This comment has been removed by the author.

    ReplyDelete
  3. Thank you for an excellent quantitative summary of the data. This is very helpful as a means of cutting through the emotionally charged, fear driven chatter.

    ReplyDelete
  4. Thank you. This is excellent quantitative analysis that makes me feel better as a neighborhood resident.

    ReplyDelete
  5. Excellent explanation. Thank you.

    ReplyDelete
  6. This is a huge help and - as you said - helps answer the main question for most of us: what does this mean in terms of health risks for my family? Some context and perspective would have been really good last, although it wouldn't have helped those folks who are too mad or frightened to take the time to understand what's really happening.

    ReplyDelete
  7. Do we yet know if these readings from October are similar to year round exposures. I was under the impression that we don't know if this set of readings is average, low or high. Until then, can we be sure that the numbers present truly reflect the risk and exposure or that the maps are accurate?
    I appreciate you breaking down the information and explaining what so far has been hard to get a grasp on, but I do still wonder if we have enough information to know the true breadth of risk involved. If anyone does know the answer to this, I would be glad to hear it.

    ReplyDelete
    Replies
    1. My understanding is that DEQ only did direct monitoring near Bullseye during October, and that they chose October because it had typically had higher emissions than other months in their long-term monitoring of Portland as a whole. But yes, right now we only know what happened right near Bullseye for October of 2015, and can just assume that those numbers are typical.

      Delete
    2. Also, the October measurement affects the absolute values of exposure in the maps DEQ is putting out, but not the relative levels. To get the scale (ng/m^3) in the maps, DEQ has combined relative Cd levels from the moss and lichen samples with absolute levels from their monitoring station at Bullseye.

      Delete
    3. Thanks for the response. This business is complicated and, understandably, troubling. Living in an urban area, we all have to accept that we are allowing ourselves/our families exposure to a certain level of excepted pollutants but I believe most people have been knocked a bit off balance by this news. I certainly was. The more clear things are made, regardless of the information being good or bad, the better.

      Delete
    4. We cannot really assume that the numbers are typical. They may be typically lower, as the DEQ rep. mentioned that we had an atmospheric inversion in October, or they could be higher, due to any number of other factors. Without data, there should be no assumptions, when there are so many variables at play. Dr. Parks, Go Bears (I hate sports, but love Cal). I graduated from Berkeley with a degree in molecular and cellular biology.

      Delete
    5. Thank you for posting this. It is very helpful. It is exactly the information I was looking for. I do not need hysteria, I need information. The cademum/arsenic comparisons were especially enlightening. I am going to sleep much better tonight having read this article.

      Delete
  8. Who exactly is the Wanton Empiricist? And what are his/her credentials?

    ReplyDelete
    Replies
    1. Hi - This is the author, Dr. Duncan Parks. I appreciate the comments below, in that I don't think you actually need credentials beyond converting metric units and being comfortable with comparing data and reference values to write the post I did. But if it makes you feel better I'm a biology prof with a PhD in Integrative Biology from Berkeley (and a BA in Bio from Reed way back when). I'm not a professional epidemiologist, which is why I'm using the EPA and CalEPA values in a straightforward way (epidemiologists construct those models and references based on their particular expertise and knowledge of the relevant literature).

      Delete
  9. This comment has been removed by a blog administrator.

    ReplyDelete
  10. he cited his sources and explained his reasoning. does it matter if he's the pope or a beggar?

    ReplyDelete
  11. I appreciate the quantitative analysis you've done and your effort to put known/assumed health risks of residents’ cadmium/arsenic exposure in perspective (emphasis on "known/assumed," because there's a lot we don't know about the risks). That kind of analysis is essential to this conversation, and it would be a great outcome if we were able to conclude that no significant health risks have occurred as a result of Bullseye’s 45 years of operations in the neighborhood. At the same time, I think you’re misunderstanding the source of public anger toward DEQ. DEQ hasn’t been a good detective. More like a really inept perpetrator. DEQ has known about Bullseye’s toxic emissions for decades. Bullseye has, for much or all of its existence, possessed a DEQ permit to emit this stuff. (And recall that though Bullseye has voluntarily halted use of cadmium and arsenic, the factory, DEQ is so far permitting the factory to continue emitting substances understood to be even more dangerous, such as chromium.) Putting up monitors to assess a public health risk DEQ has known about for decades, only because a Forestry Service scientist forced attention on the matter based on his accidental finding in a study about moss, is not good detective work, at least not on DEQ’s part. It’s more like letting your dog take a shit in the park for 45 years, then hiding behind a tree and counting how many people step in it. The only good detectives in this story, so far, have been the Forest Service and the Mercury. The big remaining mystery is how to fix an environmental protection system so broken that those in charge have forgotten the purpose of their jobs.

    ReplyDelete
  12. Has there been any testing related to the train yards? Diesel exhaust contributes the same toxins into the air. Both of these glass factories are located along the train lines and specifically where trains idle. Don't get me wrong, I'd love to have the cleanest air to breath, but pointing the finger all at one business doesn't seem like the answer. These glass producers are in very industrial areas of town and without direct testing of their stacks I find it difficult to solely condemn them. Further testing should be done for all industries that are in this loophole and air regulations created to avoid further pollution of our planet.

    ReplyDelete
    Replies
    1. Given that Bullseye directly adds arsenic and cadmium to their furnaces, and the chart shows a bullseye (get it?) pattern around their facility, it is obvious that they are the major contributor to the problem. Diesel emissions are likely a smaller contributor.

      Delete
    2. Yes I get it and I am sure they contributed to the problem, but complete blame is shortsighted. What I am saying is ALL the industries should be checked and regulated.

      Delete
  13. This comment has been removed by the author.

    ReplyDelete
  14. The studies that established the benchmarks for arsenic used data from chronic exposure through ingestion in Bangladesh...and the state toxicologist noted that the liver is involved in detoxification through this route of exposure. He stated that the liver is not as involved in detoxification through an airborne route of exposure, so the risk could be significantly higher...something to think about. That being said, I have a toddler at the Fred Meyer CCLC, and am not worried enough to yank him out, since the exposure was not 24/7, and the numbers were not totally off the charts. Still, there is very little data (one month of measurements? hard to draw any conclusions from that) and the toxicological studies are quite limited in scope. Many little erroneous assumptions have a way of magnifying exponentially. I need more data before the matter is settled in my mind.

    ReplyDelete
  15. Just want to add a couple of things. One, despite our differing takes on the politics of this situation, I just want to underline how much I appreciate what you've done in terms of crunching and interpreting these numbers, Duncan. I would have no idea how to do this myself. Two: Having started to make a closer study of what you wrote, along with lung cancer info on the ALA website, I have a question I wonder if you could weigh in on. You write that if a population of 1 million were exposed to cadmium at concentrations of .06 mg/cubic meter (the "1 in 10,000" level or about twice the ambient average in the inner ring of the Bullseye hotspot), you would expect about 50 cadmium-related lung cancers vs. about 70,000 lung cancers from other causes. Makes the cadmium cases sound fairy negligible. But what if you consider this info from the ALA: "It has been estimated that active smoking is responsible for close to 90 percent of lung cancer cases; radon causes 10 percent, occupational exposures to carcinogens account for approximately 9 to 15 percent and outdoor air pollution 1 to 2 percent. Because of the interactions between exposures, the combined attributable risk for lung cancer can exceed 100 percent." I'm not a math/science person, but wouldn't this suggest that of those 70,000 lung cancers, we'd expect about 63,000 to come from smoking, and 700-1,400 to come from air pollution? Relative to only those 700-1,400 lung cancers caused by air pollution, doesn't an increase of 50 start to look rather high? What's your take? Thanks.

    ReplyDelete
    Replies
    1. Granted, that inner ring is pretty small, and the number of people exposed at that level relatively few. As you say, if you go two rings out, cadmium levels are at the CapEPA threshold below which no health effects are expected.

      Delete
    2. Carmela, I think your line of reasoning is valid, and it looks quantitatively good to me. It's simply a question of what comparison is meaningful. If you want to say that living 24/7 within 120 meters of Bullseye adds 7-15% to the environmental component of your lung cancer risk, I think you're making a reasonable claim. But it's 7-15% because the (non-occupational) environmental exposure is so small. To me, comparison to at least the non-active-smoking part is more meaningful; personally, I want to know if those 50 cases/million amount to much compared to the other ways to get lung cancer, and even if we exclude smoking, it's still a small change (that would be 50/7000 or 0.7%). Getting your house radon-remediated is way more important, and to me that's a meaningful comparison in setting priorities. The other factor (which you brought up) is how many people actually live close enough to Bullseye to be exposed at those levels (24/7, lifetime), which I'm going to ballpark estimate is around 100. If the cancer models are right, 100 people adding a 1 in 10,000 risk means there are likely to be zero additional lung cancers (actually, in 99 out of 100 situations). Whether that sounds risky is a matter of opinion, but to me it really doesn't.

      Delete
    3. To put it another way, if the contribution from Bullseye pollutants made environmental exposure risk equal to radon risk, then I'd be worried; it would have become a significant factor in my mind. But it's orders of magnitude lower, even close to Bullseye. Again, that's about my threshold of concern, but I do think it's reasonable to argue that we should prioritize the largest non-smoking risk, and my opinion flows from that priority.

      Delete
    4. Oops...caught a math error. That should read that the (non-occupational) environmental component of lung cancer risk is increased 3.5-7% by the exposures measured at the DEQ monitoring station. Sorry!

      Delete
    5. It seems to me there is a real qualitative difference between risk you accept upon yourself (activities such as smoking or choosing to live in an area with some radon) and risks forced upon you by an individual or industry. There is a real moral difference and I believe (though I may be wrong) that our laws also reflect this difference.

      Delete
  16. What a relief that the chances of developing cancer are so low! But if one out of two men and one out of three women are already expected to develop cancer in their lifetimes, could the exposure to the high levels recorded of these toxics result in potentially developing these expected cancers at an earlier age? Also, caner is not the only concern here. There are many chronic illnesses including auto immune and developmental disabilities that can result from exposure to these materials. So far I have not seen any analysis of risk of developing other diseases associated with this uncovered exposure.

    ReplyDelete
    Replies
    1. Joshua - The CalEPA reference exposures are attempts to capture those risks. They may not be perfect, but they are evidence-based thresholds below which the risks of non-cancer disease are small, within the expertise of the epidemiologists who set them.

      Delete
  17. This comment has been removed by a blog administrator.

    ReplyDelete
  18. This comment has been removed by a blog administrator.

    ReplyDelete
  19. This comment has been removed by a blog administrator.

    ReplyDelete