Coronary Heart Disease: Neither Degenerative Nor Man-Made?

In "On to a Fifth Age? How About We Finish the Second?" we discussed a JAMA editorial wherein Dr. Michael Gaziano asserted we may be entering a fifth age of the so-called epidemiologic transition. These transitions are claimed to be changes in the primary causes of morbidity and mortality and Dr. Gaziano opined that we are moving into an era in which obesity and inactivity will drive preventable illness. We discussed the origin of the idea of epidemiologic transitions and questioned  whether we'd ever finished the second age which would have required the conquest of infectious diseases.

The so-called third age was supposed to be the "age of degenerative and man-made diseases" but it keeps turning out that many illnesses thought to be due to wear and tear, lifestyle or pollutants actually have an infectious disease process at their core. Now there's growing evidence that coronary heart diseases (CHD) may in many cases have more to do with a number of infections, including influenza, than with lifestyle or the environment.

Here's a link to a letter published in the Reflections section of The Lancet: Infectious Diseases that nicely summarizes the pre-1970 thinking that pointed to infections as the cause of CHD, the subsequent predominating narrative of chronic diseases not being caused by infections, and the new evidence that chronic diseases are in fact often caused by previously undetected infectious processes: "Inflammation as the Cause of Coronary Heart Disease". And here's a link to a written debate about "this nascent field associating chronic diseases with infections" from 2002 with the author of the recent Lancet paper cited above: "Debate on the Paper by Maria Ines Reinert Azambuja & Bruce B. Duncan".

Given the enormous renewed interest in infections as a possible cause of chronic illness and the ease with which scientists can now find traces of bacterial, fungal and viral DNA (or RNA) at the scene of the suspected microbial crime it's fair to assume that we'll be seeing many more such stories in the future.

You Know Those Mass Screenings for Prostate Cancer? Nevermind.

According to the Houston Chronicle the American Cancer Society has finally come to grips with mounting evidence that indiscriminate screening for prostate cancer causes more harm than good thanks to (a) the inevitable morbidity resulting from needless biopsies and surgeries due to false positive tests; (b) the realization that an awful lot of people who consider themselves "cancer survivors" would never have known they had cancer but for the screening test as their cancers would have gone away on their own or would have grown so slowly that they'd have died of something else before the prostate cancer became threatening; and, (c) the unfortunate fact that early detection, despite what everybody has been led to believe, does not mean that aggressive cancers can be cured - it just means that we get to be treated for them, and worry about them, longer.

Here's a link to the new screening recommendations: "Revised Prostate Cancer Screening Guidelines: What Has -- and Hasn't -- Changed"

Also of interest may be the readers' comments over at the Chronicle and elsewhere. Predictably there are two dominant camps. One sees this change as a nefarious plot by Big Pharma and Big Medicine to prevent early detection so they can make more money by making people wait until they need more expensive medicines and surgeries. The other one sees the new guidelines as a nefarious plot by Big Government to save money by preventing early detection so it can save money on treatment and hasten the deaths of Americans thereby saving money on Social Security payments as the cherry on top. I've run across veniremen able to hold both views simultaneously. But that's a discussion for another day.

Loves Safranin, Hates People

The New York Times has an alarming article about the rise of gram-negative bacteria as the cause of a staggering amount of morbidity and mortality among hospital patients. The rogues' gallery of gram-negative "wretched beasties" includes Klebsiella pneumoniae, Legionella pneumophila, Pseudomonas aeruginosa, Escherichia coli, Enterobacter cloacae and Helicobacter pylori.

For more on the problem see: "How Can We Stem the Rising Tide of Multidrug-Resistant Gram-Negative Bacilli?" wherein you'll read: "... we know much less about how best to detect and prevent transmission of multidrug-resistent gram-negative bacilli. These organisms are relative newcomers to the healthcare-associated infections arena ..."

Does Estrogen + Progestin Increase the Risk of Lung Cancer?

According to "Lung Cancer and Hormone Replacement Therapy: Association in the Vitamins and Lifestyle Study" published in the Journal of Clinical Oncology, long term use of estrogen and progestin (> 10 yrs) produced a 48% increase in lung cancer and typically advanced disease at that. Interestingly, those taking estrogen alone experienced no increase in risk.

The hazard ratio was only 1.48 with a confidence interval from 1.03 to 2.12 meaning the reported result suggests at best a hypothesis for further testing and not a presumed causal relationship.

Possible Link Between Low Level Benzene Exposure and Birth Defects, Miscarriages

In this article in last month’s Envrionmental Health Perspectives, the authors reported an increased incidence of sperm aneuploidy (abnormal numbers of chromosomes) in men exposed at air levels close to the current 1ppm Permissible Exposure Level (PEL) recognized by EPA. The chromosomes studied have previously been associated with birth defects and miscarriages. This study is the first to associate sperm aneuploidy with levels of bezene exposure below the PEL. It may be pointed to as some indication that the current PEL in the United States insufficiently protective against reproductive harm.

Household Exposures to PAH from Coal Tar Driveway and Parking Lot Sealants

In a paper published last month in Environmental Science and Technology, the authors measured polycyclic aromatic hydrocarbons (PAH) in settled house dust (SHD) in 23 apartments in Austin, TX. Half of the apartments had parking lots sealed with coal tar based parking lot sealcoat and half of them did not. The authors report 530 times more PAH in SHD in the apartments with parking lots sealed with coal tar based sealants than in the other apartments. These findings suggest that coal tar based sealants may play a significant role in household exposures to PAH, notwithstanding the fact that PAH are ubiquitous in urban environments.

The High Cost of Infections Acquired in the Hospital

According to "Clinical and Economic Outcomes Attributable to Health Care-Associated Sepsis and Pneumonia" just published in Archives of Internal Medicine just two types of preventable hospital acquired infections, sepsis and pneumonia, killed an estimated 48,000 Americans in 2006 and added more than $8 billion to the cost of care. The article details the alarming number of infections due to simple lapses in sanitation, from poor hand washing habits to dirty surgical devices, and the spread of so-called "superbugs" which have acquired immunity to common antibiotics.

Another Meta-analysis of Benzene and NHL Studies Finds No Independent Association

In "Benzene Exposure and Non-Hodgkins Lymphoma: a Meta-Analysis of Epidemiologic Studies" published in the Journal of Occupational and Environmental Medicine the authors looked at 8 cohort and 14 case-control studies of benzene-exposed workers and their risk of developing non-Hodgkins lymphoma (NHL). Whether in the aggregate or looking just at those with the highest reported exposures the summary relative risk estimate remained almost exactly one. The authors conclude that these studies at least do not demonstrate that benzene is a risk factor for NHL

Determining Causation After an E. Coli Outbreak

In a case involving lots of infected meat all from the same plant and all infected with the same pathogen cultured from the victims and known (thanks to the deployment of Koch's postulates) to cause the illness in question, well, tracing the source isn't too hard. But that's not how it usually goes.

After an outbreak and after all the food has been tossed and the food areas washed down with Clorox how could you possibly find the source? Thanks to the robust sales data kept by more and more businesses it may be possible to construct a "virtual cohort" and thereafter through the usual epidemiological techniques trace an outbreak all the way back to "cooked beef topside" at a single delicatessen. The method is discussed in "The Use of a New Virtual Cohort Study Design to Investigate an Outbreak of E. coli O157 Linked to a Supermarket Delicatessen".

Six new Stevens-Johnson Syndrome Papers in a Month

Among the conclusions: mycoplasma pneumoniae causes Stevens-Johnson syndrome (SJS) but symptoms were less severe than those cases produced by drugs and biopsy specimens were distinguishable; treatments for vision-related pathologies are available if administered timely; NSAIDs are only weakly associated with SJS and perhaps not at all; the percentage of total body surface affected by SJS/TEN (TEN is toxic epidermal necrolysis) is the best predictor of whether a potentially deadly blood stream infection will occur thus informing and the types of bacteria typically responsible may guide clinicians regarding antibiotic therapy; and, Tamiflu, SJS and you.

Does Hormone Replacement Therapy Prevent Colon Cancer?

A study of 57,000 California teachers found that those on hormone replacement therapy (HRT) had a 36% lower risk of colon cancer than women who never used HRT. The study followed California teachers for a over a decade. The study, "Menopausal Hormone Therapy Use and Risk of Invasive Colon Cancer: The California Teachers Study" is published this month in the American Journal of Epidemiology.

The effect, and by the way a 36% decrease in one of the leading causes of cancer is a pretty big deal, held whether the women were active or not, obese or not, taking NSAIDs or not, consumers of alcohol or not and was especially strong among women with a family history of colorectal cancer. Interestingly, the protective effect disappeared after 15 years.

So what should a perimenopausal/postmenopausal woman do? Is it just a matter of picking her poison? I've no idea. But until someone figures out how cancer really works it looks like we're stuck in a "damned if you do, damned if you don't" dilemma.

Soft Drinks and Pancreatic Cancer: A Plausible Link?

Late last night the press began reporting on a press release from the authors of "Soft Drink and Juice Consumption and Risk of Pancreatic Cancer: The Singapore Chinese Health Study". By this afternoon the claim that drinking as few as two soft drinks a week may nearly double the risk of getting one of those "go get your affairs in order" cancers had appeared on nearly every major news site. So what to make of it?

First it's probably wise to read up on the Singapore Chinese Health Study. How was dose (number of soft drinks per week) determined? It looks as though it was done via questionnaire rather than any objective measurements. And how many cases were there? 140; which isn't many thereby pushing the reported hazard ratio of 1.87 to the "slight" end of the range of the scale used to weigh evidence when making causal inferences.

Second let's look at the other claims on which the authors rely to support their causal hypothesis: specifically, a) Increased sugar intake may stimulate tumor growth through effects of insulin; and b) Pancreatic cancer rates increased nearly twofold over the past several decades.

As we reported recently there is indeed evidence tending to show that chronic moderate to high levels of glucose in the blood are associated with an increased risk of pancreatic cancer. Sounds supportive, right? Only if drinking sweetened soda chronically elevates blood sugar levels. And here's an article that says that it doesn't - not in lean men, not in obese men and not even in type 2 diabetic men; and not even at 75 grams of sucrose per day (seven times the purported risk elevating level referenced in today's article under discussion). So maybe it's like the fat follies of the past few decades; eating fat makes you fat, right? Wrong. Similarly, maybe the association between consuming sugar and hyperglycemia isn't so simple. And isn't that doubt furthered by the finding that sugary fruit drinks were NOT associated with pancreatic cancer?

Finally, here's the SEER Stat Fact Sheet for cancer of the pancreas. As noted there are lots of ways to look at trends but from the SEER perspective the upward trend in pancreatic cancer incidence and mortality is quite recent and looks to be driven by an increase among women. Doesn't that suggest smoking? It is after all the most consistently identified causative factor in cancer of the pancreas.

New Meta-Analysis Finds No Causal Association Between Benzene and NHL

A meta-analysis of 22 studies found no evidence of a link between benzene and non-Hodgkin's lymphoma (NHL) even at the highest levels of benzene exposures. Sliced and diced three different ways the data all revealed a summary relative risk estimate of almost exactly one every single time.

The study is titled "Benzene Exposure and Non-Hodkin Lymphoma: A Meta-Analysis of Epidemiologic Studies" and will be published in February's Journal of Occupational and Environmental Medicine.

Does Trichloroethylene Cause Parkinson's?

"I think people will really move on this as quickly as possible now" said Dr. Samuel Goldman of the Parkinson's Institute in Sunnyvale, CA according to the LATimes. Goldman was interviewed about a soon to be released study indicating that while other industrial solvents like xylene, toluene and n-hexane were not associated with Parkinson's, trichloroethylene (TCE) was strongly and significantly associated with the malady.

The results are to be presented in April at a meeting of the American Academy of Neurology and so there's not much to assess at the moment save the large grain of salt to be taken with any science that makes its way into the popular media before it lands in a journal. Nevertheless, the finding, if borne out by subsequent population studies, would be significant. Many substances have been blamed for Parkinson's but none with the 5.5 relative risk reported in this, well, press release.

Speaking of Saturated Fat

Yesterday we reported on a new paper in JAMA counseling caution in the face of activist efforts to impose universal sodium salt reduction on the American populace. We reminded our readers that once upon a time activists demonized saturated fats and stampeded the country towards greater consumption of trans fats. Now there's a new study looking at the health outcomes of 384,000 Americans to see whether those who consumed lots of saturated fats were more likely to develop heart disease than those who avoided saturated fats.

Reuter's Health reports that the study,published in the American Journal of Clinical Nutrition, involved the combination and assessment of 21 large studies of people, their diets with regard to the consumption of saturated fat and whether or not they developed heart disease. It turns out there's no demonstrable association between saturated fat and your risk of heart disease. Eat a lot or eat a little, your risk remains the same. Somewhere my late great grandmother, who lived to 102 and who made the best pies and pastries thanks in part to lard, is smiling.

Does Chronic Elevated Blood Sugar Cause Cancer?

It very well might, according to "Diabetes Mellitus Type 2 - An Independant Risk Factor for Cancer?" More alarming is the finding that significant risk of cancer is incurred even at blood glucose levels simply on the high side of "normal". A doubling of the risk for the following cancers has been found among those with type 2 diabetes: colorectal, breast, endometrial, renal, liver and pancreatic.

On to a Fifth Age? How About We Finish the Second?

In a 1971 paper that profoundly influenced how scientists and policy makers approached public health issues Abdel Omran set out his theory of "The Epidemiologic Transition". He hypothesized that societies went through three different ages, or phases, that defined their experience with regard to mortality and life expectancy. In the first, the "age of pestilence and famine", life expectancy is low and episodes of widespread death are common. In the second, the "age of receding pandemics", infectious diseases are overcome and life expectancy increases dramatically. Finally, in the third, the "age of degenerative and man-made diseases", diseases of aging and self-inflicted suffering becomes the predominant determinant of mortality. Eventually others, noting the dramatic increase in life expectancy due to the rapid decline in deaths due to heart attack and stroke, posited a fourth age; essentially the same as the original third age but with cardiovascular disease removed from the "degenerative disease" category.

Now in an editorial in this month's JAMA  Dr. Michael Gaziano asserts that we may be entering a fifth phase, or age, of the epidemiologic transition. We are now, he writes, entering the "age of obesity and inactivity" in which ailments due to gluttony and sloth predominate on death certificates. The editorial references two new articles in the same issue purporting to show Americans are fat and getting fatter; especially the children.

But wait a minute. The age of man-made diseases barely materialized. Certainly there have been many many cases of people suffering terribly as a result of some man-made health hazard. Look no further than the cases of mesothelioma among the men who served aboard amosite laden Navy ships. And smoking continues to exact its terrible toll. Yet if you throw all the deaths due to occupational diseases and every last lung cancer/COPD death into the same category you can't get to 10% using worst case estimates. More sober estimates put the percentage of deaths due to man-made diseases at considerably less than one. Nevertheless, this powerful meme - that most of our woes are self-inflicted and due to some failure to live in a natural way - still propels not only mass tort litigation but also much scientific and political thinking.

However, there's more than just AIDS to demonstrate that we never really saw the "disappearance" of infectious diseases. Go to www.pubmed.gov/ and do some searches on helicobacter pylori and humanpapilloma virus and you'll see just how many cancers are now being attributed to just these two organisms. Investigate mollicutes and you'll find that all sorts of microbes are suddenly being found associated with disease and they're only now being found because the technology to identify them is only now being refined.

Finally, remember to read the fascinating journey of Barry Marshall and Robin Warren from authors of an abstract rejected as one of the year's worst to winners of the Nobel Prize in Medicine for the very same work. In the end, the view, supported by the work of one of the world's preeminent public health researchers, that peptic ulcers were caused by that most modern of man-made insults, stress, only gave way to the understanding that the cause was in fact a bacteria when the evidence was irrefutable.

A Critique: Recent Epi Studies of Motor Skills and Manganese

In "Risk Assessment of an Essential Element: Manganese" Annette Santamaria and Sandra Sulsky of ENVIRON critically review recent epidemiological literature associating a variety of abnormal psychometrics with relatively low levels of manganese exposure.  

The authors conclude that the available epidemiological data is generally flawed and unreliable at least for the purpose of doing risk assessment. Furthermore, they demonstrate that some exposure levels claimed to pose a risk of neurobehavioral injury produce effective doses well below the amount of manganese recommended in a healthy diet; they also elaborate on the adverse health effects of manganese deficiency. Santamaria and Sulsky conclude by suggesting that more accurate and defensible risk assessments for manganese will have to come from objective data such as the determination of manganese dose via inhalation and the subsequent development of physiologically based pharmacokinetic models to predict the consequences of exposure at various levels and by various routes.

An Explanation for the Quebec/South Carolina Chrysotile Conundrum?

Why is there such a disparity between health outcomes of the workers who mined chrysotile and the textile workers who used it? In "Comparing Milled Fiber, Quebec Ore, and Textile Factory Dust: Has Another Piece of the Asbestos Puzzle Fallen into Place?" Berman reports that the dusts encountered at the facilities vary in two important ways. First, whether assessed by PCM or TEM, samples revealed that mine and mill dusts contain only 67% asbestos. On the other hand, typical feedstock for the textile facility was 100% asbestos. Second, the feedstock asbestos fibers were significantly longer than those typically encountered by workers exposed to mined or even refined asbestos.

Thus, for a given quantity of dust textile workers would have an approximately 50% greater exposure to asbestos and that exposure, on a fiber per fiber basis, would be significantly riskier. The author also suggests a method for reconstructing past exposures which would more accurately estimate the risk posed by each of those exposures.

Does Education Cause Autism?

If a strong and consistent association between autism and a single chemical or vaccine were found in ten separate clusters around California you'd expect to read about it; and you'd expect the researchers to infer that the chemical or vaccine was, in fact, the cause of autism. But what if the only strong and consistent association wasn't between autism and exposure to some substance but was instead between autism and something complex, like "high parental education"? Well, you'd probably expect a lot of frustration; and you'd be pretty sure autism wouldn't be laid at the feet of higher education.

So what should we make of "Geographic Distribution of Autism in California: a Retrospective Birth Cohort Analysis" ? Why would the children of parents with college degrees be at a 400% increased risk of autism? Why aren't chemicals to blame?

The authors concede in interviews with Scientific American that the results tend to undermine claims that industrial pollution causes autism. Yet rather than explore the possible reasons why college-educated parents might be significantly more likely to have autistic children the authors speculate that their results may be biased. Perhaps, they suggest, the less educated aren't educated enough to take their autistic children in to be diagnosed. Or maybe, they wondered, the educated are much more likely to use some unknown household chemical which does, in fact, cause autism. Scientific American hastens to add that there's an unpublished study that shows a doubling of the risk of autism in mothers who use pet flea shampoos.

But why point to a much smaller effect allegedly to be found in an unpublished paper when a much larger effect published in a peer reviewed journal is in front of you?

First, people tend to prefer simple answers with readily implemented solutions, the blame for which, and the costs of which, get imposed on someone else. 

Second, epidemiology, at least as practised in the West, tends to be reductionist. The triumph of linking tobacco to smoking led many epidemiologists to believe, or at least to want to believe, that indeed there are simple solutions to big problems and that silver bullets are out there just waiting to be found.

Finally, it's becoming increasingly clear that all sorts of ailments, from cancer to obesity, are highly complex and result in large part, from lifestyle, cultural and economic choices. Take breast cancer for example. One of the very best ways to reduce the risk of breast cancer is for a woman to bear children and to begin doing so in her late teens or early 20s. Yet no one wants to say that putting off children to get a degree and start a career is a cause of breast cancer.

And so, for now, the search is still on for the usual suspects.

 

 

Are Michigan Courts Shifting the Burden of Proof to Defendants?

One explanation for Genna v. Jackson (a new opinion out of the Michigan Court of Appeals) and Gass v. Marriott Hotel Services, Inc is that the courts thought about it and decided to shift the burden of proof regarding causation to the defendant any time a plaintiff suffers an injury that is consistent with the possible harm, as set out on a warning label or MSDS, associated with a potentially toxic material. In Genna the court reversed summary judgment in favor of the defendant in part because "[h]ere, like in Gass, defendant has not submitted any scientific evidence that the mold in her condominium could not have cause [sic] plaintiffs' injuries." In Gass the 6th Circuit noted "Defendants have offered no evidence to refute the MSDS's representation of Demand CS as a chemical which could have caused Plaintiffs' symptoms". The court later continued saying "[p]laintiffs are not required to produce expert testimony on causation where Defendants have failed to offer scientific evidence regarding the effects of Demand CS or Suspend SC."

The Genna court concluded, on the issue of causation, "[t]his is not a complicated case: the children were sick, the children were removed from the home, the mold was discovered, and the children recovered." Noting evidence that there was lots of mold and that mold in general had been reported to cause some of "the types of symptoms suffered by the children" the court concluded [i]t does not take an expert to conclude that, under these circumstances, defendant more likely than not is responsible for plaintiffs' injuries." (citing Gass). "Here, there was ample circumstantial evidence that would "facilitate reasonable inferences of causation, not mere speculation."

Under this view of an unannounced burden-shifting approach taken by these two courts it appears that in Michigan if a plaintiff develops ailments consistent with a those listed on a product's label or MSDS the jury is free to infer causation even in the absence of evidence of what dose produces those ailments and what dose plaintiff suffered and even in the absence of an expert to opine that the product in question caused the harm at issue.

Another explanation for these cases is that Michigan does not require toxic tort plaintiffs to show what some call specific causation - a concept common in cases in which causal inferences are derived from epidemiological evidence. The idea is that when dealing with an illness that has been associated with multiple causes, among which is the chemical at issue, the plaintiff must show that the putative cause was more likely than not the cause in fact of his illness.  Evidence for this view can be found in the following passage from Genna: "Defendant urges this Court to adopt the requirement that, in order to prove causation in a toxic tort case, a plaintiff must show both that the alleged toxin is capable of causing injuries like those suffered by the plaintiff in human beings subjected to the same exposure as the plaintiff, and that the toxin was the cause of the plaintiff's injury. They urge this Court to find that direct expert testimony be required to establish the causal link, not inferences. We decline to adopt this requirement. There is no published Michigan case law on the subject."

A final explanation is that more attention to Aristotle's Rhetoric is needed. As Chief Judge Boggs, dissenting in Gass wrote the problem here is the logical fallacy behind post hoc, ergo propter hoc causal inferences such as those suggested by the plaintiffs in Genna and Gass. "It is the fallacy of saying that because effect A happened at some point after alleged cause B, the alleged cause was the actual cause. Such logic has never been enough to survive summary judgment. See, e.g., Abbott v. Federal Forge, ("[P]ost hoc, ergo propter hoc is not a rule of legal causation.") 912 F.2d 867, 875 (6th Cir.1990).

Equally importantly Chief Judge Boggs understands that lay juries are in no position to judge whether there has been a breach of the duty of care without evidence of what levels are harmful and the levels to which plaintiffs were actually exposed. If you don't know the dose you can't know what risk, if any, the defendant imposed on the plaintiff. He wrote: "Thus presented, the question is whether the plaintiffs needed expert testimony in this case to prove how much chemical exposure is too much chemical exposure or to prove whether the amount of exposure actually caused the alleged harmful consequence. In my view, the majority pays too little attention to this issue, rushing from the fact of exposure and odd symptoms to the legal conclusion of fault." He continued: "As I understand it, these cases require expert testimony in complex, professional, or scientific-based negligence cases in order to limit the dangers associated with indulging the post hoc impulse: it is too easy to charge an uncommon harm to the presence of a mysterious substance. Properly credentialed expert testimony operates as a bulwark against such fallacious attribution of guilt. As in the Daubert context, our concern in applying these cases should be to "assure that the powerful engine of tort liability ... points towards the right substances and does not destroy the wrong ones." General Electric v. Joiner, 522 U.S. 136, 148-49, 118 S.Ct. 512, 139 L.Ed.2d 508 (1997) (Breyer, J. concurring).

Whatever the explanation these cases will be a boon for Michigan toxic tort plaintiffs.

Don't Pitch the Water Softener

Have you been worrying that your water softener is significantly increasing your risk of dying from a heart attack? I didn't think so. But just because you haven't been feeling vulnerable around your water softener doesn't mean the WHO hasn't been fretting for you.

Thanks to epidemiological studies going back a decade or more (e.g. "Magnesium and Calcium in Drinking Water and Death from Acute Myocardial Infarction in Women") a worry arose that we were killing ourselves by eliminating the minerals naturally found in most drinking water. Yet subsequent studies have failed to confirm the finding including the just published "Effect of water hardness on cardiovascular mortality: an ecological time series approach". So what gives?

Well, what gives is that most of what gets published in peer reviewed journals is probably false; and when it comes to causal inferences drawn from epidemiological studies "the apparently indiscriminate indentification of particular aspects of daily life as dangerous to health" is, as witty programmers say, a feature, not a bug.

 

Is Your Drinking Water Safe? If Not, Why Not?

The New York Times has run an extensive article about the nation's drinking water claiming that our public water is contaminated with thousands of chemicals, hundreds of which are "associated with a risk of cancer". It even has a link to the articles on which the claims are based. Unfortunately, or fortunately depending on your point of view, the evidence cited for the proposition that tap water is putting the citizenry at risk of cancer is pretty thin if that's all there is.

For example, under "Studies Regarding Illnesses and Drinking Water" (of which there are only eight) the only one to make a broad claim of tap water carcinogenesis is a 28 year old study titled "Cancer and Drinking Water in Louisiana: Colon and Rectum" which used the 1970 census to compare 692 rectal cancer deaths from 1969-1975 by where along the Mississippi River they and controls got their drinking water. The authors noted a small increase in risk as the source got closer to the Gulf of Mexico and suggested that the finding may have something to do with an increasing concentration of industries along the river as it approaches the Gulf. More importantly they wondered whether by-products of chlorination might have something to do with the finding.

The link suggests just four other papers that have cited the study and of those only two studied drinking water. Of the two, the first is a Canadian paper from 2000 (the first true Y2K victim I've ever run across - note: "Received 1999 Accepted 1900") which found no association with rectal cancer but a small one for colon cancer among males who drank "chlorinated surface water for 35-40 years".

Following the papers that cited the Canadian paper you quickly find another drinking water paper that finds a small protective effect for all leukemias combined, a large protective effect for chronic lymphocytic leukemia and a small but significant association with chronic myeloid leukemia.

The other papers are similarly all over the place and there appears to be no consensus that U.S. drinking water is a cause for concern about cancer from the perspective of chemical contaminants.

On the other hand, there's a growing body of literature associating drinking water contaminated by microorganisms with cancer. There's a new one discussing the waterborne transmission of helicobacter pylori to be published in next month's Journal of Water and Health and then there's this alarmingly titled paper in the same journal: "Free-living amoebae, Legionella and Mycobacterium in tap water supplied by a municipal drinking water utility in the USA

It's unclear why the NYTimes focused on trace levels of chemicals as a cause for concern when there does appear to be something to be worried about when it comes to bugs in our water.

Cancer: Annual Report to the Nation

There was more good news last week about Americans' incidence and rates of death due to cancer. The rates of new cases and the rates of death from all cancers combined are dropping and dropping steadily. For example, cancer deaths for all forms of cancer collectively dropped 1.6% per year for each of the last six years for which data is available (2001 - 2006). The progress being made against colorectal cancer was highlighted in the report.

While men continue to have higher rates than women their decline in rates was also higher. Decreases in deaths and incidence for lung, prostate and colorectal cancer drove the decline for men.

Among women, breast and colorectal rates declines lead the overall improvement.

The news was not uniformly good. In men, rates increased for myeloma, kidney, liver and esophageal cancer as well as multiple myeloma and leukemia. In women the expected increase in lung cancer rates was found along with increases for thyroid, pancreatic, bladder, melanoma and kidney cancers as well as for non-Hodgkin's lymphoma and leukemia.

The NIH press release can be found here.

IARC Assesses Evidence for Carcinogenicity and Genotoxicity of Formaldehyde

IARC has determined that there is strong evidence that formaldehyde is the cause of nasopharyngeal cancer and moderate evidence that it causes leukemia, especially of the myeloid variety. Here's the ungated table from "A review of human carcinogens - Part F: Chemical agents and related occupations".

Formaldehyde: A Cause of Leukemia?

In a new article in the Journal of the National Cancer Institute an association between embalming, as a proxy for exposure to formaldehyde, and myeloid leukemia was found to be strong, significant and dose dependent. In coming days we'll be discussing the controversial claim of a causal link between leukemia and formaldehyde and IARC's recent musings about it.

The article is Mortality From Lymphohematopoietic Malignancies and Brain Cancer Among Embalmers Exposed to Formaldehyde .

Interstitial Fibrosis Among Farmworkers Due to Agricultural Dust

A series of lungs from California farm workers who died in accidents or from illness were examined for evidence of pneumoconiosis.   Despite being young these Hispanic farm workers had significantly more evidence of interstitial fibrosis than did non-farm workers.  A variety of analytical techniques demonstrated that those with pneumoconiosis had significant exposure to crystalline silica and aluminum silicate.

The article is "Pneumoconiosis from agricultural dust exposure among young California farmworkers".  As for the source of mineral dust exposure the authors note "most agricultural soils are composed largely of silicate materials (e.g., feldspars, mica, clay minerals) and crystalline silica (CSi) (quartz)".

Are Big Punitive Awards in HRT Cases Justified?

Law.com is reporting that Philadelphia juries have awarded a total of $103 million in punitive damages alone to two women in separate breast cancer product liability trials. The women claimed that hormone replacement therapy (HRT) was responsible for their subsequent development of breast cancer.

In light of the recent controversy over the use of Bayesian decision-making approaches to mammography and Pap testing in which probabilities of outcomes are estimated and benefits are then weighed against costs (including other bad outcomes) I thought it might be of interest to see if such an approach had been applied to HRT. Sure enough, "Bayesian Meta-analysis of Hormone Therapy and Mortality in Younger Postmenopausal Women" was just published in The American Journal of Medicine.

So what does it show? It shows that across a number of randomized controlled trials of HRT in postmenopausal women under 60 those women had a reduced overall mortality compared to those postmenopausal women under 60 who weren't on HRT.

As is often the case in these modern times science does not yield a cure but does allow one to pick one's poison as it were; not to avoid death but to influence the odds of whether you die of stroke instead of breast cancer.

O-toluidine and Aniline Again Associated with Bladder Cancer

A recently published study in Occupational and Environmental Medicine reanalyzed the results of a 1991 investigation conducted by NIOSH of 1749 workers potentially exposed to o-toluidine and aniline at a chemical manufacturing plant and the incidence of bladder cancer. In the reanalysis, the investigators used updated exposure categories based on additional information ascertained through plant visits, employee interviews, documents, and answers provided by company and union officials on specific questions. Increased risks of bladder cancer were observed the longer the exposed employee worked at the facility and the longer the latency period from the date the employee initially worked in a department in which exposure occurred.

The investigators concluded that the findings were comparable to the results reported in 1991 by NIOSH that workers at the plant have an increased risk of bladder cancer.
 

Sulfonamides and Nitrofurantoin Associated with Birth Defects

The New York Times is reporting on this study by Crider KS et al. in November’s Archives of Pediatrics and Adolescent Medicine which appears to demonstrate an association between certain antibiotics taken for urinary tract infections (used just prior to pregnancy or during the first trimester of gestation) and several common birth defects.

The antibiotics associated with the most pronounced increases in birth defects were  nitrofurantoin and the sulfonamides. Caution is in order when interpreting these results. First, it appears that the exposures were extrapolated from memory rather than measured quantitatively. Questionnaire epidemiology is notoriously prone to error. Second, there’s the question of where does the arrow of causal suspicion point. Maybe it’s the infection that’s responsible for the birth defects. Finally, this particular cohort would not appear to be a cross section of pregnant American women but instead is drawn from a narrower slice and is significantly homogeneous. Social structures and interactions are increasingly recognized as important independent causes of certain morbidities.
 

Good News on Styrene

A review of the studies of workers exposed to styrene found no evidence of a causal link between styrene exposure and cancer. The literature contains numerous studies of workers with exposures to styrene during its manufacture or use in synthetic rubber or plastics production. Though a couple of studies have found an increase in either non-Hodgkins lymphoma or esophageal cancer many others did not and no dose response was seen in the outliers. The authors conclude that the available epidemiologic literature does not support a causal inference for styrene and any form of cancer.

The paper, published in the October issue of Journal of Occupational and Environmental Medicine is titled “Epidemiologic Studies of Styrene and Cancer: A Review of the Literature” and was authored by Boffetta P, Adami HO, Cole P, Trichopoulos D and Mandel JS.

Mortality Among 127,266 Petroleum Company Employees

The Journal of Occupational and Environmental Medicine is publishing in its October issue the results of an epidemiological study of 127,266 men employed in the U.S. by a major petroleum company with at least one day of employment from 1979 through 2000. Standardized mortality ratios were calculated for 94 causes of death. Overall these workers had risks below, and often well below, those of typical American males of the same age. However, the mesothelioma rate was triple that expected (CI=1.15 – 1.90) but only for men hired before 1960. Overall, the rate for acute non-lymphocytic leukemia (including AML) was essentially as expected though a subgroup in the chemicals portion of the business did show a statistically significant increase. Interestingly, there was a statistically significant increase in mortality from melanoma among truck drivers.

As is typical of refinery worker studies, these employees fared quite well as a group. Whether that’s because, as a plaintiff attorney friend of mine speculates, the oil companies are especially good at spotting people who’ll be less likely to die of say stomach cancer or whether it’s because these workers tend to be better paid, have better access to health care and so tend to live healthier lives remains to be seen.

The article is titled “Mortality Patterns and Trends Among 127,266 U.S.-Based Men in a Petroleum Company: Update 1979-2000” and was authored by Huebner WW, Wojcik NC, Jorgensen G, Marcella SP and Nicolich MJ.

Trichloroethylene: A Risk Factor for Cancer?

US EPA has been working on a risk assessment of trichloroethylene (TCE) for some time now. Here’s a link to the EPA Issue Papers through 2005. Now a comprehensive review of the issues has been published in Critical Reviews in Toxicology. The article is entitled “Trichloroethylene risk assessment: A review and commentary” and it provides an excellent overview of the developing molecular biological and molecular epidemiological approach to causal attribution and risk; one we’re sure to see increasingly in asbestos, benzene and other mass tort litigation.

If a Lead, Must It Bleed?

Over at Mind Hacks there’s a fun discussion about the media’s propensity to take epidemiological data regarding say the risk of some dread disease, to note the uncertainty inherent in such data and to then conclude that the risk could actually be much higher. Why not also note that the uncertainty also means the risk could be much lower? Is it just a matter of hyping a story to sell more newspapers?

I think it may be something else. I think it may have more to do with the product of the news industry – a narrative that customers want to buy. Accordingly, when the news seems not to fit with the readers’ worldviews (e.g. a report about the disparate outcomes between children in daycare versus those who stay home with their mothers) you can count on the reporter to note that the surprising result may be overstated. Thus, it may not be a matter of reporters not understanding such data or of wanting to fan the flames of a story but rather to pound the day’s news into a narrative to the liking of its readership.

Next week in Munich, Germany - The 2009 Benzene Symposium

Here's a link to the program and abstract book: http://www.tum-benzenesymposium.de/AbstractBook_FINAL_082509.pdf

Given the current state of benzene litigation among the most interesting presentations will be those pertaining to (1) latency (whether attributable cases drop dramatically after 10 years or so); (2) disease endpoints (just AML, just certain subtypes of AML, or NHL, MM and even ALL); (3) molecular biology/epidemiology (biomarkers and pathways) for assessing past exposures; and, (4) risk assessment.

Details to follow.

Diagnosing Mesothelioma

Distinguishing a mesothelioma from a lung adenocarcinoma is critical in asbestos malignancies. Over time, diagnoses made on the basis of morphology and the presence of asbestos bodies gave way to immunohistochemistry. Immunohistochemical staining panels keep changing and debates often rage over whether say a positive calretinen, etc. and negative CEA, etc. is enough or whether they merely show for example an adenoma of mesothelial origin. Now there's a new paper out discussing the use of epigenetic (your genes aren't so deterministic after all) profiles. Its title is "Differentiation of lung adenocarcinoma, pleural mesothelioma, and nonmalignant pulmonary tissues using DNA methylation profiles" and you can buy a copy of it there.

Of course all these new methodologies raise the obvious question of "What happens when you try to draw causal inferences about a case diagnosed today from epidemiological research conducted at a time long before these diagnostic techniques existed?" Will these then be, at least with regard to the litigation, distinctions without a difference?