First things, first, apologies for the lack of posting of late, which is due to a number of ongoing real life issues that I can’t discuss for the time being, but which I may be able to clear up in due course.
Now, having got that out of the way, let’s get down to business… and the business for today is yet another piece of piss poor science/health journalism by the Independent:
Children who live in homes with vinyl flooring have double the chance of being autistic, research has discovered. The finding – which amazed even the scientists conducting the study – provides one of the first clues as to a possible cause of the condition.
Shall we run down the checklist?
Scary sounding and, for preference, incurable/untreatable/terminal (delete as applicable) but poorly understood illness/condition (autism)? Check!
Sympathetic ‘victims’ (childrens)? Check!
Common household item (vinyl flooring)? Check!
Meaningless statement of relative risk derived from a wholly speculative and unverified epidemiological correlation (‘doubles chances’)? Check!
Right, so this is all a load of bollocks? CHECK!
But don’t just take my word for it, let’s take a look at the rest of the article, by Geoffrey Keen and Nina Lakhani, and explain exactly how we arrive at that last conclusion, starting with…
The study, by scientists in Sweden, Denmark and the United States, stumbled across the connection almost by accident.
Well that’s a bad sign for starters.
Yes, science is no less prone to throwing up its fair share of serendipitous accidental discoveries than another field of intellectual endeavour but for every such discovery (electric current, radioactivity, pencillin) there’s are thousands of other unlooked for ‘discoveries’ that rapidly fall by the wayside and into well-deserved obscurity for the simple reasons that further investigation shows them to be complete and utter rubbish – and this is particularly true of apparent epidemiological ‘discoveries’ arising from post hoc analyses derived from data that was originally collated for an entirely different purpose.
This is precisly the situation we have here. The main study, from which the data used in this paper was abstracted, was primarily concerned with an altogether different set of common childhood ailments and conditions:
The present paper describes the results of a study undertaken as part of a larger project devoted to the connection between properties of the indoor environment and asthma and allergy in young Swedish children. The larger project, The Dampness in Buildings and Health (DBH) Study, began in the year 2000 with a questionnaire distributed to parents of all children 1–6 years of age in one Swedish county (DBH-I).
So, we start with an asthma/allergies study and now, nine years down the line, we have an alleged ‘connection’ between autism and vinyl flooring with the researchers getting from A to B as follows:
The original survey collected information about the child, the family situation, practices such as smoking, allergic symptoms, type of residence, moisture-related problems, and type of flooring material, which included polyvinyl chloride (PVC). The 2005 survey, based on the same children, now 6–8 years of age, also asked if, during the intervening period, the child had been diagnosed with Autism, Asperger’s syndrome, or Tourette’s syndrome.
There is a screamingly obvious problem with this particular methodology.
The main study was designed specifically to look for possible causes and contributory factors relating to asthma and allergies and, consequently, would have collated information relating to a range of factors that the researchers thought might be relevant to the study. It also means that there’s a wide range of other factors relating to the families included in the study which were excluded from consideration because they were thought not to be relevant to the study’s main purpose any one (or collection) of which could also be correlated with the prevalence of autism found by the follow-up study AND also correlated with the prevalence of vinyl flooring. That’s one of key reasons why one cannot infer causation from correlation, because an apparent correlation between two factors may, quite easily, be the product of individual correlations between each of these two factors and third, unknown, factor (factor X).
The classic example of a false assumption based on a ‘factor X’ correlation was the miasmatic theory of disease, which was popular from the Middle Ages right through to the 19th Century and which was, at one time, thought to explain why diseases such as cholera and the ‘Black Death’ were endemic in areas where the water supply was unfiltered, untreated and foul-smelling. Ironically, the wholly mistaken belief that cholera was transmitted by ‘bad air’ did lead to improvements in public sanitation that did result in fewer outbreaks of cholera, all of which appeared to support the validity of the miasmatic theory and it was only when John Snow traced the cause of an outbreak of cholera in Soho (in 1854) to the public water pump in Broad Street, that miasmatic theory was disproven.
Despite this very obvious problem with the articles headline assertions, the Indy goes on to add:
It is being taken seriously because autism has long been thought to result from environmental factors.
All of which completely misrepresents the current state of autism research.
As most people are, I think, perfectly aware, we don’t know exactly what cause autism or even if there is a single common cause for the condition – increasingly the prevailing view of the condition is that its an extremely complex disorder in which its core aspects have distinct but co-occuring causes. There is no single cause for autism, rather there are a number of complex, overlapping causes in which environmental factors may play some part, even if it well established that autism has a strong genetic component, one that is equally complex to the point where its not clear whether it occurs as a consequence of complex multi-gene interactions or due to rare mutations with major effects. There is also a growing body of evidence linking the condition to ‘de novo mutations’ (mutations found in the child that are not present in its parents) which suggests that autism is heritable but not inherited, i.e. that what children may actually inherit from their parents is a genetic susceptibility to autism that may become active due to the influence of anything from environmental factors to simple random chance.
So, while there are many environmental factors that have been put forward as possible causes for autism (including certain foods, infectious disease, heavy metals, solvents, diesel exhaust, PCBs, phthalates and phenols used in plastic products, pesticides, brominated flame retardants, alcohol, smoking, illicit drugs, vaccines and prenatal stress) there is no solid evidence of causation for any of these candidates as yet and if de novo mutations are, indeed, part of the causal mechanism, no guarantees that eliminating any of the alleged causal factors will actually have any impact on the prevalence of the condition
Between 133,000 and 200,000 British children are thought to be autistic, but nobody knows for sure, or whether their numbers are increasing, because they are not counted. But the numbers of babies born with the condition in California has risen more than seven times in the past two decades, convincing scientists that pollution must be to blame.
Convincing some scientists but not all, not be a long way and, again, the article woefully misrepresents the findings of a 1999 report by the California Healthy and Human Services Agency to support its already poor line of argument as the National Autistic Society point out in its own, excellent, overview of the existing evidence for the prevalence of autistic spectrum disorders.
At present, it is not possible from the California study to draw any definite conclusions concerning the apparent rise in rates. The author of the California report comes to the same conclusion. The report emphasises the complexity of the problem and the need for properly designed research.
Quite… and please do read the NAS’s article in full, it does provide a very good overview of both the current state of knowledge and understanding of the prevalence of the condition and the difficulties that arise when attempting to calculate accurate incidence/prevalence rates for autism and autistic spectrum disorder.
Skipping quick over the next bit, which give only a bit of basic factual information about the study…
The new research, which traced nearly 5,000 Swedish children from infancy to at least six years old, set out to investigate links between air pollution and asthma and other allergies. The scientists – from Karlstad University in Sweden, the universities of Rochester and Texas in America, and the Technical University of Denmark – identified the type of flooring in each home at the start of the study, but only started to look at autism later.
…the articel goes on to state that:
Their paper, published in the journal Neurotoxicity, describes the findings as “puzzling, even baffling, and not readily explicable at this time”. But it adds: “Because they are among few clues that have emerged about possible environmental contributions to autistic disorders, we believe that they should be weighed carefully and warrant further study”.
Oops – looks like its the sub-editor’s day off. The article doesn’t appear in a journal called Neurotoxicity, in fact you’ll be hard pushed to find such a journal as it doesn’t exist and the actual journal it appears in is called Neurotoxicology, but having tracked down the abstract, what we find is not only that the researchers have been, as always (and as the article notes, belatedly), much more circumspect in their conclusions but that vinyl flooring is only one of five factors where the study found an apparent correlation with the incidence of autism in the study group:
An analysis of the associations between indoor environmental variables in 2000 as well as other background factors and the ASD diagnosis indicated five statistically significant variables: (1) maternal smoking; (2) male sex; (3) economic problems in the family; (4) condensation on windows, a proxy for low ventilation rate in the home; (5) PVC flooring, especially in the parents’ bedroom. In addition, airway symptoms of wheezing and physician-diagnosed asthma in the baseline investigation (2000) were associated with ASD 5 years later.
So, in theory, the article could have been headed up as “Maternal smoking ‘doubles chances of autism’, say study”, or maleness, poverty or even poor ventilation and condensation, but it doesn’t because all this is leading up to a specific point… this:
A possible explanation, they suggest, is that vinyl, or PVC, flooring produces dust full of phthalates, which are then breathed in. Experts who have reviewed the study believe that, if this is the cause, the children may have been most vulnerable when their brains were developing in the womb.
That’s a big if there and all the more so for having tracked down the researchers previous study on phthalates and asthma, which was found to be subject to an important limitation, the data on exposure to phthalates was too imprecise to draw any kind of solid conclusions from the study.
Nevertheless, it seems that phthalates, a plasticising agent used to increase the flexibility of plastics, are set to become the ‘next big thing’ in ‘environmental’ autism scare stories, with this latest paper following closely on heels of an epidemiological study by researchers at UC Davis, which revisited and updates the statistical data used in the 1999 California prevalence study and claims to show a substantial real increase in the prevalence of autism in the state, although its not clear from the abstract alone whether this paper has adequately addressed the methodological issues that the National Autistic Society highlighted in their article (linked earlier). It may well be that the researchers, here, have hit upon an interesting line of future enquiry but, and its a very big but that the Indy have chosen to ignore, its far too early to say whether this will lead anywhere in particular or prove to be yet another blind alley. Other than the two papers already linked in this article I can find only two other research papers on PubMed which reference both phthalates and autism, a 2004 animal study (using rats) which looked at the possible connection to motor hyperactivity in autism and ADHD and a new study, published in January this year, which postulates that a transient thyroxine deficiency in pregnancy might be a causal factor for autism. This last, and seeningly new, hypothesis has yet to be tested but, if the abstract is anything to go by, could well lead, in the fullness of time, to a rich vein of autism scare stories relating to what are, otherwise, very humble and mundane vegetables of the brassica family (cabbage, cauliflower, broccoli, etc.) all of which contain natural chemicals (goitrogens) which suppress the function of the thyroid gland.
But that’s almost certainly part of the appeal, to press, of this kind of story, as the articles final paragraph hints…
If this is so, the threat may come from more than just vinyl flooring. Californian research has also found high levels of the chemicals in wall-to-wall carpeting.
Quite – and if one does nothing more than consult Wikipedia, one also finds that…
Phthalates are used in a large variety of products, from enteric coatings of pharmaceutical pills to viscosity control agents, gelling agents, film formers, stabilizers, dispersants, lubricants, binders, emulsifying agents, and suspending agents. End applications include adhesives and glues, agricultural adjuvants, building materials, personal care products, detergents and surfactants, plastic objects, paints, printing inks and coatings, pharmaceuticals, food products and textiles.
Phthalates are also frequently used in soft plastic fishing lures, nail polish, adhesives, caulk, paint pigments, and sex toys made of so-called “jelly rubber.” Phthalates are used in a variety of household applications (shower curtains, adhesives, perfume), modern electronics and medical applications such as catheters. The most widely-used phthalates are the di-2-ethyl hexyl phthalate (DEHP), the diisodecyl phthalate (DIDP) and the diisononyl phthalate (DINP). DEHP is the dominant plasticizer used in PVC, due to its low cost. Benzylbutylphthalate (BBzP) is used in the manufacture of foamed PVC, which is mostly used as a flooring material. Phthalates with small R and R’ groups are used as solvents in perfumes and pesticides.
Phthalates are pretty much everywhere in any modern work or home setting, all of which makes this story perfect fodder for the jobbing but terminally lazy journalist – its one of those gift stories that just keeps on giving because there’s always going to be another scare story just around the corner… just pick a product off what is an already very long list and run the same basic story over, and over, and over, and over…
Bags I get the ‘Jelly Dildos Cause Autism’ headline.