"You were sold a bill of goods" was one of the colorful phrases that I picked up from my late and much lamented mentor when I went from Olde to New England. It means that someone is trying to convince you that something is worth more than the paper (the bill of goods, shipping docket) it's written on, when it's not. I've been meandering through the backlog of Nature now that I have some Summer space to indulge in such luxuries. This can throw up some neat juxtapositions: Nature covers the whole of science and wouldn't dream of boring and bemusing their astrophysics readers with stem cell nonsense week after week. But if the stack of unopened Natures is well shuffled you can read 19th June right after 1st May and the old One-Two makes an impact.
On 19 June 2014, Elena Cattaneo and Gilberto Corbellini were given three pages and colleagues Paolo Bianco and Douglas Sipp were given the following two pages to inveigh against the far-too-early commercialisation of stem cell therapy. The first of these articles reads like a cross between sour grapes and a jihad against a particular company - The Stamina Foundation - which is using stem cells to cure a range of life-threatening conditions. The problem is that the company has proffered no peer-reviewed science to show that their product is efficacious against Parkinson's disease or muscular dystrophy. What Stamina offers instead is anecdotes about particular cases where their injections have made patients better. Q: "What is the singular of data?" A: "Anecdote". The panoply of saints in the Catholic church relies on anecdotes of a few (or even one) miraculous cures rather than data showing that prayer or veneration or a lit candle is significantly better at achieving what is desired than just allowing time to pass. But I'm not here knocking faith, it has currency other than delivering the goods. What Cattaneo et al. object to is that a company is turning a handsome profit in Italy selling desperate and desperately sick people and their relatives hope rather than something that has been shown to work.
The therapy under scrutiny hinges on the idea of stem-cells, either pluripotent or at least multipotent. The process of development from a single fertilised egg cell to 100 trillion cells of 100, or 1000 or 10,000 different types can be viewed as a series of gates which are more or less one way. That original zygote has the capacity to become any of those 100 trillion. But early on, the embryo differentiates into ectoderm (which goes on to make skin and nervous tissue), endoderm (ultimately lining the gut and lungs) and mesoderm (everything else). Once a cell is committed to the endoderm option it will not, cannot, ever become a neuron. Some stem cells (the zygote as an extreme example) can become any cell and they are called pluripotent, other have a more limited (multipotent) set of possible fates. A haematopoietic stem cell, apart from cornering a helluva lot of vowels, is able to become any sort of blood cell but no matter how much it strains, it's not going to become a neuron either. Tricking about with new haematopoietic stem cells is what happens when you do a bone marrow transplant after killing all the existing white blood cells, some of which are cancerous, through radiation and/or chemotherapy. It is the one tried-and-true situation where stem cell therapy has been shown to work . . . more than it screws up the patient with side-effects and new and more immediately fatal diseases.
In the capitalist West we like to believe that the Market is always right, or at least will find its level. If a wheel-barrow doesn't work or breaks after a week's normal use, then nobody (else) will buy it and the producer will go bust or re-jig the product so it is fit-for-function. But governments have long recognised that consumers are often as thick as pig dribble and gullible to boot, and need to be protected from the chicanery of entrepreneurs, shop-keepers and insurance-salesmen. In the arena of medicine this is particularly important because the sick may not be at their swiftest and the downside of a medical intervention may take years to become manifest. Therefore the Irish Medicines Board, the FDA in the USA and the European equivalent require that all medicines are licensed and that they can only be approved if it can be shown that a) they work and b) they do no harm. Of course it's a bit more complex and utilitarian than that - interferon alpha therapy is pretty good, for some people, at damping the symptoms of hepatitis C, and there are side-effects (nausea, headaches) but they are nowhere near as bad as untreated hep C. To be licensed you first have to show that that your novel drug works in animals, then that it has no bad effects (ooops) on a small sample of healthy people (usually young males), then on a small sample of sick people to check that people are essentially the same as lab rats (often not true), then on a larger sample of patients under regular medical scrutiny and check-up. All this takes time, and a failure at any stage means you have to start again. The usually quoted figure is $1 billion to bring an interesting biomedical finding to market. When our noble and selfless BigPharmaCorp is a $billion$ down before the first packet of pills is shipped there is a certain pressure to cut corners consciously or unconsciously. When you look at the data you find that a disturbingly large proportion of biomedical science research is a bill of goods (cannot be repeated and replicated) because of inadequate statistics, inadequate sample size, wishful thinking, or unconscious bias.
Back in the 1st May 2014 edition of Nature there was an editorial and a News in Focus piece reporting the results of a meta-analysis in the British Medical Journal BMJ of all the studies that had investigated stem-cell therapy for heart disease. When you have a heart attack or heart failure, some of the cardiac tissue dies never to be regenerated. A few years ago, some bright spark had the idea that injected mesenchymal stem cells MSCs, derived like their haematopoietic cousins from the bone marrow, would be able to regenerate into heart cells or, failing that, at least stimulate an inflammatory response to encourage the development of new life-giving blood vessels. The logic seemed sensible or at least possible and, as a lot of fat white men with money get heart attacks, it has been widely investigated. The BMJ study, fronted by Prof Darrel Francis of Imperial College London, looked carefully at 49 separate studies. The studies that passed muster under this intense scrutiny and reanalysis of their materials and methods showed that there was no clinical effect. I've previously channelled Trisha Greenhalgh's rule for reading scientific papers: if the M&M are wonk then you don't need to read the rest of the paper because the results, let alone the conclusions, cannot be relied on. The Imperial group found that only methodologically flawed papers showed any positive effect. And the flaws were many and various: tables and figures with conflicting data; inappropriate statistics or statistically impossible results; dead people continuing to report their symptoms; patients coded both as men and women. Now these-all could be mere typos, or have an allowable explanation, or be trivial given the sample size of the study, but it doesn't inspire confidence in the conclusions if nobody has caught these rather obvious errors before it was too late - getting into print is too late. It means that 2 or 3 referees and an editor have nodded it through without doing their job; it means that the Principal Investigator hasn't really scrutinised the work of the team; that they haven't sought effective statistical advice; that they can't add; that they can't use ExCel properly.
One of the perps, a famous cardiologist from the famous Mayo clinic offers the fact that his paper was peer-reviewed as an excuse as if science ran a system of double-jeopardy - if you have gone through the peer-review process once then you and your data are home-free. Another offers "We strongly believe in the science and results we have seen with adult stem cell therapy for coronary artery disease" which may remind you of the Catholic catechism "We firmly believe and confess without reservation that there is only one true God, eternal infinite (immensus) and unchangeable, incomprehensible, almighty and ineffable, the Father and the Son and the Holy Spirit; three persons indeed, but one essence, substance or nature entirely simple."
If someone offers your Dad stem-cell therapy after his heart-attack and you, or your insurer, are going to pay big money for the privilege, Just Say No (thank you). If your government funds a much larger Phase III clinical trial based on these sort of data, tell them you'll be saying No at the next election. If you're a shareholder with BigPharmaCorp sell Now because the company is going to be $billion$ in the hole in five years time . . . unless they massage their data too.
Updated Aug 2015 to correct spelinge: s/Catteneo/Cattaneo/
No comments:
Post a Comment