Richard Lehman’s journal review—12 November 2012

Richard LehmanJAMA  7 Nov 2012  Vol 308
1747   Big trials: don’t you love ‘em? James Penston doesn’t, arguing in his book stats.con (2010) that we have been duped into adopting interventions with small but statistically significant effect sizes that merely prove that most people receiving the treatments derive no benefit from them. And this is true: but I believe that just goes to show that people should be properly informed and make their own choice. Moreover it will soon be possible to do randomized trials at the switch of a button in general practice. As this Viewpoint states, “With a learning healthcare delivery system in place, when effect sizes for an intervention are less than an odds ratio or relative risk of 1.6 (in the range of most effective therapies), randomization could occur as an intrinsic element of an intervention at the individual or neighborhood scale, increasing the ability to generate high-level evidence by an order of magnitude at a much lower cost. That dream, described as an aspiration in the 2002 NIH Roadmap, is now within reach.”

1751   Vitamins are vital amines; you need them to stay healthy; therefore the more you take, the healthier you will be. It’s amazing how pervasive this delusion is: it led even Linus Pauling, two times Nobel prizewinner, to make an ass of himself. And it sells billions of dollars’ worth of useless tablets. Did I really see a two-storey emporium called Ye Olde Vitamin Shoppe in New York City, or had someone slipped some absinthe into my Starbucks? Anyway, you are a doctor, so you should know better than to take this rubbish: here is the Physicians Health Study II. “Among this population of US male physicians, taking a daily multivitamin did not reduce major cardiovascular events, MI, stroke, and CVD mortality after more than a decade of treatment and follow-up.” Antioxidants, antischmocksidants.

Of course there are sound mechanistic reasons why antioxidants should prevent cardiovascular disease; it is entirely perverse that that they don’t. And there are sound mechanistic reasons why prasugrel should lead to better outcomes than clopidogrel when used with aspirin to prevent ischaemic outcomes following ACS without ST-elevation. In the TRILOGY ACS trial, patients who had not undergone revascularization had their platelet reactivity measured, and sure enough prasugrel was uniformly more effective than clopidogrel at inactivating platelets. But O perversity! “Among those in the platelet substudy, no significant differences existed between prasugrel vs clopidogrel in the occurrence of the primary efficacy end point through 30 months and no significant association existed between platelet reactivity and occurrence of ischemic outcomes.” Sound mechanistic reasons count for nothing in real life, yet again.

NEJM  8 Nov 2012  Vol 367
1783   Practically every week I report to you about a cancer trial and practically every time this reports an advantage of about 3-4 months of progression-free survival in exchange for pretty awful side effects and unknown (but presumably massive) cost. Mostly these are industry-funded trials in patients with terminal disease. I guess the logic of these trials is not just that they can provide extra life for patients nearing the end, but that they may provide mechanistic insights into what could work in earlier disease and perhaps even be curative. Siddhartha Mukerjee’s sprawling account of progress in cancer, The Emperor of All Maladies, contains some instances where this has happened. Here is a trial of an antibody-drug conjugate which goes some way to prove that this approach can work, and that trastuzumab emtansine can usefully extend life in patients whose disseminated HER-2 positive breast cancer has already escaped control with trastuzumab and a taxane. And in fact the side effects of this new compound, given intravenously every 21 days, seem to be fewer than those from combined treatment with oral lapatinib plus capecitabine.

1792   “Statin use in patients with cancer is associated with reduced cancer-related mortality. This suggests a need for trials of statins in patients with cancer.” Denmark stars again: the investigators assessed mortality among patients from the entire Danish population who had received a diagnosis of cancer between 1995 and 2007, with follow-up until December 31, 2009. They found a 15% reduction in overall mortality in people with 13 types of cancer taking statins of any kind and any dose, driven by a reduction in cancer-related mortality. So their conclusion is absolutely spot-on: we need trials of statins in cancer.

1803   Interventional trials which are not conducted by industry are such a rarity that it is always interesting to look how they differ from the usual marketing-oriented study design. This trial very usefully compared an anticholinergic drug (solifenacin) with a single intradetrusor injection of onabotulinumtoxin A in women with moderately severe urge incontinence. “The group receiving onabotulinumtoxin A was less likely to have dry mouth and more likely to have complete resolution of urgency urinary incontinence but had higher rates of transient urinary retention and urinary tract infections.” So this trial—which was truly double-blinded, with dummy injections as well as dummy pills—gives us all the information needed for women to choose between these treatments. Bravo. This is comparative effectiveness research at its best. Compared with most industry-funded trials, the design was both rigorous and parsimonious, involving ten centres and 249 women. The Patient Centered Outcomes Research Institute (PCORI)—given a new lease of life by Obama’s re-election—should support many more such studies on condition that full anonymized individual patient data will be released on completion. That way doctors might finally be able to learn which treatments work best for which patients.

1821   I do not like pay for performance. If you dangle carrots in front of doctors, you risk donkey-like behaviour. If you succeed in increasing some kind of clinical activity, it is likely to be at the expense of some equally worthwhile but less measurable activity, and is certain to harm the whole ethos of professionalism and patient-centred care. That may just be an old-fogeyish view, but it is broadly supported by a recent King’s Fund report on incentives in the National Health Service. This paper describes Reduced Mortality with Hospital Pay for Performance in England, comparing the mortality rates for patients admitted with pneumonia, heart failure, and myocardial infarction in hospitals which were given incentives compared with historical rates in the same hospitals, and with rates in other English hospitals where incentives were not introduced. There as a small but statistically significant improvement in pneumonia outcomes which drove a 1.3% drop in overall mortality for the three conditions. Hmm. The authors claim that this is “clinically” significant too, but I beg to differ.

Lancet  10 Nov 2012  Vol 380
1649   The Chronic Kidney Disease Prognosis Consortium is given two market stalls in this week’s Lancet, and guess what—the consortium uses them to sell the importance of kidney function as a prognostic marker. As readers will know, I take a dim view of prognostic markers unless they are (a) clearly an improvement on the markers we already use and (b) can be placed in a useful clinical context. And I take a dim view of “chronic kidney disease” mongering because I don’t see what we can do about CKD that we aren’t already doing for other reasons. The first paper trawls through data from more than a million subjects to determine the prognostic importance of decreased renal clearance with or without high blood pressure. “Chronic kidney disease should be regarded as at least an equally relevant risk factor for mortality and ESRD in individuals without hypertension as it is in those with hypertension. Funding: US National Kidney Foundation.” OK: so what should we be doing about it? The evidence for all interventions directed at CKD is weak, and observational studies seem to show that the sooner you start dialysis, the sooner your patients die.

1662   The other million-subject data trawl seeks to discover whether CKD outcomes differ with and without diabetes: “Despite higher risks for mortality and ESRD in diabetes, the relative risks of these outcomes by eGFR and ACR are much the same irrespective of the presence or absence of diabetes, emphasising the importance of kidney disease as a predictor of clinical outcomes. Funding: US National Kidney Foundation.” Same questions.

1684   At least we know roughly what to do for acute community-acquired bacterial meningitis but getting to the diagnosis quickly and accurately can be a challenge. This is discussed at great length by four academic authors in this clinical review who use terms like “hypoglycorrhacia” (meaning low CSF sugar) and examine in detail the evidence for every clinical and CSF finding. They should ask themselves what they are trying to achieve: clinicians worried about meningitis don’t have time to look up obscure terms in large medical dictionaries, and they need quick accurate rules of thumb (or “heuristics” if you prefer Greek), not complex prediction tools. But maybe this is not what Lancet reviews are about. The same four authors make an equal meal of describing advances in the treatment of meningitis in the paper that follows.

BMJ  10 Nov 2012  Vol 345
The World Health Organization has stated that glycated haemoglobin (HbA1c) can now be used to diagnose diabetes and this very useful editorial by Andrew Farmer explains why and also how HbA1c levels can mislead in some clinical situations, usefully summarized in a box.

A large Dutch study shows that women who have been treated for cervical intraepithelial neoplasia  with three subsequent normal cervical smears are still at four times the risk of cervical cancer compared to women whose smears have all been normal. To make sense of this finding in the context of a national screening programme, turn to the editorial: the message seems to be (and who are we to disagree?) that combined cytology and HPV testing at six and 24 months should be sufficient.

The most interesting article in this week’s BMJ is a discussion paper in which “Peter Gøtzsche and John Ioannidis argue that it is not always sensible to include subject experts as authors of systematic reviews and meta-analyses.” I’m not sure “sensible” is the right word there: maybe “appropriate” would be better. What these two excellent gentlemen argue is that subject experts—especially older ones—tend to be wedded to the prevailing view of a disease area and what constitutes appropriate management. It often takes an outsider to realise the weakness of the current model and to look at what the evidence actually shows. Selecting studies for systematic reviews can be a highly subjective process, as I have noted time and again especially in the fields of heart failure and diabetes.
“I’m a specialist and famous—
My name is Sir Trevor Dense.
If your paper doesn’t name us
Then it doesn’t count as evidence.”

Ann Intern Med  6 Nov 2012 
601   After the fall of the Roman Empire, the developed world entered centuries of intellectual darkness marked by minimal scientific progress, a period often called the “Dark Ages.” After many centuries, progress resumed and eventually accelerated during the Renaissance. In a similar fashion, knowledge about the comparative effectiveness of drugs to treat type 2 diabetes is finally beginning to emerge from 40 years of stagnation. This period of darkness and the current reawakening provide critically important lessons for contemporary medicine about the use of surrogate end points in drug development, regulatory oversight, and the hazards associated with reliance on commercial funding for pivotal clinical trials. OK, that’s the standard Lehman rant about diabetes trials. But it wasn’t written by me: there should be quotation marks around those opening sentences, which were actually written by Steve Nissen. That’s slightly ironic, given how much commercial funding he has received for conducting pivotal trials: but his editorial is still well worth reading. He is commenting on a large cohort study which compares the effects of sulfonylureas and metformin on CVD outcomes (acute myocardial infarction and stroke) or death. It’s nothing so ground-breaking as a randomized trial, and it simply confirms that use of sulfonylureas compared with metformin for initial treatment of diabetes was associated with an increased hazard of CVD events or death, of about 20%. It doesn’t tell us if metformin is beneficial, or sulfonylureas are harmful, or a bit of both. And these drug classes appeared half a century ago…

632   Physiotherapy is greatly valued by many patients because it is the only hands-on therapy available through the NHS; it is greatly valued by doctors because it buys time to see whether musculoskeletal conditions are going to get better, stay the same, or get worse. In most conditions, there is not just absence of evidence for the benefit of physiotherapy, but evidence of absence. This systematic review shows that this may be true for osteoarthritis of the knee, but at the same time it says that better trials are needed.

Fungus of the Week: Armillaria mellea

I am driven to write again about this dreadful parasite because its fruit bodies have just appeared in abundance on a tree stump just below one of our most treasured showpiece plants, Cornus controversa “Argentea.” This probably spells the doom of this beautiful small tree, which we have planted high on a bank where all the village can enjoy its lovely tiered branches and foliage. So far on the same bank we have lost two fine examples of Wisteria venusta (aka W brachybotris “Shiro-kapitan”) and an irreplaceable cream-flowered magnolia called “Yellow River.”

This shows rank denial of reality: the honey fungus will always win. It spreads by black bootlace-like strands of mycelium which are so active metabolically that they glow in the dark. It can parasitize almost any plant, but some are more resistant than others, so we continue to offer our children to this Moloch, hoping that one day he will show mercy, or they will defeat him. It does not help to remember that this fungus forms the largest single living organism on earth—underneath thousands of square miles of Canadian forest, where its fruit bodies have been found to have an identical genotype as they sprout from millions of doomed trees, hundreds of miles apart. This is the way the world ends—not with a bang but an underground glowing bootlace.