Richard Lehman’s journal review—18 February 2013

Richard Lehman JAMA  13 Feb 2013  Vol 309
559    Last week I had a go at the editors of the NEJM and The Lancet  for publishing misleadingly reported pharma funded trials, in contradiction to their own idealistically stated views. Now, on p 607, the editor of JAMA maps out the moral high ground which he thinks necessary for restoring confidence in the pharmaceutical industry. Howard Bauchner is no Ben Goldacre, and readers of his book Bad Pharma will find this piece wishy-washy by comparison. Nonetheless he calls for independence in analysis: “the data analysis should be performed by academic investigators who are not employed by the company sponsoring the research.”

Secondly, he calls for transparency about authorship: “preparation of the manuscript reporting the study results should primarily be the responsibility of the academic investigators.” “Third, data from clinical trials could be made publicly available to qualified investigators for analyses of important research questions.” 

Now let’s see what JAMA actually prints by way of a pharma-funded study this week. It’s a head-on comparison between linezolid and a new antibiotic in the same group called tedizolid, used as first line treatment for acute skin infections. You have to look right the end of the paper to realise who conducted and wrote this trial: Trius Pharmaceuticals, the manufacturer. I can’t see any sign of independent academic input at all, except in the data reanalysis arranged for JAMA. Despite its close involvement with everything in the trial, Trius also employed a writing agency.

Clearly no expense was spared – 83 sites were asked to enrol patients, but only 54 succeeded, and of those, 8 sites enrolled just one subject apiece. “Conclusions and Relevance  Tedizolid phosphate was a statistically noninferior treatment to linezolid in early clinical response at 48 to 72 hours after initiating therapy for an ABSSSI. Tedizolid phosphate may be a reasonable alternative to linezolid for treating ABSSSI.”

So this new drug may be as good as the old one, though no better: but to my mind the study design was a poor mimic of real life. For cellulitis and skin abscesses, I would almost always use a standard agent first while awaiting bacteriology.

So why print this study in JAMA, contradicting the principles stated by its own editor in the same issue? Could there be some reprint income at stake? Naturally, that issue doesn’t get a mention in this editorial. “If integrity and trust in industry-sponsored research are restored, and rigorous studies demonstrate that a new product substantially improves the health of patients, physicians will be more likely to have confidence in prescribing that product.” Indeed. And this one doesn’t.

570    The incidence of autism has gone up markedly in the last two decades, and so has the use of mobile telephones, water sold in plastic bottles, and preconceptual folic acid. So if you did an association study with any of these things, you would probably find quite a good correlation. Actually what these Norwegian investigators did was to compare the rates of autism in the offspring of women who did or did not take preconceptual folic acid, and found that there was 40% less in those who took it.

But there was no difference in the rate of Asperger’s syndrome, which is odd. What I’m trying to say here is that you cannot trust observational associations even if there is a plausible underlying hypothesis. Of course every woman intending pregnancy should take folic acid to prevent neural tube defects, and I am all for putting it into bread: but I would still be surprised if this led to a fall in diagnoses of autism.

587    Speaking of hypothesis generation, here’s a little brain teaser: if you look at risk-adjusted hospital figures for death and readmission for the commonest medical conditions – myocardial infarction, heart failure and pneumonia – would you expect to find a correlation between the two? And which way would it go? Would hospitals where more patients die have fewer readmissions? Or would they have more, because they are so bad at managing acute illness? 

Well, don’t spend too long over this, because this immense study of outcomes in over 4,500 American hospitals shows that there is no correlation between risk-standardised mortality and risk-standardised readmission in these conditions, with the possible exception of some grades of heart failure. You may recollect the BMJ editorial a couple of weeks ago from the chief investigator of this study, Harlan Krumholz: nothing is straightforward when trying to assess the quality of hospitals, even using such apparently robust data. A timely lesson, as the NHS goes tipping the barrel in search of rotten apples.

NEJM   14 Feb 2013  Vol 368

599    The devastation and misery that followed the earthquake in Haiti three years ago was followed nine months later by the worst cholera epidemic in living memory. This paper gives the figures: 604,634 cases reported and 7,436 deaths. I remember reading elsewhere that the mean time from onset of symptoms to death in this epidemic was less than twelve hours, and in some regions the case-fatality rate was over 4%. So an overall case fatality rate of 1.2% must reflect well on the heroic efforts of medical teams working in the worst of conditions with little more than portable beds, buckets and bags of fluid.

610    Many a pleasant evening has ended with me watching a crazed doctor, usually with a thick German accent, inserting electrodes into the brain of a hapless young victim, who then usually ends up strangling somebody, often the doctor himself. The collection of B movies on the American Netflix site is really quite remarkable.

However, this cultural stereotype does few favours to German doctors trying to use such methods to help younger patients with Parkinson’s disease. Neurostimulation of the subthalamic nucleus is a method of proven usefulness and proven risks in PD with severe motor complications resistant to dopaminergic drug treatment. This trial looked at the earlier use of the technique in order to prevent the worsening of such complications in patients aged 60 or younger, and overall it was a success. It is especially good to see that the outcome measures used were completely patient-centred, and most of the subjects noted a marked improvement. Previous studies of subthalamic stimulation have shown an increased risk of suicide: this study was too small and too short, but there were two suicides in the stimulation group and one in the medical therapy group.

633   Tonight I am doing an out of hours primary care shift and I will surely be repeating the familiar words, “He’s got quite a nasty virus, but he’s holding his own well and I’m sure he’ll be better in the next two days.” It is very rare for us to know which virus we’re talking about, of course, especially in children under five with upper respiratory symptoms. Human metapneumovirus was first recognized in 2001: it does very similar things to influenza virus, rhinoviruses and respiratory syncytial virus, but according to this US survey, kids with it are more likely to be diagnosed with asthma or pneumonia. It is found in about 7% of children attending out of hours.

Lancet  16 Feb 2013  Vol 381

537    Most doctors are very uncomfortable with the fact that people over 65 who are overweight or obese live longer than those who are of “normal” weight. Not only does this run counter to the deep puritanism of medical culture, but it also flies in the face of logic, because such people are much more likely to have diabetes, heart failure and hypertension.

And yet obese people with hypertension have the best outcomes, in trial after trial. And if you give them thiazide diuretics and so increase their insulin resistance, they do even better. And if they get heart failure, they will greatly outlive their thinner peers. Here is an analysis of the ACCOMPLISH trial – don’t even try to remember which one that was – which clearly shows that thiazide treatment gives better outcomes in hypertensive fat people.

By all that’s holy in mechanistic reasoning, this should not be true, and it is all too much for the authors of the accompanying editorial. They list their objections and state:” Therefore, we reject the conclusion of Weber and colleagues that diuretic-based regimens are a reasonable choice in obese patients. On the contrary, we surmise that thiazide diuretics are contraindicated in obesity, relatively speaking.” So surmise trumps evidence? I don’t think so.

546    According to widely held belief, if you are involved in a road traffic accident caused by somebody else, you are almost certain to have whiplash neck injury and must make an immediate claim for private physiotherapy at the expense of the other driver’s insurance. Here’s a nice trial called MINT which tests this hypothesis in drivers with “whiplash” from mishaps on the very roads that I drive along myself. Giving whiplash sufferers special consultations in A&E and an early physio package has comfort value, but provides no long term benefit. “Usual consultations in emergency departments and a single physiotherapy advice session for persistent symptoms are recommended.” That’s for the NHS. If you can claim for private treatment, get all the comfort you can.

557    The Lancet loves cutting-edge gizmology, and here we have a case report from the US describing a tetraplegic lady who had two arrays of 96-channel microelectrodes placed in her motor cortex, which enabled her to use one arm within two days. In fact, they had better keep a close eye on her, or she may disappear beyond the boundaries of space-time: “After 13 weeks, robust seven-dimensional movements were performed routinely.”

That’s way too many dimensions: what they mean are degrees of freedom: three-dimensional translation, three-dimensional orientation, one-dimensional grasping. A terrific technical feat, but just how easy will it be to convert this into a generalizable intervention. The editorial sounds a sombre note: “This brain—machine-interface technology remains confined to the controlled environment of laboratories where highly skilled engineers tune decoding algorithms daily. Dissemination of such brain—machine-interface-controlled devices is contingent on a number of non-trivial technical issues, the resolution of which remains uncertain.”

BMJ   16 Feb 2013  Vol 346

This is a good week for the Less is More principle. We saw that it applies to whiplash, and this Swedish study shows that it applies to acute anterior cruciate ligament tears in fit people. “In this first high quality randomised controlled trial with minimal loss to follow-up, a strategy of rehabilitation plus early ACL reconstruction did not provide better results at five years than a strategy of initial rehabilitation with the option of having a later ACL reconstruction. Results did not differ between knees surgically reconstructed early or late and those treated with rehabilitation alone.”

Half the patients randomized to the latter group ended up having surgery. Here is an area of preference-sensitive choice. Patients with these injuries need to know the evidence and make their own choice. Months of hobbling about on an unstable knee may not make any difference in long term functional outcome, but may severely affect activity and employment.

And Less is More even applies to calcium intake in older women. Many osteoporosis-fixated “women’s health experts” have recommended high calcium intake for decades, well in excess of anything that is plausibly available from dietary sources. But this mammography cohort study from Sweden followed up 39,000 women for 19 years and found that high calcium intake – over 1400mg daily as judged purely from two dietary assessments – is associated with a 40% increase in all-cause mortality and a doubling of death from ischaemic heart disease. “Harmless” chalk as sold by thousands of health supplement vendors should hereafter carry a health warning.

Almost all over-the-counter drugs are more expensive in the US, with the exception of acetaminophen, which you can buy in cheap tubs of 100, most useful for a frequent migraineur such as myself. Acetaminophen is what we call paracetamol, and can now only buy off the shelves in measly packs of 16 (or 32 if you ask a pharmacist). Patients are forever moaning that this can’t possibly be useful as a way of preventing suicide, but in fact the figures suggest that it is. There has been a 43% reduction or an estimated 765 fewer deaths over the 11¼ years after the legislation, and a big drop in liver transplantation for paracetamol-induced hepatic necrosis too.

Plant of the Week: Abeliophyllum distychum

In a month or so, many a suburban garden and hedgerow will be screaming with the yellow of forsythia, and I will wonder yet again why no hybridist has managed to give the world a cream-flowered, scented version of this coarse and repellent shrub. Perhaps the trick would lie in getting a late-flowering Abeliophyllum to pollinate an early-flowering forsythia, because the two are sufficiently close in kinship for this to have a chance of producing a bigeneric hybrid.

Until someone has achieved this, go for the Abeliophyllum and shun the forsythia. Here is an unpretentious, straggly medium sized shrub which will merely fill space for eleven and a half months of the year. But for a fortnight in late winter, it will be your dearest garden possession, covered in white or pinkish forsythia-like flowers with a wonderful scent of honey mixed with Nivea.