Richard Lehman’s journal review—30 August 2016

richard_lehmanAfter a month’s break, I’m catching up with articles of interest in the main non-BMJ journals throughout August. Normal service will be resumed next week. 

NEJM  Aug 2016  Vol 375

No parasites for five years

405   Following its famous “data parasites” editorial last January, the NEJM has struggled to find a comfortable position on the issue of opening up trial datasets for independent analysis. In its first print issue for August it hosted four Viewpoint articles offering a range of options, mainly for the sharing of trials conducted by academic centres. Completely open sharing does not get a mention—the idea that researchers should make their de-identified data available without restriction. The NEJM spectrum instead begins with the Yale Open Data Access project, of which I’m a proud founder member. We pioneered the “academic intermediary” model and still run one for those who wish to use it. If researchers want to use more direct methods of sharing data, we’re hardly going to stop them. But alas, the very opposite tends to be true. A piece from a hitherto unknown body called The International Consortium of Investigators for Fairness in Trial Data Sharing argues that researchers should retain sole use of their databases for up to five years. I’m genuinely sad and baffled that people with prosperous careers supported by public funding can argue that the research they have performed through the altruism of unpaid people taking voluntary risks is somehow their private property.

Amish, Hutterites & asthma in children

411  This is a fun paper which mixes various methods to propose that old-fashioned horse-based farming fills kids’ lungs with antigens that stop them developing asthma. Same applies to transgenic mice. North America contains two entirely separate historical Anabaptist communities who only marry within their own groups and speak their own South German dialects. The Amish practise traditional farming, live on single-family dairy farms, and use horses for fieldwork and transportation, while the Hutterites live on large, highly industrialized, communal farms. Strikingly, the prevalence of asthma in Amish versus Hutterite schoolchildren is 5.2% versus 21.3% and the prevalence of allergic sensitization is 7.2% versus 33.3%. Read on. No doubt some entrepreneur will soon market an aerosol of farmyard microbiome for use by all new parents.

Adalimumab for hidradenitis suppurativa

422   Hidradenitis suppurativa is a horrible condition to live with, and any new treatment is likely to be seized on eagerly. Adalimumab is an anti-TNF antibody with a basic cost of £352 per 40mg vial in the UK, and it is already quite widely used for this condition, amongst several others. Apart from costing a minimum of £50 per day, it carries a risk of serious infection and non-melanoma skin cancer. The results of two 12-week manufacturer-supported placebo-controlled phase 3 trials called PIONEER are moderately positive, but don’t give a non-expert reader much idea of its relative effectiveness among the treatments available. Above all, they don’t give any clues about long-term safety.

Thymectomy for myasthenia gravis

511   “The thymus gland seems to have been unknown to Hippocrates and Aristotle, and was first mentioned by Rufus the Ephesian in his enumeration of discoveries in anatomy made at the School of Alexandria under the Ptolomies.” Thus Sir Geoffrey Keynes begins his groundbreaking Hunterian Lecture on the Surgery of the Thymus in 1945: it is somehow comforting that he managed to spell “Ptolemy” wrong. He was right about almost everything else, including the texts of William Blake, blood transfusion during field service in the Great War, breast-conserving surgery in the 1920s, the Works of Sir Thomas Browne, and the organization of surgical services for the RAF in the second war. And here is a trial showing that he was right about the thymus and myasthenia gravis. I have given away my copy of his wonderful autobiography The Gates of Memory, but I seem to remember that he first described the blood supply of the thymus in living adults, and he certainly pioneered the use of thymectomy as a treatment for myasthenia gravis in the UK (Blalock was the US pioneer). Even in the era of immunosuppression and corticosteroids, the first randomized trial of surgery shows that it improves outcomes in most patients with myasthenia.

Paracetamol or ibuprofen for wheezy kids?

619   Probably most of you will have read about this already. “In a multicenter, prospective, randomized, double-blind, parallel-group trial, we enrolled 300 children (age range, 12 to 59 months) with mild persistent asthma and assigned them to receive either acetaminophen (paracetamol) or ibuprofen when needed for the alleviation of fever or pain over the course of 48 weeks. The primary outcome was the number of asthma exacerbations that led to treatment with systemic glucocorticoids.” There was no difference. A basic common question is laid to rest. How many trials can you say that of?

Breast cancer signatures & chemo

717   It’s undeniable that as a result of mammographic screening, too many women receive a diagnosis of breast cancer, and that among those who do, chemotherapy is widely used without survival advantage. So anything which helps select the women who might actually benefit from chemo is to be welcomed. There are clinical scoring systems which take us some way, and now there are a number of competing commercial tests that rely on genomic markers within individual biopsy specimens. To validate them, you need to subject these cancer “panel” tests to prospective study, which can’t be done in less than five years. Here is just such a study using the 70-gene signature called Mammaprint. The abstract suggests that using this test could avert the need for chemo in nearly half of women at high clinical risk of distant recurrence. The design of this study is complex and confirming this claim is not easy. Here’s an area which cries out for good infographics to help women make a personal choice in a very difficult area.

JAMA  Aug 2016  Vol 316

Diabetes & the kidney 1988-2014

602   If you can get access to this paper, take a look at Table 1 which gives the characteristics of four separate cohorts of people with diabetes sampled by the US National Health and Nutrition Survey between 1988 and 2014. You’ll see that their HbA1c percentage drops slightly over that period, from 8.1 to 7.6. Over that period, the use of renin-angiotensin-aldosterone system antagonists increased by a factor of 2.8 and the use of lipid lowering agents by 4.2. So what effect did that have on the point prevalence of severely reduced eGFR (30 or below)? It increased by a factor of 2.8, despite a considerably lower duration of diabetes in the final cohort. There’s a funny coincidence between the RAAS figure and the severe CKD figure. These agents reduce albuminuria, and that reduction was observed in these cohorts. But could they be doing more harm than good for diabetic kidneys? It’s looking like a real possibility. Even more, it’s looking like we don’t really have a clue about treating most aspects of type 2 diabetes, because we have built the entire towering edifice on a marshland of surrogates.

Abaloparatide to prevent vertebral fractures

722   Women over the age of 60 have a 44% chance of suffering a fracture over the rest of their life, which of course increases if they have osteoporosis. In fact 63% of the women with osteoporosis (mean age 69) recruited to this trial had had previous fractures. And yet a third of them were assigned to placebo (bisphosphonates and denosumab were not allowed). The active alternatives were teriparatide 20mcg or abaloparatide 80 mcg, both activators of the parathyroid hormone type 1 receptor. In the 18 months of the trial, 30 women in the placebo group had vertebral fractures, compared with 6 using teriparatide daily and 4 using the new drug. On the basis of this report, it seems to me that untreated women suffered fractures that could have been prevented, and that the trial wasn’t powered to detect a significant difference between teriparatide and its would-be successor. Is this good enough?

How many pairs of genes does a couple need?

734   Somewhere in your genome, there allegedly lurk between 2-6 lethal recessive genes. If you chance to mate with a carrier of these same genes, then your baby might die. That’s the scary sales pitch which has already led to over 400,000 young Americans having a preconceptual expanded-carrier gene test for over 100 heritable conditions. But the same pitch can be used to argue against preconceptual testing, because our ignorance about the predictive value of these tests is profound, and their cost in health resource use and parental anxiety is almost limitless (see the excellent editorial). Even the authors (some of whom are from the manufacturers of the screen) conclude that “Prospective studies comparing current standard-of-care carrier screening with expanded carrier screening in at-risk populations are warranted before expanded screening is adopted.

Lancet  Aug 2016  Vol 388

Brachial block for AV fistula

If you’ve ever done a minor surgical procedure, you’ll be aware of the way that local anaesthetic can distort the operative field and make your results even worse than they would otherwise be. Creating an arterio-venous fistula for patients needing haemodialysis is not a procedure you want to mess up on. Renal patients in three large Glasgow hospitals were randomly allocated to regional anaesthesia using brachial plexus block or to local anaesthesia at the time of fistula creation. Brachial nerve block also helps to dilate the vessels. The odds of a functioning fistula at 12 weeks were significantly higher than with locally infiltrated bupivacaine + lidocaine (84% v 62%).

Death after atrial fibrillation

Everybody knows that the main cause of death attributable to atrial fibrillation is stroke. Except that across the world, it isn’t. The RE-LY group keep a prospective registry of patients in 47 countries who present to hospital emergency departments with atrial fibrillation or atrial flutter as a primary or secondary diagnosis. Of 15,361 patients registered from 2008 to 2011, 1758 died in the first year: 8% from stroke and 30% from heart failure. Death rates were double the North American level in South America and Africa.

Vertebroplasty within 5 weeks

Many years ago I had a patient with an agonizing vertebral fracture which made it impossible for her to sleep or move. I looked up percutaneous spinal vertebroplasty—then a new treatment—and according to case reports it seemed to produce worthwhile results which were time dependent: the sooner you did it, the better, though there were no randomized trials that I could find. I made an urgent referral, but it was about three months before she had the procedure, which made no difference. In 2014, a Cochrane systematic review of percutaneous vertebroplasty for osteoporotic fracture showed that overall it provided no more benefits than a fake procedure. Now we have a well-conducted Australian RCT which allocated 120 patients with new or recurrent vertebral collapse to real or simulated vertebroplasy within 6 weeks. Of patients in the active group, 44% had a marked reduction in pain at 14 days, compared with 21% in the sham group. It looks as if the original case reports were right: timing matters with this procedure. But a larger confirmatory trial would also be good.

Plant of the Week: Salvia x jamensis “Melen” 

This is a lovely little shrublet which dies back in the winter and then – if the frosts haven’t been too hard – reappears and bears snapdragon flowers of primrose yellow edged with brown from May to October. These are truly charming and often abundant. If you stoop to smell them you will be overpowered by the aroma of any leaves that you disturb. They’re worth rubbing each time you pass the plant.

Ours looks a little old and woody after three years so I took small cuttings throughout July and put them into pots of compost. Each is now a sturdy plant with its own flowers.