JAMA 24-31 Oct 2012 Vol 308
1660 Medicine is full of surprises, but sometimes things are just the way you thought they were. Back in 1973, I was taught that the causes of peripheral arterial disease in men were smoking, diabetes, hypertension and high cholesterol. Then in 1986, the Health Professionals Follow-up Study recruited 44 985 men without known cardiovascular disease and followed them up for 25 years to discover the causes of peripheral arterial disease. This week you can discover that they are: smoking, diabetes, hypertension and high cholesterol.
Medicine is full of surprises, but sometimes they are just artefacts due to poor methodology. The authors of this study looked through the entire Cochrane Database of Systematic Reviews to find randomized controlled trials with huge effect sizes, and discovered that these are mostly small first studies looking at surrogate outcomes. Beware odds ratios of 10 or more: these usually diminish to something less than 4 in subsequent trials. That still makes them significant, provided always that the outcome measure they use is clinically relevant. But in most cases it is not. Most of the medical literature is dross.

NEJM 25 Oct 2012 Vol 367
Even the New England Journal is not an entirely dross-free zone, but a number of papers in this week’s issue show why every doctor should try to read it regularly. We’ve all wished that atrial fibrillation could be abolished, and indeed this week’s Lancet makes the broad and not quite intelligible claim that “new technologies have propelled arrhythmology into an effective modern therapeutic arena.” Here an important trial tests the truth of this claim in relation to one such new technology – Radiofrequency Ablation as Initial Therapy in Paroxysmal Atrial Fibrillation. The Danish Heart Foundation and others paid for a direct comparison trial of primary pathway ablation versus standard drug therapy in 294 patients with newly presenting AF. There were more major complications in the ablation group and overall at 24 months there was no significant difference in the overall burden of atrial fibrillation between groups.
I like the way that cancer genomics – i.e. the genetic subclassification of individual tumours – is beginning to tell us which cancers will respond to what. This is real science that works – a far cry from the wild claims of the whole-person genome overlords who have yet to come up with anything of generalizable value for their billions of pounds of preferential funding. Take the simple question of which colon cancers will show a markedly reduced recurrence rate in response to daily aspirin. Fortunately there is a big database of tissue samples and treatment information in the Nurses’ Health Study and the Health Professionals Follow-up Study. This admirable investigation used these to demonstrate that regular use of aspirin after diagnosis was associated with longer survival among patients with mutated-PIK3CA colorectal cancer, but not among patients with wild-type PIK3CA cancer.

For my last thirty years as a general practitioner, the challenge which I’ve sought hardest to meet was to help people come to terms with dying. This never gets easier, especially as you get to know people over decades, and your own turn looms. But I thought that the medical community as a whole had made a lot of progress over that time, and that I had been privileged to share in that. In fact by 1992 I thought that palliative care had so transformed patient communication and humane care in cancer that it was time to move on and apply the same lessons to other disease areas, especially end-stage heart disease. Here’s a study to pull us all up short and ask ourselves what we have really achieved. The Cancer Care Outcomes Research and Surveillance (CanCORS) study (a national, prospective, observational cohort study) questioned 1193 patients who were alive 4 months after diagnosis and received chemotherapy for newly diagnosed metastatic (stage IV) lung or colorectal cancer. I have read this paper twice and still find it hard to believe the bottom line: in response to a carefully worded and well-administered questionnaire, 69% of patients with lung cancer and 81% of those with colorectal cancer did not report understanding that chemotherapy was not at all likely to cure their cancer. This paper is mandatory material for every journal club run by oncologists and palliative care professionals. The study needs replicating in other health systems. The irony here is that the USA actually has the best system in the world for producing and updating materials for decision support in cancer, hidden away on the NIHR website and used by far too few patients and clinicians. There is a superbly written editorial on this study, which begins, “Self-deception is a valuable personal coping tool. It allows us to aspire to significance, strive for new knowledge, and yearn to make a lasting contribution to the world despite the certainty of our inevitable end.” So I’ll keep on trying to deceive myself, then; until the time comes for my own chemotherapy.

A generally very good review deals with the “peripheral mechanisms in irritable bowel disease”. This should really read “local bowel mechanisms in IBS” since the bowel does not lie on the periphery either physically or conceptually. In the interesting and comprehensive discussion I was occasionally struck by the mention of unusual agents e.g. “Treatment with intestinal secretagogues (i.e., lubiprostone and linaclotide) or prokinetic agents (e.g., tegaserod) is effective in relieving constipation and associated IBS symptoms such as pain and bloating.” That led me to look at the author’s declarations of interest: “Dr. Camilleri reports receiving consulting fees from Takeda Pharmaceuticals USA, Albany Molecular Research, BioKier, Theravance, Alkermes, ARYx Therapeutics, AstraZeneca, Domain Therapeutics, Ironwood Pharmaceuticals, Tranzyme, and NPS Pharmaceuticals”; (there is further section on institutional payments). Now it so happens that lubipristone is made by Takeda, and linaclotide is made by Ironwood: many other agents are mentioned too, so this might be a coincidence. But authors who accept so much industry funding should not be invited to write reviews for leading medical journals of record. And once again, journals should be obliged to disclose which drug companies order reprints, and to what value.

Lancet 27 Oct 2012 Vol 380
1462 The most readable thing in this week’s Lancet – a feeble compliment, I’m afraid – is a two-page review of Ben Goldacre’s Bad Medicine by the outgoing chairman of NICE, Sir Michael Rawlins. It begins with a statement about people dividing into two camps about the pharmaceutical industry, and criticizing Ben for a lack of equipoise. The strapline complains that “It is this balance, between the beautiful and the ugly, that Goldacre fails to achieve.” The problem for Ben, as for me, is that the more you look, the uglier pharma becomes, and that equipoise would be ridiculous when there are several instances of admitted massive fraud for which no individuals have been held accountable. On the side of beauty, Rawlins cites no convincing examples. In fact, NICE itself has had to accept a gradual dilution of its influence due to successful lobbying by the UK pharmaceutical industry. As things stand, it will have ceased to operate as an effective assessor of new interventions within a year of Rawlins’ departure.

A large American observational study looks at outcome differences between hospitals with different duration times for cardiopulmonary resuscitation. The hospitals where CPR is recorded as lasting longest tend to have slightly higher CPR survival rates. The moral seems to be: don’t give up before 25 minutes.
Hard on the heels of last week’s episode of Stent Wars, here comes another: Everolimus-eluting stent versus bare-metal stent in ST-segment elevation myocardial infarction (EXAMINATION): 1 year results of a randomised controlled trial. The trial had a “patient-oriented combined endpoint” of all-cause death, any recurrent myocardial infarction, and any revascularisation at 1 year. On that basis it was a draw.

“If he doesn’t ease up, he’ll have a heart attack.” How true is it that job strain is a risk factor for coronary heart disease? Slightly true, according to a massive survey of 13 European observational studies in which a total of nearly 200,000 participants were asked about stress in the workplace. The authors conclude that the effect size is very small in comparison with conventional coronary risk factors.

BMJ 27 Oct 2012 Vol 345
Two editorials feature amongst the best pickings in this week’s BMJ. Ironically (in view of the NEJM study) the first is a plea from a US professor of primary care for better end-of-life communication with people who have chronic diseases. “The best doctors know when to help patients understand that it is time to stop active treatment. They also know when to have gentle, caring, supportive conversations to realise a new set of shared goals. At times these conversations may be easy and at times they are hard, but they shouldn’t be avoided because goal oriented person centred care is the only sensible approach to caring for those with multiple chronic conditions who are approaching the end of life.”

The second editorial is by Peter McCulloch and is about devices regulation in the European Union. This follows a piece of agent provocateur style investigation by Deb Cohen and the Daily Telegraph which showed that it is very easy to get the European regulators to accept a new medical device (in this case a non-existent hip prosthesis) on the basis of spurious data. The editorial suggests that “This may also be the right time to propose the radical idea that devices, like drugs, should require evidence of efficacy as well as safety before being given a full licence.” How odd that this should be considered a radical idea in 2012.

The BMJ’s research section often includes studies that address important questions but fail to resolve them. “In this prospective population based study, new use of benzodiazepines was associated with increased risk of dementia.” In fact the association barely reached statistical significance, and it is quite plausible that people with anxiety due to early dementia receive more benzodiazepine prescriptions, especially in prescription-loving France, where this study was conducted.

But a Dutch prospective cohort study of how to exclude pulmonary embolism in primary care reaches a much more robust conclusion. “A Wells score of =4 combined with a negative qualitative D-dimer test result can safely and efficiently exclude pulmonary embolism in primary care.” The so-called “qualitative” D-dimer test was a new point-of-care test which was not interpretable in 6% of patients with suspected PE. This is much better than previous tests I have tried to use, and is probably as good as one can get. So this is a definite move forward, consistent with many other

Arch Intern Med 22 Oct 2012 Vol 172
All the acupuncture trials that I have read over 14 years – with one recent Japanese exception – have shown equivalence between sham and traditional acupuncture for chronic pain, and I have had fun exhorting readers to buy suitable robes and needles and make easy money by becoming sham acupuncturists. This individual patient data meta-analysis does not change my view, though it shows a slight advantage for traditional acupuncture – which I think is probably driven by the greater confidence exhibited by traditional practitioners. If you think you have what it takes to become a sham practitioner, don’t let this article puncture your confidence.

Your patient is taking warfarin and has a gastrointestinal bleed, so you stop the warfarin. How long for? Probably the best answer is for as short a time as possible, since for most patients the risk of a thrombotic event (and death) is greater than that of a further GI bleed.

Plant of the Week: Rhus copallinum

We’re back in New England for the very end of the “fall colors” season, though we are mainly on errands other than just “leaf-peeping”. Travelling north to Vermont, we noticed the wooded hills becoming barer as each mile passed, whereas travelling down to Connecticut we could still enjoy beautiful miles of yellow, purple, red and green. Now we are staying with friends who live close to West Rock near New Haven, a place famous for sheltering three of the judges who had signed the death warrant of Charles 1st, Whalley, Goffe and Dixwell. They fled to America when Charles 2nd returned to claim the British throne, and the places they haunted are known to this day as the Regicide Trail.

We have not yet tackled the full Trail, but we have had a pleasant wander round Bishop’s Loop, which lies in the foothills. This is an area of regenerative woodland with young trees which have already lost their leaves. In amongst the tall bare trunks are wonderful sparse shrubs covered with leaves in various shades of light-to-darkish pink, glowing like flamingos of the understory. The effect is quite magical. They are known as shining sumacs. A small hidden river flows swiftly beside them.

It is some years since anyone offered the shining sumac for sale in the UK. It would grow readily in England, but in truth the two other ornamental sumacs, Rhus typhina and R glabra are better garden plants, particularly in their cut-leaved versions. Unfortunately they do tend to sucker freely, so you can end up with more autumn colour than you bargained for, in places you little expected.