Richard Lehman’s journal review—15 June 2015

richard_lehmanNEJM 11 June 2015 Vol 372
2307 Here at last is a study that shows some benefit from out of hospital cardiopulmonary resuscitation. It’s not a randomised trial, since that would be considered unethical, or at least heretical. Instead it comes from interrogating a big Swedish database of outcomes following cardiac arrests outside hospitals. “CPR was performed before the arrival of emergency medical services in 15 512 cases (51.1%) and was not performed before the arrival of EMS in 14 869 cases (48.9%). The 30 day survival rate was 10.5% when CPR was performed before EMS arrival versus 4.0% when CPR was not performed before EMS arrival.” Meanwhile, a big Japanese survey shows that there has been an improvement in outcomes there since resuscitation in the community dropped the “kiss of life” and moved to a strategy based on chest compression only.

2316 There seems to be great enthusiasm for CPR in the Swedish population. Three million Swedes have undergone training, and in Stockholm there were nearly 16 000 citizens willing to join a trial of being on call to rush to a nearby incident should the need arise. “A mobile phone positioning system that was activated when ambulance, fire, and police services were dispatched was used to locate trained volunteers who were within 500 m of patients with out of hospital cardiac arrest; volunteers were then dispatched to the patients (the intervention group) or not dispatched to them (the control group).” Sure enough, those randomised to the intervention got their CPR just a bit sooner than the control group, but “the arrival of these volunteers had little effect on clinical outcomes.”

OL Great news for everyone who has diabetes or treats it. Sitagliptin, which is becoming the most popular add-on drug for type 2 diabetes, has undergone a massive multi-centre trial paid for by Merck Sharp & Dohme and you can read all about it for free. They assigned 14 671 patients at 673 sites in 38 countries to add either sitagliptin or placebo to their existing therapy. On the evidence of the ACCORD, ADVANCE, and VADT trials, there was no need for these people to have add-on treatment since their HbA1c levels were already between 6.5 and 8%, their mean age was 65, and they had had diabetes for a mean of 11 years. But not to worry: the fact is that sitagliptin did them no harm! It did absolutely nothing for their risk of cardiovascular events over three years and it lowered their HbA1c from a mean of about 7.2% to 6.9%. You too can prescribe this drug to do this at a cost of little more than £1 a day.

JAMA 9 June 2015 Vol 313
2244 Limb prostheses don’t feature much in these reviews because they don’t feature much in the general medical journals. But of course they are pretty important to those who wear them. I was reminded of this recently on a visit to Plas Newydd on Anglesey, a lovely Georgian mansion built by the Earl of Uxbridge after the battle of Waterloo, when he became First Marquess of Anglesey. During the battle, he cried out, “By God, sir, I’ve lost my leg!”—to which Wellington, who happened to be nearby, replied, “By God, sir, so you have!” Uxbridge/Anglesey was famous for his amatory pursuits and fathered 17 children by his two wives, plus an uncounted number by others. As his leg was being sawn off in the battle dressing station, he is reported to have smiled and remarked, “I have had a pretty long run. I have been a beau these 47 years and it would not be fair to cut the young men out any longer.” Actually, he fathered six of his legitimate children after he had been fitted with a new articulated prosthesis, now known as the Anglesey leg. Meanwhile, his original leg was kept and mummified in Waterloo itself: perhaps the only known example of the same sided leg becoming a tourist attraction in two different places. Time to leg it back to JAMA. You can read the article and watch the videos for free, whereas to see the Anglesey leg you have to pay an entrance charge or join the National Trust. This new kind is hi-tech, put to the test in a crossover trial involving six participants with above knee amputation and one with a knee disarticulation. The prosthesis learnt its job as it went along, using electromyographic feedback. This allowed it to be controlled intuitively: “Inclusion of EMG signals and temporal gait information reduced classification error across ambulation modes and during transitions between ambulation modes. These preliminary findings, if confirmed, have the potential to improve the control of powered leg prostheses.”

2263 “Non-alcoholic fatty liver disease” is a bizarre diagnostic category, which has sprung up in the last 20 years or so. Between 75-100 million Americans have this “disease,” according to this single author review. “Two large cohort studies demonstrated that 66% of patients older than 50 years with diabetes or obesity had nonalcoholic steatohepatitis with advanced fibrosis on index liver biopsy.” Help! “Advanced fibrosis” sounds terrible. And she suggests that “any patient with suspected steatosis and the metabolic syndrome, or who has metabolic risk factors, particularly diabetes, has a high risk for nonalcoholic steatohepatitis and advanced fibrosis and should be considered for biopsy.” One hundred million liver biopsies, most of which will show fibrosis! And which will no doubt need further biopsies for follow-up. It seems to me that the main risk of this condition is falling into the hands of an enthusiastic hepatologist. Otherwise, just do the usual. Lose weight, get more active, eat fruit and vegetables, and forget your liver, since no-one knows what to do for it and you are far more likely to die of something else.

JAMA Intern Med June 2015
OL A systematic review of tests for the detection of dementia comes at a time of renewed turmoil about this in the NHS. It’s useful, but it can’t be stressed too heavily that figures for sensitivity and specificity depend entirely on the population studied and the context in which the test is to be used. Furthermore, reproducibility is absolutely key here. So when the abstract tells you that “The combined sensitivity and specificity [of the MMSE] for detection of dementia were 0.81 and 0.89, respectively,” don’t go off saying, “well, that sounds pretty good.” It’s lousy for something that gives a person a label for the rest of their lives. Perhaps, given the absurd imposition of copyright on the MMSE, a more important message is, “Besides the MMSE, there are many other tests with comparable diagnostic performance for detecting dementia. The Mini-Cog test and the ACE-R are the best alternative screening tests for dementia, and the Montreal Cognitive Assessment is the best alternative for mild cognitive impairment.”

OL The craze for dementia detection in the UK has also had the rather amusing effect of quadrupling the incidence of dementia in the 30 days following prescription of statins for the first time in British general practice. Mr Prendergast, aged 77, comes along to see his GP for an itchy ear, and has his cardiovascular risk calculated and his MMSE done by the practice nurse in accordance with national guidance. Poor Mr P leaves with some ear drops, a box of statins, and a label of cognitive impairment. So if you look at The Health Improvement Network database from UK primary care, you can do some rather fancy crossover analyses that show an increase in dementia diagnoses in all people given lipid lowering drugs (LLDs) for the first time, whether statins or others. “Thus, either all LLDs cause acute memory loss regardless of drug class or the association is the result of detection bias rather than a causal association.”

Lancet 13 June 2015 Vol 385
2363 Dear Dr Withering, it is with regret that we must inform you that we have read your manuscript An Account of the Foxglove and some of its Medical Uses with interest, but find ourselves unable to accept it for publication. We suggest that you seek the offices of Mr Swinney, a printer in Birmingham, or wait for two centuries to publish it on a most estimable website, to be known as the James Lind Library. It comes as rather a shock, 230 years later, to see a paper suggesting that “A randomised trial of digoxin in treatment of AF patients with and without heart failure is needed.” This conclusion is drawn from a post-hoc analysis of the ROCKET-AF trial, comparing rivaroxaban with warfarin for stroke prevention in people with atrial fibrillation. Over a third of the participants were taking digoxin. “Digoxin treatment was associated with a significant increase in all cause mortality, vascular death, and sudden death in patients with AF.”

2371 Just as an illustration of various meta-analytic strategies, you might consider a glance at this paper with the title, “Mortality in patients treated with extended duration dual antiplatelet therapy (DAPT) after drug-eluting stent implantation: a pairwise and Bayesian network meta-analysis of randomised trials.” Then again, you may not, and I won’t blame you. It is chewy stuff, and may not have much relevance to you, but it does illustrate that more is often less and that you can only discover that by looking at data in the right way: “Although treatment with DAPT beyond 1 year after drug-eluting stent implantation reduces myocardial infarction and stent thrombosis, it is associated with increased mortality because of an increased risk of non-cardiovascular mortality not offset by a reduction in cardiac mortality.”

2383 Now if a stepwise Bayesian approach is needed anywhere, it’s in the assessment of people with suspected angina. Here the diagnostic gold standard is demonstration of coronary artery stenosis, preferably with some measure of fractional flow. Already we are getting into somewhat murky territory. The SCOT-HEART study was designed to get simple answers from a patient and clinician perspective. It randomised patients attending Scottish clinics with chest pain suspected to be angina into two groups: one would get CT angiography at first visit and the other would not. At the outset, 36% of these patients were thought to have angina, and this turned out to be about right, although CT resulted in reclassification of a quarter of patients. These patients went on to have more angiography, not less. Despite pretty good power (n=9849), there was no statistically significant difference in the risk of myocardial infarction at three years. I could use this study as a teaching paper for about 20 topics, but I will desist.

The BMJ 13 June 2015 Vol 350
There are several articles in this week’s The BMJ commenting on the NEJM‘s articles signalling a dilution of its conflict of interest policy. In addition to contributing to a splendid editorial led by Elizabeth Loder, Fiona Godlee herself opens with a piece suggesting the way that The BMJ will proceed. “The same person or people shouldn’t be asked to produce the evidence and appraise it. As our editorial says, ‘These are different professional responsibilities, and they clash.'” The BMJ’s new policy on conflicts of interest among authors of educational articles seeks to make clear this distinction. “Ironically, we took as our model the New England Journal of Medicine’s former policy, which set similar rules. These are hard to implement, but we are determined to push on, evaluating as we go.” Bravo.

All four research articles in The BMJ this week are from the US, and all of them are critical of what goes on there. Fair enough, but I’m not sure this is going to wean BMA members from their alleged habit of only reading the obituaries and occasionally Minerva. The pick of the bunch for me is Rita Redberg et al’s look at the licensing of high risk cardiovascular devices. “Many clinical trials for high risk cardiovascular devices approved by the FDA [Food and Drug Administration] remain unpublished. Even when trials are published, the study population, primary endpoints, and results can differ substantially from data submitted to the FDA.” So the consent form for these procedures should really read, “I hereby consent to have XXX fitted into my arteries/heart (delete as applicable), which was approved for this purpose on the basis of information which later turned out to be incomplete/misleading/made up as they went along (delete as applicable).”

I cast a brief eye on a clinical review of sepsis in children, a “never miss this” topic which still scares me even though I’ve given up assessing febrile children. The thing that struck me most was its unhesitating advocacy of rapid fluid resuscitation, despite the Fluid Expansion as a Supportive Treatment (FEAST) trial, which found that it increased the risk of death in African children. I’m no expert, but I’d have thought that this deserved a mention rather than just a general caveat about the possibility of fluid overload in these critically ill kids.

Plant of the Week: Lupinus arboreus

This shrub does not get much mention in polite circles because lupins are not posh and its habit is rather straggly and vulgar. It is even considered a weed in parts of Australia and New Zealand. But staked properly and in full flower, a good one is a very splendid sight.

The irascible Captain Vancouver was sent to survey the Pacific coast of America and find the North West Passage in 1795, carrying with him the excellent Scottish physician and botanist Archibald Menzies. It was Menzies who first spotted the lupin tree on the coast of California. He collected seed of this and innumerable other species, and when he found that they were not being looked after properly by one of Vancouver’s personal servants, he protested to the captain. Vancouver flew into a rage, put him under arrest, and confined him to his cabin for the remainder of the voyage. It was little bigger than a cupboard. Menzies managed to get some of his collection home on other ships but much of it perished.

There are various forms of this plant, some of them with washy blue flowers. Ignore these and go for the type which flowers a creamy yellow. It will carry more and more flowers for about three to four years, by which time it will bear a hundred or so luminous spikes. Then, tired of the English climate, it will die.