JOURNAL SNIPPETS November 3

NEJM 25 Nov 2010 Vol 363
2091 Quality improvement is actually, of course, a good thing in itself, and we need better ways of doing it and better ways of studying it. High quality outcomes research, carefully reflected on, is one essential input, and there are two good examples in this week's New England Journal. A general survey of North Carolina hospitals (see below, p.2124) produces a rather gloomy view of overall improvements in patient safety, but this study of mortality from allogeneic haematopoietic-cell transplantation is very much cheerier. The period 2003-7 showed an overall mortality fall of over 40% compared with a decade earlier, driven by significant decreases in the risk of severe GVHD; disease caused by viral, bacterial, and fungal infections; and damage to the liver, kidneys, and lungs. Further details are strictly for the haematologists: but it seems they have much to congratulate themselves about.
http://www.nejm.org/doi/full/10.1056/NEJMoa1004383

BMJ 27 Nov 2010 Vol 341
1144 Thrombolysis for acute occlusive stroke has been shown to be marginally beneficial in several RCTs, but the number of people over 80 in these trials is minuscule, whereas in real life, 30% of strokes occur in this age group. There is a presumption that the hazards of thrombolysis will be greater and the outcome difference less. This European registry study indicates that neither is true: thrombolysis remains beneficial for stroke beyond the age of 80.
http://www.bmj.com/content/341/bmj.c6046.full
1146 Any enthusiastic regular drinker of wine, will be delighted to note the PRIME study which confirms that by doing so you halve your chance of myocardial infarction. I suppose you also increase your chance of pancreatitis, cancers of the GI tract and stroke. Perhaps liver disease too, though the literature is surprisingly obscure at levels of intake below about 100u/week. The thing not to do is binge drink, which is a common pattern in Northern Ireland, and probably increases your baseline risk of MI. I think the further north you travel, the more dysfunctional alcohol use becomes, as warm oblivion becomes ever more desirable. As if to illustrate this point, a review of frostbite on p.1151 finds that nearly half of it is associated with alcohol use. I bet that means vodka or whisky in most cases, and wine alone hardly ever.
http://www.bmj.com/content/341/bmj.c6077.fullhttp://www.bmj.com/content/341/bmj.c5864.extract

Arch Intern Med 22 Nov 2010 Vol 170
1926 In studies of drugs that put people into hospital, warfarin usually comes near the top. This study looks at how combined platelet inhibition with aspirin plus clopidogrel compares in emergency department visits for haemorrhage-related events. The score is 2-1: 2.5 events per 1000 prescriptions of warfarin as compared with 1.2 events for aspirin/clopidogrel.
http://archinte.ama-assn.org/cgi/content/abstract/170/21/1892

JAMA 17 Nov 2010 Vol 304
2129 Like all doctors who survived their hospital jobs in the 1970s, I have some shocking memories. Oddly enough, though, some of them are happy too, as the shocks saved lives. The woman dragged out of a freezing canal with a core temperature of 28ºC who survived intact after 16 defibrillations; the 43-year old man with chest pain who went into VF just as we were putting the leads on his chest: all of us can still remember these kinds of event, while our futile attempts go forgotten days after. Surely an automated defibrillator must beat a sleep-deprived, dishevelled house doctor at achieving survival following in-hospital cardiac arrest? Actually no: another massive US cardiac outcomes study looks at the results of introducing automated defibrillators on to the wards of 204 hospitals and finds that results actually tend to be worse.
http://jama.ama-assn.org/cgi/content/abstract/304/19/2129
2137 The harmful effects of low-dose ionizing radiation are not well understood, but from about 100mSv upwards we are no longer talking about low doses, but the kind of exposures about which we have data from Hiroshima and Nagasaki. Alarmingly, such doses were received by a third of patients in this study of repeated myocardial perfusion scanning. OK, the majority of these people were over 60 and had heart disease, and would escape long-term harm: but it suggests that we are getting too gung-ho about exposing people to high energy photons from X-ray machines and unstable isotopes, and the cumulative damage which they cause.
http://jama.ama-assn.org/cgi/content/abstract/304/19/2137

Lancet 20 Nov 2010 Vol 376
1741 Many doctors in the 1990s went through a phase of taking low dose aspirin and recommending it to many of their patients with high blood pressure and/or type 2 diabetes. Then came a series of trials which showed that it doesn't work for primary prevention of cardiovascular events, even in groups who are at increased risk. But it does prevent about 25% of bowel cancer, according to this long-term follow up study of participants in 5 large aspirin trials, matched at a median of 18.3 years with mortality registers. The results suggest that you need to take about 75mg of aspirin for at least 5 years to achieve such protection, and the effect may be specific to the proximal colon. Thus in theory universal aspirin consumption, combined with a universal programme of screening sigmoidoscopy, could prevent most bowel cancer. However, an analysis like this can tell us little about adverse events, and we will only know for certain after a prospective trial lasting at least ten years.
http://www.thelancet.com/journals/lancet/article/PIIS0140-6736(10)61543-7/abstract

SNIPPETS FROM JOURNALS

The following are all from www.cebm.net

Lancet 27 Feb 2010 Vol 375
727 Not long ago, someone had a myocardial infarction on a transatlantic flight. So what does a professor of surgery do under these circumstances? Possibly take an aspirin and pray a good deal. He should have squeezed his arm repeatedly for periods of five minutes or so at a pressure above systolic. Believe it or not, this simple manoeuvre can reduce the area of myocardial damage, as proved in this Danish trial where patients with presumptive MI were randomised to have the squeezing done (or not done) by a sphygmomanometer in the ambulance conveying them to hospital. This is known as ischaemic preconditioning, though in such circumstances it should perhaps be known as simultaneous ischaemic conditioning. There were no hard end-points in this trial but a convincing reduction in damage on myocardial perfusion imaging at 30 days. More trials are needed, but meantime there seems no possible reason not to give it a try.
http://www.thelancet.com/journals/lancet/article/PIIS0140-6736(09)62001-8/abstract

752 This is a rather rambling 10-page review of dilated cardiomyopathy. The Panel of Mechanisms covers just about everything except interference by aliens, and there's a panel of gene loci too, but no panel of relative frequencies and prognoses for each aetiology. There is no mention of spontaneous recovery, which can't be rare if it has been seen twice.
http://www.thelancet.com/journals/lancet/article/PIIS0140-6736(09)62023-7/abstract

763 Say you had stable coronary artery disease: would you want an angiogram? And if the cardiologist saw a stenosis, would you want a stent put in? A year or two ago, these questions would have seemed like no-brainers, but then along came COURAGE and BARI-2D showing that medical treatment is as good as percutaneous intervention. Do you truly and deeply believe this, though? See how you feel when you read this review by two Swiss and an American cardiologist. It presents enough evidence to allow a tailored approach, and says that it "proposes a treatment algorithm that is applicable to daily clinical practice." The word "algorithm" is to be avoided, but people who use it generally refer to a flow chart, but there isn't one. http://www.thelancet.com/journals/lancet/article/PIIS0140-6736(10)60168-7/abstract

BMJ 27 Feb 2010 Vol 340
459 A useful systematic review compares the results of carotid endarterectomy vs. carotid stenting in 11 randomised trials. On the face of it, endarterectomy wins, because the risk of periprocedural stroke is less; in the longer term there is little difference. Techniques and experience increase all the time - it may happen that the guy who puts in stents near you does that better than the guy who scrapes arteries. So more studies are justified and the matter is not quite settled yet.
http://www.bmj.com/cgi/content/full/340/feb12_1/c467

JAMA 17 Feb 2010 Vol 303
631 The Women's Genome Health Study is a prospective cohort of 19,313 women followed up for a median of 12.3 years, during which they experienced 777 cardiovascular events. In these women, 101 single nucleotide polymorphisms were added with one or two other genomic factors to create a genetic risk score. Surely this would usher in a new era of refined cardiovascular risk prediction? Well, actually it showed no significant association with the incidence of total cardiovascular disease: a simple family history alone was more predictive. On the other hand, there is so much anonymized data about the participants that you could probably find out the full disease status of any individual if you could identify their genome from some other source. This is discussed in a fascinating commentary on p.659. Genomic studies seem almost disconcertingly useless at the population level, but if you know 35,000 gene variants in a single individual, you can measure their left ventricular mass more accurately than if you had an echocardiogram.
http://jama.ama-assn.org/cgi/content/abstract/303/7/631

NEJM 18 Feb 2010 Vol 362
590 Some people like to see evidence from randomised controlled trials with hard end-points before a computer prediction that a certain intervention will reduce new cases of CHD in America by up to 120,000 annually, stroke by up to 66,000, and death by up to 92,000. The editorial on p. 650 suggests a saving in health costs of $10-24 billion. Aha, we save health costs by keeping older people alive longer, do we? Apart from that basic point, there is also the problem that the evidence for salt reduction is - as far as I can tell - nowhere near as strong as the computer model in this economic simulation suggests. The evidence we have is about a surrogate marker - blood pressure - which can be reduced slightly by the sort of salt reductions proposed here and already in force in the UK for prepared foods. On the balance of probabilities, I'm happy to support salt reduction, as I am carbon emission reduction; but that doesn't mean swallowing every extrapolation that zealots come up with. A paper like this doesn't really belong in the world's leading medical journal.
http://content.nejm.org/cgi/content/abstract/362/7/590

Ann Intern Med 16 Feb 2010 Vol 152
211 The Women's Health Initiative trial was an RCT of hormone replacement therapy which brought about a volte-face in clinical practice but which is described as "far from impeccable" in a letter in this week's BMJ (p.382). Peccability is openly confessed in this Lenten analysis of the effect of continuous combined HRT on coronary heart disease. They more or less admit to residual confounding and small subgroup sizes. The bottom line message is that continuous HRT may confer added risk of CHD in the first years, then decreased risk after 6 years. Which is not quite what we were all initially led to believe.
http://www.annals.org/content/152/4/211.abstract

218 Initiated permanent anticoagulation for severe recurrent superficial thrombophlebitis with the reluctant concurrence of the local haematologist. Such events are benign and self-limiting and do not herald serious thromboembolism, but this French study casts doubts on that. In fact 25% of subjects with superficial phlebitis of 5cm or more had or went on to develop deep vein thrombosis in this series of 844 consecutive cases in a specialist referral centre. We need some primary care studies, quite urgently.
http://www.annals.org/content/152/4/218.abstract