Journal Watch - November 2008

JAMA 5 Nov 2008 Vol 300

2012 The Great Vitamin B Disappointment continues. Folic acid, pyridoxine (B6) and cobalamin (B12) all lower levels of homocysteine, for which reason they were thought likely to prevent cardiovascular disease, dementia and perhaps also cancer when given as supplements. But they don’t. The Women’s Antioxidant and Folic Acid Cardiovascular Study was, as its name suggests, designed primarily to look at cardiovascular end-points, but at the same time it measured the occurrence of invasive cancer in its cohort of 5442 US health professionals. While it was “on the air”, so to speak, everybody in the USA was given extra folic acid anyway as an additive to bread flour to prevent neural tube defects. What we can conclude from this study, therefore, is that vitamins B6 and B12 do not prevent cancer in women, since there was no difference between the active treatment and the placebo group; both groups ended up getting plenty of folic acid.


2022 Before your post-electoral brain forgets all its American geography, turn your attention to Olmsted County in Minnesota (Obama, 54%). This far northern county has been the focus of epidemiological research since 1979, because it has a conveniently small number of medical care providers well isolated from most others. Nearly 30 years ago, sudden death following myocardial infarction was nearly 40% commoner than it is now. But the 30 days following infarction remain a high-risk period for dropping dead (about four times the background risk); this risk declines steadily unless you show evidence of heart failure, in which case it stays quadrupled.

2051 A commentary piece to tag along with last week’s weightier offerings on glucose lowering to control macrovascular disease in type 2 diabetes. Clearly glucose itself is an imperfect target, since drugs like sulfonylureas can lower it while having no benefit on cardiovascular outcomes. So should we be trying to lower insulin instead? Probably, but I’m not convinced by the argument here that this is the real target for treatment, rather than insulin resistance, which in most people is linked mainly to abdominal fat.

NEJM 6 Nov 2008 Vol 359

2005 This trial of transdermal testosterone for low libido in postmenopausal women is given the rather witty acronym APHRODITE, after the Greek goddess of love. Although a little less of a man-eater than her Mesopotamian precursor Innana, or her Roman successor Venus, she nonetheless showed every sign of retaining her libido for the duration of her divine existence. Greek statues of Aphrodite do not show her wearing a 300mcg/day patch of testosterone because goddesses presumably do not experience a menopause. This is the dose which was required to make a noticeable difference to the waning sexual satisfaction of the 814 postmenopausal women enrolled in this trial. They were not taking any oestrogen, but worryingly four cases of breast cancer were diagnosed in the testosterone group during the 52-week trial, compared with none in the placebo group. A larger longer trial would be needed to establish whether this is a significant risk.

2018 The biggest favour you can do for a patient is to help them to stop smoking, and the best drug for the purpose may be varenicline. The evidence that it is better than slow-release bupropion is not huge and it is slightly more expensive, comparing the doses usually prescribed. But there seems no reason not to use it as one first-line drug for breaking this deadly addiction, and it is now approved for this purpose in the USA. This is a comprehensive account of its interesting pharmacology: it’s a modified plant poison derived from laburnum which binds to the same brain receptors as another plant poison (nicotine) derived from tobacco. I have spent many weary hours over the years listening to “smoking cessation facilitators” describe the various hoops through which they expect addicted patients to jump in order to get their ration of ineffective nicotine replacement; whereas the imperative here is to give the person at risk as much as they need of whatever it takes to keep them away from cigarettes, as many times as necessary. They are hardly going to sell it on the streets. For a female smoker (caught young), you will be adding 14.5 years of life expectancy; for a male, 13.2 years.


2025 Pregnancy is a pro-thrombotic state and venous thromboembolism is the leading cause of maternal death in the developed world. Most venous clots develop in the second and third trimesters, but the majority of pulmonary embolism occurs in the puerperium, especially after caesarean section. This is a complicated subject which is dealt with very thoroughly in this review. Remember that D-dimer testing in pregnancy is liable to false positives, though a good quantitative assay (as opposed to a near-patient stick test) still has some utility. Stay DVT aware while you go through the usual antenatal motions.

BMJ Journals Nov 2008

ADC 918 This leading article has the apposite title “why is the evidence not affecting the practice of fever management?”, referring of course to children. It contains a very useful summary of the evidence and comes down heavily in favour of the physiological usefulness of fever in most infections and against the use of antipyretic drugs in most circumstances. This is a culture shift we all need to encourage, as Anthony Harnden said in his recent BMJ editorial.


Gut 1545 The alarm features for colorectal cancer form the subject of a rather disappointing systematic review here: the truly alarming ones are a palpable low abdominal mass and dark red rectal bleeding, but we didn’t really need to be told that. The selected secondary care studies indicate that other clinical features, such as change in bowel habit, rectal bleeding, weight loss and anaemia all have very low predictive value. The problem is that they all need to be investigated nevertheless.


Heart 1419 In 1995, primary care clinics for the secondary prevention of coronary heart disease were a new idea, and the GP department of Aberdeen University was keen to set up a controlled trial of such nurse-led clinics, commendably randomised by individual rather than practice. The ten-year follow-up data show a definite survival advantage to those who attended the clinics, but oddly enough the rate of coronary events began by favouring the clinic group but converged at ten years. Make of it what you will; everybody gets these interventions now, and a good thing too.

Lancet 8 Nov 2008 Vol 372

1633 Generosity makes an unusual appearance as a scientific term in this Swedish study extolling the benefits of a Scandinavian policy which supports dual-earner families. This may well be a good idea but I’m not sure that it is conceptually possible to argue a causal effect between helping both parents to stay at work and lower infant mortality, though it is possible to construct a hierarchy of countries which may indicate an association. This kind of study – like the one on p.1655 purporting to measure greenness – makes the scientist in me feel a bit uneasy, while the social welfarist in me ardently wants at least some of it to be true.


1641 This issue of The Lancet is devoted to health equity, mostly within highly developed but highly unequal societies like the UK, where the Sure Start programme was launched to help children in disadvantaged neighbourhoods. This got a thumbs down for effectiveness in a preliminary study published by these same investigators in the BMJ two years ago, but now they find definite evidence of benefit in the behavioural outcomes they measure.


1648 My greatest medical hero is Rudolf Virchow (1821-1902) who was not only the founder of cellular pathology but also a passionate advocate for the underprivileged. He expressed the view that “medicine is a social science, and politics is nothing more than medicine on a large scale” (see pp.1609, 1626) and his ardour nearly led to public fisticuffs with Bismarck in the Reichstag. This study of what is needed to reduce socioeconomic inequalities in coronary disease mortality in the UK is based on data from the Whitehall studies, which were led by the Virchow of our times, Michael Marmot (see p.1625). His zeal for social equality has been rewarded by a knighthood rather than fisticuffs, though I imagine that Margaret Thatcher’s handbag sometimes developed a menacing swing when he was nearby. The medicine on a large scale that is needed here is just that: current best-practice interventions could practically eliminate social differences in cardiovascular disease, and it’s our job to see that they do.

Ann Intern Med 4 Nov 2008 Vol 149

601 The first great work of literature contains an ironic description of Gilgamesh finding the plant which provides eternal youth and then losing it. What pharmaceutical marketing opportunities there might have been had he not let a snake carry away the pukku plant! One modern attempt to imitate pukku is an oral grehlin mimetic (MK-677) which provokes a pulsatile flow of growth hormone from the pituitary glands of ageing men and women. The trouble is that it doesn’t make them feel any younger, and they become more insulin resistant and get higher glucose levels, which can’t really be the secret of eternal youth.


612 Which would you rather have – hospitalisation or death? Whatever your choice, I doubt whether you would give these two events quite the same weighting; but that is exactly what happens in 53% of cardiovascular trials which report composite outcomes. And nearly three-quarters of randomised trials in cardiovascular medicine indulge in this pernicious practice of lumping together up to four disparate outcomes such as death, coronary intervention, angina or stroke. Moreover they can double count and jumble at will. It is time that leading journals took notice and put an end to it.


659 I have an innate bias against screening for anything, especially when it involves sticking things into people’s bottoms, but there is no getting away from the facts that (a) colon cancer is increasing, (b) the clinical warning features are inadequate (see Gut above), so (c) it usually presents at Dukes’ stage C. The US Preventive Services Task Force has devoted enormous effort to selecting the best test strategies for colorectal cancer screening, but I’m not convinced that anything but universal colonoscopy between the ages of say 55 and 70 will really cut the mustard.

An Eclogue for Obama

I am not a great fan of the Latin poet Publius Vergilius Maronis, usually known as Virgil. On the whole, I think it would have been better for the world had his friends carried out his instructions and burnt the manuscript of his epic poem, the Aeneid, following his death in 19 BCE. On the other hand, when he is not glorifying Roman militarism in the guise of imagined Bronze Age heroes but praising the rural pleasures of peace, he can be delightful (especially on the subject of bees).

From about the fourth century CE onwards, while the Aeneid was rejected as too pagan by many of the early Church Fathers, Virgil’s fame came to rest increasingly on a short poem – the Fourth of his Eclogues (or Bucolics) – written after the Peace of Brundisium between Mark Anthony and Octavian in 37 BCE. This fourth eclogue came to be seen as a prophecy of the birth of Christ:


Now comes the last age of the Cumaean song;
The great order of the ages arises anew.
Now the Virgin returns, and Saturn’s reign returns;
Now a new generation is sent down from high heaven.

A Christian African of the second century, Lactantius, was the first to make the connection, but he was keen to point out that the Sibyl of Cuma, and not Virgil himself, was the instrument of this prophecy. Later Christian commentators did not maintain this distinction, and so Virgil appears in Dante as the noblest of pagans and the poet’s guide through hell, while two centuries later Michelangelo has the pagan Sibyls partnering the Hebrew Prophets around the Old Testament scenes on the ceiling of the Sistine chapel.

The poem is addressed mainly to a consul called Pollio, during whose term of office the sacred child will be born:

And indeed, Pollio, during your consulship
This glory of the age will enter in,
And the great months will begin to advance;
While you lead, if any stains of our sins still linger,
Their removal will free the earth from endless fear.

Two phases of the golden age will then follow under the new-born child: an age of trade and plenty, though marred by last traces of heroic war, and then a final phase during which trade will cease because every land will be self-sufficient, and nature so bountiful that even farming will no longer be needed.

JAMA 12 Nov 2008 Vol 300

2123 In 1997, over a quarter of a million US male physicians were invited to participate in a trial intended to settle the question of whether vitamins E and C are cardioprotective, as observational evidence and antioxidation theory suggested strongly that they were. From this colossal congregation of medical men, a mere 14,641 aged 50 or over were eventually randomised into this factorial study, which then ran for a mean follow-up of 8 years. The 1245 cardiovascular events which occurred were fairly evenly scattered across the four groups who received placebo only, vitamin C only, vitamin E only, or both vitamins, though there was a just significantly higher incidence of haemorrhagic stroke in the vitamin E group. So, taken together with other C and E studies and the various B vitamin trials, the Physicians’ Health Study II just about rules out all vitamins for cardiovascular protection; with the notable exception of vitamin D which hasn’t been put through any adequate randomised trials.

2134 This week’s BMJ contains a nice little piece by a pair of Queenslanders who were sceptical about the evidence base for giving aspirin to all people with type 2 diabetes, and by coincidence JAMA here publishes a Japanese study which very much justifies their scepticism. 2539 people with type 2 DM but no clinical cardiovascular disease were randomised to get either open-label aspirin or non aspirin, and their cardiovascular outcomes did not differ significantly over 4 and a half years.

2142 We do a lot of fasting lipid profiles on patients, often without any good reason (e.g. because they are taking statins), but the lipids best measured without fasting are the triglycerides. I tend to ignore them except as an indicator of impending diabetes, but the Copenhagen City Heart Study (14,000 men and women of all ages, 1976-2007) shows that increasing TG levels are associated with increasing risk of stroke, with a doubling of risk in the highest quintile compared with the lowest. The trouble is, I don’t think our local lab will even measure triglycerides on a non-fasting sample, and I’m not sure either what I’d do about them if they came back high.

2161 In one of the various rounds of tweaking that the UK Quality and Outcomes Framework has undergone, mainly in an attempt to reduce GP earnings, we have been called on to screen patients with known cardiovascular disease for depression. We know that patients with CVD fare a lot worse if they are significantly depressed, but this systematic review attempts to discover some evidence for the notion that screening for depression will improve outcomes. Far from unearthing any such thing, it discovers that (a) there are no clinical trials whatever that measure the effect of depression screening in this group, either on depression or cardiovascular outcomes and (b) trials of antidepressants and cognitive treatment show modest improvements in depression but none on cardiac outcomes.

2172 Are you confused by dementia? Do you find it difficult to make plans? Sometimes forget your car keys? Don’t worry: a leading expert on dementia, Vladimir Hachinski (of the Hachinski score, 1975) says that the distinctions between Alzheimer’s and vascular cognitive impairment are mainly artefactual and clinically of little importance, since the main way to prevent both is to improve vascular health. Golly, did I take my statin last night? And just what is the evidence about daily aspirin and cognitive impairment?

NEJM 13 Nov 2008 Vol 359

2105 An Italian friend tells us that her father (84 and going strong) finds his abdominal bulge very useful for resting a glass of wine while he reads his newspaper. Such secrets of contentment were not bruited in the press coverage of this paper about adiposity and the risk of death in Europe, which was full of alarm at the idea of a connection between tummy girth and dying; and I must confess to going in search of a tape measure while I read it. In fact it only tells us what we knew already, but with tighter confidence intervals since it was a study of EPIC proportions (n=359,387). I suppose we’ll go on measuring BMIs for years to come because they are in so many guidelines and they pop up on our computers whenever we weigh and measure a patient: but belt length matters more, of course. Keep drinking the wine, but buy a side table.

2121 I suppose all advances in invasive brain procedures involve a period of trial and error and the risk of major adverse events, but all the same this trial of subthalamic nucleus stimulation for severe obsessive-compulsive disorder makes disturbing reading. Fifteen serious events in sixteen patients is not good news. All patients underwent subthalamic electrode insertion (with one intracerebral haemorrhage and two cases of infection) and were then randomised to a period of stimulation followed by a period of sham stimulation, or vice versa. During the real stimulation periods their OCD improved a bit but there was no alleviation of anxiety or depression. Not much to show for a very risky bit of brain surgery.

2135 A tear of the anterior cruciate ligament often leads to osteoarthritis within 10-20 years, and because these injuries frequently occur in young people, that can mean knee misery in middle age or sooner. Moreover, females who indulge in team sports are at higher risk than males. Beware the hobbling hockey mom. This clinical review is a reasonable guide to diagnosis of ACL – listen out for a history of twisting injury without contact but with a distinct feeling of something snapping, followed by instability and swelling. You can try Lachmann’s test, but it isn’t reliable: with a history like that you might as well go straight for MRI. As in most orthopaedic surgery, a variety of repair techniques is used, without any real evidence to guide a choice between them (this sentence is a grammatical teaser for pedantic readers).

2143 If you are looking for a good update on what is currently known about inherited susceptibility to common cancers, here it is. Little bits of risk from common alleles you can’t possibly remember, or big lumps of risk from rare alleles you’ll probably never encounter. Prime time for this stuff has not arrived, despite the salesmanship which is already going into genetic cancer profile testing, which threatens to discredit the whole enterprise. The author wistfully observes that “it would be a great loss if policymakers and the public were to tire of genes and gene-based medicine as a whole, because genetics has shown itself to be an enormously powerful tool in the discovery of new knowledge.” If only it were not so enormously boring and inaccessible for jobbing doctors and nearly everyone else.

Lancet 15 Nov 2008 Vol 372

1746 The onset of type 1 diabetes is heralded by the appearance of autoantibodies to islet cells, protein tyrosine phosphatase-like protein (sic) and glutamic acid decarboxylase, mainly in individuals with HLA-DR or HLA-DQ haplotypes. You need to remember this nerdish sentence in order to understand the current direction of research into this devastating auto-immune disease of children and adolescents. Otherwise the idea of getting children to sniff insulin to prevent or postpone the onset of diabetes might seem barking mad, rather than merely unsuccessful, as it proved in this trial. Finnish children with the above-mentioned HLA haplotypes were identified at birth and followed up with auto-antibody tests every 3-12 months. Those who showed positive to two different autoantibodies were then randomised as soon as possible to nasal insulin, in the hope of inducing immune tolerance, or placebo. The study was terminated early when it became clear to the steering committee that the insulin wasn’t preventing diabetes.

1756 The study which ended the career of rofecoxib was actually a success in terms of its original end-point, which was Adenomatous Polyp Prevention on Vioxx (APPROVe). Vioxx did indeed suppress the recurrence of neoplastic polyps of the colon, but, as we all know, at the expense of an increase in serious cardiovascular events. The original trial protocol stopped counting events 14 days after discontinuation of study treatment, but this final analysis looks at what happened for at least a year afterwards. The message is much as we have come to expect: rofecoxib carries a CV risk which persists for about a year after cessation, which may be greater than with other NSAIDs but not by very much.

BMJ Journals Nov 2008

BMJ 1159 Here’s a review of statin induced myopathy, a subject which is always worth reading about, though this is not one of the better accounts. It mentions but doesn’t attempt to explain the puzzling difference between muscle side-effects in randomised trials (1.5-3%) and prospective clinical studies (10-13%); ditto the general dissociation between symptoms and levels of creatine kinase. “Myopathy” can mean anything from vague muscle fatigue to life-threatening rhabdomyolysis. The evidence base for some of the authors’ recommendations is distinctly shaky – for example, switching from a lipophilic to a hydrophilic statin, trying an alternative lipid-lowering agent (though none of them has been shown to lower cardiovascular risk) or trying coenzyme Q on the basis of a tiny trial. Since the entire population (over ?55) should be on a statin, these are important matters.

ADC 936 When Prof Sir Roy Meadow stood up in court and said that the chance of two cases of Sudden Infant Death Syndrome in the same family was one in seventy-three million, he virtually assured the conviction of Sally Clark, who later died from an alcohol overdose following her subsequent acquittal. It is a terribly sad story which drives home the importance of a grasp of basic statistical concepts, and not just by doctors: as Ben Goldacre points out in one of the best sections of Bad Medicine, lawyers and journalists also failed to spot an obvious error, even in the appeal hearing. The choice facing the court was between double SIDS (very rare) and double child murder (even rarer). That said, exactly how rare is SIDS in a family where it has already happened? Paediatricians and Bayesian statisticians from Sheffield here do a follow-up to their earlier systematic review of the evidence. Since the same risk factors usually apply to the second child as to the first, the risk is always higher than in the general population: about 1:456 in the typical high-risk scenario of a smoking mum under 27 with several children and no income.

Heart 1485 The wonders of cardiac MRI (CMR) are showcased in well-illustrated review here. A single examination can give you an accurate picture cardiac anatomy, function, perfusion, viability and physiology . The trouble is that it remains very expensive and needs a whole team to perform.

Arch Intern Med 10 Nov 2008 Vol 168

2183 A couple of weeks ago I was musing about thrombolysis for pulmonary embolism – a topic of huge potential importance about which we don’t know nearly enough. There were some randomised studies in the 1970s which were too small to reach significance, and this analysis of outcomes from admissions for PE in Pennsylvania does not get us much farther. The patients with sub-massive PE who were given thrombolytics had higher mortality, after adjusting for a number of variables; but the accompanying commentary (p.2191) casts doubt on the accuracy of the propensity analysis and hence the significance of the differences. There is almost certainly a group of patients with major PE who need thrombolysis, but we don’t know who they are.

2261 We’ve become used to the idea that wrist fractures are a predictor of other osteoporotic fracture, and so they are, but perhaps less strongly than we supposed. Here are ten-year follow-up data for 2652 women in the Manitoba Bone Density Program: all of them had a primary fracture prior to bone density measurement, and in nearly half this was a wrist fracture. Such fractures did not predict the risk of further fractures nearly as much as a primary fracture of the hip, humerus or vertebrae. For example, a woman who has had a vertebral fracture has a 25% risk of a further osteoporotic fracture whereas a woman who has broken her wrist has a 14% ten-year risk (compared with a 10.8% risk in those without previous fracture).

Fungus of the Week: Pleurotus ostreatus

The main fungus season has now ended, but the main season for this attractive, edible fungus is just beginning. You can buy beige, yellow and pink species of oyster mushroom in supermarkets all the year round, and it would be nice to claim that this blue-grey capped one was markedly superior, but I’m afraid it isn’t. The one which grows every year near us is decidedly inferior, in fact.

We have no means of knowing when fungi entered the human diet, except perhaps by the study of prehistoric coproliths (fossilised poo), though I am not sure how feasible it is to look for large aggregations of edible fungus spores a few microns across in the middle of a poo-stone. I have just read nearly 500 pages of Steven Mithen’s lively and informative book on the transition from hunter-gatherer to farming societies, After The Ice: a global human history, 20,000-5000 BC, in which scenes of fungus foraging are frequently described, but he doesn’t cite the evidence. If you look up “Paleolithic (or Palaeolithic) diet” you get a lot of websites written by people wearing anoraks made from animal skins or bark.

Palaeolithic hunter-gatherers certainly knew a lot about the hazardous business of distinguishing edible plants from poisonous, and which roots and seeds could be made edible by prolonged cooking. It seems highly likely that they collected fungi and occasionally died as a result of their experiments. One principle they are likely to have discovered early on is that in subarctic and temperate woods, no fungus that grows on wood is likely to kill you, with the possible exception of the sulphur tuft, which is said to be very bitter anyway. Moreover wood-growing (lignocolous) fungi are more likely than ground-growing fungi to fruit in the same place and in useful quantities every year. Our rude forebears may even have practiced a bit of fungiculture by pulling other dead trunks on to those bearing oyster mushrooms, or scattering wood chippings around them. Nowadays it’s done around the world on an industrial scale.


JAMA 19 Nov 2008 Vol 300

2253 The heyday of the ginkgo tree was the Jurassic period, when forests of Ginkgo adiantoides towered over even the largest dinosaurs. The whole genus was thought to be extinct until the eighteenth century, when G biloba was discovered in remote eastern China; following which ginkgos have enjoyed a second golden age, as widely planted ornamental trees around the world. Moreover this ancient tree is widely believed to contain a mystical elixir which prevents dementia. If only. This US trial in 523 people aged 75 and over had a follow-up period of six years and in that time slightly more of those randomised to ginkgo extract developed Alzheimer’s or other dementia.

2277 Bevacizumab is a monoclonal antibody directed against vascular endothelial growth factor, and for mab-challenged readers I should mention that it is widely used as a last-ditch agent against a variety of cancers. It has acquired a reputation for increasing the risk of venous thromboembolism, and this meta-analysis wrestles with the question of how big an added risk this might be in patients with a variety of malignancies which can themselves increase the rate of VTE, and which often kill patients in the time scale of the trials. Bevacizumab probably increases VTE by about a third overall: what we can do about it is uncertain.

2286 We know that morbid obesity is bad for fertility, so does bariatric surgery improve pregnancy rates? Unfortunately this systematic review finds that existing studies don’t permit a definite answer, though it’s likely that women who have lost a large part of their excess weight following bariatric surgery have fewer adverse maternal and neonatal events.

NEJM 20 Nov 2008 Vol 359

2195 In the JUPITER trial, the manufacturers of rosuvastatin recruited 17,802 subjects with healthy levels of LDL cholesterol – men aged 50 and over and women aged 60 and over – with slight elevation of high-sensitivity C-reactive protein (>2mg/dl) and gave placebo to half of them and 20mg of their product to the rest. Although healthy individuals to meet these criteria abound in every location, AstraZeneca involved 1315 sites in 26 countries – perhaps in honour of the god Jupiter, who was a well known practitioner of seeding. Indeed, this was one of his main roles as father of the gods (for a naughty illustration, see Wikipedia). Another of his roles was to make a big noise as the thunder god Juppiter tonans. This trial certainly did that, in exactly the way its sponsors would have liked. But does it really prove, as all the newspapers reported, that we should give rosuvastatin to everybody from middle age onwards? That would certainly reduce major adverse cardiovascular events, but the strange design of the study makes it hard to extrapolate to the whole population, or indeed any definable clinical population (how often do you measure high-sensitivity CRP?): an editorial later in the journal (p.2289) raises some pertinent questions, but for once the editorial in the BMJ (p.1182) is even better, by Jove.

2208 In this week’s Lancet, a Profile of Sir John Bell, Regius Professor of Medicine at Oxford and President of the Academy of Medical Sciences, has him confidently asserting a future for “personalised medicine”, meaning medicine in which you don’t so much encounter persons as predict illnesses from their genome. To you and me, personalised medicine for type 2 diabetes means meeting a lot of mostly fat people over 60 whose arteries tend to fur up, whose feet go numb and septic, and whose eyesight often fails, and helping them cope as best they can. But the task for prestigious doctors is to predict type 2 diabetes from a genotype score. To judge from the study here and the one that follows on p.2220, they are rubbish at it

2245 A recent study showed that dropping dead after myocardial infarction has become less common, but it is still a considerable risk if you develop left systolic dysfunction with adverse cardiac remodelling, or if you show a high rate of ventricular arrhythmias on ambulatory monitoring weeks after the event. In such cases you might do well to be fitted with an implantable cardioverter-defibrillator, unless your heart failure has reached the point at which you would prefer to drop dead anyway. This clinical review discusses the indications for ICDs following MI in detail.

Lancet 22 Nov 2008 Vol 372

1835 The best thing – in fact the only thing of general interest – in this week’s Lancet is this seminar on age-related macular degeneration. It is a disease which is commonest in white people and smokers, and of course it increases with age. No preventive treatment has been shown to be effective, and regimes which include beta-carotene carry an increased risk of lung cancers. Only a few individuals will benefit from laser treatment, though lots benefit from photodynamic therapy with verteporin, except for an unfortunate 3-4% who experience severe deterioration. As for the place of intravitreal becavizumab (relatively cheap but unlicensed) versus the vastly more expensive ranibizumab, there are no current trial data, but a head-on randomised trial has at last begun.

BMJ Journals Nov 2008

ADC 952 Here are the results of a national UK audit of children admitted with acute wheeze/asthma during the month of November in the years 1998-2005: interesting reading for anyone who looks after wheezy kids. There are wide variations in practice between hospitals in the UK, but the most consistent finding is a reduction in the use of nebulised therapy over recent years in favour of therapy delivered through spacers. Too few children admitted with asthma go home with clear treatment plans or with their inhaler technique properly checked.

Heart 1364 Just now, aspirin needs all the friends it can get, with John Cleland on his familiar anti-aspirin warpath in the BMJ correspondence and the Polypill in danger of dwindling down to an evening dose of statin. This editorial revisits the evidence for giving aspirin to all over the age of 50. It is curiously non-committal though it makes the important point that although aspirin carries a risk of bleeding, such events are usually much less serious than the events which aspirin prophylaxis may prevent, particularly stroke.

Heart 1407 Back in the 1970s when I was a house officer, medical takes consisted of a stream of men in their forties and fifties who had myocardial infarction and who would have had a lower mortality had they been kept at home. Now that hospitals deliver treatment that actually helps patients rather than kills them, how do most people with ST-elevation MI fare in the UK? In this national audit, 90% survived their stay: those who were given aspirin immediately and had out-of-hospital thrombolysis had the best outcomes.

Ann Intern Med 18 Nov 2008 Vol 149

698 Following deep vein thrombosis, many people get varying degrees of leg discomfort. This Canadian study set out to investigate the size and time-course of the post-thrombotic syndrome, including swelling, cramping and heaviness as well as pain. The bigger and higher the clot, the more severe the symptoms, which did not change very much over the 24 months of the study: mild in 30%, moderate in 10%, severe in 3%.

708 The default setting in type 1 diabetes is death, and most GPs will recollect some young patient who couldn’t come to terms with the condition and succumbed to insulin coma or, less commonly, to ketoacidosis. It can be something of a nightmare trying to improve motivation in such patients, and that is confirmed in this British study of motivational enhancement therapy with or without cognitive behavioural therapy versus usual care for poorly controlled type 1 diabetics. Nurse-delivered motivational therapy on its own had no effect, while combined with CBT it delivered a drop in %HbA1c of 0.46. Not much to show for an unfeasible amount of encouragement.

720 You may be a doctor, but do you have healing skills? This study looked at the common ground between 50 practitioners considered particularly good at relating to their patients: 10 of them were from the “alternative sector”. If you’ve ever done any communication skills training, you won’t be surprised at most of the themes that emerge, though you might prefer them served up with a smaller helping of corn syrup. If you start with a real interest in your patient, most of the rest will follow.

734 It’s a rare day in general practice when you don’t have to decide on which second generation antidepressant to prescribe, and here is a massive evidence base to help guide your choice: a 37-page analysis of 299 studies, prepared for the American College of Physicians. The result: they all work equally well, as far as anyone can tell. Mirtazapine probably starts working faster than most others: bupropion (which for some reason isn’t licensed for treating depression in the UK) has the fewest sexual side-effects, paroxetine the most. And about the important issue of habituation and withdrawal, we don’t seem to know nearly enough.

751 We spend a fair amount of time trying to share decision-making with patients, and I suspect that if we had twice as much consultation time, we would do it twice as well. As it is, I’m constantly on the search for short cuts to explain risk to people in a meaningful, non-biased way, so I pounced on this paper eagerly. Alas, it provides a long cut rather than a short one: risk stratification tables may be good for academics wishing to refine their methodology, but of little use to jobbing doctors.

Plant of the Week: Jasminum nudiflorum

There isn’t much to say about winter-flowering jasmine: it is common, propagates with the greatest ease, provides a bit of yellow cheer through December and January, and is best planted where it can sprawl at will. But what a fantastic plant it could be with only the simplest bit of genetic engineering. A slight change to its colour genes and it could be a soft cream, or even pink; a gene splice from almost any other jasmine, and it could carry the scent of summer evenings into the depths of winter; another might make it evergreen. Come on, all you gene gnomes: we need you in our gardens.

JAMA 26 Nov 2008 Vol 300

2379 We know that depression worsens outcomes in people with coronary heart disease, but we don’t really know how or why. The Heart and Soul Study followed up just over a thousand people with CHD for just under five years in order to find out. The answer is quite basic and physical: if you are depressed, you do less. Lack of exercise is the main factor, and the rest of the added risk is accounted for by other simple physical causes, such as more smoking and a poorer diet. In other words, depression in itself doesn’t cause cardiovascular risk, but it drives behaviours that do.

2398 If you are unfortunate enough to reach the end stage of heart disease, you will either die or – in rare cases – be given a ventricular assist device. Even in the USA, the number of Medicare beneficiaries getting these devices was less than 3,000 between 2000 and 2006. About half of these patients had devices fitted acutely when cardiac surgery went wrong, and 60% in this group died in hospital. If they survived to leave hospital, however, they did slightly better than patients who had elective VAD insertion. And although these devices are often thought of as a “bridge to transplantation”, fewer than a quarter reached transplantation; but about half of those who made it out of hospital were still alive five years later.

2407 Chronic obstructive pulmonary disease involves chronic inflammation, which for a long time we have sought to suppress with inhaled corticosteroids. But – and here this systematic review ties in nicely with NEJM’s review of the role of infection in COPD – what if the inflammation is driven by bacterial colonisation and we are simply suppressing a protective response, a bit like treating skin infection with Betnovate? Looking at the randomised trials, it seems that we may indeed be doing more harm than good: rates of pneumonia were a third higher in those randomised to inhaled steroids.

2417 This two page commentary on the benefits and risks of drug treatments deserves to be read by everyone who prescribes anything – a neat explanation of why randomised trials are the best way to measure the intended benefit of a drug treatment, but a poor way of measuring adverse effects. What we need is a cycle of drug evaluation which measures and integrates the benefits and the risks throughout the entire market life of each drug. Adverse effects may be uncommon and unanticipated (or brushed under the carpet by trial sponsors) and are best picked up by observational studies in large populations.

NEJM 27 Nov 2008 Vol 359

2305 Innovation in Primary Care – Staying One Step ahead of Burnout is a title that’s bound to catch the eye of every UK GP, however many steps past the threshold we may already be. It’s an account of the joys and griefs of working in 114-doctor “medical home” run by Kaiser Permanente, hyped by Richard Smith as the fountainhead of all good medical management until he left the BMJ to work for United Health. Here are the travails of primary care doctors who see 15-20 patients a day in 20 minute appointments, and have to do more paper work all the time. A quarter of patient contacts are already via e-mail.

2324 Do you want to know how healthy your coronary arteries are? The cheapest way is to fly off to India and get a 64-row CT coronary angiogram. According to the most popular Indian website, it has a 99% negative predictive value, which may or may not be true of the kind of people who fly to India to have coronary CT. In this study the negative predictive value was 83%, against the gold standard of conventional angiography in a population of patients referred because of symptomatic suspected coronary artery disease; missing 17% is not really good enough in this situation. And for asymptomatic patients, knowing about coronary lesions is seldom worth a hefty dose of radiation (not mentioned on the website) and the subsequent anxiety and interference: as a letter in last week’s NEJM memorably observed, “Coronary stenoses are like chocolate- they are very difficult to leave alone.”

2347 What is it about ophthalmologists and the furthest reaches of the Greek tongue? I did O-level Greek but I had no idea what “rhegmatogenous” could possibly mean, though the author of this otherwise exemplary review of primary retinal detachment assumes that we all know. At least he is kind enough to explain aphakia and pseudophakia. A rhegma is a rip or tear or rupture, and retinal detachment commonly occurs due to a retinal tear and a build up of fluid behind it. So the treatments consist of putting a gas bubble into the vitreous, or occasionally some silicone gel; or else buckling the posterior eyeball inwards behind the detachment site. Scleroplasty, vitrectomy or pneumatic retinopexy: these are your options for rhegmatogenous detachment in plain ophthalmological English. But be specially careful in cases of pseudophakia.

2355 The “British hypothesis” of progression in chronic obstructive pulmonary disease in the 1950s and 1960s was that it was due to infection – not surprising considering we all called it “chronic bronchitis” in those days and defined it as three or more chest infections a year. All this changed in the 1970s, when the role of smoking was emphasised. It became bad form to prescribe mucolytics and long courses of preventive antibiotics. Now things have come full circle, and inhaled steroids and cholinergics are under a cloud, while mucolytics have proved a boon in a long Chinese study and antibiotics may be due for a comeback in the wake of this superb review of the role of infection in COPD. Smoking paralyses mucociliary clearance, allowing bacterial colonisation to occur in the smaller airways, which are normally sterile. This sets up an inflammatory process which is then driven further by bacterial exacerbations: Haemophilus influenzae is the biggest culprit, with Streptococcus pneumoniae and Moraxella catarrhalis in joint second place and Pseudomonas aeruginosa bringing up the rear in advanced disease.

Lancet 29 Nov 2008 Vol 372

1881 Gloom has descended on the prospects of vaccine against HIV-1 over the last year, and it will only be deepened by the failure of the Step Study to show any benefit from a cell-mediated vaccine – in fact the vaccine may have done harm. The Lancet sounds a chirpy scientific note on its front cover: we have gained knowledge, this may be a step to future success. But if I were the health minister for say Malawi or Namibia, that might not be enough to cheer me up just now.

1906 Meanwhile in the rich world, obesity looks set to become a major killer, and an effective anti-obesity drug is probably the biggest jackpot any pharmaceutical company can hit. Tesofensine inhibits the pre-synaptic uptake of noradrenaline, dopamine and serotonin, and was first tried out in patients with Parkinson’s and Alzheimer’s diseases: the fat ones lost weight, and this phase II double-blind RCT in 200 otherwise healthy obese Danish adults quickly followed. Bingo! Twice as much weight loss as with sibutramine (now the only licensed drug) and fewer adverse effects – perhaps. We’ll have to be patient and await the phase III trials.

BMJ 29 Nov 2008 Vol 327

1272 Most of my generation learnt Bayesian diagnostic thinking from the chapter in Clinical Epidemiology about exercise electrocardiography; and for some days, nay weeks, later we nervously fiddled with the little nomogram ruler hidden in a marsupial pouch inside the back cover of the book. This study shows how far the march of evidence-based medicine has advanced. We do more treadmill ECGs than ever – in fact they are done on every patient we send to the rapid chest pain clinic, and all these patients need to come armed with a mandatory resting ECG too. Now this study proves what we long suspected – that the resting ECG adds no useful information at all, and the exercise ECG is nearly as useless for the great majority of patients. And still the editorial accompanying the paper tells us that we should keep doing them. Why bother with EBM at all if it’s never going to change practice

1287 For another example of how hard it is for evidence to pierce the armour of medical habit, take melanoma prevention. We all know that the key is reducing sun exposure, don’t we? Actually it’s extremely difficult to prove any such thing, although melanoma rates are certainly higher in areas where pale-skinned people have colonised sunny places. But studies trying to correlate individual sun exposure and risk of melanoma draw a blank. And contrary to hype, melanoma mortality rates are stable or falling, while the diagnosis of early melanoma keeps rising due to earlier detection.

Arch Intern Med 24 Nov 2008 Vol 168

2311 The Nordic countries were the first to introduce whole-population mammographic screening for breast cancer, and it was from there that the first strong expressions of scepticism were heard. Better still, a group of Norwegian investigators decided to find out about the natural history of screening-detected lesions while their national mammography programme was being introduced a few counties at a time. If this is a subject that interests you, you really need to read the whole paper, and also the excellent editorial about it on p.2302. Essentially, screening every two years detected 22% more lesions than were present in an exactly matched cohort of women who had a single screen at the end of the same 6 year period. The presumption is that a proportion of screening-detected “invasive cancers” regress spontaneously though there is of course no way to tell which they are, or to test this hypothesis by failing to remove them all.

2368 We’ve pretty well all stopped prescribing rosiglitazone, but most of us continue to prescribe pioglitazone, keeping our fingers crossed that its cardiovascular safety is appreciably better. The matter is not settled with this study of elderly Medicare beneficiaries, but at least it points in the right direction. Over 28,000 of them received either of these two drugs between the start of 2000 and the end of 2005. After much adjustment for patient characteristics, rosiglitazone was associated with a very modest increase in heart failure and mortality (13-15%) but strokes and myocardial infarcts were the same in both groups.

2377 Many cancer patients become anaemic and until recently we happily treated them with either erythropoietin mimetics or blood transfusion. The erythropoietic drugs are now known to increase mortality, so what of the transfusions? This study of patients in hospital with cancer also shows a significant mortality hazard (OR 1.34) and in particular an increased risk of thromboembolism, both venous and arterial.

Plant of the Week: Daphne bholua

This is scarcely a time to be writing of plants, when you can barely see the trees as they hide in the fog, and the ground is all freezing mire. But what is this perfume that hangs so richly on the dank hyperboreal air? None other than the breath of this daphne that I always tell you about as winter sets in, a shrub which grows larger and more floriferous with every year. If I haven’t persuaded you to get your own by now, I must despair of you.






Page last edited: 13 March 2009