Journal Watch - August 2009

JAMA  26 Aug 2009  Vol 302

849    What’s the connection between the pharmacogenomics of clopidogrel and religious dissent amongst devout Protestants in rural seventeenth century Switzerland? The answer is the Amish people of North America, a gift to genomics because they have only married amongst each other since they came to America around 1700 to pursue a godly life and produce 6.8 children per family. Although they shun the trappings of modern civilisation they do allow their blood to be taken in the interests of medical science, and this has allowed confirmation of the link between polymorphisms of the CYP2  gene and clopidogrel platelet activity. This was then confirmed as clinically important by a study of patients receiving percutaneous coronary intervention at Sinai Hospital in Baltimore, with a twofold event rate in patients who can’t metabolise the pro-drug clopidogrel into its active moiety. The investigators must have gnashed their teeth when the NEJM published a similar study earlier this year. We are supposed to interest ourselves in the question of whether genotyping should be done on patients receiving clopidogrel, but most of us would just like to know if prasugrel is going to replace it, and how soon.

866   If I grab a can saying “Baked Beans” I expect some nice slimy tomato-covered things to pour on buttered toast or eat with sausages. Similarly, if I read a paper called “Hormonal therapy use for prostate cancer and mortality in men with coronary artery disease-induced congestive heart failure or myocardial infarction”, I expect to learn something about the risks of hormone treatments used mainly for advanced prostate cancer. I don’t expect to be told about a subset of patients who have received prophylactic neoadjuvant treatments following brachytherapy for localised prostate cancer and their survival stratified by coronary disease history. The study shows that these treatments, including goserilin and flutamide, provide no survival benefit generally to men receiving radioactive prostate implants (rare in the UK) and specifically kill off men according to their degree of coronary disease – which is interesting, but not what it says on the can. JAMA needs to get its labels sorted: Baked Beans, please, not lima beans in mushroom sauce.

883    We are living longer, which may be a mixed blessing, but surprisingly the rate of hip fractures is falling in Canada, which is an unmixed blessing, especially for old Canadians. Men as well as women. There has been a steady decline from 1985 onwards, more rapid in recent years. The investigators don’t know why. Keep your BMI around 28, go for walks and take vitamin D, say I.

NEJM  27 Aug 2009  Vol 361

849   Last week I promised you another mega-study with Harlan Krumholz among the authors, and here it is: a survey of exposure to low-dose ionizing radiation from medical procedures in the Great American Public aged 64 or under. There is a lot of it about. The brilliant Perspective piece on p.841 shows how easily one investigation can follow another in a seemingly logical progression which adds nothing to clinical judgement but exposes the patient to a lot of radiation, especially from CT scanning and nuclear imaging. We base our assessment of the risk from this on an assumption of continuously distributed and cumulative effects, carefully examined by various scientific heroes of mine like Joseph Rotblat, Alice Stewart and Ernest Sternglass in the 1950s and 60s. But it is an inexact science, in part based on extrapolation from accidents and atrocities such as nuclear plant leakages, Hiroshima and Nagasaki. We don’t know how many people will get cancer from medical radiation but the numbers are bound to go up until we are more careful.

859   I can remember being dazzled by the appearance of cardiac troponin assays about ten years ago: it was obvious from the start that they would transform the diagnosis and management of myocardial infarction, and it took very little time – about two years in the UK – for them to be universally adopted. But the problem from the start has been that they can only provide a diagnosis after the main opportunity for treatment. Current assays measure the main surge of troponin released by dead or damaged cardiac myocytes, 6-12 hours after the event: this study and the one after show that sensitive troponin 1 assays can measure troponin release as early as three hours after the event. Still a bit late for the ideal interventions, but bound to improve clinical management.

Lancet  29 Aug 2009  Vol 374

This issue of Britain’s senior medical journal is devoted to chronic obstructive pulmonary disease and showcases everything that is wrong with our understanding of COPD and everything that is wrong with The Lancet. For me the best sentence comes at the end of an editorial written by four New Zealanders: “We propose that, as with asthma, it is time to abandon COPD as a disease concept, and to define, identify, and treat the different disorders that make up this complex syndrome.”
685   But hang on, why redefine a syndrome complex when there is so much money to be made from selling new drugs for it? In real life, we know that COPD is largely irreversible lung damage caused by smoke inhalation, and we treat it by getting rid of the smoke and by various weakly active symptom-relieving drugs, with antibiotics for infective exacerbations. The pompous cover of the journal asks whether phosphodiesterase 4 inhibitors might be the Holy Grail for COPD treatment and actually reduce mortality. There is nothing whatever in the journal to suggest this is true, and the two studies of roflumilast published here (both paid for by its manufacturer, Nycomed) simply show that it might have some effect on the rate of exacerbations, depending on how you define them.

695    The fact is that we are desperate to try anything that might help people with moderate-to-severe COPD, so any tin mug can pass for the Holy Grail. Roflumilast in combination with salmeterol or ipratropium can improve FEV1 by all of 49ml and improve some patient-reported outcomes at the expense of diarrhoea, nausea and weight loss. A Holy Grail from Monty Python rather than Parsifal.

704   But what we label COPD is a life-shortening disease, and if we are to classify it more intelligently and treat it more effectively, we need better prognostic scoring systems. In themselves, these achieve nothing, but at least they can point the way to a better analysis of the disease process. But I don’t much like the ADO index which is promoted in this study from Spain and Switzerland. It performs better than the BODE index but that is because it incorporates data about drug treatment and hospital admissions. In other words, it depends on a mixture of objective characteristics such as BMI and airflow obstruction, and doctor behaviour in the shape of treatment decisions. This may predict prognosis but it doesn’t help us to understand the determinants of mortality. I fear it’s a case of much ADO about nothing.

712    At one time, our local chest physician (one of the best and kindest doctors I have ever met) advised us to try out a course of oral steroids on patients with bad COPD to see which ones responded, and then convert them to inhaled steroids at high dosage. Then things went quiet as studies seemed to show that this increased the rate of community-acquired pneumonia. Several meta-analyses of the randomised trials of inhaled steroids in COPD have confirmed this, but this meta-analysis of individual patient data in trials using budesonide finds no additional risk. In fact the data concerning mortality and inhaled steroids are reassuring and if your patient functions better when using them, I personally wouldn’t stop them.

721   Earlier in the year I had a go at the Quality and Outcomes Framework for diabetes, and one respondent who defended the QOF pointed out plaintively that there were many sillier targets in it than the HbA1c of less than 7. Indeed there are, and some of them concern the detection and monitoring of COPD using spirometry. Because we get paid according to our prevalence figures, this amounts to an incentive to screening, though it doesn’t feature in the “British experience” section of this article about screening for COPD. The main argument in this rather diffuse paper is that objective proof of lung damage is an incentive to people to stop smoking. Fair enough: but beyond getting them to stop, we have no disease-altering interventions to offer them. Though we might have in the future, perhaps, according to a wide-ranging survey of new drugs for exacerbations on p.744. Arise, Sir Perceval, for yonder gleam may be the Holy Grail, I wot.

733   Then again, not all COPD is related to cigarette smoke: dung smoke will do nearly as well, if you live in a house full of it. This is politely referred to in this article as the burning of biomass fuel. There are other environmental associations with COPD in non-smokers too, and in many ways this review of COPD in non-smokers is the most interesting paper in this week’s Lancet.

BMJ  29 Aug 2009  Vol 339

488   If you’ve ever had a sleepless night with a raging sore throat, gagging on your own saliva, then next time you might do well to take a handful of corticosteroids. Although a specific dosage isn’t given in the printed text of this meta-analysis of the RCTs, the trials typically used generous amounts like 8mg of dexamethasone or 60mg of prednisolone. Nobody knows how safe this strategy is from a clutch of small studies, but I think I’ll do this for my next patients with really horrible, raging pharyngitis. Twenty-four hours of relief are worth a lot in this situation.

491   As readers of Easily Missed will know, I’m keenly interested in timely diagnosis in general practice, as indeed every GP should be. But there is so little literature to guide us in a practical way. This interesting paper about alarm symptoms in general practice again needs to be read in the full text version rather than the printed one, which is hard to follow due to compression. Most patient s with haematuria, dysphagia, haemoptysis or rectal bleeding do not have a cancer diagnosis, but that should not lessen our alarm. Some of them do, and many others have diagnoses that need timely intervention. We need further, and more up-to-date, database studies of this kind.

494    As an old lag GP living on the margins of academic life, I found this paper about novel methods to deal with publication biases inordinately difficult to follow. From reading the full text, I think I can divine what these new shaded funnel plots are meant to do. To me, though, it’s more interesting that the analysis of FDA data from RCTs of antidepressants confirms that these drugs are very effective indeed, more rather than less so once you have done the proper correction for publication bias. Or have I completely misunderstood? It’s always a lively possibility.

498   It’s a truism of orthopaedics and general practice alike that radiographic features of osteoarthritis and patient symptoms often bear little relation to each other. That applies strongly to backs, less so to hips, and according this study, hardly at all to knees. In fact radiographic features of knee OA, especially loss of joint space, are highly associated with knee pain, according to this pair of cohort studies.

502   It’s the smoke that kills; the tobacco itself is harmless, is what we often think. Except for oropharyngeal cancer. But here’s a systematic review showing that smokeless tobacco increases the risk of stroke and myocardial infarction. There should be a worldwide ban on the planting of Nicotiana species except for their evening scent in private gardens.

506    After tobacco smoke, the most dangerous twentieth-century threat to the lung was asbestos. On p.511, Kieran Sweeney, professor of General Practice at Exeter, gives a heart-rending account of what it is like to be dying of mesothelioma with a young family and full knowledge of the disease process. Asbestos, that magically heat-resistant mineral, can cause lung cancer as well as pleural cancer, and of course progressive interstitial lung disease too. Yet a lot of asbestos-related lung disease is benign. Here’s a nice clear guide to all of it.

Ann Intern Med  18 Aug 2009  Vol 151

229    I seem to remember that about 10 years ago, a team from Hong Kong looked through 10,000 papers about traditional Chinese remedies and found them all to be worthless. The papers, that is: and that probably applies to most of the remedies, though I think the study was prompted by the introduction at that time of phenomenally effective artemisinin-based treatment for malaria, based on Chinese folk practice. Will this now be followed by trypteriginin-based treatment for rheumatoid arthritis? Apparently the herb Tryperygium wilfordii Hook F has long been used by the Chinese for a variety of inflammatory conditions, and in this study it proved better than sulfasalazine in the treatment of RA. Interesting, but then a whole lot of drugs are better than sulfasalazine for rheumatoid.

264    Systematic reviewing is boring work, and hard work if you want to do it well. Bad systematic reviews appear by the score every week, often in the most prestigious journals, lumping or splitting in a random fashion to achieve their unconvincing conclusions. The Chief of Police in this field is Doug Altman who has helped to develop the PRISMA statement promulgated here in the Annals. Read, mark, learn and inwardly digest before starting your systematic review, or Doug will be on to you.

274   “But I have so much left to do!” was Mozart’s despairing cry as he realised he was dying at the age of 35 in 1791. Imagine what music we would have had Mozart lived another 40 years. And Weber, and Schubert, and Keats and Shelley. The world would be a different, kinder, infinitely more beautiful place. Ah me, I fondly dream. So, less fondly, do the many doctors who set about giving Mozart’s last illness a modern diagnosis. Here’s the latest attempt, based on an analysis of the official death register of Vienna between 1791 and 1793. Mozart died of “oedema” and so did an unusual number of young Viennese men that winter. Did they all have post-streptococcal nephritis? We will never know. Perhaps we will never care. Better to get out Bruno Walter’s pre-war performance of the Requiem and hear Anton Dermota cry “Recordare!” with the help of Elisabeth Schumann and Alexander Kipnis. I go faint even thinking about it.

JAMA   19 Aug 2009  Vol 302

741    “There are few randomized controlled trials on the effectiveness of palliative care interventions to improve the care of patients with advanced cancer.” Very true. This statement opens the abstract of the ENABLE II randomised trial of a “nurse-led palliative care-focused intervention addressing physical, psychosocial and care co-ordination”. Not quite English as we speak it but you get the general drift. In England, we call this Macmillan nursing, after the main charity which has supported it since the 1930s. Here comes an American study which gives a rare insight into what such nursing actually achieves when introduced for the first time into a community. The interventions – more structured than real life, perhaps, but similar - produce higher scores for quality of life and mood. But no improvement in symptom scores and no reduction of days in hospital or ICU or emergency department visits. A boon for dying patients, but a problem for health economists.

750    Quadrivalent human papillomavirus recombinant vaccine has been given to lots of American girls over the last three years with a great deal of hype about its ability to prevent cervical cancer (see p.781 for an analysis of the marketing strategy). We don’t have certainty on that issue, but in the mean time we can get some idea of its adverse effects from a voluntary system of reporting in the USA. These are about the same as for any other vaccine, except that syncope and deep vein thrombosis are a little commoner. Which suggests that teenage girls are likely to faint easily, and that many take oral contraception when they become sexually active.

758    Primary care doctors all over the world are naughty and prescribe too many antibiotics for acute respiratory infections. Naughty doctors in European countries often comfort themselves with the thought that American doctors are even naughtier. However, levels of naughtiness are dropping on both sides of the Atlantic, probably because patients and parents are getting better educated and demand fewer antibiotics. In America, visits to doctors for otitis media and other acute URTIs are decreasing, and so too did antibiotic prescribing for these conditions in the decade from 1995/6 to 2005/6. However, new forms of naughtiness are creeping in, with more prescribing of azithromycin and ciprofloxacin. GPs will always provide plenty for bacteriologists to tut about.

767   From time to time I have wondered in these columns if Harlan Krumholz, professor of cardiovascular medicine at Yale, ever sleeps; and since I have begun an e-mail correspondence with him, I’ve concluded that the answer is rarely. Give him a mass of data and the prospect of tight confidence intervals on an important topic, and I bet his office light stays on longer than you can stay up to watch. Tonight he’s up looking at 3,195,672 discharges following myocardial infarction in Medicare recipients aged 65 and over between 1995 and 2006, using a validated risk model. The news is good. Not only has the survival rate at 30 days gone up by about a sixth, the between-hospital variation has gone down by about the same. And this is in patients of mean age 78, when co-morbidity comes with the territory. OK, team; our next paper on a different topic is in next week’s New England Journal.

774    How much estradiol should you give to women with advanced breast cancer? If your reflex answer is “none at all”, you’ll be largely right, but for a selected few, the correct answer may be 6mg daily. I defy you to read this paper and the accompanying commentary on p.797 without getting completely confused. You see, high-dose oestrogen was used as a treatment for breast cancer in the 1960s and 70s until the arrival of tamoxifen, an oestrogen receptor blocker and failed oral contraceptive: then along came oestrogen receptor detection in breast cancer and the aromatase inhibitors, which block oestrogen synthesis; and now, you see, we are starting to use oestrogen again to stabilise or inhibit advanced oestrogen-receptor positive cancers which have become resistant to aromatase inhibitors. In fact low doses work as well as higher, with far fewer adverse effects, and this may permit the successful reintroduction of aromatase inhibitors. I’m reminded of the old mediaeval symbol of a snake swallowing its own tail.

792    A lot of our UK primary care Quality and Outcomes Framework is based on the assumption that if you improve process indicators you are bound to improve outcomes. An example is the section on heart failure, which includes indicators like LV function measurement and prescription of ACE inhibitors. These were also specified in the 1996 core quality indicators of the Centers for Medicare and Medicaid, together with smoking cessation and proper discharge instructions. The graph for US hospitals from 2002 to 2007 shows a steady rise in implementation to nearly 90%. At the same time mortality and rehospitalization rates for heart failure showed no change whatever.

NEJM  20 Aug 2009  Vol 361

745    Denosumab is a fully human monoclonal antibody against receptor activator of nuclear factor-?B ligand. So is it a cure for:

  • von Waldenstrøm’s macroglobulinaemia
  • dandruff
  • osteoporosis
  • obsessive-compulsive disorder
  • all of these?

Its makers, Amgen, would naturally prefer the last answer, but for now they will have to content themselves with number 3. But the osteoporosis market is already crowded with effective competitors, so how is denosumab going to make its way to the front? Answer: compare it with placebo. In this first study, the investigators could argue that this is ethical because nobody has worked out the best and safest way to prevent bone mineral loss in men receiving androgen-deprivation therapy for prostate cancer. This was not in fact an industry sponsored trial, but it gave the answer required: denosumab subcutaneously every 6 months increases BMD whereas placebo does nothing to stop the slight decrease in density associated with androgen deprivation. By the way, the handy abbreviation for what this drug targets is RANKL.

756    Now for the biggest osteoporosis market of all: postmenopausal women. We have lots of effective preventive options for these women, many of them about to come off patent: and yet Amgen persuaded ethics committees in the USA, the Czech Republic, New Zealand, Italy, France, Denmark and Argentina to approve a comparison with placebo in women with a bone mineral density score of less than -2.5 but more than -4. During the 36 months of the trial, 7.2% of the women in the placebo group suffered vertebral fracture, a terribly painful and disabling condition: in the treatment arm the percentage was 2.3. That equates to nearly 200 extra spinal fractures in the placebo arm, not to mention fractures at other sites. The authors put the blame on the FDA for requiring placebo-controlled three year trials for new osteoporosis agents. “Some observers have raised concern about the enrolment of subjects with osteoporosis in placebo-controlled trials, although there is no consensus about an allowable risk for inclusion.” Well there jolly well needs to be; and every new agent should be compared with an existing one, e.g. IV zoledronate once a year.

777   Opioid addiction is medically nearly harmless, though the behaviours associated with getting hold of illegal opioids can be medically very harmful and socially disastrous. By painful degrees, we are moving towards ways of legalising opioids for those who have become addicted to them, and this Canadian trial of injectable heroin versus oral methadone will hopefully prove a milestone in the humane and rational treatment of addicts in North America. Loose methodology comes with the territory, but this study of addicts (or narcomanes as they are called in French-speaking Canada) shows that relapsing opioid users do better across a variety of end-points, including employment and family stability, if they are given diacetylmorphine by injection rather than oral methadone; though this has to be done under medical supervision due to the risk of seizures and/or overdosage.

795    Here’s as good a short review of sudden infant death syndrome as you are likely to find. Progress has been slow over the decade since the 50% reduction in SIDS which resulted from a change in sleeping position advice. A few rare syndromes have been identified which are associated with repeated infant deaths, and we also have a few more clues about the cardiorespiratory events which can predict the fatal sequence of apnoea and anoxia that lead to sudden death in babies with immature protective reflexes.

Lancet  22 Aug 2009  Vol 374

609    The Lancet tends to regard primary care as an embarrassing oddity, poking at it nervously from time to time like a jellyfish which has washed up on its exclusive beach. Mix it up with a bit of mental health and you can just about fill an August issue, at a time when important doctors are away in Tuscany. The opening paper is a meta-analysis of studies concerned with the clinical diagnosis of depression in primary care, conducted by a clutch of psychiatrists from Leicester. There is an awesome spread of results on the Forest plots showing that something is wrong. It seems that GPs do not share the diagnostic infallibility of psychiatrists. But the glory of this issue is an editorial with the title “Are general practitioners really unable to diagnose depression?” This pours gentle scorn on the whole idea of a gold standard for the diagnosis of depression, concluding that “one can only hope that the new revisions of the International Classification of Diseases and the Diagnostic and Statistical Manual of Mental Disorders revise the nosology of mood disorders in such a way that current labels can be cast into oblivion.” One can only hope indeed.

620    A disease register study from Finland looks at the changes in mortality in patients with schizophrenia over the last 11 years. There have been no changes in mortality. But there are some interesting insights into the possible harms of second-generation antipsychotic drugs. Quetiapine carries the highest hazard ratio for death, and clozapine, seemingly the most toxic, carries the lowest compared with a phenothiazine. It was the first of the “atypical” antipsychotics and is still the most effective; if only it didn’t carry a risk of agranulocytosis and need monitoring so closely.

628    Let’s face it, fellow oldies: we’re going to have to use webcams and Facebooks and even learn to send text messages before we die. Especially if we want access to cognitive behavioural therapy. Your real-time therapist can visit you in your home, but only one cup of tea is needed - for yourself, since the therapist can remain in Baghdad or Montevideo. You still get to chat and reflect and do all the things that one-to-one CBT can do. Internet psychotherapy seemed to be effective in a few English counties near to me, according to this study.

635    When I reported that there has been no change in the mortality of schizophrenia in Finland, this was not particularly good news, because schizophrenia carries a 12-15 year penalty in life expectancy. Moreover, this seminar on the condition doesn’t persuade me that we understand it or treat it any better than we did 30 years ago.

BMJ   22 Aug 2009  Vol 339

431    The BMJ also enters the murky waters of psychiatric science this week with an analysis of data submitted to the US Food and Drug Administration from 372 double-blind trials of antidepressants in adults. A clear picture emerges of suicidality risk, which is increased under the age of 25 but decreases with increasing age. But as the editorial points out, the trials excluded those at highest risk of suicide. Also, different SRI drugs carry markedly different risks of suicidal behaviour, with sertraline seemingly the safest. I just take comfort from the fact that overall rates of suicide have fallen everywhere at the same time as SRI drugs have been increasingly prescribed as an effective treatment for depression.

435    When I was a houseman in the 1970s, Distalgesic was the standard mild analgesic prescribed in British hospitals. Even then, when you could prescribe just about anything, we were being told that paracetamol was just as effective and a lot cheaper, but nobody listened, and when generic co-proxamol became available, the cost argument diminished and we carried on. Unfortunately the dextropoxyphene element in this combination was (a) very effective for some people and (b) very dangerous in moderate overdosage. Now co-proxamol is expensive and hard to get hold of, but several of my patients swear that nothing else works, and argue that if people want to kill themselves they will always find other ways to do it. But this study shows they don’t. Restricting the drug probably prevents 300 deaths per year.

449   Various friends and acquaintances of mine could be heard on British radio channels a week ago, following their systematic review showing that oseltamivir makes a lot of children vomit and may not afford them any useful protection against influenza. None of the studies is sufficiently powered to give a clear answer in relation to the serious consequences of flu. My sympathy goes out to anyone who tries to explain the concept of scientific uncertainty on the live media. I went to find the authors in the Oxford Department of Primary Care in the wake of this and nobody was there. Spooky.

450    I’ve long advocated the use of D-dimer to exclude deep vein thrombosis in low risk patients in primary care. Finally the day came when we were given our own point of care test kits. But they are so fiddlesome to use that we’ve taken to sending blood to the hospital across the road, and now the kits are being taken off us. Good riddance, though this paper argues otherwise.

Plant of the Week: Ugni molinae

This is a dull looking little shrub with very dark little leathery leaves, but lovely once its little pink bell flowers appear and even more when they ripen into globular crimson fruit. It was classed as a myrtle until recently (Myrtus ugni), but its fruits are far better to eat than those of the common myrtle, just as aromatic but tasting like intense wild strawberries. Plant your ugni in some small sunny spot where you can wait for it to perform.

JAMA  12 Aug 2009  Vol 302

627   So, dear middle-lifers, is it really a good thing to do physical activity and eat a Mediterranean diet? Well, I’m glad you all put your hands up. It means that you are full of good middle-class values and you will not only live longer but also reduce your risk of Alzheimer’s disease, like two cohorts from New York (a city in the USA, not one of three villages of that name in eastern England). The same message is carried in the subsequent study of a French cohort from Bordeaux, where the local wine is better. “For now, it is reasonable to nibble on these findings and savor them, but not to swallow them whole.” concludes the editorial (p686) in this mid-August issue of JAMA,which trots out familiar lessons with a minimum of conviction.

649    And does aspirin improve survival after colorectal cancer? Once again the answer is a rather feeble yes, according to this prospective study of a cohort of American health professionals. The benefit is confined to people who started taking aspirin after the diagnosis, and is only significant in those whose tumours over-express cyclo-oxygenase 2.

660    It’s odd to find perinatal mortality statistics from Scotland tucked away in JAMA, as if a slice of haggis had found its way into a Chicago hamburger. British obstetrics is undergoing something of an upheaval, with many small units being shut down in favour of midwife-led units and large hospital centres. The statistics from 1988 to 2007 don’t yet reflect this trend. There was a 37% drop in perinatal mortality during this period, entirely due to a reduction in birth anoxia. At the same time, there was a rise in rates of Caesarian section, but the two do not appear to be causally related.

NEJM  13 Aug 2009  Vol 361

653    The recent debate about mammographically detected breast cancer is just one illustration of how little we know about the natural history of this major killer. One recent advance has been the use of axillary node sampling, which has become the most important prognostic tool in breast cancer, though current routine techniques may be inadequate to detect micrometastases or isolated tumour cells. To do this, you need to examine all sections of the sentinel node with immunohistochemical staining. This Dutch study looked at the effect of adjuvant therapy on those with micrometastases compared to those with such metastases who did not receive adjuvant chemo, and also those without micrometastases. The presence of single cell metastases is equivalent to the effect of cell clusters, and raises the risk of recurrence by about 50% compared to having no detectable lymph node metastases at all. There is a big advantage in receiving adjuvant therapy if you have these tiny metastases – hazard ratio 0.57 compared with no chemo.

664    Lymphoedema of the arm is common in women who’ve had breast cancer treatment, and apparently it’s routine to advise them against weight lifting. By this the authors mean the regular lifting of weights as a form of exercise, which I have not found particularly popular among English ladies of my acquaintance. But if they want to do it after breast surgery, they can go ahead quite safely, according to this randomised trial. It does not exacerbate lymphoedema and may even have some benefits.

674    The reason we got scared at the Mexican swine flu outbreak was its initially high reported mortality, due to severe respiratory disease. Even in the UK, which can claim to be the home of infectious disease epidemiology, tracking H1N1 influenza seems like tracking enemy air attack in the days before radar: a patchwork of observers on the end of telephones reporting to WAFS who move wooden counters around on a big table map. In Mexico, they seem to have fewer observers and hardly any WAFS, but they’ve managed to establish a temporal association between the first identification of the virus in April this year and a big surge in severe pneumonia in people aged 5-59. It seems that if you were exposed to the strains of flu in circulation before 1957, you are relatively protected. Excellent.

680   Another report delineates the clinical course of pneumonia and respiratory failure in 18 cases from the respiratory unit in Mexico City. Seven of them died, and there were 22 secondary cases in health care workers who came into contact with them, but none of these were serious. Severe H1N1 (or S-OIV) respiratory disease is commonly accompanied by a rise in LDH and often also of CK.  Only 8 of these patients had any pre-existing medical condition, and the majority were under 47.

690    A couple of years ago, a team of Scottish gynaecologists managed to hit the top spot in the New England Journal with their study of uterine fibroid embolization. This happened on Burns’ Night so I felt compelled to celebrate the occasion in suitable Caledonian verses, notable for their attempt to rhyme “gynaecology” with “follow ye”. For which I make apology. Because it’s really radiology. To do this procedure you have to cannulate the uterine artery by quite a complex route and then embolize the fibroid circulation with some kind of particulate matter. But were I a woman with fibroids reading this review, I might still be inclined to opt for open myomectomy, as embolization can be followed by days of severe pain and generalised reaction to tissue necrosis.

Lancet   15 Aug 2009  Vol 374

511    In an issue largely devoted to stale cardiology, the most interesting item in The Lancet is this short editorial about something else – the attempt to raise a scare that insulin glargine might cause cancer. Here’s a nice brief summary which shows how such hypotheses can be generated by enough post-hoc analysis if you wish to do so: refuting them is more difficult and in the mean time there’s no particular need to worry.

519    A recent medical biography by a noted Swedish man of letters (Axel Munthe: the road to San Michele, by Bengt Jangfeldt, 2008) had me marvelling at the author’s infinite capacity to get medical facts wrong. When lay people write about medicine they often need medical proof-readers, which even The Lancet seems to lack, judging by this piece describing the sudden death of John Hunter in a fit of rage: “The autopsy, done by Hunter’s brother-in-law, Everard Home, revealed that the carotid arteries around the heart were thick and ossified and the heart itself was shrunken.”  Hmmm.

525   Gosh, an open-label study claiming to show that the more you lower blood pressure, the more cardiovascular events you prevent. There is such abundant evidence for this that this Italian study (Cardio-Sis) needed something special to add. In fact it contributes nothing at all, since its main end-point, electrocardiographic left ventricular hypertrophy, is the crudest of surrogates. Another unaccountable inclusion in The Lancet.

534   The thrombi that cause stroke in atrial fibrillation form within the left atrial appendage. There are currently three ways to deal with this: try and stop the atrial fibrillation (though even successful cardioversion isn’t a guarantee against embolic stroke); go on warfarin for life; or, most recently, close off the atrial appendage. This is quite a tricky percutaneous procedure and there were a number of periprocedural strokes and air embolization events in the interventional group of this 59-centre trial. But at three years on, LAA closure yielded better results in high-risk patients than continued warfarin. This was a randomised non-inferiority study: expect many more and longer studies before this procedure becomes routine.

543   If there’s one cardiovascular surrogate I hate even more than ECG-LVH, it’s albuminuria. Sure, it’s easy to measure and albumen is not a good thing to be peeing. If you have chronic heart failure, it’s a sign that your kidneys are protesting and that you have a worse prognosis. Therefore it has prognostic importance, claim the CHARM investigators. No it doesn’t. It would only have importance if it were better than other markers of prognosis in heart failure, like BNP, and it certainly is not. In fact it isn’t compared here with any other meaningful markers. You wouldn’t claim that leeches had therapeutic importance in heart failure without comparing them with diuretics and ACE inhibitors: likewise you don’t say that this and that has prognostic importance in heart failure until you’ve compared it with something important.

BMJ   15 Aug 2009  Vol 339

374   We’re supposed to counteract the overuse of antibiotics in childhood respiratory tract infections by gentle persuasion directed at worried parents. There are times we do; and other times, I must confess, when I don’t even try. I frequently take the “deferred prescription” route, though I’ve never monitored how many of those prescriptions are cashed in on the same day. Well, here’s a trial that surprised me by showing that parents respond to interactive written material. Get this booklet on your computer database and print it out as soon as you hear a child sniffle. The main outcome measure was not in fact antibiotic use but re-attendance for URTI – even better.

377    We had to go to an American journal for data about Scottish perinatal mortality but here’s the British Medical Journal telling us about England. A bit. It’s an attempt to explain the variance between various Primary Care Trusts in perinatal and infant mortality – both are charted, but they look exactly the same. 80% of the variance can be attributed to familiar factors including maternal age and deprivation but also notably Pakistani ethnicity. And then there are two PCTs which are far outliers with double the expected mortality – West Devon and Wyre Forest. Both have created large remote midwife-led units in the recent past. Next month there will be a report on the safety of such units, which will be important for the future of small hospitals throughout the UK.

390    The debate about euthanasia grumbles on the British press and in the Rapid Responses of the BMJ. Opponents come up repeatedly with arguments based on slippery slopes, the vulnerability of the old and frail, or if all else fails, rhetoric about agents of death and Hitler. But the level of debate is improving and will be helped by this Belgian study which shows that physician assisted death remains a rarely used option and is not associated with poorer palliative care but rather with increased use of spiritual care aspects of palliative care. The associated editorial by a palliative care doctor (p.357) also marks a welcome shift in tone, conceding the rarity and validity of decisions about assisted death, and arguing not for total rejection but only for a clear separation between palliation and assisted death.

396    I’m delighted to see the whole Practice section of this BMJ devoted to a concept which I was among the first to propose – a series on primary care diagnosis with interlocking sections on quality improvement, a specific clinical situation, and a theory piece. Here it’s the Ottawa ankle rules as an illustration of how to manage the injured ankle and foot using a clinical prediction rule. I’m glad I didn’t have the hard work of commissioning these pieces and making them fit together. But if this series keeps delivering the goods like this, we might eventually have a textbook based on real life in primary care; and who knows, one day we may even be able to teach our trainees some real life diagnostic strategies that are based on evidence gathered in the right context.

Arch Intern Med  10 Aug 2009  Vol 169

1364    An interesting spin-off from the vast Nurses’ Health Study II (1997 to 2005) shows that breast-feeding reduces the risk of breast cancer. But – and here’s the interesting bit - this only applies if you have a family history of breast cancer. In this case it halves your risk, but women without such a history get no protection.

1389   As the debate on Obama’s health reforms brings out the worst in the American Right, I hesitate to suggest that the NHS can learn anything from US medicine, but I remain fascinated with the idea of hospitalists as a bridge between primary and secondary care. That’s something that both systems could do with a lot more of. “Hospitalists are physicians whose primary professional focus is the general medical care of hospitalized patients and whose activities include patient care, teaching, research, and leadership-related inpatient care.” And this survey of US hospitals with and without these paragons concludes that their presence is associated with higher quality of care. But the verdict of the accompanying editorial (p. 1351) is “not proven”. If you’re trying to integrate your local primary and secondary care services, this is a debate you need to follow. Whereas the rest of the American health debate is just too depressing.

1393    Thiazolidinediones ought to be good drugs for type 2 diabetes, improving insulin sensitivity without carrying much risk of hypoglycaemia. But luck is just not with them. The first one, troglitazone, was withdrawn for causing liver failure; the second one (rosiglitazone) is under a cloud for causing cardiac events; and just as we are switching to pioglitazone, here comes this prospective cohort study showing that it’s probably the worst for increasing fractures in both men and women. The data for a definitive answer on this association are probably already out there, in clinical trial databases and the UK GP Research Database.

1411   Recent analysis of the British QOF system shows how easy it is to skew the direction of care in one direction to the detriment of other aspects. The most intensively studied process measurement in acute cardiology is the door-to-balloon time in myocardial infarction. Before that, it was the door-to-needle time. But what of the other things that constitute good care in acute infarction? This study shows that door-to-balloon time shows dissociation from other measures of quality, including in-hospital mortality.

Plant of the Week: Rosa “Desprez à fleur jaune”

Although I’ve mentioned this wonderful climbing rose before, I can’t resist doing so again as I look through the study window and see it covered with flower. O that I were on the bench beneath, sipping a glass of white Burgundy; as indeed I will be very soon. All about me will be the rich peachy scent of this marvel amongst French roses, raised by a nurseryman called Desprez in 1830. No relation to the greatest early Renaissance composer, I suppose, though we do call the plant “Josquin” from time to time, so elevated are our tastes.

The flowers have only a hint of jaune about them: they are really a kind of apricot pink. They start to appear in July and never stop until mid-October; sometimes they go into November. “It is so powerfully fragrant that one plant will perfume a large garden in the cool weather of autumn” says Plant in his Manual of Roses, 1846. This quotation is cited by the greatest of modern rose buffs, the late Graham Stuart Thomas, who also says, ”I think it is wonderful that this rose, raised so long ago and without probably any idea in the raiser’s mind as to what might materialize, should remain one of the most perpetual of all roses”. (The Graham Stuart Thomas Rose Book 1994).

JAMA  5 Aug 2009  Vol 302

517    Here’s a look at the sadder aspects of old age in the USA, as in most other places – elder self-neglect and elder abuse. It’s a bit odd to lump the two together, and of the two, self-neglect carries a far worse prognosis than deliberate abuse. Except that we don’t really know the true extent of either, and this study only looked at “elders” who were reported to the social service agencies of Chicago. A society that really wanted to help the elderly would encourage active case-finding, which would then result in higher expenditure. Somehow I don’t see it happening.

537    In this gloomy issue of JAMA, we’re invited to look at the effects of deliberate violence on health. Sixty-five years ago, almost the whole population of Europe was witness to mass displacement, torture, and traumatic events. Now researchers usually head off to Africa to study such things. This meta-analysis looked at 161 studies with wildly different results – for example the prevalence of post-traumatic stress varied between 1% and 99% in refugee populations and depression varied from 3% to 85.5%. With baselines like these, does it make any real sense to estimate the additional effects of torture and/or traumatic events? The authors try their best, but I don’t know what in the end we can say, other than deliver us from evil.

550    Another meta-analysis looks at the literature relating sexual abuse and lifetime diagnosis of somatic disorders. Here the 23 studies are reasonably consistent. There is a relationship between sexual abuse and a lifetime diagnosis of functional gastrointestinal disorders, non-specific chronic pain, psychogenic seizures and chronic pain – in every case the odds ratio is between 2.2 and 3. But there is no association with fibromyalgia, obesity or chronic headache.

NEJM  6 Aug 2009  Vol 361

557    Those of you who turn to the Rapid Responses of the BMJ for occasional entertainment will have enjoyed a feast of British tribal conflict over the last two weeks as NICE and the British Pain Society shout war cries and brandish spears at each other over the issue of facet joint denervation and the banishment of the Pain Society Chieftain. I expect there will be similar yells and dances in response to this paper and the almost identical one which follows it, demonstrating that vertebroplasty for painful osteoporotic fractures works about as well as sham vertebroplasty for painful osteoporotic fractures. In fact, short term improvements were rather better in the sham group, who were treated to an elaborate ritual in the operating theatre, including the preparation of PMMA (polymethylmethacrylate) “so that its smell permeated the room.” The lesson here – yet again - is that you have to carry out properly blinded trials if you want to know whether an intervention really works. Thanks to unblinded trials and biased reporting, vertebroplasty has become a highly popular procedure for a common and highly painful condition, for which there is now no alternative treatment. So watch out for displays of pain and anger from its practitioners, who know from experience that this is not only an effective and life-enhancing procedure in skilled hands (i.e. their own), but also a nice little earner.

594    Here’s the latest bulletin from the Clot Wars, a nakedly commercial trial of apixaban paid for by Bristol-Myers Squibb and Pfizer. The battlefield is thromboprophylaxis following major lower limb surgery and the paper doesn’t pull any punches in naming the enemy – Xarelto from Bayer and Pradaxa from Boehringer Ingelheim. No polite messing about with generic names in this New England Journal paper. To put you in the picture, Xarelto is rivaroxaban, an orally available direct factor Xa inhibitor, like the drug in this study. Pradaxa is dabigatran, a direct thrombin inhibitor. All these drugs can be given at fixed dosage and none of them requires INR monitoring. So the battle is to take over the entire heparin and warfarin market and who knows what else beside – all ischaemic stroke and myocardial infarct patients, perhaps? No wonder there is a bare-knuckle feel about some of this competition. Desperate for a result, the manufacturers of apixaban devised a massive composite end-point in this comparison with the fractionated heparin, enoxaparin: asymptomatic plus symptomatic DVT, nonfatal pulmonary embolism, and death from any cause during treatment. But they shot themselves in the foot and managed to show a slight superiority for enoxaparin (8.8% versus 9%). Still, there’s no doubt that apixaban is much easier to use and causes less bleeding. In exchange for 5% of the profits, I will allow B-MS and Pfizer to use my Love Song of Venous Clot in their promotional campaign:

I’m small and dark and slimy
But people think I’m hot:
I make them shout gor blimey!
My name is Venus Clot:

I’m getting used to hollers
Of praise at every spot;
It’s worth a billion dollars
To date with Venus Clot.

Xarelto and Pradaxa
Joust at my Camelot:
But suitors such can’t tax her
Whose name is Venus Clot.

Right now there’s just a single man
Who makes me lose the plot:
His surname is Apixaban
And he’ll have Venus Clot.

For another 5% I will write them a similar song about Venus’s sister, Arty Rielle Clot.

Lancet  8 Aug 2009  Vol 374

451    The most worrying aspect of H1N1 influenza is its capacity to kill women in pregnancy, especially at the time of labour and the puerperium. This study from the US Centers for Disease Control and Prevention looks at a few (34) pregnant American women with this infection up to May 18thand at six who died up to June 16th. The cause of death in each case was pneumonia and adult respiratory distress syndrome. Since the current strain is but another variant of the 1918-9 virus, you would expect a similar pattern in previous pandemics, and the interesting editorial on p.429 finds some evidence for this. We urgently need to know more about the safety and effectiveness of oseltamivir in pregnancy and the puerperium.

459    Rheumatology was a sleepy backwater until ten or so years ago, when the arrival of tumour necrosis factor antagonists started a slow revolution towards the aggressive treatment of early rheumatoid arthritis. The evidence base for this is accumulating in fits and starts, with good study design at a premium, since it is very easy to manufacture improved results for new drugs if the comparator is inadequate. Here 15 rheumatology units in Sweden collected patients presenting with early RA and gave them all methotrexate as initial therapy. Those who could tolerate methotrexate but weren’t showing much response were randomised to receive either sulfasalazine plus chloroquine orally or infliximab intravenously. There was adequate randomisation but no blinding, and the infliximab group did better. However, as the editorial (p.430) points out, the main benefit in rheumatoid arthritis is likely to come from making sure that the drugs are working early in the disease process for each individual, rather than from using a particular drug class.

477    A seminar on gastric cancer sums up current knowledge about this malignancy, once the main cause of cancer deaths in Europe and the US. We know a lot about its epidemiology but we can only explain a small part of it. Current ideas centre on the genotyping of H pylori. There is no evidence to support screening and mass H pylori eradication in countries of low prevalence. Nor is there really much hope that improved surgical technique will save most patients with invasive disease. It’s still a disease with more questions than answers and more victims than survivors.

491   I suppose that writing comments on the main medical journals every weekend for 11 years counts as a form of obsessive-compulsive disorder, and I certainly frequently feel the urge to wash my hands of it. During that period I haven’t noticed any major advance in the area of OCD, except for trials showing the serotonin reuptake inhibitors and cognitive behavioural therapy are about as effective as each other but are no more effective in combination than singly. This article makes claims that D-cycloserine might be the drug to enhance the effects of CBT, but the evidence is tenuous to say the least. There are lots of biochemical pathways which may be involved in OCD, as there are in every condition which we don’t really understand.

BMJ  8 Aug 2009  Vol 339

327    If you want to effect a quiet revolution in England, August is a good time to do it. Nobody is at home, and when they come back their In Box is full and they have to come to terms again with how awful their job is instead of worrying about wider issues, such as whether the BMJ is right to shrink all its research articles into pico format. This means that the summary sections are now piccolissimo and I could almost save myself a job by simply typing them out for you: “About 20% of women with one year HPV persistence and 40% with one-year HPV 16 persistence will develop cervical precancer or cancer in the subsequent three to five years.” I don’t know about the pico format, but this is important stuff by any standards. In fact there really is an embarrassment of riches in the one-page reports about cervical screening and cancer in this issue of the BMJ. Ironic that the research section should be the best yet this year just as it’s being picolysed in the interests of making the magazine look livelier.

328    Our practice nurses assure us that in the wake of Jade Goody’s death at 27 from cervical cancer, attendance rates for cervical screening have shot up. At the same time, Karol Sikora, a Briton who was once a London professor of oncology, produced a sob-film for the Republican campaign against Obama’s health reforms, featuring a woman dying from cervical cancer because she was denied screening by the socialist dictators who run the British NHS. In fact there were 350 cases of invasive cervical cancer in this UK-wide case-control study, but the conclusion is robust and clear: cervical screening below the age of 25 does not prevent cervical cancer. The NHS may not get all its public health policies right, but it has in this case. And it’s free. And universal.

329    Here’s another really important study looking at the best strategy following a “mildly abnormal” smear – borderline nuclear abnormalities or mild dyskaryosis. Send them for immediate colposcopy and you end up diagnosing more grade II changes and doing more procedures than if you just do cytological surveillance. Harder outcomes – CIN III and worse – are the same either way.

330   The same group (who call themselves TOMBOLA, but we won’t hold it against them) now examine whether it’s better to do punch biopsy with selective recall or large loop excision of the cervix at colposcopy for low grade abnormal cervical cytology. You get more bleeding and discharge with the latter and no medium term gain.

337    Call something interstitial cystitis and you imagine it’s a tissue diagnosis: call it painful bladder syndrome and you begin to realise that we don’t have a clue. And that’s largely the case with this sometimes crippling chronic condition which affects many more women than men. You take the history and make the diagnosis: cystoscopy makes little difference. Nor do most treatments, if you look at those with adequate trials. If you are more easily satisfied you can try anything from amitriptyline to ciclosporin. But though I’m taking a rather weary tone I do think the authors have produced a very good update on a very difficult subject.

Ann Intern Med  3 Aug 2009  Vol 176

176   “Does My Patient Have Clostridium difficile Infection?” sounds like a title from the brilliant but spasmodic series The Rational Clinical Examination in the senior sister journal of the AMA. This paper is shorter and just as useful. If like me you don’t really have a clue about modern diagnostic tests for C diff here’s the rule: try to access a PCR stool test for a gene that codes for toxin B and/or its regulators. These tests are quick and accurate and a negative one does not need repeating. Enzyme immunoassays for stool toxin or glutamate dehydrogenase are also quick but less reliable.

191   What end-points would interest you if you had atrial fibrillation? I guess death, stroke, heart failure, quality of life and long-term anticoagulation for starters. Here’s a systematic review of 108 studies of radiofrequency catheter ablation for AF. There is evidence that it is effective for up to 12 months of rhythm control when used as second-line therapy in younger patients with intact cardiac function. For the important end-points, we need more studies: numbers 109 and rising.

Plant of the Week: Chitalpa tashkentensis

In this section I almost always discuss plants I’ve grown at some time or other; very occasionally ones I’ve seen and wanted to grow; and only on one previous occasion a plant I’ve read about but never seen. That was Acer diabolicum purpurascens, whose name alone made me long to have it, but which probably doesn’t exist.

Chitalpa tashkentensis has existed since 1964, when a nurseryman in Uzbekistan succeeded in hybridizing the well-known bean tree, Catalpa bignonioides,with the desert tree Chilopsis linearis. The result is a smallish tree which is both very hardy and very heat tolerant, and which flowers from late Spring to October. The flowers are scented and azalea-like, pink or white with a frilly edge and a yellow throat. It seems popular in the USA and is beginning to appear on nursery lists in England.

For ages I’ve wanted a tree that flowers all summer long, but the only one I know of is Cladastris sinensis which is (a) almost unobtainable and (b) looks manky and sad in England on the rare occasions I’ve seen it. I’ve a feeling that the chitalpa might take exception to the English climate too, and produce a mass of leaf and few flowers, or just rot and die. If anyone knows, please get in touch.


RSS Subscribe to the RSS Feed for this page





Page last edited: 08 September 2009