Richard Lehman’s journal blog, 5 July 2010

Richard LehmanNEJM  1 Jul 2010  Vol 363
11    So far, the trials of carotid stenting versus endarterectomy have sent out mixed messages, but the CREST study sends out a message of equipoise. The triallists decided to recruit asymptomatic patients mid way through, to bump numbers up to 2502, but the main population was fairly homogeneous and had stenoses of 70% or more which were symptomatic. In the periprocedural period those who had endarterectomy had more strokes, and the stent group had more myocardial infarcts. But at a median follow-up of 2.5 years, there was no difference in major outcomes and a pleasingly low incidence of ipsilateral stroke at 2-2.4%.
24    Most of the 200 human papillomaviruses are harmless, but type 16 can get nasty and is found in many genital and oropharyngeal cancers. This study looks at the influence of HPV-16 on survival in treated oropharyngeal cancer. If you adjust for other factors, tumours are much more likely to respond to radiotherapy and platinum-based chemotherapy if they contain HPV-16 DNA.
36    I’m no great fan of implantable cardioverter-defibrillators, not least because they often go wrong due to lead failure, and they can lead to shockingly bad ends in heart failure. This trial assesses a new type of ICD which does not rely on venous access but is entirely subcutaneous, delivering shocks to the thorax close to the heart. Its success depended a lot on accurate positioning, and over the ten months of the trial it worked well and appropriately, though it’s too soon of course to say anything about long-term reliability, let alone long-term mortality benefit.
63    If you get regular migraine, you probably use a triptan from time to time. I was delighted when sumatriptan injections first became available about 25 years ago but rather disappointed by the results, on myself particularly. Since then there have been many more triptans with better oral availability, but surprisingly little evidence to guide any choice, or even to prove that they work better than analgesic/metoclopramide combinations. Although I no longer need to take triptans myself, I know many others who do and I fell upon this review article with keen interest. But I’m afraid I learnt very little.

Lancet  3 July 2010  Vol 376
20    I don’t normally comment on the letter pages, but the letter here from Mike Clarke, Sally Hopewell and Iain Chalmers is the most important thing in this week’s Lancet, and may even have achieved a change in the policy of this august publication (see editorial on p.10). The Cochrane stalwarts simply restate a truth that all researchers should always heed; clinical trials should begin and end with systematic reviews of the evidence. That’s the main part of the title of their letter: the other bit is 12 years and waiting. They have gone through the same journals as I go through each week and state that “there is no evidence of progress between 1997 and 2009 in the use of updated systematic reviews in discussing the findings of trials published in these five medical journals.” And these are the best journals: when I wander into the rest I am dismayed and appalled at the even greater extent of repetition, muddle, and futility. At least The Lancet promises to pull its socks up. I’ll believe it when I see it.
23    Tranexamic acid has been around for ages and is mostly used in gynaecological and urological practice, though it’s recently also been used to reduce bleeding in elective surgery. The CRASH-2 trial tests the hypothesis that it might reduce bleeding and improve outcomes in trauma victims with significant haemorrhage. Over 20 000 such patients were recruited in 40 countries, and the result is a very narrow win for tranexamic acid over placebo: a 9% reduction in mortality which just reaches significance.
49    Antiviral therapy and management of HIV infection. That’s a nice straightforward (if ambitious) title for a seminar. And a very good seminar it is: one every doctor should read, since as the opening sentence states, “Advances in understanding of HIV biology and pathogenesis, an in application of that knowledge to reduce morbidity and mortality, rank among the most impressive accomplishments in medical history.” Not just that but the story continues to be largely optimistic, despite the awesome adaptability of the virus.
63    As I said a couple of weeks ago, no child should die of diarrhoea, and in many European countries none do, but worldwide the death toll is 1.5 million. The treatment is usually nothing more than an appropriate oral rehydration solution, with added zinc. I was quite unaware that zinc has marked benefits in the acute phase and even months later, and that added to the standard sugar and electrolyte solution in the poor world it can halve overall mortality from diarrhoea.

BMJ  3 July 2010  Vol 341
30    Steroid injections remain popular with patients who have moderate to severe shoulder pain, and the figure from this study shows why: there is substantial added relief beyond the effect of exercise alone in the first three months. Thereafter there is no significant difference.
31    A team of epidemiologists and statisticians has worked on national databases of childhood cancer and found that there is no relation between these and proximity to mobile phone radio masts. This should logically put an end to such claims, but I wouldn’t put any money on it.
32   Proton pump inhibitors are associated with an increased risk of postoperative pneumonia in elderly patients. These people then have to have antibiotics and get Clostridium difficile. Hence naughty doctors should prescribe fewer PPIs. So runs the usual logic: but it isn’t quite right. This Canadian study found that elderly people taking acid suppressants (mainly PPIs) had more major risk factors for pneumonia than those who did not at the time of operation. Adjust for those, and it seems that acid suppressants confer no extra pneumonia risk at all.
34   A Clinical Review article about Huntington’s disease maintains the generally very high standard of current BMJ reviews, as it should, since it comes from the UK’s leading centre for treating the condition. It’s well worth reading through for general interest, and keeping if you have a family with the condition in your practice. There is also a readable and rather chilling historical account of Huntington’s discovery and early medical attitudes to it in this week’s Lancet (Stigma, history, and Huntington’s disease by Alice Wexler, p.18).
41    A short piece in the much more variable Rational Testing series deals with investigating symmetrical polyarthritis of recent origin. Its main virtue is in emphasising the urgency of referring such patients (especially with hand joint involvement) to a rheumatologist straight away, since the diagnostic tests available in primary care can be entirely misleading. The 32 year old woman in the vignette has a normal CRP, ESR and negative rheumatoid factor but nevertheless has markedly elevated anticyclic citrullinated peptide antibodies which are highly predictive of progressive rheumatoid arthritis. Rheumatoid factor is negative in 38% of patients with RA, especially in early disease: conversely 20% of healthy over 65s have a positive RF. It’s a bad test and anti-CCP looks set to replace it.

Arch Intern Med  28 Jun 2010  Vol 170
1013   Last week I grumbled that the Archives and Annals had yielded me nothing to write about for three weeks: well this week the Archives take their revenge. I was trying to take a break from thinking about medicine, and here is more thinking matter than a week can easily contain. Let’s begin with diagnostic adverse effects. Looking back over your distinguished career, how many patients have you harmed by late diagnosis or wrong diagnosis? OK, let’s not go there. And how many systematic papers have you read about this topic? None at all, in my case, if you exclude essentially anecdotal material like Lesson of the Week in the BMJ and similar. Well, here is a rare and exemplary example of the genre, based on a review of nearly 8 000 patient records randomly gathered from Dutch hospitals. This is extraordinarily labour-intensive work with a low yield: the team found that only 4 hospital patients in every thousand were harmed by diagnostic errors, and in almost every case the error was caused by human judgement. The mortality from these misjudgements was 29%. For those with time enough, and a big project grant, the entire field of primary care lies open for further studies. They need to be done.
1024    Next on to statins, my favourite drugs, which I have defended against all comers. If in doubt, prescribe, said I. Among patients with known cardiovascular disease, I was right, although it is very hard to find any benefit once heart failure has set in. But prescribing statins to high-risk patients for primary prevention may be futile, according to this literature-based meta-analysis. It is a very hard paper to follow, however, with a fairly heterogeneous mix of studies which are not adequately characterised or analysed in these six pages: to do that would require twice the length, or ideally an entire database, which could then be analysed on an individual patient basis…
1032    The JUPITER trial of rosuvastatin was stopped early and has been a source of controversy ever since. The acronym stands for Justification for the Use of Statins in Primary Prevention, but when JUPITER’s data are fed into a meta-analysis like the one we’ve just seen, there is no such justification. In fact the data of this trial are internally contradictory in a way that strongly suggests manipulation, according to this critical reappraisal, which suggests that Jove’s ire should be directed at the role of commercial sponsors. I can hear the distant peal of thunder across the Atlantic: Jupiter tonans.
1037   The well-conceived new Archives series called LESS IS MORE here lives up to its radical credentials: we are giving diabetic patients too many drugs for cardiovascular protection. Again, this flies in the face of what we have been taught over the last few years. It also seems to fly in the face of the calculation done by these authors that treating to targets for LDL-cholesterol and blood pressure results in gains of 1.5 and I.35 quality-adjusted years respectively. But they demonstrate that these overall gains are largely accounted for by the treatment of a small number of very high-risk individuals, and that the more drugs you put in, the more you are likely to achieve minimal benefit or actual harm. A key paper in the continuing debate about targets in type 2 diabetes.

Plant of the Week: Oenothera odorata

Last year our small back garden was made even smaller by building work and had to have large parts dug over and replanted: an ideal opportunity, you might have thought, for a careful reshaping of the whole. Well, it didn’t quite happen like that: we bought nearly a hundred new plants, but dug in most of them in a hurry as they threatened to die in their pots.

We now have a miscellaneous riot of bought-in perennials in flower, but they don’t dominate the scene. The masters of the garden are self-sown beauties that we can’t find the heart to pull up: white mallows, great flopping opium poppies in shades of pink and red; but above all, beautiful pale yellow evening primroses, fading to coral pink on long red stems. By day they look like weeds, but as evening falls they come out like a forest of creamy lamps.

They are scented, as their correct botanical name implies; but often sold as “sulphurea” or “stricta”. I have no idea where our first one came from, but once you have this plant, you are likely to enjoy its descendants in perpetuity, since it seeds itself freely. By day, you will be tempted to pull these children out of the ground as they appear randomly; but by evening you will wonder how you could ever have harboured such strange and cruel thoughts.