JAMA  3 Apr 2013  Vol 309
1355    With the runaway success of the Alltrials petition, it may seem as if everyone in the world has now agreed on the need to share every bit of data relating to every medical device and product used on millions of patients every day. In reality, this is going to be a very slow process, involving hard work over many years. Nobody is more aware of this than Joe Ross and Harlan Krumholz, whose YODA project is pioneering the methodology needed to do the job properly, in a way that few others have even considered attempting. The imperative to do this work is absolute, and is beautifully set out by them in this Viewpoint article. But the editor of JAMA, Howard Bauchner, announced in Oxford that he is planning to sit on the fence about Alltrials a while longer, consulting his editorial board in a few months’ time. In the meantime we can look forward to a piece on the “unintended consequences” of data disclosure by Robert Califf some time soon.

1359   I continue to puzzle over what journal editors think they are actually for. Here, in open access format, is a trial of duloxetine for painful chemotherapy-induced peripheral neuropathy. The comparator was placebo, on the grounds that there are no known treatments for this condition. Really?

The authors set out by telling us that “painful chemotherapy-induced neuropathy can persist from months to years beyond chemotherapy completion, causing significant challenges for cancer survivors due to its negative influence on function and quality of life.” They then report a trial where there was a run-in period of one week, followed by four weeks on treatment or placebo. “The minimal clinically important difference in pain severity was defined for the current study as a 0.98 difference in mean average pain severity between the duloxetine and placebo groups. The observed mean difference in the average pain score between duloxetine and placebo was 0.73 (95% CI, 0.26-1.20).”

Just what is this study telling us? I writhe at the thought that doctors will read “Conclusion and Relevance:  Among patients with painful chemotherapy-induced peripheral neuropathy, the use of duloxetine compared with placebo for 5 weeks resulted in a greater reduction in pain,” and that they will now be prescribing long-term duloxetine – a drug with the potential for dependence and a host of adverse effects – on the back of this painfully inadequate study which failed to reach its own primary end-point.

NEJM  4 Apr 2013  Vol 368 
1279   The most important event in England in 1950 (apart from my birth) was the publication of a slim volume called A Book of Mediterranean Food by Mrs Elizabeth David. In this book the British public, then thin and pale from ten years of food rationing, caught a vision of unlimited delicious sun raised produce, simple to cook and heavenly to eat. Unfortunately it never found its way into my parental home. The chapters of Mediterranean Food dealt with: soups; eggs and luncheon dishes; fish; meat; substantial dishes; poultry and game; vegetables; cold food and salads; sweets; jams, chutneys and preserves; and sauces. Growing up as the child of immigrant benefit scroungers living on national assistance, I rarely saw any of these. I have tried to catch up ever since, and suggest that you do the same.

As for what is called a Mediterranean diet by American researchers, I have no strong views. This seems to be the same as Elizabeth David’s diet, but with many of the nice things omitted. Compared with a standard low-fat diet, the PREDIMED diet achieved a reduction in cardiovascular events of about 30% over a median of 4.8 years, at which point the trial was stopped.

1303   All the time I’ve been writing these reviews, drug companies have been trying to come up with new platelet inhibitors to break into the immense market dominated by aspirin and clopidogrel. Intravenous cangrelor was compared with an oral dose of clopidogrel (which could be 300mg or 600mg) in this study of 11,145 patients undergoing percutaneous coronary intervention for a variety of reasons. And cangrelor definitely won, with a small but statistically significant reduction in the composite end-point (an absolute reduction of 1.2% in immediate thrombotic events or death). Overall, there was a 0.6% difference in stent thrombosis.

1314   It you have metastatic prostate cancer, you generally have about two years for hormone based treatment to work, after which you can but hope that some of the other treatments now being trialled may come to your rescue. Much the same applied in 1993, when this trial of intermittent versus continuous LHRH agonist therapy was devised. It was hoped that men whose androgen activity was allowed to return from time to time might show better survival than those who were under continuous blockade. But if anything, the opposite proved true, though even in this big trial with long follow-up statistical significance was not reached. Men on intermittent therapy had a slightly better quality of life.

Lancet  6 Apr 2013  Vol 381
1185   Golly, I’m not sure this makes me proud to be British. More than 70,000 barium enemas were carried out in England alone during 2011- and here was me thinking the last one must have been done around 1995. In this part of the SIGGAR trial, adults aged 55 or over with symptoms suggestive of colon polyps or cancer were randomised to have a barium enema or computed tomographic colonography, from 2004 onwards. The authors state that this is the first randomised trial comparing CTC and BE for diagnosis of colorectal cancer or large polyps in symptomatic patients. The predictive value of the two tests seems to have been calculated from the number of missed cancers presenting over the next 3 years: 7% for CT and 14% for BE. The barium enema is an obsolete investigation, and I can’t think of any good reason why barium should ever enter the British fundament from this day onwards.

1194   The first SIGGAR trial makes you wonder if the patients shouldn’t all have been offered colonoscopy, and indeed the next SIGGAR trial compares CTC with colonoscopy in a similar group of patients. Here CTC failed to spot only one cancer out of 29, whereas colonoscopy missed none out of 55. The authors conclude in favour of CTC as a less invasive investigation. I think patients should be offered a choice. I’m telling you, colonoscopy isn’t as bad as you might think.

1211   We sometimes hear it claimed that the so-called Mediterranean diet is good for you because it represents the historical ideal for human beings; arterial atheroma is our fault because we no longer follow this lost Edenic model. In reality, humans throughout their history have eaten a huge range of food, both as hunter-gatherers and as agricultural pastoralists – and whatever they have eaten, they have always developed some gunk in their arteries as they grew older. This wonderful study of atherosclerosis in mummies from around the world only goes back about 4,000 years, but the message is clear. The mean age of the bodies was just 36 years, but probable or definite atherosclerosis was noted in 47 (34%) of 137 mummies and in all four geographical populations: 29 (38%) of 76 ancient Egyptians, 13 (25%) of 51 ancient Peruvians, two (40%) of five Ancestral Puebloans, and three (60%) of five Unangan hunter gatherers (p=NS). Eat whelks. Or elks. Or nuts, or grains. Butter, olive oil or seal blubber. Beer. Mealy worms, peaches, samphire and locusts. Wine. Goats, mammoths, aurochs, whales, snails and anchovies. All these are good for you and bad for you.

BMJ  6 Apr 2013  Vol 346
There is one snag to the human diet, however, and that is how difficult it is for people to get enough vitamin D once they have travelled beyond a certain latitude. Exactly what the “ideal” level of vitamin D might be is still a matter of hot debate, but certainly as mankind fanned out beyond Africa towards the vast northern glaciers a few tens of thousands of years ago, it quickly became a matter of importance.

Even today, vitamin D insufficiency is very common in pregnant women throughout northern Europe and America, and this systematic review shows that it is associated with a variety of adverse outcomes in pregnancy. Limited evidence suggests that vitamin D supplementation in pregnancy may be something we’ll see more of in the future.

A few years ago, a BMJ cover showed a junior doctor with an ultrasound scanner dangling round her neck, asking if this was the future of bedside diagnosis. And why not? I would rather the SHO knew how full my bladder is before putting a catheter into my urethra, or just how high my pleural effusion extends before sticking a needle between my ribs. And many studies have shown that doctors trying to give intra-articular injections usually only reach the joint space if they use a bit of ultrasound guidance. This particular study shows how ultrasound improves the accuracy of lumbar punctures and epidural catheter placements. Get those machines to everyone please, before my turn comes.

About fifteen years ago, there was a lot of interest in the possibility that long term macrolide antibiotics might prevent coronary events by killing off Chlamydia pneumonia in arterial plaque. The ACES trial, published in 2005, finished off that hypothesis, and now the argument has turned full circle in a study which finds a higher rate of cardiovascular events in patients given clarithromycin in hospital for community-acquired pneumonia or infective exacerbations of COPD. The harmful cardiovascular effect of the macrolides seems very persistent in this and other studies, and takes some explaining. The drugs of greatest risk seem to be erythromycin and clarithromycin.

Ann Intern Med  2 Apr 2013  Vol 158
505   A nice little study randomised workers at the Children’s Hospital Philadelphia to 24 weeks of monthly weigh-ins, either offering individuals $100 a month for achieving a weight-loss goal, or sharing the prize money among groups of five. The sharing groups achieved better levels of weight loss: so altruism rules.

526   Once you have started taking a statin, is there ever any good reason to stop? I can’t think of any, except for intolerable side effects which do not respond to a change of agent. Yet more than half of this cohort of patients started on statins by doctors affiliated to the Massachusetts Hospital or Brigham & Women’s stopped their medication at some point. Fortunately this study found that contrary to widespread belief, “most patients who are rechallenged can tolerate statins long-term. This suggests that many of the statin-related events may have other causes, are tolerable, or may be specific to individual statins rather than the entire drug class.” The paper is behind a paywall, but the guide for patients is not.

544   I became interested in the comparability of diagnostic tests in the early 1990s, and went to the first conference on systematic reviews in diagnosis run by Jon Deeks about 17 years ago. Boy, has progress been glacial. Poor Jon is still reduced to asking for some proper basics: “Evidence derived from noncomparative studies often differs from that derived from comparative studies. Robustly designed studies in which all patients receive all tests or are randomly assigned to receive one or other of the tests should be more routinely undertaken and are preferred for evidence to guide test selection.” Doctors are for the most part blithely unaware of how utterly inadequate the basis for most clinical decision-making really is.

Plant of the Week: Skimmia “Rubella”

Last week, shivering with flu in a biting northerly wind, I refused to comment on any plant. Today the sun is shining, the wind has dropped, and the state of the garden is similar to early March, with a few new deaths due to unremitting frost. There is little colour and no scent.

So what are we looking forward to most as the soil of Britain’s gardens returns to warmth and fecundity? I think my personal choice would be the scent of skimmias, given that our supreme scent-giver, Daphne odora, is dead. Skimmias seldom die: in fact if any bit falls off a skimmia, it will generally persist and grow into a new plant after months or years. They are shade-tolerant little shrubs, and grow to every size from less than half a metre in the case of Bowles’ Dwarf to about two metres in the case of something like “ew Beauty” in full sun and a fertile soil.

All skimmias have a tantalising habit of producing masses of flower buds in neat pyramids from about October onwards, which refuse to open until late March when the daytime temperature usually gets above 10 degrees. I’ve chosen “Rubella” here because it is widely available and grows well, with good dark green leaves in all but the limiest soil. The pink tinge in its beautiful fragrant flower-heads must be why it shares the name of a once-common childhood illness. But do not be deceived by the “ella” ending. This clone is not a young lady at all, but a male, and will bring forth no berries. To me this is no great loss, but if you must have your berries, go for Skimmia “Reevesiana” instead.