Richard Lehman’s journal review—14 July 2014

richard_lehmanNEJM 10 July 2014 Vol 371
107  I was very confused by this paper. It describes two trials of three drugs for premenopausal breast cancer with various permutations, and the bottom line is that all the interventions give the same result. Or, if you are a sponsor of the trial, you can report: “In premenopausal women with hormone-receptor–positive early breast cancer, adjuvant treatment with exemestane plus ovarian suppression, as compared with tamoxifen plus ovarian suppression, significantly reduced recurrence.” That is just about true: 88.8% of the women in the tamoxifen group were recurrence-free at five years, compared with 92.8% in the exemestane group, but there was no difference in mortality. This figure was reached by pooling the two trials, though they were not identical. One of them used a variety of means for ovarian suppression and the other only triptorelin. OK, enough of this. If you are a woman with premenopausal breast cancer, demand an option grid to explain the possible adverse effect of these regimens, as well as their minutely different long term benefits. And if you are unhappy at the summary reporting, just think of the bad old days when that might have been framed as a “30.5% reduction in recurrence at five years.”

119  The next study would also be a nice teaching paper to demonstrate all the various ways of presenting data: absolute percentages, rate ratios, how to use confidence intervals, tabulated data, Kaplan-Meier charts, and so on. One evening when you are in a medical library, in the middle of an essay crisis, pick up a few old medical journals and you will see how papers used to look before these things were standardised. Quite a mess. But that won’t get your essay written, will it? This one describes a trial of letrozole versus clomifene to treat infertility due to polycystic ovarian syndrome (PCOS). If you can get hold of a copy in colour, see which format you think the data stand out best in. For me, the Kaplan-Meier charts win hands down. You can see at once that letrozole produces more ovulation and more babies, especially in women of BMI 30-39. It’s hard to imagine medical papers without these simple graphics, which only came along in 1958. And letrozole seems to be a real advance in the treatment of PCOS related infertility.

130  And now, for even more people, something even more promising. In treatment-resistant eczema, “Patients treated with dupilumab had marked and rapid improvement in all the evaluated measures of atopic dermatitis disease activity. Side-effect profiles were not dose-limiting. (Funded by Regeneron Pharmaceuticals and Sanofi).” Golly, this could be big. It could cure eczema and bankrupt health systems. Or it could make limbs drop off, though that is less likely as it has been around for a few years. It blocks signalling from both interleukin-4 and interleukin-13, and it has already been shown to work for asthma. Although rhinitis was more common in the treatment groups, skin infection was more common in the placebo groups.

OL  Next, psoriasis. And secukinumab, which targets interleukin-17. According to two phase 3 trials, it works very well indeed, with a PASI 75 score at 12 weeks of 81.6% with 300 mg of secukinumab, 71.6% with 150 mg of secukinumab, and 4.5% with placebo in the first trial (ERASURE). This score is unusual because it does not apply to individuals but to the group: how many had skin clearance of more than 75%, plus a few other things. Now this is pretty amazing: the two major skin diseases conquered in a week. We must arrange trips to the moon for people with eczema and psoriasis, because I hear these drugs are affordable there.

JAMA 9 July 2014 Vol 312
145  Antibiotics before, during, but not afterwards: that seems to be the FRENCH rule when operating for mild-moderate acute cholecystitis due to gallstones. The patients in the FRENCH trial had co-amoxiclav 2G tds until surgery with one dose during, and were thereafter randomised to more or none postoperatively. There was no difference in postoperative infection rates. The trial was conducted in France. Three years ago, I was hunting for the first trial to mention the word “outcome,” and I hit upon a trial of acute surgery for gallstones in Massachusetts in 1904. The outcome? Death in 54% of cases.

155  Nicotine is a nerve poison that plants produce to kill the insects that would like to eat them. So is laburnum toxin, which binds to some of the same receptors. Varenicline is made out of laburnum toxin and when taken orally it binds to the nicotinic cholinergic receptor, and specifically the α4β2 receptor subtype, with a more powerful affinity than nicotine. So we use varenicline to help people stop smoking because it reduces their craving for nicotine, and, at the same time, means that they get little effect from smoking. But now children, what happens when you use both varenicline and nicotine replacement therapy (NRT) at the same time? Theoretically, there should be no extra effect, but somewhere in the biochemical tussle that results, people find benefit. This was a double blinded trial that achieved very high abstinence rates at six months: 65% in the varenicline and NRT group and 47% with varenicline alone.

171  “How to Read a Systematic Review and Meta-analysis and Apply the Results to Patient Care.” Golly, I sometimes wish I had been born a fox huntin’ man. No meta-analyses for them. In bygone Oxfordshire on hunt days, some country GPs would turn up to work in hunting pink and boots, and after talking to a few adoring old ladies and writing a couple of prescriptions, would leap on their horse and ride with the hounds. There is only one such left now. Out in the wilds of Minnesota, they don’t hunt foxes because the winter is so harsh that they can rarely appear above ground. And that’s just the humans. I imagine most of the authors of this essential paper to have been spending the winter in some kind of warm burrow thinking their great thoughts and communicating via underground wires, except for the two in Chile and India. I know that some of them made it through, because I saw Thomas Agoritsas alive in the spring. OK: I have two conflicts of interest. Some of these authors are friends. And part of my job is to tell UK Cochrane how its systematic reviews are relevant to primary care. But readers lacking these motives should still try and get hold of it, because it is a landmark summary: it marks where the land of medicine presently stops, and where the new land lies.

Lancet 12 July 2014 Vol 384
OL  The Lancet has published some interesting papers this week, but none of them are in the paper journal. Two vaccine trials stand out. The first is of a tetravalent dengue vaccine given as a course of three immunisations to children in the Asia-Pacific region. So far as I can see, it does not have much effect. Two hundred and fifty cases of virologically confirmed dengue took place more than 28 days after the third injection: 47% in the vaccinated group and 53% in the controls. This satisfied their primary endpoint but only because it was so unambitious; they wanted “the lower bound of the 95% CI of vaccine efficacy to be greater than 25%.” The vaccine may have saved 23 children out of 10 275 from getting dengue, while failing to protect 117. I must be missing something, because they describe this as efficacious.

OL  One vaccine that does work is oral polio vaccination (OPV). Without the Taliban, we might now be mopping up the last traces of polio from the face of the earth. But alas, it will be some time yet. At some point in the process, we may need to boost the immunity of those already protected by oral vaccination because of new imported outbreaks. This Gates funded trial tested the booster response to intramuscular inactivated polio vaccine given to children aged from one to four years from Chinnallapuram, Vellore, India, who were healthy, had not received IPV before, and had had their last dose of OPV at least six months before enrolment. I don’t know why these children from India were chosen: they could equally well have been from Portugal or Canada as far as I can see. Anyway, the booster dose worked well, so this may be a useful trick to have in reserve. You can’t be too careful when sharing a world with men who feel it’s a divine mission to kill those who are trying to protect their children.

The BMJ 12 July 2014 Vol 349
This issue of The BMJ makes me wonder if the words “telomere” and “Mendelian randomisation” are just ways of making readers’ eyes glaze over. Telomeres are the things at the end of your chromosomes that show how much longer you have left to live. They do not cause anything. A study here finds that “available observational data show an inverse association between leucocyte telomere length and risk of coronary heart disease independent of conventional vascular risk factors. The association with cerebrovascular disease is less certain.” Is this interesting? Well, maybe, in a spooky kind of way. Short telomeres=earlier death. Man with sickle has many blades. Hah!

Someone recently told me that the link between alcohol and reduced coronary disease is purely observational, and that therefore we should not recommend alcohol as part of the “Mediterranean” diet. I didn’t want to argue, but you could say much the same about smoking and cardiovascular disease. The evidence of benefit from alcohol is solid, robust, and repeatedly found wherever you look, but almost impossible to replicate experimentally for the very good reason that people who drink do so as part of their daily pleasure. Yet the several hundred authors of this paper have tried to do something even more impossible: make this evidence disappear by a Mendelian hat trick. I am completely baffled that they should (a) want to do it and (b) think this is good enough: “Individuals with a genetic variant associated with non-drinking and lower alcohol consumption had a more favourable cardiovascular profile and a reduced risk of coronary heart disease than those without the genetic variant. This suggests that reduction of alcohol consumption, even for light to moderate drinkers, is beneficial for cardiovascular health.” No it doesn’t.

Plant of the Week: Magnolia delavayi

This is one of those great slothful beasts from primeval antiquity. It lies there, twisted and vast. You tap it: nothing happens. Then, just as you are turning, you mark the opening of an eye. No, of course not: it was just the light playing on a knot of wood. Magnolia delavayi sleeps on, shrouded in vast leaves of olive green. The light plays on the knots and the leaves, twisting sometimes.

Then one morning in July, a flower appears. There was no flower here before. A bulge perhaps, of olive green. Now this flower, 20cm wide, of olive white. You smell it: skies of Avalon, steel of Excalibur, honey of Mordred. There was no flower here, says tomorrow. A bulge perhaps, corrugated.

Climb me, says the tree. No one dares.