JAMA  13 Mar 2013  Vol 309
997   Obstructive sleep apnoea is very, very common. In the preamble to this Australian study, a third of the adult population is accused of having symptoms that might be OSA. Previous studies have shown that polysomnography has no significant advantages over a home video, reporting by sleeping partners, or a trial of self-titrated continuous positive airways pressure. So the question arises whether specialist sleep clinics are really needed, or whether a sleep service of equal effectiveness can be provided in primary care. In this randomized non-inferiority study, the results of the two modes of care were equally good. So come on, enterprising commissioning GPs: buy a few CPAP machines and offer a local service. Nobody seems to care about conflicts of interest any more, and I guess if this coalition lasts a year or two more, you’ll be able to charge patients at the door on the grounds that OSA is “just snoring” and not worthy of NHS provision. In fact it is a powerful cardiovascular risk factor and a major prognostic indicator e.g. in heart failure.

1005   Fervent advocates of breastfeeding have suggested that one reason that so many working class children are fat is because their mothers feed them from the bottle. This interventional study from Belarus gives the lie to this middle-class myth. “Among healthy term infants in Belarus, an intervention that succeeded in improving the duration and exclusivity of breastfeeding did not prevent overweight or obesity, nor did it affect IGF-I levels at age 11.5 years.”

1014   It’s a troublesome fact that most people who give up smoking (and come off nicotine replacement) gain weight. Together with the desperately addictive qualities of nicotine, this is a major cause for people to continue smoking. I am sure that you already tell them that the health benefits of giving up far outweigh the health hazards of a few extra pounds in weight; and the Framingham Offspring Study proves you right, with 6 years of follow-up showing a halving of cardiovascular events in the quitters, irrespective of weight gain.

NEJM  14 Mar 2013  Vol 368
987   Last week, I lambasted the NEJM for publishing a truly terrible pharma-funded trial, but this week there are none of these, presumably because the feeding needs of the publishers have been satisfied for the time being. I await next week’s issue with interest. Meanwhile, enjoy this excellent study based on long-term data from all women in Stockholm and Denmark who underwent radiotherapy for breast cancer under the age of 70 between 1958 and 2001 in the case of the Swedes, and under the age of 75 between 1977 and 2000 in the case of the Danes. Their radiation doses were calculated using original records where possible, and they were compared with carefully matched controls for the occurrence of cardiac events. These themselves were carefully validated from hospital records and autopsies. “Exposure of the heart to ionizing radiation during radiotherapy for breast cancer increases the subsequent rate of ischemic heart disease. The increase is proportional to the mean dose to the heart, begins within a few years after exposure, and continues for at least 20 years. Women with preexisting cardiac risk factors have greater absolute increases in risk from radiotherapy than other women.” Observational science at its best.

999   Doctors love new diagnostic technologies, and I’m no exception. The greatest diagnostic advance in my working lifetime has been magnetic resonance imaging and I still stand in awe when I look at slices through people showing all the structures I merely guess about. For example, when someone comes along with pain all the way down their leg accompanied by a bit of tingling and reduction of straight leg raise, I deduce that they have S1 nerve compression due to disc prolapse. How satisfying when they have their MRI and that is exactly what is reported: not like the bad old days when you only had a plain X-ray to guess from. But actually, this Dutch study shows that MRI has no predictive value whatsoever in sciatica. Remember your ROC curve? You’ll be aware that an area under the curve of 0.5 equates to pure chance. Here is a follow-up study of a randomized trial of surgical treatment versus conservative management in 283 people with sciatica. “At 1 year, 84% of the patients reported having a favorable outcome. Disk herniation was visible in 35% with a favorable outcome and in 33% with an unfavorable outcome (P=0.70). A favorable outcome was reported in 85% of patients with disk herniation and 83% without disk herniation (P=0.70). MRI assessment of disk herniation did not distinguish between patients with a favorable outcome and those with an unfavorable outcome (area under ROC curve, 0.48).” And you thought that improved physical diagnosis would always mean better patient outcomes? Think again, doctor.

1019   If you’re looking for great outcomes research in cardiology, you could find most of it by going to PubMed and searching under the names of Harlan Krumholz, John Spertus, and increasingly Brahmajee Nallamothu. Not too surprising, then, that all three of these names appear on this paper about long-term outcomes in elderly survivors of in-hospital cardiac arrest. The figures are much better than you might surmise: “Among elderly survivors of in-hospital cardiac arrest, nearly 60% were alive at 1 year, and the rate of 3-year survival was similar to that among patients with heart failure. Survival and readmission rates differed according to the demographic characteristics of the patients and neurologic status at discharge.”

1033   Antiphospholipid syndrome is the unsatisfactory general term given to a range of disorders, from lethal multi-organ failure due to small vessel thromboses to a tendency to miscarry repeatedly. Do we know what causes it? Well, it has to be something to do with antiphospholipid antibodies or it would have a different name. Delve deep into this extremely learned article and you are bound to conclude that we don’t know a great deal more. My general rule for everything from string theory to prions is that if I can read a whole page without anything happening but my eyes glazing over, then it’s not worth persevering with.

Lancet  16 Mar 2013  Vol 381
907    The Lancet specializes in finding ways to exasperate readers. One of its great themes is global health, and the way that violence interferes with achieving it. That’s noble, and good, and I am glad there is a medical journal that takes on these issues. But then you look at Richard Horton’s Offline, which this week contains sentences like: “We should also consider the enabling environment for universal health access. Governance is a critical enabler that should be given the highest priority. The global community must commit to health information as an indispensable part of the governance of sustainable development. Only with reliable and transparent data can full and participatory independent accountability (a further enabler) be realised.” Boy does this man need a good English teacher with a red pencil to strike through it and write “WHAT DOES THIS MEAN?? SEE ME”. And it’s the same with his choice of papers (assuming he is the one who chooses them).

“Violent offending by UK military personnel deployed to Iraq and Afghanistan: a data linkage cohort study” is a worthwhile study conducted by worthwhile medical authors, which shows that if young men come forward in their teens to be trained to kill and are sent to wars which are widely held to be illegal and/or unwinnable, they come back with high rates of psychological dysfunction and alcohol dependency: they find it hard to adjust to normal life, and are much more likely to commit crimes of violence than men who have opted to become bus drivers. Is this the kind of thing we turn to a leading medical journal to discover?

918   But wait: here is that rarest of beasts, a pharma-funded study which provides useful clinical knowledge as well as providing potential reprint income for The Lancet. A lot of people with rheumatoid arthritis achieve a low level of disease activity by taking methotrexate by mouth combined with weekly injections of etanercept, a fusion protein developed by Pfizer to mop up tumour necrosis factor α from the circulation. The aim of the PRESERVE trial was to assess whether low disease activity would be sustained with reduced doses or withdrawal of etanercept in patients with moderately active disease: an important clinical question affecting a lot of patients. It is also a question which affects Pfizer, since a year’s standard dosing with etanercept costs about $20K and brings the company an annual income well in excess of $3bn. And hey presto, this trial finds that patients who carry on taking etanercept with methotrexate have far fewer relapses than patients continuing on methotrexate alone. Great news for Pfizer, then: maybe great news for some patients with RA; terrible news for cash-strapped health systems until etanercept comes off patent. And the thoughtful editorial picks up on another issue: if long-term treatment with etanercept is going to become routine, what do we know about the long term effects of etanercept on other outcomes, such as infections and cardiovascular events? The drug licensing authorities are finally waking up to the fact that all drugs for long term use need long term safety trials—before and not after licencing for use on the public.

BMJ   16 Mar 2013  Vol 346
Interventions for depression tend to have effect sizes in direct proportion to the severity of the original depression. Thus serotonin reuptake inhibitors, for example, show some evidence of effectiveness in moderate-to-severe depression but little for mild depression. This meta-analysis of individual patient data (more of these, please!) shows that the same applies to low-intensity non-pharmacological interventions too: though the absolute difference is very small, and I don’t know what the take home message for clinicians might be.

Reader, I am a terrible person: I have never used a computerised clinical decision support system. I began to practise medicine in the middle of the 1970s, when far-sighted medical thinkers predicted that before the end of the millennium, most diagnosis would be carried out by computers. It is now thirteen years into the new millenium, and I am old and bald and grey: and although I see young doctors using hand held devices all around me, something in their demeanour tells me that they are not using them primarily for the purposes of diagnosis. This big Canadian meta-regression study led by Brian Haynes looks at 162 randomized trials of electronic clinical decision aids to determine the characteristics of those that work least well, or best. Systems more likely to succeed provided advice for patients in addition to practitioners and required practitioners to supply a reason for over-riding advice. Developers who tested their own systems also tended to come up with favourable assessments. This paper is a great example of the use of simple and multiple logistic regression models to test characteristics associated with system effectiveness, employing several sensitivity analyses. Now what we need is some good qualitative research on barriers to use of these systems, and some realist evaluation of both the qualitative and the quantitative data. Who knows, then perhaps computer support may come to play a significant role in diagnosis before the end of the current millennium.

μῆνιν ἄειδε θεὰ Πηληϊάδεω Ἀχιλῆος
οὐλομένην, ἣ μυρί’ Ἀχαιοῖς ἄλγε’ ἔθηκεν,
were the most famous lines in ancient poetry, beginning Homer’s Iliad. They curse the rage of Achilles, which brought great suffering to the Greeks; though ironically the lasting effect of the epic was to glorify military values, so causing suffering to countless millions more over the next 3,000 years. Anyone who has tried to make their way through the whole Iliad will remember skipping many pages describing interminable military games on the windy plains of Troy; so it is very appropriate that this clinical review of Achilles tendon disorders is co-authored by a director of military sports medicine. You’d never have thought that a single tendon could have so many disorders, and so with little evidence to guide their treatment. However, it seems that nowadays very few of them are caused by fickle goddesses or poisoned arrows, and none are fatal.

Plant of the Week: Crocus vernus

We have all loved crocuses since we were kids, and in the end the best ones are the big fat purple, yellow and white kinds that form part of the earliest memories of anyone growing up in northern Europe. These crocuses that we rejoice in for a week or two each year all have ancestors from the mountains of Anatolia and the Near East, and have since been have been fattened up, hybrized and selected by Dutchmen over several centuries. No garden is too small for a few crocuses, and some varieties will of course grow and multiply in lawns. As lawns are generally such pointless things, this is worth trying.

For determined collectors, it is possible to find and cultivate rare species of crocus, but these all tend to be sickly plants with washy mauve flowers, so I can’t much see the point of crocus snobbery. Crocuses are meant to shout and sing from lawns and borders and roundabouts and supermarket car parks. Leave the runtish ones to the mountains, and rejoice that we have the loud plump Dutch ones to gladden our little children.