Richard Lehman’s journal review—8 October 2012

Richard LehmanJAMA  3 Oct 2012  Vol 308
1333    Can vitamin D prevent the common cold? The answer is almost certainly yes, depending on baseline levels. If you run a trial in a place where the sun shines, good dairy products are abundant, and the ocean teems with oily fish, you might get a negative result. Such a paradise exists in the form of New Zealand, where this trial was done, and it was indeed negative for a monthly dose of 100,000 units. On the other hand, trials of supplementation in places where vitamin D deficiency is rife, such as Mongolia and Afghanistan, have shown a protective effect against colds, and also more serious infections. To inform practice in the United Kingdom, we need to set up a study in a region where the weather is bad and the diet is deplorable. There are just four countries to choose from.

1340    This week’s most fascinating and practice-changing paper must be this large observational study of beta-adrenergic blocking drugs and clinical outcomes in three classes of patients: those with a remote history of myocardial infarction, those with stable coronary artery disease (CAD) and those with risk factors for CAD. The primary outcome was a composite of cardiovascular death, nonfatal MI, and nonfatal stroke. The result of this propensity-matched analysis was even more startling than the authors let on in their abstract: not only was there no significant benefit to patients with past MI, there was statistically significant harm to those with CV risk factors but no known CAD (see Figure 1: HR 1.18 (95%CI 1.02-1.36). When I first started doctoring, ß-blockers (a brand new British invention) were considered a panacea for everything cardiovascular, except heart failure. Now it seems they are useless for most things cardiovascular, except heart failure.

NEJM  4 Oct 2012  Vol 367
1287    Cardiogenic shock is the kind of situation where people want to rush in and do things. It is not for me to criticize such people, as I am rarely faced with patients suddenly on the cusp of death; but reading the literature, it does strike me that their intentions are often better than their results. Case-hardened, cash-starved British cardiologists often refer to intravenous inotropes as “embalming fluid,” but they still sometimes use them: intra-aortic balloon support, on the other hand, has never really caught on over here. Nor should it, according to this admirable German trial which randomized 600 patients in cardiogenic shock following myocardial infarction to receive intra-aortic balloon counterpulsation or none, prior to intended revascularization. The 30-day mortality in both groups was the same, at 40%. The study had more than 240 primary end-points—an insurance policy gone mad—but it still could not prove that the procedure has any real value. These balloons should be filled with helium and sold at street corners.

1297   More disappointment comes in this long-awaited trial of prasugrel versus clopidogrel for acute coronary syndromes without revascularization. Since a substantial proportion of people lack the ability to convert clopidogrel to its active metabolite, the newer drug prasugrel (which doesn’t need this conversion) should be much the better platelet inhibitor. But this just doesn’t seem to happen in real life. Eli Lilly and Daiichi Sankyo spent tens of millions of dollars recruiting 9326 patients at 966 sites in 52 countries and following them up for two and a half years—all to prove that their new drug is no better than the old one which costs twenty times less. Much as we would welcome a better drug for cardiovascular protection, this is good news for cash-strapped health systems.

1310    Erasmus of Rotterdam visited Cambridge in England in the 1510s and was struck by its polyphonic music (which he couldn’t follow) and its futile scholastic learning, both of which he went on to ridicule in Encomium Moriae (literally Praise of Fools, but also containing a friendly pun on the name of his friend Thomas More). How little has changed in 500 years. The Cambridge (UK) authors of this study demonstrate that if you add high-sensitivity C-reactive protein to the risk score of a European population of mean age 60, you can reclassify enough intermediate risk individuals to make them worthy of a statin and so prevent one cardiovascular event (non-fatal or otherwise) in 10 years for every 400-500 people screened. That’s assuming you follow the current rules and don’t just lower the general threshold for statin prescribing to age 50, or something like that. You cannot help feeling sorry for the Emerging Risk Factors Collaboration which did this study: for all its clinical relevance it might as well be the St Thomas Aquinas Institute for Angel-Counting. But there are still a lot of great choirs to be heard in Cambridge, and someone could write a motet for them about it: O sensitivissimum protein C-reactivum, imple in cordibus servorum tuorum etc.

1353    What can we do about the problem of missing data in studies? America puts its best minds to the task, but they don’t really get as far as Donald Rumsfeld:
“Missing data threaten the validity of many clinical trials. In this issue of the Journal, the members of an expert panel convened by the National Research Council (NRC) provide recommendations regarding the design, conduct, and analysis of studies to minimize that threat. The authors define missing data as “values that are not available and that would be meaningful for analysis if they were observed.” They find that there is no analytic approach that can assuredly produce unbiased estimates of treatment effects when relevant data are missing and therefore recommend that investigators place increased emphasis on strategies for designing and conducting studies to minimize missing data.” Ah, yes, good. And what about all the other data that are missing, because nobody has ever published the studies?

Lancet  6 Oct 2012  Vol 380
Britain’s premier medical journal gets more bizarre with each issue. Thrombectomy devices for acute stroke have yet to earn any place at all in mainstream clinical practice, but the Lancet gives all its research space this week to commercial studies comparing one device with two others. There’s a bit of global stuff of specialist interest, and an editorial with the intriguing title Ten Types of Breast Cancer which concludes—if I understand the last paragraph—that there can be a great deal more than ten types even in a single tumour. I am puzzled. I would like to like the Lancet, but it’s very difficult.

BMJ  6 Oct 2012  Vol 345
I don’t find it at all difficult to like the BMJ, though it tends to err in the opposite direction: it tries almost too hard to be readable and to grab headlines. Is the sunbed really a carcinogen in the same league as tobacco and asbestos? There is certainly an association between sunbed use and squamous cell cancers of the skin, as you’d expect. Sun-lovers and tan freaks get these tumours, and dermatologists and plastic surgeons make a living removing them. This is all in the settled order of nature. Sun-lovers and tan freaks also use sunbeds. The weakness of any association with malignant melanoma is striking, and the meta-analysis (which lacks mortality figures) of the observational studies is in line with other evidence that ultraviolet exposure in adulthood is only a minor factor in the pathogenesis of this cancer in its lethal form.

Tranexamic acid is a great drug for saving lives after major trauma – that much was clear from the CRASH-2 trial, published 2 years ago in the Lancet. A really important trial, that: and the CRASH investigators could be forgiven for trying to get a few more papers out of their study. For example, they prestratified their patients so that they could determine if there were subgroups who would benefit more than others. So when somebody else came along and asked to do this analysis, did they tell them to get stuffed? Did they argue that the data were theirs alone to keep, to shed future academic glory on the triallists? No: they just handed over their data. Hurrah – this is how science should work! And then when the analysis was complete, the BMJ made it available free to everybody. Hurrah – this is how publication should work!  A great example of open science, in every sense.

Ann Intern Med  2 Oct 2012  Vol 157
471    Whenever you read a paper about diagnosis, you need to ask two questions. Firstly: will this help me treat patients? and secondly: if there is a gold standard test, why don’t we use this instead of the comparator? (A corollary is that if there is not a gold standard test, what does this diagnosis really mean anyway?) This study of renal dysfunction in elderly people is a good illustration. Will a more accurate test for renal impairment in people over 70 help me to treat them more effectively? You can think up your own answer to that: it will probably vary between “not really” and “sometimes.” Let’s say we want to do it anyway: what’s the best method known to man? I guess it’s some very invasive method of measuring glomerular filtration rate directly under lab conditions. Here it is said to be “GFR measured as the plasma clearance of the endogenous marker iohexol.” I was really surprised by that: I had no idea that iohexol is an endogenous marker; and in fact it is not. It is an iodine-containing substance which we know better as Accupaque, and you have to inject 3235mg of it into people who have fasted for 4 hours and then take 8 blood samples over 5 hours. This study really had to be done in Germany: where else would you find 610 people aged 70 or over with the fortitude to go through with this? But for clinical purposes, it is much easier to measure plasma cystatin C, which truly is an endogenous marker. Or, for a more approximate idea, you can just measure plasma creatinine. Or you can combine the two and come up with the Berlin Initiative Study 2 score, which is within 12% of the iohexol GFR. Are you getting a bit tired of all this? So am I. But I may soon be wanting to measure renal function as a prognostic factor in elderly people with advanced heart failure, and the BIS2 score could come in handy for that.

498    Now for a real downer: Antipsychotics in Adults With Schizophrenia: Comparative Effectiveness of First-Generation Versus Second-Generation Medications: A Systematic Review and Meta-analysis. “Conclusion: Clear benefits of FGAs versus SGAs for treating schizophrenia remain inconclusive because of variation in assessing outcomes and lack of clinically important differences for most comparisons. The strength of evidence on safety for major medical events is low or insufficient.” Thirty years of psychopharmacology and this is where we are. The whole appalling tale is best told by David Healy in Pharmageddon (2012), esp. chapter 5.

Plant of the Week: Clematis campaniflora

This is the very last of the clematis species to flower, providing a mass of very pale blue flowers with prominent stamens on vigorous woody stems which go off in every direction. I call it a species, although some botanists demur and refer to it as C viticella var campaniflora. I can’t really believe that the plant we have by this name is just a variety of viticella: it’s just too vigorous and the leaves are wrong. Perhaps ours is actually a hybrid with our native old man’s beard (C vitalba).

It’s a very welcome plant for its generosity at this late stage of the season, but you need to watch out. At the very first frost, it turns into an ugly mess of blackened leaves. Also, it is too vigorous for most small gardens. However, you can remedy both these faults by cutting it right down to the ground in November. That way it won’t be an eyesore once it has flowered, and you can keep it under control.