NEJM  29 Aug-12 Sep 2013  Vol 369
799   When comparing a new fixed-dose anticoagulant with conventional warfarin based anticoagulation, what do you look for? Thrombotic events, death, major bleeding? Yes, certainly. But what makes the comparison meaningful? That depends largely on how well the warfarin group were kept within the target INR range. In the AMPLIFY trial, 5395 patients with new episodes of venous thromboembolism were randomised to apixaban twice daily for 6 months or to conventional treatment with enoxaparin followed by warfarin. The context of the trial is not described in the main paper, but from the appendix it seems to have been conducted in 28 countries, and it takes 14 pages to list the contributing centres which recruited patients. Half of them recruited fewer than ten subjects, and 31 centres recruited only one. There were numerous exclusion criteria. Overall, these hundreds of centres only managed to keep their patients within the INR target area of 2-3 for 61% of the time. So if you’re happy to accept that as your comparator, then apixaban is non-inferior to warfarin and causes fewer major bleeds. Otherwise, you might want to ask yourself why the drug companies who paid for this trial chose this way of doing it, and whether that 61% figure is what you hope for when you give warfarin to your patients.

901   Over the fifteen years I’ve been watching the journals, I’ve had a grandstand view of the shift from thrombolysis to immediate percutaneous intervention as the treatment of choice for myocardial infarction. And still the shift goes on towards ever-quicker PCI, because there is good evidence that time means myocardium. This important study shows that in the USA, median door-to-balloon time in hospitals fell from 83 minutes to 67 minutes over a three year period (the years July 2005 to June 2006 compared with July 2008 to June 2009). But in-hospital mortality did not change. This is interesting but hard to interpret: much depends on what happens to the patient before reaching the hospital door. My take on all the studies—including those of thrombolysis given by paramedics—is that the sooner you reopen the coronary supply, the better. What matters critically is blockage to reopening time, not any subset of that.

999   Moving swiftly on, we get to this week’s issue of the mighty journal. Here the first paper concerns itself with a comparatively small issue in the treatment of myocardial infarction: the timing of prasugrel administration in relation to percutaneous intervention for non-ST-elevation MI. Do I see your eyes glaze over? OK, we are not all interventional cardiologists, but for the record, if you give the prasugrel before coronary catheterization, you get more bleeds but no extra benefit.

1011   And now for something completely different: male oestrogen deficiency. Bet you haven’t heard of that before. Men produce quite a lot of oestrogen by aromatizing testosterone, but we have little idea of what it might do for good or ill. Investigators at the Massachusetts General Hospital set out to find out by recruiting 400 healthy men between the ages of 20 and 50 and subjecting them to a 16-week period of gonadotrophin suppression (using goserilin) together with various degrees of testosterone supplementation, or none. Crucially they gave some participants anastrozole to stop them aromatizing testosterone, so that they became oestrogen deficient. This allowed them to distinguish between the effects of male oestrogen deficiency and testosterone deficiency. “The amount of testosterone required to maintain lean mass, fat mass, strength, and sexual function varied widely in men. Androgen deficiency accounted for decreases in lean mass, muscle size, and strength; estrogen deficiency primarily accounted for increases in body fat; and both contributed to the decline in sexual function.” Gosh, I can see this becoming an endless playground for the pharmaceutical industry and andrologists with big cars.

JAMA  28 Aug -11 Sep 2013  Vol 310
797   IDLE is the brilliant acronym someone has come up with for those “early neoplasms,” “precancerous growths,” and “carcinomas-in-situ” which scare the living daylights out of healthy people, who become instantly converted into cancer patients as the result of some test or biopsy. It stands for “indolent lesion of epithelial origin.” For more discussion, read this dry but comprehensive short piece about Overdiagnosis and Overtreatment in Cancer.

812   It stands to reason that if you visualize the whole common bile duct during cholecystectomy, using intra-operative cholangiography, you will help to avert common bile duct injury. But things that stand to reason in medicine are often wrong. The authors of this study of bile duct injury rates in Texas find that patients who had cholangiography during cholecystectomy had a CBD injury rate of 0.21% compared with a rate of 0.36% in those who did not. But then they put the figures through some statistical software. First of all some logistic regression analysis controlling for patient, surgeon, and hospital characteristics: the difference remains significant, in favour of cholangiography. Then some instrumental variable analysis “to control for confounding”—and the difference becomes non-significant. To check this out you need to read two sections of the paper (one of which is only comprehensible by stats geeks who are familiar with the Mahalanobis distance method) plus a whole supplement. But ultimately all talk of confounding is just guesswork about what you think might matter. “Intraoperative cholangiography is not effective as a preventive strategy against common duct injury during cholecystectomy,” the authors tell us. Well, maybe: I haven’t a clue about the Mahalanobis distance method, but I can see that any difference it makes is very small.

JAMA Intern Med  9 Sep 2013  Vol 173
1506  “Those who live to an older age are likely to be disabled, and thus in need of caregiving assistance, many months or years prior to death. Women have a substantially longer period of end-of-life disability than men.” This is the conclusion of a study of people who died while enrolled in the Health and Retirement Study between 1995 and 2010. I don’t think it will come as news to most of my readers, unless they happen to include the current Secretary of State for Health for England.

1514   Osteoporotic vertebral compression fractures are awful: patients sit painfully in their chair by day and night for weeks, drugged with half-effective morphine, constipated, anorectic, and dismayed at the prospect of future disability. Vertebroplasty and kyphoplasty offer a surgical route out of their hell, and it’s no wonder that many patients clutch at it. In the USA, where these procedures are much more accessible, there are observational data to show that patients who undergo these procedures have a lower mortality than those who do not. Now why should that be, ask the investigators of this Medicare database study. And the answer is clear: these patients are fitter to start with. “After accounting for selection bias, spinal augmentation did not improve mortality or major medical outcomes and was associated with greater health care utilization than conservative therapy. Our results also highlight how analyses of claims-based data that do not adequately account for unrecognized confounding can arrive at misleading conclusions.”

1547   I got interested in heart failure in the early 1990s, and helped to draw up the first Oxfordshire guideline on the subject. Avoid beta-blockers, we said in 1993. Then, oops, in came the first carvedilol trials, and a couple of years later we said introduce them cautiously in all stable patients without contra-indications. Now it is a matter of dogma that all patients with systolic heart failure should be on one of the beta-blockers shown to reduce mortality in a randomized trial. No choice in the matter: you must take these because you will live longer. But most patients dying from heart failure don’t want to live longer but to feel better. And in unstable heart failure, going on a beta-blocker can make you a whole lot worse. Here is a short report on the way that beta-blockers are handed out in US hospitals to heart failure patients who are at risk of clinical instability. The lead author is Kumar Dharmarajan, a brilliant young researcher who is interested in all the ways we mishandle heart failure patients when they come to hospital. He will know he has succeeded when he has nothing left to write about.

BMJ  7-14 Sep 2013  Vol 347
Influenza is rarely out of the medical journals in some guise or other, and you would think that after one recent pandemic, and seasonal epidemics since records began, we might at least know who is most likely to suffer harm from it. But we only have the vaguest idea, according to a trawl through the vast literature carried out by a team from McMaster. Ethnic minorities and pregnant women—no clear evidence. Women in the first four weeks post-partum—yes. Some evidence for obesity, cardiovascular disease, and neuromuscular disease. “Rigorous and adequately powered studies are needed.”

Data from the Nurses’ Health Study show that “Greater consumption of specific whole fruits, particularly blueberries, grapes, and apples, is significantly associated with a lower risk of type 2 diabetes, whereas greater consumption of fruit juice is associated with a higher risk.” If you are an American nurse, eat blueberries! Let me peel you a grape. I shall dash that glass of orange juice from your hand, for you must never become diabetic, my darling. Why, I could write a whole novel about this paper.

For many years now, Paul Glasziou has been trying to compile a “Non-Pharmacopeia”—an evidence-based collection of alternatives to drug treatment. But as he and two colleagues explain in this paper, it is very easy to track a pharmacological intervention through the literature but much harder to track a non-pharmacological intervention, because they tend to be labelled and characterized in a variety of ways. Only 39% of such papers contain an adequate description of the procedure, which can be increased to 59% by the laborious process of contacting the investigators.

Lancet  31 Aug – 14 Sep 2013  Vol 382
769   The disclosure of concealed harms that led to the withdrawal of Vioxx (rofecoxib) in 2004 suddenly made us aware of the cardiovascular risks of non-steroidal anti-inflammatory drugs, and from then on league tables of harm have appeared regularly. Not that they have had all that much effect. Diclofenac, which carries a risk at least as great as Vioxx, remains a highly popular drug that is still prescribed long-term to innumerable high-risk patients. Now, from the Coxib and traditional NSAID Trialists’ (CNT) Collaboration, comes what is probably the definitive examination of the relative risks of the various NSAIDs. As usual naproxen comes out best for lack of cardiovascular risk, while the others jostle for position in the harm league.

780   Business as usual: “Ustekinumab significantly improved active psoriatic arthritis compared with placebo, and might offer an alternative therapeutic mechanism of action to approved biological treatments” says the conclusion of this paper, half of whose authors were employed directly by Janssen, the manufacturers of ustekinumab. So why, pray, was the comparator placebo rather than an approved biological treatment? And why did this trial get ethical approval? And why is it in The Lancet?

941   And more business as usual: “The sponsor of the study (Janssen) had a role in study design and conduct; data collection, analysis, and interpretation; and writing of the Article.” The title of the Article? “Efficacy and safety of canagliflozin versus glimepiride in patients with type 2 diabetes inadequately controlled with metformin (CANTATA-SU): 52 week results from a randomised, double-blind, phase 3 non-inferiority trial.” The results? Well, efficacy was measured—you guessed it—by a reduction in HbA1c. And safety—in a drug which might be used by millions for up to half a lifetime—was what happened in a year to 968 patients. Yes, and it took 157 centres in 19 countries to recruit the 1452 participants. Does business as usual give us the slightest idea of how to treat patients with type 2 diabetes? You might like to ask any Janssen rep who comes along with a reprint of this paper.

951   Now—relief—back to another useful survey of a dangerous group of drugs. You will recollect that the life expectancy difference between people who take drugs for schizophrenia and the general population is fifteen years and rising. Traditionally, these drugs have been classed as “typical” or “atypical” antipsychotics, but this classification is absurd and gets short shrift in this paper. Clozapine, which first appeared in 1971, is clearly the most effective of all the drugs for relief of psychotic symptoms, but may be the worst for causing death. Olanzapine is top of the league for weight gain. And so forth. Even if you don’t prescribe half the drugs in this analysis (sertindole? asenapine?), you need to read it for what it tells you about those you do. It’s not one bit comforting.

Plant of the Week: Gentiana asclepiadea

The willow gentian is a plant of mountains and woods but does well in most gardens, except ours, of course, where it died long ago of neglect/slugs/rubbishy soil/lack of moisture/poor climate (underline as many of these as you think apply).

Unlike the spring gentians, these autumn-flowerers don’t mind a bit of lime in the soil, but they do need moisture. They are invaluable—along with ceratostigmas—for providing flashes of rich blue flower among the yellowing and reddening leaves of the declining garden.