Richard Lehman’s journal review—27 December 2012

Richard LehmanJAMA  19 Dec 2012  Vol 308
2469    Most healthy people don’t take aspirin, so if you look at a long-term study of aspirin in relation to some clinical event, such as macular degeneration, you have to be very careful to allow for confounding by indication. And what is the main indication for aspirin? Why, it is cardiovascular disease. And what is the main cause of cardiovascular disease? Why, it is smoking. And what is the main cause of age-related macular degeneration? Why, it is also smoking. So you can go round some nice logical short-circuits if you just go comparing the amount of long-term aspirin use with the amount of macular degeneration fifteen years later. The Beaver Dam Eye Study provides a longitudinal data set to interrogate for associations between AMD and a number of factors. We don’t know how many hypotheses these authors pre-specified before they put the data through their stats programme, but we do know they came up with a positive relationship between aspirin use and AMD ten years later. Do I believe it? No: and nor do the authors, really, because they ask for more studies. Their dataset just doesn’t seem good enough to me to support this level of beavering. There is only point-of-time information every 5 years on smoking and on aspirin use: nothing you can really build a dose-related relationship out of, or even (in my view) a plausible hypothesis.

2489    The Look AHEAD (Action for Health for Diabetes) study is perhaps the largest randomized controlled trial of an intensive lifestyle intervention among adults with type 2 diabetes to date. And boy is it intensive: weekly coaching sessions for the first six months and a diet which induced weight loss of 8.6% in the first year. This achieved biochemical remission in 11.5% of subjects at the end of the first year, falling to 7.3% at the end of year 4. As with all diabetes interventions, it’s the cardiovascular outcomes in the long term that will matter most, and we won’t know those for a while yet. In the meantime, here is yet another great data mine for diabetes researchers—if they can get access to it. O that all studies of this kind were part of a global collaboration: if that happened we might even find out how to treat type 2 diabetes (at last).

2497    Two weeks ago, I took the Lancet to task for using most of its research article space to publish two Amgen-funded phase 2 trials of AMG145, its new monoclonal antibody to plasma proprotein convertase subtilisin/kexin type 9 (PCSK9). And do you recollect what does this does to people? Yes, it reduces their low density lipoprotein cholesterol. And what else does it do?—the girl at the back. “We don’t know, sir!”—excellent, that is the correct answer. We really have no idea what this drug does to people in the long term. But here is yet another paper in a leading journal describing a 12 week phase 2 trial called GAUSS, showing that it reduces LDL-C in those who are intolerant of statins. This wouldn’t be a way of softening up the FDA to make it available before the long-term safety and effectiveness data are in, would it? I do hope not, because all good children know that would be naughty, and dangerous, and wrong.

NEJM  20 Dec 2012  Vol 367
2370   “Genomewide Association Studies and Common Disease—Realising Clinical Utility” Ah, a rare slip in the copy-editing of the NEJM—in the last word, they have missed off the initial F.

2375   A few years ago, somebody gave me a book of cardiovascular trial acronyms. It lay in our downstairs loo for a while, and I seem to remember that it had 700+ pages, one for each trial. Now, for your seasonal delectation, comes the FREEDOM trial which compares coronary artery bypass grafting with percutaneous coronary intervention in people with type 2 diabetes and multivessel coronary artery disease. And it comes out in broad agreement with BARI, CARDia and SYNTAX. Those of you who take your acronyms seriously (or spend a lot of time on the loo) will remember that each of these showed an overall advantage from CABG rather than PCI in these patients. So does FREEDOM, with an absolute reduction of 7.9% in major CV events. Since all the subjects in this trial were given maximal cardioprotective drug therapy as well, you could argue that in the light of COURAGE, there should also have been an arm in which there was no invasive treatment. Then we could really share decision making with this high-risk group of patients. But to be fair to the FREEDOM triallists, they recruited before COURAGE was competed, so they couldn’t have known. Time for another trial, perhaps: but please no more stupid acronyms; there just isn’t space in the loo.

2385   I don’t know if Dickens tells us why Tiny Tim had crutches, but maybe he had systemic juvenile idiopathic arthritis. I’m no expert on this condition, but I can’t believe that there aren’t half a dozen or more third-line drugs for JIA once non-steroidal and steroidal anti-inflammatories have failed. I really don’t understand why Hoffmann-LaRoche were allowed to run a placebo-controlled trial of the latest “biological,” an antibody to the interleukin-6 receptor called tocilizumab. It lasted 12 weeks, and after that all the kids could get put on to open-label tocalizumab, which provides a high rate of remission at the cost of a significant incidence of neutropenia and infection. So how does this help us decide on the place of tocilizumab for JIA among all the rest? This is a long-term condition, and the disease is nasty but the risks of treatment are high. Long-term head-to-head trials are the only ethical answer: the rest is just marketing.

2396   “And oh look Tim, here is another Christmas present for you—a selective, fully human, anti–interleukin-1β monoclonal antibody called canakinumab, brought to you by nice Mr Novartis. You will thank the gentleman, won’t you Tim?” “Well indeed thank you sir, but begging your kind pardon, I don’t see how you can tell if it will do me more harm than good until you’ve done them phase 3 trials and compared it with other agents targeted at the inflammatory pathways in systemic juvenile idiopathic  arthritis if you catch my meaning sir. So I bid you a Merry Christmas if I may make so bold, and good night and all sir.”

2407   Competitive blockade of the vasopressin V2-receptor is a great idea, but it’s taken ages for a real use to emerge for it. Tolvaptan is sometimes used to treat hyponatraemia, and it also has been shown in preclinical studies to reduce progression in autosomal dominant polycystic kidney disease. This trial takes that idea a lot further, though not to the point of the hardest outcomes such as death and need for dialysis. What it shows is that tolvaptan slows the increase in kidney volume by nearly a half over 3 years, and also slows the decline in kidney function: but there were enough adverse symptoms (thirst, polyuria) and biochemical changes (hyponatraemia, hypokalaemia, and liver enzyme elevations) to lead to a discontinuation in about a quarter of the subjects.

BMJ  22 Dec 2012  Vol 345
A worthy Christmas edition of the BMJ containing many delights. Two adorable dogs, including Cliff the poo-sniffer, who is the best known diagnostic agent for detecting C difficile; some wonderful historical items including an exhaustive account of the surgical thimble, and how Hitler introduced gastroscopic art to British medicine; suggestions for a medical uniform which deserves wide adoption; and a warning on the dangers of the tooth fairy (and GPs who can’t perform otoscopy).

I’m delighted to see that Anthony Harnden has taken up my suggestion and investigated the predictive characteristics of the speed bump test for suspected appendicitis with a team of Oxford colleagues. It’s actually amazingly good as a rule-out test, with a sensitivity of 97%. It has a poor specificity (30%) because any inflamed viscus will hurt when you go over a speed bump in a car. Now combine it with the Mars bar test, as I suggested (read Hershey bar or similar in the USA)—if your offer of a chocolate treat is firmly refused, and the speed bump test is positive, you need to send that child straight off to the surgeons.

Ann Intern Med  18 Dec 2012  Vol 157
878   How to keep Cliff away from your bed: take probiotics. The evidence isn’t first class, say these meta-analysts, but it all points one way: that if you eat lactobacilli while taking antibiotics, you are much less likely to get Clostridium difficile afterwards. Woof!

Plant of the Week: Viscum album

The practice of kissing beneath mistletoe apparently only grew popular in the sixteenth century, and nobody really knows why. Any excuse will do, I suppose. At junior school we were told that Yule was a pagan festival when Druids would cull mistletoe from trees with a golden knife because they believed it was a magical evergreen emanation of the tree-spirit. That cut little ice with me, because I had never seen either mistletoe or Druids. As a result, when we were asked to draw such scenes with our crayons, my efforts lacked conviction. I much preferred doing baby Jesus in his manger, though my cows were notable more for imagination than for perfect verisimilitude. I find that this is a trait that I share with many other artists including Fra Angelico.

Few gardens contain mistletoe as a deliberate ingredient, but we have a great opportunity. A person whose memory we adore once helped her granddaughter to plant an apple seed in a yoghurt pot many years ago, and now we have a fast-growing apple tree which from time to time bears small scabby apples reminiscent of Golden Delicious. On this sacred tree we have sent up Clematis armandii to flower in March and Clematis viticella to flower in July. As the tree continues to grow, I think I might make a cut in one of its boughs and dress the wound with bird dung and mistletoe seeds. Later perhaps, when the mistletoe has grown and I have acquired a flowing white beard, I shall celebrate Yule by putting on robes and flourishing a knife covered in gold leaf; then I shall mount a ladder to cut the mistletoe from the tree while intoning nonsense in a loud Welsh accent.