Richard Lehman’s journal review – 7 May 2012

Richard LehmanJAMA 25 Apr 2012 Vol 307
1809    Among the many virtues of JAMA, one cannot number a strong sense of the ridiculous. The poetry and medicine section is the world’s most reliable source of po-faced bad verse, this week’s example being an invective against Decadron; and the first research paper this week is a study called FISH.

This is short for Fish Oil Inhibition of Stenosis in Hemodialysis Grafts; and believe it or not, this is a study which aims at finding out whether a fish oil capsule produced by Ocean Nutrition Canada Ltd prevents stenosis in newly formed arteriovenous shunts in haemodialysis patients over a 12 month period. Or, as a typical JAMA poet would put it:

Dragged in shoals from the Canadian oceans, unwillingly their flapping silver bodies yield
Oils rich in omega-3 fatty acids with antiproliferative, antioxidant, and vasodilatory effects
Reeking with promise to keep open life-giving shunts in the renal unit.
Only they don’t.

1817   For mere bystanders, the world of interventional cardiology never ceases to amaze with its profusion of trials seeking to compare one odd sounding intervention with another. The bottom line of this paper about the INFUSE-AMI is that intracoronary abciximab is superior to manual aspiration thrombectomy in patients with large anterior myocardial infarction.

I guess you have to be either an interventional cardiologist or a trial methodology dweeb to find this very interesting. Bear with me: I am becoming somewhat of the latter, and to me this looks like a complex marketing trial, and intrigues me like a rare fungus (you can skip to the next section at this point).

“Atrium Medical was involved in the design and conduct of the study and site selection and had the right to a nonbinding review of the manuscript.”

Atrium Medical makes coronary artery catheters for drug delivery, so they have a natural interest in promoting the use of intracoronary drug delivery. And one way to do that is to compare it with a technique that has already been shown to fail in the majority of trials – coronary clot aspiration. The thing not to do is compare it with intravenous abciximab delivery, because that might prove just as good.

It pains me to list all the other potential defects of this trial, but basically it had a 2×2 factorial design, was single blinded, enrolled only 7.2% of MI patients presenting at 37 sites in 6 countries, used an end-point (infarct size at 30 days) which was not accurately determined in a significant proportion of patients, with a difference in pooled groups of small statistical significance, plus four other disadvantages you can read about from the authors themselves in the comment section.

And by the standards of the trials which regularly appear in the top journals, it is not even particularly bad.

1838    When I was a GP in the prime of life, Muir Gray bade me read a book called Clinical Epidemiology written by a number of Canadians. Reading Sackett et al by day and by night, I suddenly became aware of the wonders of randomised controlled trials, the potential of the internet, the intellectual challenge of using diagnostic tests according to Bayesian principles, the need to seek evidence, evidence, evidence.

My heart still aches for the simple optimism of that vision. Instead, most of what we deal with in medicine is complexity made worse by inadequate knowledge. Surely the way forward must lie with further clinical trials.

These days, if you want the results of your trial to appear in a peer-reviewed journal, it has to be registered; and the US registry is known usefully as Here the past and present custodians of this site look at the quality of the trials registered between 2007 and 2010. They “are dominated by small trials and contain significant heterogeneity in methodological approaches, including reported use of randomisation, blinding, and data monitoring committees.”

In other words, these trials are never going to yield clinically dependable data; most of them are futile, and therefore by definition unethical. Something is terribly wrong with the system which governs clinical trials: it is failing to protect patients and failing to generate useful knowledge. Most of what it produces is not evidence, but rubbish. And with no system in place to compel full disclosure of the data, it is often impossible to tell one from the other.

NEJM  3 May 2012  Vol 366

1663   Now when we doctors seek certainty, the person we ask is the histopathologist. Thyroid cancer, however, is notoriously difficult to define, even under the microscope. Since a lot of it is actually benign, it is also difficult to know how to treat it. In this French study, they went for total thyroidectomy followed by the use of recombinant human TSH (thyrotropin) plus one of two doses of radioiodine (131I). The outcome measure was total thyroid ablation as assessed by neck ultrasound and TSH-stimulated thyroglobulin production. The lower dose of radioiodine (1.1GBq) proved as effective as the higher (3.7 GBq). By this short term measure, 95% of patients in the trial seem to have been cured.

1674   A very similar British trial using the same method of thyroid stimulation and the same doses of radioiodine produced cure rates which were slightly lower – 85-89%.

1686   The Placement of Aortic Transcatheter Valves (PARTNER) trial did what it says on the can: placed a lot of new aortic valves via a catheter (TAVR), so sparing patients the ordeal of open surgery. One part of the trial randomised patients at high risk of decompensation to one or other procedure, and at one year the two procedures yielded very similar results for mortality, symptoms and haemodynamic measures. This report gives the two-year results: there is now a divergence towards paravalvular regurgitation in the TAVR group which is associated with higher mortality.

1696   Another part of the PARTNER investigation randomised patients with inoperable decompensating aortic stenosis to either TAVR or standard management (which often included balloon valvuloplasty). In these unfit elderly patients (mean age 83) with severe heart failure (mostly NYHA III-IV) survival for 2 years was 56.7% in the TAVR group and 32% in the others. So worth a go in “appropriately selected patients”, as the usual phrase has it. The fittest survive the best, even at this level.

1705   And if you are a breathless 83-year-old with aortic stenosis, it may be best to go and have your TAVR procedure done in France. Here are some registry data from that country, fresh as a newly baked baguette -3195 patients enrolled between January 2010 and October 2011 at 34 centres, mean age 82.7. Rates of death at 30 days and 1 year were 9.7% and 24.0%, respectively. Formidable!

1716   The article that this issue of the NEJM will be best remembered for, however, is Atul Gawande’s wonderful Two Hundred Years of Surgery. The whole piece is open access so there is no need for me to quote extensively. My favourite bit is in parentheses half way through:

“Liston operated so fast that he once accidentally amputated an assistant’s fingers along with a patient’s leg, according to Hollingham. The patient and the assistant both died of sepsis, and a spectator reportedly died of shock, resulting in the only known procedure with a 300% mortality.”

Lancet  5 May 2012  Vol 379

1705   “Point-of-care genetic testing for personalisation of antiplatelet treatment (RAPID GENE): a prospective, randomised, proof-of-concept trial.”

Spartan Biosciences has developed a bedside test which allows rapid genotyping for the CYP2C19*2 allele. The simple story is that patients with this allele don’t respond to clopidogrel but do better with prasugrel, and that you can test this by measuring platelet reactivity, which was the outcome measure for this trial.

But the evidence from randomised trials of clopidogrel v prasugrel is a good deal less clear. Moreover, I can’t see why a bedside test is needed for patients who have just had percutaneous coronary intervention and are lying in a hospital bed close to a proper pathology laboratory. So the concept that this trial proves is that you can get a Lancet paper out of a genetic test that predicts a surrogate marker, immediately and expensively.

What it does not prove, thankfully, is the concept that the future of medicine lies in double rip-off – ripping off for the genetic test and then ripping off for the expensive drug.

1721   Although obstetrics led the rest of medicine in auditing its outcomes (and thanks to Iain Chalmers, in systematically reviewing its evidence base), a lot of obstetric practice remains based on tradition rather than evidence. Clamping of the cord is one example: controlled cord traction in the third stage is another.

This huge trial was designed to put the question to rest for ever, with nearly 12,000 women in each arm, randomized to active or expectant management. In the end, the trial fell victim to its own power: the results almost prove noninferiority for expectant management, but with so many subjects in the final count, a slight advantage for cord traction reaches statistical significance. The main message is to give oxytocin promptly if you want to avoid haemorrhage.

1739   In my youth it was an article of faith in the north of England that wearing glasses meant you were a swotty weakling, and I suffered deep chagrin when I acquired the nickname “Goggles” at the age of 9.

I know I didn’t acquire myopia from too much reading because I first got glasses at the age of 6: destiny had clearly determined at an early stage that I would be a swotty weakling with glasses and so I have proudly remained to this day.

The very idea that you could become myopic through too much close eye-work always struck me as fanciful, but this Lancet review points to a huge increase in myopia among school-leavers, especially in east Asia (80-90% prevalence). To dismiss the possibility that this is caused by reading, iPads, computers, and television, and lack of outdoor activity would definitely be short-sighted.

BMJ  5 May 2012  Vol 344

At the end of his essay on 200 years of surgery, Atul Gawande endearingly admits that he has no idea what “nanotechnology” means but predicts that it will have replaced surgery within 100 more years. Operations like appendicectomy will have vanished as minute robots or something do unknown things in ways we cannot yet predict.

Or do we already have enough evidence to relegate surgery for appendicitis to a second-line place in management? This systematic review of 4 studies makes the case for using antibiotics first: that way you may be able to avoid surgery in two-thirds of cases. I am not wholly convinced, but this is certainly a debate to watch.

Another surgical debate to watch will be the aftermath of the metal-on-metal hip debacle. This study comparing hip resurfacing with total arthroplasty is already obsolete: it was a randomized trial, the first of its kind, but the resurfacing was all metal-on-metal, and so now banned due to long-term harms. For the record, the two procedures had identical outcomes at one year.

Ann Intern Med  1 May 2012  Vol 156

618    Here’s a study of hospital atrategies for reducing risk-standardised mortality rates in acute myocardial infarction: an impeccable paper about an important topic, and moreover one which marries the insights of qualitative research with rigorous quantitative analysis of observational data.

Am I praising this too much? I don’t think so, though I have to confess that I’ve been working for the best part of a year among the team that produced it. I thought they were incomparable before, and now I know it for a certainty.

So call me biased, but do read this paper: it shows that virtuous care is rewarded, but sadly not by very much. Standardised MI mortality in the 537 US hospitals examined varies by about 10% in a near-normal distribution, but only just over 1% can be accounted for by the factors which emerged from qualitative interviewing of hospital staff: a culture that encouraged physicians to solve problems creatively, physicians and nurses, acting as quality-of-care champions, hospital and emergency department clinicians meeting at least monthly to review care, cardiologists always being present in the hospital, and not cross-training nurses to work in both intensive care and cardiac catheterization settings.

Fungus of the Week: Pleurotus eryngii

I had hoped that spring in New Haven would bring large crops of morels on the flowerbeds of Yale, which are all heavily mulched with shredded bark, a favourite substrate for these delicate fungi. But the only place I ever saw any was in a delicatessen, at $69.99 per pound. The same deli also offers Pleurotus eryngii at $19.99 per pound, while another Italian deli nearby charges $9.99 and a local Chinese supermarket has huge ones at $3.99.

These “king oyster mushrooms,” or whatever you want to call them, have a flavour highly reminiscent of the much-prized Boletus edulis. You are most unlikely ever to find them in the wild, even in their Mediterranean habitats. As organs of reproduction, they are singularly ill-adapted, having but a few ineffective gills high on an enormous fruiting body. But by the same token, they provide a lot of meat for those who eat them. They are very easy to cultivate, though I have seen them but rarely in British shops. In the peak of condition they are probably the finest of the cultivated mushrooms.

I think the best stage to buy them at is about 12cm long, still with a distinct though small brown cap. Larger ones can be a bit rubbery. You can eat them raw, thinly sliced with olive oil and shavings of parmesan. You can slice them more thickly and fry in butter with shallot, and decorate with parsley. Or in olive oil with garlic, à la bordelaise. You can put them in a pan with rashers of bacon and they will cook nicely together.

You can no doubt use them in most cep recipes, such as truite aux ceps, though I have not yet tried this. Worth seeking out.