Richard Lehman’s journal review – 9 January 2012

Richard LehmanJAMA  4 Jan 2012  Vol 307
37    The gradual makeover of JAMA takes a further step with the introduction of a series of Viewpoints in the opening section. Quite nice, and very like the NEJM: it’s good to see some of America’s liveliest minds at work here thinking about medicine in general and their chaotic health economy in particular. On the latter topic, there are good contributions from Zeke Emanuel, Robert Brook and Jon Skinner, though nothing quite as trenchant as the article in last week’s New England Journal on The Savings Illusion which points out that most costs in the US system are fixed in people and institutions, and hence very difficult to shift by any means at the government’s disposal. I’m also keen to flag up the contribution here by John Ioannidis and colleagues on how slow doctors are to abandon discredited treatments: they cite PCS for stable angina, vertebroplasty for vertebral fractures, and bevacizumab for breast cancer. And they point out how crazy the current system for developing new ones is:

“Asking corporate sponsors to conduct pivotal trials on their own products is like asking a painter to judge his or her own painting so as to receive an award. If a manufacturer can be allowed to manipulate the system to create a blockbuster product from an ineffective drug, the temptation is hard to resist.”

47    It’s not just useless interventions that persist in medicine either: useless measurements seem to be even more intractable. Things like IQ or BMI that simply don’t measure anything you need to know. The body mass index takes quite a hammering in two editorials in this week’s JAMA, and this study of weight gain following overeating helps to show why. Subjects who were overfed with carbohydrate in this study could drop their BMI while gaining fat: this was corrected by adequate protein intake which promoted gains in lean mass. So people overfed with protein put on weight, but this is “good weight” in the form of muscle, associated with reduced insulin resistance and higher rates of metabolic activity. Just measuring BMI says nothing about all this: in fact here it sent entirely the wrong signal as to which form of over-eating is “healthier.” I do think the one thing we know about diet, amongst the huge accumulation of myth and rubbish, is that protein is a good form of energy intake. I’m quite in favour of fat too, but let’s not go there just now.

56    If you look at obesity measured by BMI, and then adjust for insulin resistance, hyperlipidaemia and high blood pressure, weight as such is actually a weak predictor of cardiovascular risk. Which may explain why the absolute benefit of bariatric surgery is not enormous, and is probably non-existent for obese individuals without other risk factors. This is a report from the Swedish Obese Subjects (SOS) study at a median of 14.7 years, and it shows a halving of cardiovascular events in the group treated with bariatric surgery. That sounds impressive, but absolute mortality difference actually amounts to 1.3% in favour of surgery. And the benefit was not related to baseline BMI in this study. Even longer term data are needed.

66     I have spent several afternoons over the last few months listening to presentations on readmission rates in US hospitals, wondering all the while how these compare with other health systems. For myocardial infarction, they are very high compared with other developed countries. “However,” say the authors, “this difference was greatly attenuated after adjustment for length of stay.” In other words, the facts that the US has the shortest stays for MI, and the highest readmission rates, are not unconnected.

NEJM  5 Jan 2012  Vol 366
1    The world’s foremost medical journal celebrates its 200th birthday with an account of its history which is well worth reading. It’s fascinating to note that the early decades of the twentieth century were marked by disquiet about how patients would get looked after amidst the growth of “specializm” as it was then called. “How much can the specialist know of home conditions, of family difficulties, and their relation to the case?” lamented a physician in 1923. We continue to lament. The problem of “inferior human stock” also exercised eugenically-minded physicians of those times: in 1934 an editorial declared: “Germany is perhaps the most progressive nation in restricting fecundity among its unfit. . . . In America it is probable that the sentiment of the people is not ready for the adoption of the German plan, and will be inclined to restrict compulsory sterilization to a small proportion of those who might properly be regarded as especially fit subjects of this treatment.” The subsequent history of the NEJM makes amends for this: it is even showing signs of becoming a proponent of social justice in the USA under its cloak of demure conservatism. Happy Birthday, New England Journal: do all the good you can.

9     A new era in secondary prevention after acute coronary syndrome is the headline of one birthday editorial, celebrating the success of low-dose rivaroxaban in reducing a composite end-point of death from cardiovascular causes, myocardial infarction, or stroke in 15,526 patients with recent ACS. And to be fair, this amounted to an actual (small) tally of lives saved: total mortality was reduced from 4.5% to 2.9%. It would take any competitor a comparable amount of effort to show benefit from its own fixed-dose anticoagulant, and two have already tried and failed (apixaban and dabigatran), perhaps because the dosing was wrong. Rivaroxaban also failed at a dose of 5mg b.d., but by halving this it was changed from an agent that caused too many bleeds to one which prevented more events than it caused. Good news for Johnson & Johnson and Bayer shareholders, then: but for post-ACS patients and health systems struggling to contain costs? For these, the “new era” is a mixed blessing: someone will need to interrogate the individual patient data from this vast 766-centre trial, and then we will have to wait at least a decade for evidence from trials of equal size to find out the optimal anticoagulant/antiplatelet cocktail for post-ACS patients.

20    But at least we can cross vorapaxar off the list right away: this new oral protease-activated–receptor 1 (PAR-1) antagonist that inhibits thrombin-induced platelet activation simply caused more bleeds in post-ACS patients, without any countervailing benefit. That’s a pity: I rather like the name. (I took a shine to vorapaxar: What a shame we had to axe her).

34    Another “failed” trial: a herpes simplex vaccine which was intended to protect against HSV-1 and HSV-2 genital infection in seronegative women, but which just gave partial protection against HSV-1 at the cost of a lot of local reactions. But maybe a stepping-stone to something better.

Lancet  7 Jan 2012  Vol 379
31    Deep vein thrombosis of the leg is followed by post-thrombotic syndrome (PTS)  in about half of patients with above-knee DVT, and small previous studies have suggested that some form of local thrombolysis might help to reduce this. A team from south Norway set out to put this to the test. They used a catheter with lots of side-holes in it, rather like the nose-piece worn by the Dong with a Luminous Nose: this was inserted into the popliteal vein when possible and used to feed alteplase into the immediate environment of the clot. The reduction in PTS only just reached statistical significance and there were 8 major or medically significant  bleeding complications in the 101 patients treated with catheter-delivered thrombolysis. The investigators conclude that CDT should be considered in patients with a high proximal DVT and low risk of bleeding.

39   “This large randomised study shows that targeting of the bone microenvironment can delay bone metastasis in men with prostate cancer.” Yes, by a paltry four months and at a huge cost which is not just financial: one in twenty of the subjects suffered osteonecrosis of the jaw for a treatment which makes no difference to survival. Denosumab is a fully human monoclonal antibody that specifically binds and inactivates RANKL, the agent by which osteoclasts nibble away at bone. The obvious comparator in this trial would therefore have been a bisphosphonate, but instead it was placebo. I expect that debate about the place of denosumab in the treatment of prostate cancer will continue to rankle.

47    I fell in love with Iran thirty-eight years ago, when I went there as a medical student: thereafter I always yearned to return as a rural primary care doctor, serving some of the most civilized and hospitable people on earth. It never happened; but my longing to return was rekindled recently by working alongside a delightful young Iranian doctor at Yale. Iranian medicine is still ridiculously patchy and top-down on the American model, and few good doctors want to practice in the vast mountainous rural areas of this great and beautiful country. Since the mid-1980s, much of the primary care in such areas has been done through locally recruited assistants, like the Russian feldschers, who command one-sixth of the income of doctors. This system is known as behvarz, meaning good skill. Their skills do seem to be reasonably good in looking after diabetes and hypertension, though one is slightly alarmed to note that “Respondents were classified as having diabetes if their FPG concentration was 7 mmol/L or lower.”

BMJ  7 Jan 2012  Vol 344
If you want to know my views on the papers in this themed issue on missing trial data, all you have to do is read the editorial. For a really trenchant commentary, here is Harlan Krumholz (as posted on CardioExchange):

Missing Data: The Elephant That’s Not in the Room
Harlan M. Krumholz, MD, SM

There is a problem so grave that it threatens the very validity of what we learn from the medical literature. Bad data? Not exactly. Actually, it’s missing data — information, relevant to the risks and benefits of treatments, that is simply not published. In some cases, these data would make a critical difference in the inferences that readers draw from the literature. The absence of the data renders meta-analyses, systematic reviews, and book chapters suspect. Conclusions are made on the basis of incomplete science. In short, publication bias and selective publication are impugning the validity of what we can learn from a PubMed search or even the most careful review of published studies.

This matter demands our immediate attention and speaks to the need to rethink the configuration of clinical medical science. It may be time to adopt strategies to ensure that all relevant studies, results, and supporting documentation are made publicly available. “Out of sight, out of mind” is a dangerous reality in science and medicine. It’s time for a change — and it starts with the recognition that we have a problem.

I urge you to read BMJ this week to explore the evidence of this problem. In full disclosure, the studies include one by me (with others, led by Joe Ross) showing that more than half of trials sponsored by the NIH go unpublished even 30 months after completion. The other articles reveal troubling information, including about how missing data can affect the results of meta-analyses — and how many investigators are ignoring the requirements for mandatory reporting of trial results, raising the question of what “mandatory” actually means.

It is time to pay attention to this issue — and to begin working together to solve it. Let’s advocate for open science and get all the information out in a timely way for everyone to inspect. There are many facets to this problem, and we should not look to assign blame, but we do need to change our research and publication culture. Our entire clinical research community, including those who use the information and those who contribute to its dissemination, must collectively determine how best to get beyond this period when we are working with an incomplete view of medical evidence. Let’s put everyone on notice: The era of missing data must end.

After you review the studies in BMJ, please share your thoughts here with fellow members on CardioExchange. Links to several of the articles are provided below, with key quotations from each one.

Ross et al: Despite recent improvement in timely publication, fewer than half of trials funded by NIH are published in a peer reviewed biomedical journal indexed by Medline within 30 months of trial completion. Moreover, after a median of 51 months after trial completion, a third of trials remained unpublished.

Hart et al: The effect of including unpublished FDA trial outcome data varies by drug and outcome.

Ahmed et al: Publication, availability, and selection biases are a potential concern for meta-analyses of individual participant data, but many reviewers neglect to examine or discuss them. These issues warn against uncritically viewing any meta-analysis that uses individual participant data as the most reliable.

Prayle et al: Most trials subject to mandatory reporting did not report results within a year of completion.

Wieland et al: Based on the results for 2005, at least 3000 records describing randomised controlled trials but not indexed using RCT may have been entered into Medline between 2006 and 2011.

To leave or read comments, you’ll have to visit the version that’s only open for CardioExchange members.

Ann Intern Med  3 Jan 2012  Vol 156
1     Rituals involving touch are powerful forms of social grooming, an underexplored element of the medical encounter which Michael Power drew my attention to some time ago. In any comparison of treatment for neck pain, I would have presumed that spinal manipulation would be bound to win over home exercises or medication for pain control, but in fact it proves no better than home exercise at any point in this year-long trial. Hands-on seems no better than hands-off, even as a placebo.

19    The Yale Center for Outcomes Research and Evaluation, to which I shall be returning next week, keeps an eagle eye on American hospital outcomes. This is no easy task, but Elizabeth Drye and the rest of her team have done a typically thorough and exemplary job of comparing hospital mortality figures for three conditions – acute MI, heart failure, and pneumonia – according to whether deaths occurred in hospital or within 30 days. Hospitals which discharged patients quickly tend to have an advantage in mortality statistics which disappears when you look at the 30-day mortality.

27    When I became diabetes lead for my practice a few years ago, I looked in vain for a diabetes guideline I believed in. Then came the ADVANCE and ACCORD studies, and I became increasingly puzzled and angry: how come the evidence base for treatment in type 2 diabetes was so poor, and that when large well-conducted trials were published, nobody took any notice? Why did practically every “expert” behave as if nothing had happened? Why were we adopting whole new classes of agent without any evidence about their long-term harms and benefits? I have sounded off a great deal about this since, but I am still a bit puzzled and a bit angry. So, I think, were the authors of this systematic review of  guidelines for oral treatment in type 2 diabetes. But they hide this quite well, and simply call for better standards. Yes indeed.

Ten Commandments for testing – by Michael Power
John Yudkin’s wonderful Ten Commandments for prescribing proved very popular – and they have inspired this excellent further Decalogue from Michael Power, which I send you with his permission. Please do send your feedback to, cc.

Thou shalt obey the following ten Commandments for testing, whether it be for ruling in a diagnosis, ruling out a diagnosis, assessing risk or prognosis or response to treatment, or for monitoring for adverse effects and deteriorating status.

Thou shalt understand testing in its broadest sense; it includes history, examination, laboratory tests, imaging investigations, diagnostic procedures, and therapeutic trials.
When a commandment is impractical or impossible, thou shalt treat it as an aspiration and do thy best.

For I am thy patient and client, whose interest thou shalt serve, and no other.

  1. Evidence. Thou shalt not take the evidence in vain, but test according to the best estimates of prevalence, positive predictive value, and negative predictive value. If the predictive value of a test is less than about 50%, toss a coin — it will be cheaper and as useful.
  2. Application of evidence. Thou shalt not overly rely on test results, but shalt apply your clinical judgement after clinically assessing your patient and critically appraising the evidence, taking into account its precision, risk of bias, and directness of applicability.
  3. Cost-effectiveness. Thou shalt not covet thy neighbour’s graven image technology (PET scanner, fMRI scanner, high resolution ultrasound scanner), nor his micro-array genetic tests, nor his direct to consumer testing business, nor his yacht, nor any thing that is thy neighbour’s, but thou shalt practice cost-effective testing. If a cheaper test will be as useful, use it.
  4. Patient-education. Thou shalt help thy patient understand that many diseases are gradual and progressive, analogue processes not digital events. Diagnostic thresholds and limits are chosen for convenience, but create artificial categories that may be misleading if they are misunderstood as boundaries between having and not having a disease, or having and not having a risk.
  5. Joint decision-making. Thou shalt help thy patient understand the limitations of tests. Many conditions cannot be diagnosed or excluded by tests (for example dementia, wellness). Tests can be falsely positive or falsely negative or inconclusive. No test can give a precise prognosis for survival or other probability, and interpretation of prognostic tests should consider both the average (median or mean) and the distribution in the comparator population. Thou shalt remember that test results can in themselves be distressing or harmful. For these reasons, decisions about testing are best made jointly with thy patient.
  6. Patient-centred care. Thou shalt not take thy patient’s needs in vain, but before testing help them understand what the management options are for a positive, inconclusive, or negative result, and what support is available should the result be distressing. Honour the elderly patient, for although this is where the greatest levels of risk and temptation to test reside, so do the greatest needs for avoidance of useless and harmful testing.
  7. Efficiency. Thou shalt not repeat a test when the result is already available or the result will not change (as with genetic tests or when the clinical indications have not changed). Thou shalt ensure that the results of tests you have ordered or performed are clearly recorded and available or communicated to any other physician caring for thy patient.
  8. Ethics. Thou shalt not use testing as a defence against legal action, or as a placebo, or as a delaying tactic while nature takes its course, or to avoid confronting the limitations of curative medicine when care, support, or palliation is appropriate.
  9. Education and engagement. Thou shalt help thy trainees and junior colleagues understand that they should investigate having considered the needs of their patients and the performance of the tests. The reason for testing should not be that it is routine, or policy, or what they imagine their consultant/attending expects.
  10. Gnothi seauton. Thou shalt know thy cognitive limitations and biases. Thou shalt try to avoid the fallacies of assuming that all abnormal results are important or that an abnormal result is sufficient to explain symptoms. Thou shalt consider the whole picture, and the differential diagnosis, and the possibility that tests bear false witness against thy patient.