Richard Lehman’s journal review—30 September 2013

Richard LehmanNEJM  26 Sep 2013  Vol 369
1195   Gosh, I feel vulnerable. The opening paper in this week’s New England Journal is about sources of C difficile infection in “healthcare settings or in the community in Oxfordshire, United Kingdom.” Will I be named and shamed? A few weeks ago I received a message of gentle reproof from the director of the out-of-hours service I work for, saying that I was prescribing too much cefalexin, which could encourage C diff infection amongst the residents of Oxfordshire. This paper shows that in fact local rates of C diff fell during between 2008 and 2011, despite my worst efforts; but that doesn’t let me off the hook. During this time, Oxfordshire was busy spawning new types of C diff, and by meticulous genomic analysis this study shows that 45% of C. difficile cases in the county were genetically distinct from all previous cases. I still live in fear that some new John Snow may now map them and trace them all back to my lethal prescription pad. In fact I still find myself using cefalexin occasionally as a third line antibiotic for otitis media and urinary tract infections in the young, who almost never get symptomatic C diff infection. But this study suggests for the first time that there may be C diff spores wafting all over the place from asymptomatic carriers. While ever there is that Lehman about in Oxfordshire, wear a mask.

1206   One of the few genuinely promising advances I’ve been able to track in the journals over fifteen years has been the arrival of fixed dose oral anticoagulants. My, how the drug companies have scrambled to conquer this immense market! Boehringer-Ingelheim managed to score with dabigatran for atrial fibrillation in the RE-LY trial, though doubts remain, and the cost is prohibitive for routine use. The authors of this paper describe how this success then spurred them to set up the phase 2 trial RE-ALIGN trial in 39 centres across 10 countries to see if they could find a new market in people with mechanical heart valves. In the chosen settings, patients in the comparator group receiving warfarin were in the target INR range just 50% of the time. But despite this obvious source of gross bias, dabigatran failed to show any added benefit, and showed an excess of bleeding. Goodbye dabigatran for mechanical valve patients.

1268   I’m 63 and I’ve been doing medicine for 38 years. Do I still have performance anxiety? You betcha. If I didn’t, I’d know I was unsafe. And do I resent a lot of the performance measures by which I’m judged? You betcha. If I didn’t, I’d know I was a moron. Here is the opening of a terrific essay by Debra Malina:

“When you entered this field, you considered it a calling. You had to master both an art and a science—you aimed to effect critical changes in other people, who were infinitely variable and over whom you had limited influence, but if you established relationships with them, you and they often triumphed together. Nowadays, you’re increasingly assessed on the basis of how well those people, now considered your customers, do on a few narrow tests. Although you see this as inapt quantification that breeds constricting standardization, society demands services of more consistent quality. So policymakers, applying business principles to your field, insist on measuring your performance in whatever ways it is easily measurable and then rewarding or punishing you accordingly.”

She then tells you that she is referring to US schoolteachers. It’s a gem of a piece and the full text is free for you to enjoy.

JAMA Intern Med  23 Sep 2013  Vol 173
1573    “Worsening trends in the management and treatment of back pain” is the title of a paper which maps the changes in US primary care over an 11 year period beginning in January 1999. The things the Harvard authors looked for were the use of imaging, referrals to secondary care, use of “narcotics” (they mean opioids) as markers of badness, and the use of NSAIDs, paracetamol, and physiotherapy as markers of goodness (as recommended by national guidelines). Although the trends were all in the wrong direction, I was still surprised at how little American family doctors refer for imaging—17% of low back pain patients get lumbar spine X-rays (no change over time), while MRI use climbed from a modest 7.2% to a still modest 11.3%. For all I know, this may be inappropriate (I don’t think you can tell without analysing each of the 23 918 cases), but it certainly doesn’t amount to catastrophic market-driven overuse.

1592   I was once interested in population echographic screening for systolic dysfunction and valvular disease (I was a naïve 45-year old back then), but I gradually realised the error of my ways. My interest moved to BNP screening in high risk groups—but that is another story, and another error. Here are the long-term results from 6861 middle-aged participants from the population-based Tromsø Study in Norway who were randomised 15 years ago to have echocardiography or not. “During 15 follow-up years, 880 persons (26.9%) in the screening group and 989 persons (27.6%) in the control group died (hazard ratio, 0.97; 95% CI, 0.89-1.06). No significant differences between the groups were observed in the secondary outcome measures (sudden death, mortality from any heart disease, or incidence of fatal and nonfatal myocardial infarction and stroke).”

1629   Drugs to lower blood pressure in middle-aged people typically have numbers-needed-to-treat of several hundred to prevent an adverse cardiovascular event within 10 years. So any drug you use to lower blood pressure must (a) be proved to reduce CV events and (b) be proved to be free of long-term harms. Millions of women across the world use calcium channel blockers to reduce blood pressure, and this case-control study from the Seattle–Puget Sound metropolitan area raises the unwelcome possibility that this doubles their chance of getting breast cancer over a 10 year period, regardless of the odipine, adipine or edipine used. Someone urgently needs to do a massive data trawl of the UK CPRD, to settle the matter within weeks.

JAMA  25 Sep 2013 Vol 310
1240  OK, I’m a Luddite and a sceptic, and when I see a title like “Effect of Sensor-Augmented Insulin Pump Therapy and Automated Insulin Suspension vs Standard Insulin Pump Therapy on Hypoglycemia in Patients With Type 1 Diabetes: A Randomized Clinical Trial.” I automatically think: more high tech salesmanship. But let’s try and think about it from the patient’s point of view: given that I have type 1 diabetes, is this a real move forward in my treatment? “The primary objective of this study was to determine the incidence of severe and moderate hypoglycemia with sensor-augmented pump therapy with a low-glucose suspension function compared with standard insulin pump therapy.” Does that matter to me? I think it probably does: a lifetime of hypos is not going to do my brain any good at all. And does the thing work? Yes: the adjusted incidence rate of hypos per 100 patient-months was 34.2 for the pump-only group vs 9.5 for the group with pumps that stopped the insulin for two hours once a certain low glucose threshold was reached. So this is a step forward. The trial was conducted on patients with impaired hypoglycaemic awareness in Australia: Medtronic supplied the pumps free, but was not otherwise involved.

1248   I first got the idea of the “Easily Missed” series in the BMJ after I read that British GPs missed subarachnoid haemorrhage at first presentation about 50% of the time. I have done it myself. But this really is a “never miss this” condition—one of a small handful that doctors need to pick up every time, first time. Here is a new rule from Ottawa that will help hugely if doctors actually use it. It has been validated in Canadian emergency departments, where it has a sensitivity of 100% and a specificity of 15.3%. These predictive characteristics may be slightly different in UK primary care, but basically it means that for every patient you send to hospital with real SAH you will send 5 or 6 with benign headache. That’s fine: it’s the price of safe practice. Just send up anyone over 15 years old with new severe nontraumatic headache reaching maximum intensity within 1 hour and one or more of the following features:
1.    Age ≥40 y
2.    Neck pain or stiffness
3.    Witnessed loss of consciousness
4.    Onset during exertion
5.    Thunderclap headache (instantly peaking pain)
6.    Limited neck flexion on examination

Lancet  28 Sep 2013  Vol 382
1073   Have I died and gone to heaven? Here is a Lancet full of the most wonderful things. “The wolf also shall dwell with the lamb, and the leopard shall lie down with the kid; and the calf and the young lion and the fatling together; and a little child shall lead them…They shall not hurt nor destroy in all my holy mountain: for the earth shall be full of the knowledge of the LORD, as the waters cover the sea.” If you count knowledge that helps to heal people as belonging to the Lord, then this Lancet contains a fair bit. But for me the most remarkable piece is this opening editorial in which Iain Chalmers, who for over two decades ago has condemned the selective publication of trials as a form of misconduct, lies down with Patrick Vallance, President of GlaxoSmithKline, a company with one of the worst records of selective publication. They put all that behind them and discuss how to overcome the barriers to making individual patient data available to researchers. Just two years ago this would have been unimaginable. Now real research heaven is within sight, and who knows: clean knowledge may soon begin to cover medicine, as the waters cover the sea. Iain has done more than anyone to make this possible.

1097   Orthopaedic surgeons are often thought of as wolves by the lambs of medicine, though by and large they do more good than almost any other specialty. It’s just that they drive such wolvish cars. Traditionally, they have not been notable for looking at their results or learning from their mistakes. Now, in the beckoning heaven of big data and proper outcomes research, this is changing. It’s often good to start with mortality as the outcome, and with a very common procedure, such as hip replacement. Mortality following hip replacement in England and Wales has fallen sharply from 0.56% in 2003 to 0.29% in 2011: the identifiable factors include a posterior surgical approach, spinal anaesthesia, mechanical DVT prevention or DVT prevention using anticoagulants. Being overweight is also associated with better survival.

1105   If you undergo abdominal surgery, how do you want your wound closed—with staples or with subcuticular suturing? Not that anyone is likely to ask you. In a Japanese randomised but “open-label” trial, there was no overall difference between the two in wound complications. But a good editorial criticises the use of this composite end-point: there was in fact a difference in favour of subcuticular sutures for lower abdominal surgery, where wound infection is more likely. And scar hypertrophy was less likely with the sutures. But above all, “patient satisfaction was not assessed in this trial. Future trials within this context must include the patient’s perspective, since raw surgical feasibility and efficacy is only one aspect.” Wow, is the message getting through at last? To surgeons…?

1121   We will know when the true heaven of medicine has arrived when everybody around the world has the same access to care, and the same interventions are available to everybody, using the same diagnostic criteria, and achieving the same results. But when Jack Wennberg started looking for this in New England over 40 years ago, he found the exact opposite. No medical journal wanted to publish his results, but when they finally appeared in Science and later in Scientific American, they caused a real stir, and set outcomes research on a course that it continues to follow to this day. A lot of surgical activity is supplier led—i.e. if you have a lot of surgeons, a lot of surgery gets performed. The more doctors people see, the more diagnoses they acquire. People who move from a place (e.g. a state of America) with fewer doctors to a place where there are more doctors get more tests and procedures and become sicker on paper, though their real health is unchanged. And so on. Here Jack and others take another look at the present situation of variation in surgery. Heaven is still far away: “the relative degree of variation in population-based rates of ten common surgical procedures has been remarkably stable over the past 20 years. In 2008—10, rates of hip replacement, coronary bypass surgery, prostatectomy, and many other major procedures continued to vary at least four-fold to five-fold across hospital referral regions [in the USA]. Data from the UK showed similar degrees of variation in the use of surgery in 152 primary care trusts in 2009—10.”

1130  So what is the answer? If anyone is determined to find it, it is the lead author of the next paper, Peter McCulloch, an Oxford surgeon who founded the IDEAL collaborative to improve the quality of surgical research. Jack Wennberg’s solution has always been to inform patients better and to make their informed decisions definitive (by law). We are still a long way from that; but I look forward to working on it with Peter, and to joining in celebrating Jack’s lifetime achievement at Dartmouth College in two weeks’ time.

BMJ  28 Sep 2013  Vol 347
The BMJ this week provides more evidence that complex interventions, devised top-down and rolled out with some success in research settings, usually bomb when tried out in real life. Telephone “coaching” by practice nurses in Australia failed to affect glycaemic control in people with type 2 diabetes compared with usual care. And a cluster-randomised trial in the Netherlands investigated whether an interdisciplinary primary care approach for community dwelling frail older people is more effective than usual care in reducing disability and preventing (further) functional decline. It wasn’t. Perhaps “usual care” is about all that it’s possible to provide for some aspects of the human condition without a radical reorganisation of society.

I am often quite harsh about the six journals I review regularly here, but believe me dear reader, most of the rest are much, much worse. In many articles and blogs, most recently Richard Smith has bewailed the fact that most medical journal editors have no basic training in editing and impose only the loosest quality control on what they publish. This is true: and one of the most basic things an editor should insist on before a trial is published is that it has been registered. Here Liz Wager and colleagues look at 200 journals and find that 142 (71%) did not require registration (or at least did not mention this on their website), 55 (28%) required registration, and 3 (2%) encouraged registration but did not make it a requirement for publication. The qualitative part of the study produces a mixed bag of responses from editors. What none of them say is, “Look, I took this job because it would look good on my CV. I do it in my evenings and weekends and nobody ever told me how to do it properly. I have one assistant and she is on maternity leave. We get offered a load of crap and I have to publish some of it every month.” Though in fact some of the responses come quite close to this. Boy, do paper journals need a sort out.

Plant of the Week: Rosa “Grüss an Aachen”

I’ve praised this rose several times before. It is the best. Its scent is heavenly, and its colours change form palest cream through soft pink and apricot. It starts flowering in late May and continues to November, saving its greatest beauty and abundance for October. From the late nineteenth century until recently there used to be a climbing form. If only I could find it.