Richard Lehman’s journal review—1 July 2013

Richard LehmanJAMA  26 June 2013  Vol 309
2557   Now that I’ve conceded column space to killer flu H7N9, let me put your minds at rest. Lots of people are working hard on an effective vaccine that will stop it killing people; and so far the mean age of victims is 60. But will the vaccine arrive in time for a pandemic? No chance, according to this article. Perhaps this will be the massive cull of pampered baby-boomers that so many hope for. Agh! I turn 63 this week. I must pamper myself harder while I can.

2559   Another Viewpoint piece has the tempting title “Personalised Medicine vs Guideline-Based Medicine” and discusses the uses and abuses of subgroup analysis. I think we may need more of this, once statistical science has developed sufficiently to meet the challenge of massive individual patient data sharing. This piece dwells on the counter-intuitive findings of trials of implantable cardioverter-defibrillators: they have no mortality benefit in the months of highest risk after myocardial infarction, but only show benefit much later on—perhaps as much as two years later.

2563   I wish I could point you to some research of generalisable clinical importance in this week’s JAMA, but this journal has become like an English summer: it could be so bright and lovely, but instead one dull day follows another. The burden of this study is that “Regular aspirin use was associated with lower risk of BRAF–wild-type colorectal cancer, but not with BRAF-mutated cancer risk. These findings suggest that BRAF-mutant colon tumor cells may be less sensitive to the effect of aspirin. Given the modest absolute risk difference, further investigations are necessary to determine clinical implications of our findings.” OK, we look forward to that.

2572   This one’s a bit better, and has Atul Gawande in the author list. It applies to the USA, but may be true elsewhere. The study questions the notion that most hospital admissions in high-cost patients are avoidable. In fact very few of them are: “Among a sample of patients in the top decile of Medicare spending in 2010, only a small percentage of costs appeared to be related to preventable emergency department visits and hospitalisations. The ability to lower costs for these patients through better outpatient care may be limited.” Combine this with last week’s BMJ paper showing that self-management programmes for high risk patients don’t work, and you can demolish most of the political rhetoric on this subject from the last two decades.

NEJM  27 June 2013  Vol 368
2455    Here’s a paper which shows that if your drug really works, you can be modest in your advertising. Sanofi and Regeneron funded a phase 2A trial of dupilumab in adults with persistent, moderate-to-severe asthma, an elevated blood eosinophil count (≥300 cells per microliter) or an elevated sputum eosinophil level (≥3%) at screening, and symptoms that were not well controlled with medium-dose to high-dose inhaled glucocorticoids plus long-acting beta-adrenergic stimulants. When these treatments were withdrawn gradually, the exacerbation rate in the placebo group was 87% higher than in the group receiving weekly dupilumab. So a pretty clear triumph for this monoclonal antibody to the interleukin-4 receptor α subunit that inhibits both interleukin-4 and interleukin-13 signalling. But no shouting in the paper, which ends on a note of quiet caution: “In conclusion, our 12-week study showed that in a subpopulation of patients with persistent asthma, dupilumab therapy, as compared with placebo, was associated with fewer exacerbations induced by medication withdrawal; the benefit was primarily identified by changes in peak flow and beta-agonist use. The short study period and the definition used for exacerbation may not reflect real-world asthma exacerbations. Further studies are needed to confirm these observations and better define the target population, dosing regimen, and long-term efficacy and safety.” Nicely put.

2487    Other than H7N9 flu, the recently discovered virus most likely to cull the human race is now called Middle East Respiratory Syndrome Coronavirus (MERS-CoV). This week both the NEJM and the Lancet rush to bring us case reports that may shed light on its clinical course and its transmissibility. Between last September and the end of May, 49 cases have been reported and 26 deaths. The report here is of three young men who seem to have acquired the infection from a 70-year-old relative who had been ill for months. One died. But 24 other family members living in the same household and 124 attending staff members at the hospitals did not become ill. Not many of us appear to be doomed.

Lancet  29 June 2013  Vol 381
2225   The outgoing head of NICE, Mike Rawlins, presents his wish list for his successor. High on it is the development of tools for shared decision making. He emphasises that they have to be made usable in real time, perhaps by bedding them into consultation software for general practice. This needs care: most such experiments have failed. Shared decision making does not need electronic aids so much as real buy-in from both patients and clinicians.

2255  The Dutch strike me as a rather stoical and adaptable nation, but here is an interventional study from the Netherlands that shows about 40% of those in nursing homes have depression, and perhaps half of these may have severe depression, depending on what instrument you use. I am not counting the residents of dementia units, whose figures are even higher. If you then put in a multidisciplinary care programme, involving nursing staff, activity therapists, psychologists, and physicians, you can reduce this by a very small amount. I am afraid that people congregated together to count out their time till death are not very susceptible to cheering up. When my turn comes, I shall try to be as nice as I can, but I suspect I will be cantankerous.

2265   The rest of the Lancet’s research space is taken up with killer plague items. “We report detailed clinical and virological data for two related cases of MERS-CoV disease, after nosocomial transmission of the virus from one patient to another in a French hospital.” Do you remember what MERS-CoV stands for? Look above. And what “nosocomial” means? The clue is in the sentence. The carrier of the virus was a 64 year old man taking immune suppressants following renal transplantation. Two months ago, he flew back to France from Dubai and became ill with a fever and diarrhoea. He was admitted to Valenciennes hospital without a cough and with a clear chest x-ray. He shared a room with a 51 year old man with a venous thrombosis who was taking high dose steroids for hereditary angio-oedema. Patient A developed respiratory symptoms three days after admission and died a month later. Patient B developed severe respiratory symptoms ten days after his first contact with patient A, by which time he had been discharged home. He went into acute respiratory failure but was improving at the time of this report.

2273   The next report is about 14 patients admitted with influenza A type H7N9 in Shanghai. Seven required ventilation and two died. They were all given neuraminidase inhibitors but regular swabbing revealed the rapid development of resistant viruses in two patients who were taking corticosteroids. So while oseltamivir and zanamivir might offer some protection at present, this would be unlikely to last if the virus became pandemic.

BMJ  29 June 2013  Vol 346
This week sees the publication in print of my essay on trust in the medical literature entitled Nullius in Verba—don’t take their word for it (see JAMA Internal Med below). I think I could have made more of this. While we can’t trust most of what we read, we often can’t really interrogate it either, because it would take several weeks, or statistical software and expertise we don’t possess. Take this meta-analysis of the evidence for antibiotic prophylaxis to prevent infection after catheter removal. The title doesn’t state “in surgical hospital wards,” but that is where most of the patients came from. That’s easy to pick up, but it’s pretty well impossible for the general reader to look at the quality of the studies included. Fortunately there’s an editorialist from Yale who does this for us: and if you can get past the BMJ paywall you will discover that the largest studies had huge differences in length of catheterisation between the control and the intervention groups, and that you could argue about the quality and generalizability of most of the included studies. That means that the bottom line—a spuriously exact 5.8% reduction in UTIs, with a NNT of 17—is quite meaningless for clinical decision-making. Nullius in verba—never believe the bottom line.

Those with long memories may remember that the Netherlands used to be held up in the 1980s as an example of the safety of birth at home; then the statistics got worse and it was held up as an example of the dangers of birth at home. With this study the pendulum swings back again: this large population study concludes that “Low risk women in primary care at the onset of labour with planned home birth had lower rates of severe acute maternal morbidity, postpartum haemorrhage, and manual removal of placenta than those with planned hospital birth. For parous women these differences were statistically significant. Absolute risks were small in both groups. There was no evidence that planned home birth among low risk women leads to an increased risk of severe adverse maternal outcomes in a maternity care system with well trained midwives and a good referral and transportation system.” That last sentence is critical. What is safe in a country with small distances, rapid transfers and good training may not be safe elsewhere: childbirth is inherently unpredictable and sometimes dangerous.

Data linkage studies are the evidence factories of the future—but are they able to provide reliable evidence? I was rather taken aback by the recent study that showed that the UK GPRD is terrible at recording even something as important as myocardial infarction. Let’s hope the Swedish databases used in this study are more dependable, because we really need to know if we may be harming COPD patients by giving them the wrong combination of inhaled steroid and long acting bronchodilator. The bottom line presented by these investigators is that patients inhaling fluticasone/salmeterol (Flixotide) are about three-quarters more likely to get hospitalised for pneumonia than those taking budesonide/formoterol (Symbicort). Dilemma: it would take about 5 years to do a head-to-head double blinded randomized trial. What do we do in the meantime?

JAMA Intern Med  24 June 2013  Vol 173
1049   If you use the link and have reasonable eyesight, you can read most of my Viewpoint about trust in the medical literature, which is the first of three. I spend too long going on about how I write these reviews, and the JAMA paywall means you don’t get to the end. There I talk broadly about the coming era of massive data, and the final paragraph reads: “How the medical journals of the future will cope with these profound changes remains to be seen. My hope is that readers will be able to dispense with a lot of the scepticism and the “critical reading skills” that are now required to interpret the literature. In the future, readers should be guided into the entire hinterland of the real data and its meaning for clinical practice, with each increment fitted into the whole. I look forward to a future of journal reading that is more clinically relevant, more interesting, more trustworthy—and more fun.”

1053   If you want a better argument for much the same thing, read this piece by Califf et al.

1058   Emergency rooms are dangerous places for people with heart failure. In some cases (I know from an unpublished study) the doctor is likely to put up a litre of saline and give you inhaled beta stimulants and corticosteroids before doing anything else: in many other places, your fluid intake will be restricted while you are rendered thirsty, hypovolaemic, and hypotensive with loop diuretics, and then given inotropes to bring you round and/or kill you. To try to bring some science (and humanity) into this situation, patients admitted with acute decompensated systolic HF in this trial were randomized to fluid and sodium restriction or none. “Aggressive fluid and sodium restriction has no effect on weight loss or clinical stability at three days and is associated with a significant increase in perceived thirst. We conclude that sodium and water restriction in patients admitted for ADHF are unnecessary.”

1128   Coming to hospital with non-cardiac chest pain is also extremely perilous, especially in the USA. Guidelines there mandate that you are offered an appointment with a cardiologist within three days, irrespective of the likelihood of your chest pain being cardiac. The cardiologist will then order a treadmill test, just to be on the safe side (and to earn money); and as you will know from reading Sackett et al, the false positive yield of stress tests depends on the pre-test probability, which in some of these patients will be close to zero. False positive treadmill means angiography (another nice little earner) and maybe the detection of some atheroma, sitting quietly and doing nothing. In goes the stent: the patient is grateful: she thinks her life may have been saved, and she has a label of ischaemic heart disease so every future chest twinge gets similar attention. Or, as this study observes, “In an emergency department–based chest pain unit, routine provocative cardiac testing generated a small therapeutic yield, new diagnoses of coronary artery disease were uncommon, and false-positive results were common.” But as my example demonstrates, this is only the half of it: detecting atheroma in the coronary arteries is not the same as establishing it as the cause of the chest pain. Why don’t people get this? Here is the distinguished Pat Crosskerry writing in the first article of this week’s New England Journal: “When a patient undergoes analytic assessment for chest pain in a cardiac clinic that culminates in angiography, the conclusion is invariably correct.” Oh boy.

Plant of the Week: Helianthemum “Ben More

Over the next few weeks, England puts on the brave pretence of having a Mediterranean climate, and we get what joy we can from the sun-loving shrubs from that area which can tolerate our climate. I can never remember which ones are classed as Cistus, Potentilla or Helianthemum, let alone what garden names they are given. Some will survive our coldest winters, but they all get bare and straggly with time.

They come in all sorts of exuberant colours and you just buy the ones that take your fancy. Even in garden centres they are quite cheap, and in markets you can generally get several for a fiver. “Ben More” is a gay fellow of brightest tangerine, guaranteed to clash happily wherever you put him. That is what these plants are for. Naturally, you must give them all the sun that Britain can muster.