Richard Lehman’s journal review—17 February 2014

richard_lehman NEJM

599 Most weeks I quote you the conclusion of some pharma-funded trial which overstates the benefit of an intervention. But in reality clinical trials of any kind can be a form of marketing: doing them is difficult work, there are reputations and ideas at stake, and the temptation to overstate results is always there for career academics as well as for pharmaceutical marketing departments. And I really admire the work that went in to the Multicenter Selective Lymphadenectomy Trial (MSLT-I) for melanoma, which commenced in 1994. Twenty years on, we are told that “Biopsy-based management prolongs disease-free survival for all patients and prolongs distant disease-free survival and melanoma-specific survival for patients with nodal metastases from intermediate-thickness melanomas.”

Technically, that is true: but if I were a melanoma patient deciding whether to have sentinel-node biopsy or not, I would like to be shown the Kaplan-Meyer charts on Figure 1. I’d still probably go for the biopsy, but the absolute differences in clinical outcomes are very small.

610 A title like “Oxantel Pamoate–Albendazole for Trichuris trichiura Infection” always gives me a guilty start: should I have heard of oxantel pamoate and do I know anything about Trichuris trichiura infection? I feel like the hapless Dr Watson in The Dying Detective when Sherlock Holmes rounded on him and declared “you are only a general practitioner with very limited experience and mediocre qualifications…What do you know, pray, of Tapanuli fever? What do you know of the black Formosa corruption?” It turns out that Holmes was playing tricks on poor Watson and had made up these fearful diseases, which do not exist. Trichuriasis, however, does exist; but it is not fearful and it is not an infection. It is a worm infestation, extremely common among the children of Mchangamdogo. “Now you really are raving, Holmes. You cannot expect me to believe that the earth contains any place of such name!”

“My dear Watson, here is a paper in the Journal of the Massachusetts Medical Society to prove it. Moreover, in this village in Africa, infestations of whipworm are even now being treated with oxantel pamoate.” “Holmes, you astound me! Yet I cannot deny the proof of my own eyes. I should never have believed such a thing. These strange names are nothing to me – but that the medical men of Boston should not know the difference between an infection and an infestation!”

621 There are some who say that restless legs syndrome is a non-diagnosis devised to sell drugs. And there are those who say pregabalin is just a me-again drug devised by Pfizer to be marketed once the patent on gabapentin expired. Far be it from me to give an opinion on such matters. I am only a general practitioner with very limited experience and mediocre qualifications, and this is a paper in the New England Journal of Medicine. The conclusion declares that in this mysteriously common syndrome, ” Pregabalin provided significantly improved treatment outcomes as compared with placebo, and augmentation rates were significantly lower with pregabalin than with 0.5 mg of pramipexole.” That’s enough for me. (NB. In this context, augmentation refers to a worsening of the condition thought to be due to treatment.

JAMA Internal Medicine

251 Goodness, it’s taken a long time for the diabetes community to come to terms with the obvious. A study of people over 60 with diabetes finds that the main ill effects of their condition in old age are coronary artery disease and hypoglycaemia. In people who develop diabetes after the age of 60, microvascular disease is so uncommon it is hardly a consideration: whereas the commonest treatments given to lower sugar frequently cause hypoglycaemia and have little or no effect on coronary disease. Step back, guys: treating type 2 diabetes is a whole different ball game in a 65 year old as compared with a 35 year old. One size does not fit all ages. In fact, each patient of any age needs to be treated in accordance with his or her own goals and informed preferences. People with diabetes have for so long been misinformed about the benefits and harms of treatment that I suspect we often keep them on drugs out of sheer embarrassment at admitting we didn’t know what we were doing. And it’s still going on: we “put people on” drugs like incretin mimetics when we haven’t a clue what they do to long-term outcomes, and then congratulate ourselves just because their HbA1c has gone down.

281 “All screening does harm: some screening does more good than harm,” is the sentence Muir Gray uses to begin his chapter on screening in the Oxford Textbook of Medicine. This is a huge topic, and every part of it is complex. So I’m glad to see a lucid dissection of the issues in the Less is More series, using lung cancer screening as an example. This is a must-read for thoughtful criticS of screening: “We propose a taxonomy with 4 domains of harm from screening: physical effects, psychological effects, financial strain, and opportunity costs. Harms can occur at any step of the screening cascade… the taxonomy also makes clear where … we have useful information and where there are gaps in our knowledge.”

Online Another week, another great paper from a Yale medical student. If this is beginning to sound like advertising, I don’t care: I wish every medical school had a Harlan Krumholz and Joe Ross who would encourage attached students to produce work of such quality. This time the student’s name is Kyan Safavi, and he did a massive survey of data about variation in non-invasive cardiac imaging for suspected ischaemia across US hospitals. “Hospitals with higher imaging rates did not have substantially different rates of therapeutic interventions or lower readmission rates for AMI but were more likely to admit patients and perform angiography.” So non-invasive imaging leads to invasive imaging without showing any clear benefit in patient outcomes. Those Americans, eh? But I bet you would find exactly the same variation in the UK, especially between district general and teaching hospitals.

The Lancet

603 It’s ninety years now since Geoffrey Keynes, the great surgeon and literary scholar, first tried out intra-operative local radiotherapy for breast cancer. At the same time, he pioneered breast-conserving cancer surgery. Nobody took much notice.

Now breast-conserving surgery is standard and intra-operative local radiotherapy is making a comeback. The TARGIT-A trial for invasive ductal carcinoma concludes that “single-dose targeted intraoperative radiotherapy concurrent with lumpectomy within a risk-adapted approach should be considered as an option for eligible patients with breast cancer carefully selected as per the TARGIT-A trial protocol, as an alternative to postoperative external beam RT.” Sir Geoffrey lived to the age of 95: if only he’d hung on another 40 years, he could have seen his work reach fruition. I wonder if somebody is going to rediscover his work on the thymus next. In the mean time, enjoy his magnificent editions of the works of William Blake.

614 The ARUBA trial is described in the editorial as “a valiant effort to help improve understanding of brain arteriovenous malformation natural history and treatment risks.” In fact the editorial spends most of its time showing how, with a highly selective recruitment protocol and just 33 months of follow-up, it cannot hope to give a true picture of the balance of harm and benefit between medical treatment (which could mean anything), and invasive treatment, which could mean embolization, radiotherapy or surgery. This prematurely terminated trial is reported at the end of the abstract as showing “that medical management alone is superior to medical management with interventional therapy for the prevention of death or stroke in patients with unruptured brain arteriovenous malformations followed up for 33 months.” Although this was not a pharma-funded trial, again I think this conclusion overstates the importance and generalizability of the findings. Journal editors should write these abstract conclusions themselves, and they should always contain caveats.

622 In the next paper, the abstract conclusion (called Interpretation) can afford to be laconic: “Changes in blood pressure after renal denervation persist long term in patients with treatment-resistant hypertension, with good safety.” This is indeed true of the 88 out of 153 patients who had full follow-up data at 36 months in this Medtronic-funded Symplicity HTN-1 trial. These were people whose blood pressure remained high despite treatment with an average of five different agents. And the drop in BP following percutaneous radiofrequency ablation of the renal nerve supply was little short of spectacular: a mean fall of 32 mm Hg in systolic and 14.4 mm diastolic. So something really big is happening here, and you could say that this is the kind of intervention which did not get a randomized trial because it didn’t need one. What it did need, however, was tighter follow-up. I know this is something Medtronic are keen to carry out in the future, and to be fair they didn’t design this trial themselves: it was instigated by a company they bought up half way through. But it was a missed opportunity to do better from the outset in the evaluation of a treatment which looks to have immense potential.

Online “Cognitive therapy significantly reduced psychiatric symptoms and seems to be a safe and acceptable alternative for people with schizophrenia spectrum disorders who have chosen not to take antipsychotic drugs. Evidence-based treatments should be available to these individuals. A larger, definitive trial is needed.” Now that’s the kind of Interpretation I like: modest and guarded, while optimistic about a major breakthrough. I like the prospect of a world in which people with high blood pressure can have a procedure which results in a permanent cure, and where people with schizophrenia are not universally doped with drugs which blunt their minds and shorten their lives. Perhaps medicine is turning in the right direction at last.


“Impaired first trimester fetal growth is associated with an adverse cardiovascular risk profile in school age children. Early fetal life might be a critical period for cardiovascular health in later life.” This is the conclusion of an important study of 1184 children in Rotterdam who had records of crown-rump length in the first trimester and anthropometric and blood tests at the age of 6. Ah well. If only we knew the determinants of first trimester fetal growth, we might be able to make cardiovascular disease even rarer in generations to come.

The BMJ certainly has the scoop paper this week: a 25-year follow up of the Canadian National Breast Screening Study showing that mammographic screening as practised then had no effect on breast cancer mortality and resulted in a high level of overdiagnosis. If you want to disbelieve this, you are welcome to read several rapid responses which criticize the methods and internal validity of the study. If you believe that mammography is a bad form of screening that probably does more harm than good, this is one more piece in a formidable edifice of evidence. On balance, I cannot see any justification for continuing with mammography as a whole-population screening programme. There may perhaps be a case for providing it to individual women on an informed choice basis.

“Women who carry a germline mutation in either the BRCA1 or BRCA2 gene have a lifetime risk of breast cancer of 60-70%.” This study from 12 US cancer genetics clinics followed up 390 such women for a mean of 13 years. ” At 20 years the survival rate for women who had mastectomy of the contralateral breast was 88% (95% confidence interval 83% to 93%) and for those who did not was 66% (59% to 73%).” So bilateral mastectomy improves all-cause mortality, overall.

When the Quality and Outcomes Framework was introduced into British general practice, I gave it a cautious welcome. Over the years, this has turned into deep loathing. Whatever marginal improvements it may have initially achieved in certain areas, its dominating role in primary care has totally distorted clinical priorities, disempowering both patients and doctors and leading to a complete travesty of evidence-based medicine. The sooner it disappears, the sooner medicine can move forwards to become a dialogue of choice made between informed individuals.

This study shows that when incentives are withdrawn, clinical care remains stable in those small domains where it is measurable. The immeasurable expanses of important personal care count for much more. If we could but return to those, we might become a compassionate, highly motivated profession once again.

Plant of the Week: Galanthus nivalis

Midwinter spring is its own season, Sempiternal though sodden towards sundown… I have never understood why these verses have been so widely admired, and sadly I find that I now much prefer Henry Reed’s parody Chard Whitlow to the real Four Quartets by TS Eliot. But there is no denying that at present Oxfordshire is highly sodden towards sundown. But since our copses and banks generally lie above the flood plains, the snowdrops are not affected. Wherever snowdrops appear, they give hope of new life to come. In the garden, they must have space to form drifts, which can be interspersed with hepaticas or winter aconites or early-flowering hellebores. But they are of course best seen as huge carpets in woodland. And soon after there will be wood anemones, and wild daffodils and then bluebells…In the days when he could still write poetry, Mr. Eliot declared that April is the cruellest month. It is not. February is the cruellest month. It is a pointless sempiternal time of freezing or drenching, and the snowdrops cannot fully make up for that. But at least they try.