Richard Lehman’s journal review – 30 January 2012

Richard LehmanJAMA  25 Jan 2012  Vol 307
373   Here’s the kind of study that’s all too rare in the medical literature: an important interventional trial that is not funded by pharma. The question is whether giving a proton pump inhibitor can improve outcomes in poorly controlled childhood asthma: a reasonable hypothesis to test, since a high proportion of such children have been shown to have asymptomatic acid reflux. This double-blind RCT recruited 306 children from 19 US centres, and shows that daily lansoprazole has no benefit and that it may have important harms. The active drug group had more respiratory infections and markedly more fractures (6 vs 1): not to the point of statistical significance, but definitely to the point of warranting an urgent database study of fracture incidence in children taking regular PPIs.

NEJM  26 Jan 2012  Vol 366
299     The history of breast cancer is strewn with the tragedies of millions of women made to suffer through wishful thinking. William Halsted introduced ever-more mutilating radical mastectomy in the 1890s, which remained popular in the USA for nearly a century; then came screening mammography, with its spectacular rates of overdiagnosis; and now a series of phenomenally expensive and marginally effective monoclonal antibodies against molecular targets, as illustrated by bevacizumab. Last November, the FDA revoked its approval for this vascular growth factor A inhibitor in metastatic breast cancer, but meanwhile two large trials of bevacizumab in early breast cancer had been completed. Both of these use a surrogate end-point called pathological complete response, meaning absence of invasive and intraductal disease in the breast and the axillary lymph nodes. Used in combination with neoadjuvant chemotherapy, bevacizumab has a barely discernible effect on this debatable end-point: an absolute increase in response of less than 5% in the first trial, and 2.1% in the second. In both trials, a subgroup could be found where the response rate was slightly larger, but the results seem at odds with each other: triple-negative tumours in the first, and oestrogen receptor positive tumours in the second. No matter: the cost of bevacizumab in these trials was about $50 000 per patient treated, and it will have to do a great deal better than this. Being shown to save a life would be a start.

321   But of course, nothing ever saves a life. If we don’t want to die of cancer, we might as well die of cardiovascular disease. Perhaps sooner rather than later, since “happiness (is) but the occasional episode in a general drama of pain.” (Thomas Hardy, final sentence of The Mayor of Casterbridge, 1886). Tush, I must stop reading Hardy and take some citalopram. Here is a classic paper which describes the lifetime risks of cardiovascular disease: though the very fact of its being a printed article places absurd limits on its usefulness. A meta-analysis like this, covering more than a quarter of a million people from the age of 45 till death, needs a whole website with full databases and several banks of Powerpoint slides. Traditional publishing is rubbish when it comes to a topic like this: and you don’t even get open access to these printed titbits. Suffice to say that wherever you live and whatever your ancestry, your CV risk is a simple function of four factors: blood pressure, smoking, cholesterol, and diabetes. These are all continuous variables, and we are looking at 18 cohorts studies involving 257,384 individuals. And here is an 8-page article: what a way to carry on in the year 2012!

339    Like most British general practitioners, I have never prescribed fingolimod for multiple sclerosis. In fact I had only the faintest notion of the existence of thingummybob, and only read this review article because it seemed the sort of name it would be easy to make jokes about. Well, here is a little joke for you. You are a GP who has decided (or rather been forced) to take part in your local commissioning group, aided by five tiers of NHS management (so as to save costs). Your patient Toni is a teacher with two school-age children who developed MS three years ago and has been unable to work for one year. She experienced severe adverse effects with interferon-alpha and during her latest flare-up she became incontinent and has had walking difficulties ever since. She is becoming unable to care for her children. Her neurologist has suggested treatment with oral fingolimod, subject to approval by your commissioning consortium. This will cost £15K per annum, and the consortium has been instructed to make group savings of £2M to cover last year’s deficit. You read this review article which provides a useful overview of all MS treatments and indicates that fingolimod could reduce her relapse rate considerably but has no proven effect on progression of disability. Its long-term safety is unknown and individual response is not predictable. How do you share the decision about what to do next with your patient? And your consortium? This of course is what you will end up doing all the time if the 50 signatories to the Daily Telegraph letter have their way. (And now, finally, the joke [sorry]: Q: what is a signa-tory? A: someone who signs letters in the Tory press.)

Lancet  28 Jan 2012  Vol 379
315    Following gastric cancer surgery which is intended to be curative, adjuvant chemotherapy is generally agreed to improve survival, but this has not been demonstrated specifically for D2 gastrectomy followed by capecitabine and oxaliplatin. OK, I know you aren’t interested in the details of chemo for specific cancers, but this trial raises again the issues of what end-points really matter. To a fundamentalist like me, the principal issues are all-cause mortality and quality of life. But most cancer triallists prefer recurrence-free survival, a composite outcome, which is easier to arrive at in terms of time to statistical significance. In this case, the chemo increases disease-free survival and the chart for overall survival is promising. The price is nausea, lack of appetite and neutropenia for the majority of patients.

322    I’ve already commented on this meta-analysis by Carl Heneghan et al of individual patient data relating to self-monitoring of anticoagulation, which has major implications for clinical practice. Come on, primary care academia: there is more low-hanging fruit like this if only you would get on and pick it.

335    I’m no great fan of screening in general, but most neonatal screening is self-evidently a good thing, because it detects rare but important metabolic defects which have a known natural history and a known treatment. If this knowledge did not exist, screening would simply be a recipe for lifelong distress and anxiety. In this paper an Austrian population study shows that neonatal screening is possible for several lysosomal storage disorders: Gaucher’s, Pompe’s, Fabry’s and Niemann-Pick types a and B. These proved to be commoner that expected, with an overall incidence of 1 per 2315 births, dominated by Fabry’s disease, an X-linked condition which has a late-onset prevalence in Austria of 1 per 4100 (the paper says incidence, but I assume this is an error). There is a moderately useful enzyme treatment which prevents cardiac and renal damage in Fabry’s disease, which is an argument for early diagnosis; on the other hand, this only applies to the classic phenotype. I think we need some very robust modelling and much thought before going in for this particular mode of whole-population screening.

361   For seminars this week, The Lancet can offer you cryoglobulinaemias, acne vulgaris, or thalassaemia. As we have already strayed far into recondite topics, let’s return to that most familiar of spots, acne. If you don’t know a lot about acne, this review will tell you everything you need to know. If you do know a lot, it will remind you that trimethoprim is a good alternative to the tetracyclines, and that no kind of oral contraceptive has been shown to clear acne better than any other.

BMJ   28 Jan 2012  Vol 344
I know I go on a lot about surrogate outcomes, but surely weight loss in obese people, especially if they have diabetes, must be a good thing? Well, no, in fact: sibutramine and diethylpropion both reduced weight – and thus glycaemia in type 2 diabetes – but resulted in more harm than benefit. Agonists of the glucagon-like peptide-1 receptor (GLP-1R) have a weight-reducing action in both diabetic and non-diabetic subjects, as this systematic review demonstrates; and they are as effective as any other non-insulin agents at lowering glycated haemoglobin. But that does not of itself make them either safe or useful. They should not have been licensed without clear evidence of long term benefit as measured by cardiovascular events and all-cause mortality.

My mother was diagnosed late with ovarian cancer, and died a year later. Do I blame her GP? Of course not. Unless discovered incidentally by imaging or laparotomy at an early stage, ovarian cancer is always diagnosed too late for a cure. In fact it is such an insidious and ill-understood disease that even in women with BRCA mutations who have undergone prophylactic removal of both ovaries, it can still present with omental secondaries and a pelvic effusion. Yes, studies show that with hindsight, women usually present with symptoms a few months before they are diagnosed – usually with a description of abdominal fullness that doctors confuse with bloating. Listen carefully to the patient – if she says her tummy feels bigger all the time, get an ultrasound scan. I suspect that is just as good as the computer algorithm for earlier diagnosis of ovarian cancer devised and validated here by the Nottingham group. But however you diagnose it, don’t expect to save any lives.

Arch Intern Med  23 Jan 2012  Vol 172
98     Ever since immediate percutaneous coronary intervention was shown to be superior to thrombolysis for high-risk ST-elevation myocardial infarction, health systems have been pouring huge resources into providing 24-hour immediate PCI facilities, but with inevitably mixed success. Outcomes researchers in the USA have come up with some pretty dire statistics for time to transfer from hospitals without PCI capacity to hospitals with, and in this editorial Rita Redberg grasps the nettle and urges a more realistic approach: “For low and intermediate risk patients, there is no mortality advantage to pPCI over thrombolytic therapy. Even for high-risk patients with STEMI, the mortality benefit of pPCI is frequently lost due to routine delays of 1-3 hours by transfer. It is time to reconsider transferring patients with STEMI for pPCI. Timely reperfusion by thrombolytics, not late pPCI via transfer, will save lives.” Discuss.

101    Popular myth tells us that exercise combats depression by producing endorphins – or is it serotonin? – and for all I know popular myth may be right. There’s no doubt that exercise has other benefits too, so it’s nice to know from this systematic review that the antidepressant effect of exercise in chronic illness is supported by evidence from a large number of trials in a large range of conditions.

144    I would like to claim that the population benefits of taking statins are not outweighed by any potential for irreversible harm, and that now they are so cheap there is no reason why anyone should be advised against taking a statin if they wish to reduce their cardiovascular risk. There is no threshold effect, after all. But now there is a nagging worry that statins may induce diabetes, confirmed here by data from the Women’s Health Initiative, which included over 150,000 postmenopausal women without diabetes at baseline. This needs a lot more investigation: there is a lot of scope in such studies for residual confounders, including confounding by indication; and we need to know more about the potential reversibility of statin-induced hyperglycaemia.

Poem of the Week: Her Apotheosis by Thomas Hardy

Having mentioned Thomas Hardy the nineteenth-century novelist at his most morose, I think I should give you a poem illustrating Thomas Hardy the great twentieth-century poet at his best – enigmatic and not all that morose:

There were years vague of measure
Needless the asking when;
No honours, praises, pleasure
Reached common maids from men

And hence no lures bewitched them,
No hand was stretched to raise,
No gracious gifts enriched them,
No voices sang their praise.

Yet an iris at that season
Amid the accustomed slight
From denseness, dull unreason,
Ringed me with living light.

I read the poem several times aloud, wondering at the sudden appearance of the iris, whether flower or eye, which breaks the rhythm and rings the woman with living light. Hardy’s motto for the poem, Secretum meum mihi, and his subtitle, Faded Woman’s Song, give little away.

And then I looked it up on Google and got this travesty instead:

There was a spell of leisure,
No record vouches when;
With honours, praises, pleasure
To womankind from men.
But no such lures bewitched me,
No hand was stretched to raise,
No gracious gifts enriched me,
No voices sang my praise.
Yet an iris at that season
Amid the accustomed slight
From denseness, dull unreason,
Ringed me with living light.

I assume this is the final version; but authors so seldom know best.