Richard Lehman’s journal review—27 February 2017

richard_lehmanNEJM  23 Feb 2017  Vol 376

Kallikrein rises from the footnotes

The curious word “kallikrein” first appeared in 1934, when Eugen Werle discovered an inflammatory chemical  in plasma which he thought came from the pancreas. It is supposed to be derived from the Greek word for pancreas. Anyway, the human disease most closely associated with too much plasma kallikrein is hereditary angioedema with C1 inhibitor deficiency. We had a patient in our practice with the condition, which is why I know a tiny bit about it. It has a prevalence of between 1 in every 10,000 to 1 in every 50,000 persons, so you wouldn’t have thought it was worth anyone’s time and money to develop a monoclonal antibody to kallikrein. But they have, and it’s called lanadelumab, and the NEJM has thought fit to devote some of its hallowed pages to a phase 1b trial of it in 24 patients with hereditary angioedema with C1 inhibitor deficiency, who were compared with 12 patients given placebo. From day 8 to day 50, the 300-mg and 400-mg groups had 100% and 88% fewer attacks, respectively, than the placebo group. The idea is to use this as long-term preventive therapy, which may make C-1 inhibitor deficiency into a very expensive condition to live with.

MRI & metallic cardiac devices

As everybody knows, MRI machines can send metal objects flying around and shouldn’t be used on people with certain kinds of metal inside them, because MRIs operate with a magnetic field of 1.5 teslas. I don’t quite understand what a tesla is, but this a lot of them. If you have an MRI with an implanted cardiac device inside, the worry is not that it will fly out of your chest but that you will get magnetic field–induced cardiac lead heating, which could result in myocardial thermal injury and detrimental changes in pacing properties. But here is a reassuring study from America: non-thoracic MRI was performed in 1000 cases in which patients had a pacemaker and in 500 cases in which patients had an ICD. No deaths, lead failures, losses of capture, or ventricular arrhythmias occurred. And these devices were of the “ordinary” kind, with no special protective features.

JAMA  21 Feb 2017  Vol 317

Testosterone-fuelled plaque?

This is testosterone week in the two leading JAMA journals. Seven trials called TTrials (you can guess why) tested the effect of testosterone gel on various aspects of male health over a period of one year. Reading the published findings of them, I find that my ignorance is considerably deepened. The first one concerns the effect of a year’s testosterone gel on coronary plaque formation. The volume of plaque was measured at the start and the end of the trial, and there was more plaque in the chaps who used the gel. This sounds unfortunate, but its significance is entirely unclear. A study in JAMA Internal Med below suggests that men receiving any kind of testosterone “therapy” actually have a lower rate of observed cardiovascular events. Maybe measuring the volume of plaque over a year is just a way of giving men large doses of ionizing radiation. This just wasn’t worth doing.

Testosterone is not brain fuel

The TTrials recruited men who were 65 and over who had impaired sexual function, physical function, or loss of vitality. Golly, are there any other kinds of men over 65? How can I become one? I am 66: should I have my testosterone checked? If these trials have any meaning, it can be summed up as: brother, don’t bother. The whole idea of “low T” is one of the most baseless and lucrative areas in medicine, which is saying a lot. Global testosterone sales have increased 100 fold in the last thirty years. As the editorial accompanying these articles points out, the evidence about testosterone replacement hasn’t changed at all over that period. In this part of the TTrials, men with those symptoms and a testosterone less than 275 ng/ml (9.5 mmol/L) were randomised to get testosterone gel or placebo for a year and had a paragraph memory test at the start and finish. There was no difference between groups.

Adjusting your observational analyses

Over the last few months I’ve referred several times to the village of Residual Confounding, which began as a rather weak joke about Cotswold place names. The truth I was trying to convey is that you cannot rely on multiple adjustments to “prove” causality from observational data. You can allow for anything you can conceive of, but you cannot allow for things you can’t conceive of. This inconvenient truth is explored in a terrific piece by Thomas Agoritsas. I won’t try to summarize it because it’s not a subject to take short cuts about. Simply examine your conscience: do you find yourself often wishing that Big Data and black box machines like IBM Watson will one day do away with the need for randomised controlled trials? Wake up and repeat with me: “Although propensity analysis and instrumental variable analysis can reduce the risk of bias in observational studies, none replace the balance of both known and unknown prognostic factors offered by randomization.” Rest in peace, Archibald Cochrane.

JAMA Intern Med  Feb 2017 

Testosterone continued…

Back to the men who had testosterone gel for a year. What happened to their bone density? It increased, more in trabecular than in peripheral bone, and more in the spine than hip. Well, we kinda knew that already. What we still don’t know is whether it safely prevents actual fractures in the long-term.

When they were taking blood for the TTrials, they found that of 788 participants of mean age 75, 126 had a haemoglobin of 12.7 or less, and in 62 no cause could be found. Given testosterone gel, 54% of these showed an increase in Hb of more than a gram over the year, compared with 15% of the placebo group. Not sure what this tells us.

The TTrials explored the relationship of testosterone gel, prescribed for low testosterone in men, with a surrogate measure of coronary risk. That was a randomised trial. The study here is based on observational data collected for younger men taking testosterone gel for the same reason, and the data show that they had fewer cardiovascular events over 3.4 years than men with the same testosterone “deficiency” who didn’t receive treatment. So here is the problem: do you believe a properly randomised study with a lousy surrogate outcome, or an observational study with a clinically meaningful outcome? The answer is that you don’t “believe” either. You ask for a randomised trial with clinically important outcomes over a sufficient period in a large number of men with a typical age distribution.

The Lancet 25 Feb 2017  Vol 389

BETR decontamination

One of my greatest medical heroes is John Haygarth of Chester (1740-1827)  who achieved astonishingly low rates of mortality in his fever hospital through the rigorous enforcement of basic hygiene and the scrupulous decontamination of bedlinen and hospital spaces, long before anyone believed in the microbial causes of infection. It still works. This American study tried out more rigorous cleaning of hospital rooms after occupancy by an infected patient, using much the same bleach as Haygarth, among other agents. The effect of enhanced cleaning was positive though not dramatic, probably because Haygarthian practice is already so widely implemented.

HERA: one year of trastuzumab as good as two

HERA is an international, multicentre, open-label, phase 3 randomised trial of 5102 women with HER2-positive early breast cancer, who were enrolled from hospitals in 39 countries between Dec 7, 2001, and June 20, 2005. Herceptin (trastuzumab) remains an extremely expensive drug and if I was as smart as Ben Goldacre I could tell you how much it has earned Roche since it was first marketed. HERA shows it can provide a small absolute increase in overall survival in this group of women, and the 11-year data show that it makes no difference if it is given for one year or two.

Screening with sigmoidoscopy

The median follow-up time in this analysis of the UK Flexible Sigmoidoscopy Screening randomised controlled trial was 17 years, during which incidence of colorectal cancer at all sites in screened patients was reduced by 35%, and distal cancer incidence was reduced by 56%. So this one-off intervention definitely works, and the longer it works for, the lower the number-needed-to-screen per cancer prevented drops. It now stands at 98, and I say “prevented” because most of the lesions picked up are pre-malignant polyps which can be treated on the spot rather than cancers. Does this screening reduce all-cause mortality? I don’t know. Should it be offered? Yes, with an individualised decision aid, taking account of age and sex.

Life expectancy breaks the 90 barrier

I have to say that as I get older, I value longevity less. This is probably simply an excuse to avoid activity. As you will have read in the papers, life expectancy is still trending ever upwards in most developed countries (the USA excepted) and especially in South Korea. Women in particular will face the dreadful prospect of living past the age of 90. South Korean women should form cooperatives to plant vineyards, so that those extra years are worth living in that wine-deprived country.

The BMJ 25 Feb 2017  Vol 356

Why do we bother, Fawlty?

This is a good week to be reminded of the foundational principles of evidence based medicine. First find the evidence. Make sure you have all of it before doing a systematic review. Trial registries are an obvious place to look. But this survey found that among 223 selected systematic reviews, 116 (52%) did not report a search of trial registries. This is bad. Actually, it is terrible.

Using antidepressants “off-label”

Canada falls between the UK and the USA in its medical culture. That is quite an awkward place to be. The bad habits to be found there may be those of either country, or both. One is the widespread prescribing of antidepressants for conditions for which they are not licensed, and to my mind the very worst is to prescribe trazodone for insomnia in the elderly. Trazodone is a lousy antidepressant and is as likely to cause agitation as it is sedation. It interacts with SSRIs and tramadol. Canadian GPs, like their British colleagues, are unaccountably fond of it. Like us, they also don’t pay a lot of attention to detail. “When primary care physicians prescribed antidepressants for off-label indications, these indications were usually not supported by strong scientific evidence, yet often another antidepressant in the same class existed that had strong evidence for the respective indication.”

Ultrasound does not heal fractures

A systematic review of low intensity pulsed ultrasound (LIPUS) for bone healing finds a lot of small trials (median sample size of 30) and concludes that “Based on moderate to high quality evidence from studies in patients with fresh fracture, LIPUS does not improve outcomes important to patients and probably has no effect on radiographic bone healing. The applicability to other types of fracture or osteotomy is open to debate.”

I don’t know how much LIPUS is used in practice, but it sounds like a waste of time. However, rather than rehearse the familiar arguments about absence of evidence versus evidence of absence, you can now go to a great new article and infographic in the Rapid Recommendations series, with Thomas Agoritsas starring for a second time this week. If you dabble with the interactive charts, you’ll find that there really isn’t much point to LIPUS.

Plant of the Week: Helleborus × hybridus (Spring Promise Series)”Grace”

 Another week, another hellebore. Sorry, but these magnificent perennials outclass anything else on offer at this time of year, and in some cases, at any time of the year. They grow readily on our limy clay and maintain their beauty year after year, with little need for attention except for removal of their old leaves prior to flowering.

This one caught our eye in a garden centre a couple of days ago. It has large double flowers of white, tinged with green at the centre and behind the petals, which have an exquisite fine edging of pink if you look closely. There are very few flowers to match these in any season, and you can see why the oriental hybrid hellebores are sometimes called “Lenten roses”. The flowers of the new Spring Promise series are especially rose-like and borne on long strong stems.

When they are as good as this, I think hellebores are best planted prominently and singly on a raised site, where you can admire them individually from beneath. The flowers last for weeks, unless you mistakenly cut them and bring them indoors. Never do that: settle for daffodils instead.