Richard Lehman’s journal review—13 March 2017

richard_lehmanNEJM  9 Mar 2017  Vol 376
Impressive imatinib
In my childhood, somebody behaving self-importantly was referred to as “his nibs.” Today, the nibs of medicine are very proud of themselves. Small wonder, when the New England Journal of Medicine runs an editorial with the title “Imatinib Changed Everything.” This refers to the astonishing success of imatinib in chronic myeloid leukaemia (CML), as detailed in a paper giving the long term follow-up figures from an open label trial. Among the patients in the imatinib group, the estimated overall survival rate at 10 years was 83.3%. But the upbeat editorial comes with a sting in its tail. It turns out that CML is an unusually simple kind of blood cancer, famously signalled by its Philadelphia chromosome, which disappears once you block the BCR-ABL1 tyrosine kinase. But other cancers are much more sophisticated. “We need to resist the temptation to be self-congratulatory. The prognosis for patients with common cancers is improving somewhat, but none of the new tools appears to cure a majority of patients.” Gefinitib, erlotinib, sunitinib, cabozantinib: they may think they’re his nibs, but they are not helping the majority of people with cancer. “Inib,” you may have guessed, really stands for “inhibitor.” It also stands for the Institute of Northern Ireland Beekeepers. Perhaps a good summary of the tyrosine kinase inhibitors would be “More Buzz than Honey.”

Unimpressive pembrolizumab
The racy narrative of that editorial credits imatinib with changing the whole direction of cancer drug development. Rather than developing drugs aimed at all rapidly dividing cells, laboratories started to home in on drugs that targeted specific processes within specifically malignant cell lines. Most of these drugs were not inibs but uzumabs. Humanized monoclonal antibodies, in case you’d forgotten. There are so many of these now that I defyumab to remembumab what eachumab is for. Pembrolizumab is a highly selective, humanized monoclonal IgG4κ isotype antibody against programmed-death-1, and can disrupt the engagement of PD-1 with its ligands and impede inhibitory signals in T cells. In this Merck funded open label trial, “pembrolizumab was associated with significantly longer overall survival (by approximately 3 months) and with a lower rate of treatment-related adverse events than chemotherapy as second-line therapy for platinum-refractory advanced urothelial carcinoma.” The cost seems to be the usual £10K per month of life gained.

Post-Babesiosis Warm Autoimmune Hemolytic Anemia
This is the kind of heading that makes you love the NEJM. It is a sort of Proust biscuit, bringing back memories of warm, wood floored libraries with the whirr of a photocopier in a distant corner. No electronic screens then, except perhaps a flickering greenish thing in the keep of the librarian. Here is pure exam knowledge for the world outside the northeastern and upper midwestern regions of the United States, where the incidence of babesiosis has increased dramatically in the past 10 years. This is “a tickborne zoonotic disease caused by intraerythrocytic protozoa of the genus babesia, and is characterized by nonimmune hemolytic anemia that resolves with antimicrobial treatment and clearance of parasitemia. The development of warm-antibody autoimmune hemolytic anemia (also known as warm autoimmune hemolytic anemia [WAHA]) in patients with babesiosis has not previously been well described.” Ah, and here comes the Coombs’ test! What with the Philadelphia chromosome above, I feel 45 years younger already! Time to leave the library in my purple bell bottom jeans, and go off to sit on some bean bag eating moussaka and drinking Lutomer Riesling with friends.

JAMA  7 Mar 2017  Vol 317
Diet and CVD: death by association
If I sometimes strike a note of weary cynicism after nearly 19 years of writing these reviews, much of the blame falls on the habit of all the leading journals of printing long articles and then rubbishing them in short editorials. Here’s a classic case. Six authors spend an immense amount of time trying to quantify the association between mortality due to heart disease, stroke, or type 2 diabetes and 10 kinds of dietary intake in the United States: fruits, vegetables, nuts/seeds, whole grains, unprocessed red meats, processed meats, sugar sweetened beverages, polyunsaturated fats, seafood omega-3 fats, and sodium.
They conclude that about 45% of such deaths can be attributed to diet. Their assumptions and methods are then demolished in an editorial, which consists entirely of “How do I hate thee? Let me count the ways.” Is this fair on the authors? Or the readers? I suppose it has some value as a teaching example of how not to do association studies, but to me it seems more like cruel and unusual punishment.

Pregnancy weight and cerebral palsy
The other association study on the JAMA website is very different and results in sound knowledge. It sets out to explore associations between early pregnancy body mass index (BMI) and rates of cerebral palsy by gestational age, and to identify potential mediators of these associations. To do this, it has the benefit of accurate registries covering the whole of the Swedish population, and the investigators were able to follow up children over a median of 7.8 years. The risk of cerebral palsy in Sweden is very low, at 2.13 per 1000 live births. But this increases with each level of excess BMI, doubling in the very obese category. The association was limited to children born at full term and was partly mediated through asphyxia related neonatal complications.

The Lancet  11 Mar 2017  Vol 389
Bleeds on dual antiplatelets
PRECISE-DAPT is a score that you are unlikely ever to use, unless you are an interventional cardiologist or a medical negligence lawyer. It quantifies the risk of bleeding on dual antiplatelet treatment following coronary artery stenting. I like that it was calculated from individual participant data from eight randomised controlled trials: it shows the value of making such data available. And I like that the authors suggest that this score needs to be part of larger process of clinical decision making—honestly shared with the patient by the interventional cardiologist, I hope.

Trichotillomania is when you can’t stop pulling your hair out. Polypillomania is when you can’t stop advocating for a mixture of drugs in a pill that hasn’t been shown to work. Two articles in this week’s Lancet had me in a state of polypillomania induced trichotillomania. Where before I had little hair, now I have none. Polypills are all about making the population take drugs that will reduce endpoints we can count. Hairiness, heart attacks, strokes: all these can be quantified on a population level. In theory, a mixture of a statin, a blood pressure lowering agent or two, and perhaps aspirin, could reduce cardiovascular events in whole populations. But although the concept is 15 years old, nobody has yet shown this to happen. Summarising the studies to date, the first article states: “Trials published to date have not been designed to detect differences in clinical outcomes, and thus no significant differences between polypill and comparator groups have been reported. Polypill therapy could be one of the most scalable strategies to reduce the risk of premature mortality from atherosclerosis by 25% by 2025 by improving medication adherence and access, but further trial data and clinical experience will be useful to determine how polypills can best be implemented to achieve this goal.” The Lancet has chosen the positive speculation in this passage for its front cover and left out the rest.

The second paper is about implementation science, which is not really a science but a bunch of strategies that you bring to bear once you have proved something works. Unless you intend to put the polypill into the water supply, this consists of persuading individuals to take a pill every day for life. For the prescriber, this starts with knowing the number needed to treat for benefit or harm for each ingredient, based on each individual’s modifiable risk factors. But we don’t know the NNT for any polypill. And that is just a start: the real challenge lies in making that comprehensible to every individual, and eliciting their degree of understanding about each component of treatment. Try that with the polypill and see how much time it would take. Yet this is the only way you can ethically start lifelong treatment, and in my opinion it should be ratified by a formal consent form. If this is what’s being proposed, that’s fine: anything else is a violation of shared decision making.

The BMJ  11 Mar 2017  Vol 356
RAAS drugs: long term dangers
When captopril, the first angiotensin converting enzyme inhibitor, was introduced in the 1980s, people were started on it under hospital supervision in case they fell over after the first dose. It could cause severe postural hypotension, and there were many reports of acute kidney failure in older patients. But gradually we have come to be complacent about ACE inhibitors and their cousins, the angiotensin receptor blockers. As long acting drugs introduced at low doses in healthy people with raised blood pressure, they are relatively safe. In escalating doses given to older people with heart failure, they can do a lot of harm, although one is not supposed to say so. On the contrary: not to do this is considered a grave sin. Here is a study based on the UK primary care databases looking at people who experienced an increase in creatinine after starting a RAAS blocker. They showed a steady increase in the risk of end stage renal failure, myocardial infarction, heart failure, and death according to the degree of creatinine elevation, even below the (often ignored) 30% threshold, which is supposed to trigger reduction or discontinuation.

Insulin for T2DM: encouraging the wrong Rx?
As I said in relation to the polypill, implementation science is something you use once you have established that something is a good thing. Reducing blood HbA1c  in type 2 diabetes to 7 or below is not a good thing in itself, as several randomised trials have shown. And observational studies—heavily confounded by indication, admittedly—have shown much worse outcomes in people with type 2 diabetes who are given insulin. In this Australian cluster randomised trial called “Stepping Up,” the intervention was an enhanced role for the practice nurse in leading insulin initiation, plus mentoring by a registered nurse with diabetes educator credentials. As a result, more people with T2DM were started on insulin at an HbA1c level of 8.5%. Those given insulin showed a greater fall in HbA1c than those continued on oral medication. As usual, a question hangs in the air: “And then what?” Over the rest of their lives, did they gain weight, get hypos, and die sooner of cardiovascular disease?

Plant of the Week: Lonicera strophiophora
I bought this winter flowering honeysuckle about 10 years ago, on the usual grounds that I’d never heard of it and it sounded nice. I have a whole drawer full of labels from plants bought for such reasons, most of them long dead. It can’t be because total neglect comes with our small territory, can it?

Lonicera strophiophora is a disputed name, and for a time the plant dropped out of commerce entirely in the UK. There are now two suppliers, both near Pershore. It is a close relative of L fragrantissima, and if you want to know whether it has a scent, there is your clue. Ours is a tidier plant than fragrantissima, with tiers of lovely bell shaped flowers hanging from its branches. Its leaves appear shortly after flowering: they come out pink and purple, and stay purple through the season, though with an increasing element of pale green. Like all honeysuckles, it is easy to propagate and should be seen in every garden for its scent and its dainty beauty in early spring. Not to mention its tolerance of bad soil and total neglect.