![\"richard_lehman\"](\"https:\/\/blogs.bmj.com\/bmj\/files\/2014\/01\/richard_lehman.jpg\")
NEJM\u00a0 <\/em><\/strong>8 June 2017\u00a0 Vol 376<\/strong>\nDissecting nodes in melanoma<\/u>
\nRead carefully now, because what I am about to tell you may seem odd. Say you are diagnosed with malignant melanoma, and the sentinel node is positive for cancer. What’s your first thought? Perhaps “Help, I’m going to die. Take my other nodes out while I draw up my will.” But all that is completely wrong. In this randomised trial<\/a>, disease specific survival in such patients at three years was 86% if they underwent immediate lymph node dissection and 86% if they did not. Those who had all their nodes removed had a 24% rate of lymphoedema compared with 6% in the observation group. For further enlightenment, read the editorial, which begins<\/a>: “Because of an incomplete understanding of regional lymph nodes, they have been characterized either as active filters or barriers to the distant spread of disease or as passive indicators of the risk of distant spread. The truth almost certainly resides in a hybrid model that embraces aspects of both theories. This lack of understanding has frequently led to a ‘more is better’ philosophy in the surgical management of regional lymph nodes. Yet in melanoma, as in so many other cancers, the elective removal of clinically negative nodes has rarely, if ever, been shown to increase disease specific survival.”<\/p>\n
Lithium in pregnancy and cardiac malformations<\/u> Data sharing and journal editors: sauter pour mieux reculer<\/em><\/u> Meanwhile, the editors who took a bold leap of faith last year are now moving back a step or two<\/a>, ostensibly because nobody has yet sufficiently solved the practical issues of sharing trial datasets. For heaven’s sake, I know I’m biased as a member of YODA and AllTrials, but it’s long past time that medicine got its act together on this.<\/p>\n Small breast cancers and overdiagnosis<\/u> SLGT2 inhibitors<\/u> JAMA\u00a0<\/em><\/strong>6 June 2017\u00a0 Vol 317<\/strong> Desirable weight in pregnancy<\/u> JAMA Intern Med\u00a0 <\/em><\/strong>June 2017<\/strong> Persistent pain and cognitive decline<\/u> Glucose strips in orally treated T2DM: America catches up<\/u> The Lancet\u00a0 <\/em><\/strong>10 June 2017\u00a0 Vol 389<\/strong> Brentuximab<\/u> The BMJ\u00a0<\/em><\/strong>10 June 2017\u00a0 Vol 357<\/strong>
\nLithium taken in pregnancy has been associated with cardiac abnormalities in the offspring, especially right ventricular outflow tract obstruction defects. In a massive study of over a million American births covered by Medicaid<\/a>, the rate of these was 0.60% among lithium exposed infants versus 0.18% among unexposed infants (adjusted risk ratio, 2.66; 95% CI, 1.00 to 7.06). So there almost certainly is an added risk, but notice how the 95% confidence interval includes unity.<\/p>\n
\nIt seems an age ago, but it was just January last year that the editors of the leading medical journals signed up to a bold declaration of intent to insist that authors should be prepared to share their full datasets with others. NEJM’s <\/em>editor\u00a0then began to show intense signs of unease about the statement he had just lent his signature to. This hasn’t really abated since, as shown by two contrasting perspective pieces this week. One is by a NEJM<\/em> writer<\/a> and quotes mainly from people who don’t accept the need to share data, except among themselves. The other is from a patient<\/a> who can’t believe that researchers don’t welcome the fullest possible use of information provided by patients who volunteer to undergo the risk of testing new treatments.<\/p>\n
\n“Are Small Breast Cancers Good because They Are Small or Small because They Are Good?<\/a>” Did this paper get into NEJM<\/em>\u00a0because it has such a good title, or was the paper so good that the NEJM<\/em> came up with the cool title? Or did the NEJM<\/em> come up with the cool title and then commission the special article? You can see the problem. It’s the same with cancers. We know that mammographic screening detects mostly lesions that would never grow and spread, but we also know that breast cancer can be considered as a systemic disease from the start. What we don’t know is how to handle this knowledge, and all the while it’s becoming clear that there are at least a dozen kinds of breast cancer (or is it a hundred?) with different biological and prognostic features. This paper gives a few clues from present knowledge of the relationship between cancer genotypes and size. But believe me, this is an area\u00a0where the data change week by week and interpretation struggles to keep pace.<\/p>\n
\nI’m delighted that drugs have been discovered that actually improve cardiovascular outcomes in type 2 diabetes. Inhibitors of sodium\u2013glucose cotransporter 2 (SGLT2) observationally do this as a class effect, and definitely did so (unexpectedly) in a major randomised trial of empagliflozin. But from a letter in this week’s <\/a>NEJM\u00a0<\/em>it seems that a certain amount of caution is needed. Compared with another class of drugs used as second line treatment, the dipeptidyl peptidase-4 inhibitors, SGLT2 inhibitors are associated with twice the risk of diabetic ketoacidosis. The detail of this needs more than a letter can contain, but it’s clear that these promising drugs need a watchful eye.<\/p>\n
\nHome treatment following severe COPD episodes<\/u>
\nThese 116 British patients with chronic obstructive pulmonary disease\u00a0were sick people. They had persistent hypercapnia, and of the 59 randomised to home oxygen alone, 19 were dead within a year. The other 57 received non-invasive ventilation in addition to oxygen, and of those 16 died. The main difference was in time to hospital readmission: 1.4 months versus 4.3 months<\/a>.<\/p>\n
\nThis survey of 1.3 million American pregnancies gives me the opportunity to tell you belatedly that the US Institute of Medicine (IOM) is now called the National Academy of Medicine (NAM), and has been for over two years. When it was still the IOM, the luminaries who sat there decided on the desirable range of weight gain for women during pregnancy. Roughly speaking, they seem to have got it right. Women who put on too much weight (47% of the sample) have bigger babies and more caesarean sections<\/a>, while women who put on too little tend to have smaller babies and a higher risk of premature birth.<\/p>\n
\nSalt in American food<\/u>
\nI don’t really care how much sodium there is in food, provided it doesn’t taste too salty. American food always tastes too salty. This article tells us that packaged American food is less salty now than in 2000<\/a>,\u00a0which still can’t possibly bring American packaged food to the level of edible. For that they would need to tackle the real problem both for taste and health: corn syrup. And most of the other stuff in American packaged food.<\/p>\n
\nHere’s a sad study of 10 067 Americans aged between 67 and 78<\/a>, of whom 11% suffered from persistent pain. These people showed lower baseline score for activity and higher scores for depression at baseline. And after 10 years, they also showed a higher rate of dementia.<\/p>\n
\nTo celebrate the tenth anniversary of Andrew Farmer’s landmark BMJ<\/em> paper<\/a> showing that self-monitoring of blood glucose is a waste of time for Britons with type 2 diabetes who are not on insulin, JAMA IM <\/em>publishes this North Carolina study<\/a>, which shows that the same is true for Americans.<\/p>\n
\nTildrakizumab<\/u>
\n“Tildrakizumab is a high-affinity, humanised, IgG1 \u03ba antibody targeting interleukin 23 p19 that represents an evolving treatment strategy in chronic plaque psoriasis.” For “evolving treatment strategy” please read “drug that Merck is trying to sell in a highly competitive market for new mabs.”\u00a0In two phase 3 trials, funded by Merck, it performs well against placebo<\/a>. Against etanercept, the difference is not great: it’s expressed as a difference of percentages achieving changes in psoriasis area and severity index (PASI) points and the detail is complicated and attached to a series of clinically meaningless P-values. The big question is how much value this mab adds to the treatments that\u00a0are already standard. I have absolutely no idea.<\/p>\n
\nFrom one of the commonest skin conditions to one of the rarest: cutaneous CD30-positive-cell lymphoma<\/a>. And for every disease there is now a mab that somebody wants to sell: this one rejoices in the name of brentuximab vedotin, made and trialled by a subsidiary of Takeda Pharmaceuticals. But enough of this; I fear I am losing my audience. The Lancet <\/em>may have seen fit to publish this trial, but for most of us, brentuximab ved\u00f3tin is hardly worth the n\u00f3tin’.<\/p>\n
\nReporting of RCTs: from poor to middling<\/u>
\nAt a rough estimate, I have read the abstracts of over 2000 Cochrane reviews in the past three years or so. It’s quite a reality check if you want to believe in evidence based medicine. The authors of this study<\/a>\u00a0went through about the same number of Cochrane full reviews published between 2011 and 2014, and found that they took their data from 20\u2009920 randomised controlled trials. Then they analysed the reporting of all these trials. Yes, they looked at 20 920 RCTs and they have lived to tell the tale. Over this relatively brief time span, they were able to detect an overall improvement in the reporting of summary data from RCTs, which offers a glimmer of hope. But for many, especially if reported in low impact factor journals, the standard remains dire.<\/p>\n