NEJM 3 Dec 2015 Vol 373
CPR with or without breaks
2203 Cardiopulmonary resuscitation for cardiac arrest outside hospital carries a mortality risk of more than 90%. It is usually assumed that the death rate would otherwise be 100%, but trials of CPR don’t ever include a non-intervention arm, so we don’t really know. In this huge US trial, emergency medical service groups were cluster randomised to give either continuous chest compressions without breaks for ventilation, or to have a ventilation break after every 30 compressions. There was no significant difference in outcome: 9-9.7% of patients survived to be discharged from hospital and 7-7.7% had “favourable neurologic function” at discharge.

Paracetamol and fever outcomes
2215 A double-blinded trial of paracetamol (acetaminophen) in ICU patients with fever raises the old question of whether fever helps you fight off serious infection and therefore if suppressing it is a bad idea. In these very sick febrile patients across Australia and New Zealand, paracetamol or placebo given as a regular intravenous bolus until ICU discharge resulted in identical rates of mortality or days spent on the unit. When I get my first winter bug, usually timed for 24 December, I shall take paracetamol. And, even in the absence of RCT evidence, probably some whisky too.

HIV drugs protect
2237 “We conducted a double-blind, randomised trial of antiretroviral therapy for pre-exposure HIV-1 prophylaxis among men who have unprotected anal sex with men.” I can picture some public health physicians spluttering into their salt-free muesli. Isn’t this encouraging risk-taking behaviour? Wouldn’t it be better to follow them round with bags of free condoms? Doesn’t this “normalise” unprotected anal sex? But the worldly wise French physicians proceeded with their trial, which, unsurprisingly, shows that triple therapy largely prevents HIV infection whereas placebo does not. It was probably ethical to do this trial, but I think it is now unethical not to give HIV drugs for the prevention of infection to anyone who wants them.

Diabetes drug doesn’t reduce CV events
2247 Here’s a good example of how to test a new (actually me-too) drug for type 2 diabetes according to current standards. Take a large group of people with T2DM who have had a myocardial infarction or who have been hospitalised for unstable angina within the previous 180 days. Give them your drug or placebo, plus whatever else their doctors decide. Count cardiovascular events after a mean of 25 months. Hey presto, the numbers are the same in both groups: hence your drug does nothing but reduce sugar levels a tiny bit and can be put on the market. In my opinion, it isn’t even worth mentioning its name, but perhaps worth mentioning that the trial was funded by Sanofi.

JAMA 1 Dec 2015 Vol 314
Metformin in T1DM
2241 Metformin is an old drug with an untarnished reputation. It was probably worth doing a trial to see if it improved glucose control in overweight or obese adolescents with type 1 diabetes, even though it didn’t, over a six month period. In some of the kids, it allowed a reduction in insulin dose, and in 24% of them it led to a major reduction in BMI, compared with 7% of those given placebo. Inevitably, it caused more gastrointestinal adverse effects. While the trial doesn’t define a universal role for metformin in this group, it certainly doesn’t rule out benefit in a substantial number.

Another new drug does nothing
2251 The easiest way to test a new drug for heart failure is to see what it does to levels of B-type natriuretic peptide. Measured in a single individual, readings will shoot around all over the place, but measured in a group, their mean level will give you some idea if your agent is helping to relieve ventricular strain. Vericiguat is a soluble guanylate cyclase stimulator, which theoretically might help to stabilise worsening heart failure. “Among patients with worsening chronic HF and reduced LVEF, compared with placebo, vericiguat did not have a statistically significant effect on change in NT-proBNP level at 12 weeks but was well tolerated. Further clinical trials of vericiguat based on the dose-response relationship in this study are needed to determine the potential role of this drug for patients with worsening chronic HF.” I wouldn’t hold your breath: this initial study, called SOCRATES-REDUCED, recruited 456 patients across three continents and managed to lose nearly a quarter of them over 12 weeks, all for no discernible effect. If I were Socrates, I might be asking some questions at this point.

Hail, 19% Caesar
2263 I am never quite sure what the World Health Organization is really for, but setting aside the enormous amount of practical good it might do but doesn’t, at least it seems good at issuing statements about things it feels should happen. The WHO recommends that rates of delivery by caesarean section should not exceed 10 to 15 per 100 live births to optimise maternal and neonatal outcomes. This useful study (with Atul Gawande among the authors) seeks to estimate annual caesarean delivery rates from data collected during 2005 to 2012 for all 194 WHO member states and compare these with maternal and neonatal outcomes. There was a fair amount of missing data, but the bottom line is that national caesarean delivery rates of up to approximately 19 per 100 live births were associated with lower maternal or neonatal mortality among WHO member states.

Ann Intern Med 1 Dec 2015 Vol 163
Big tummy + thin limbs = big risk
827 There are two main ways to assess “obesity”: the body mass index and the waist to hip ratio. The commonest cause for an elevated WHR is central fat deposition (“central obesity”), though I guess there can be other causes such as ascites or organomegaly. Every so often, the National Health and Nutrition Examination Survey picks out a cohort of typical US citizens and measures such things for a number of years. Looking at data from 15 184 adults (52.3% women) aged 18 to 90 years in NHANES III, investigators confirm the paradox that having a normal BMI but an elevated WHR confers a worse prognosis than being obese by the criterion of BMI alone. And this is no borderline effect: compared with simple obesity, it doubled the mortality risk in men and increased it by half in women over the 14 year follow-up period. Most of this added risk is due to cardiovascular causes.

HCV and Gilead: more trials
818 While the three Gilead funded trials of sofosbuvir with velpatasvir for hepatitis C that I commented on last week still languish on the NEJM website, two others now appear in print in the Annals. They are different but tell the same story: time is up for HCV infection once these drugs become affordable. I’ve seen various tweets indicating that Gilead is squirming a bit under political pressure to reduce its $80K+ price tag for this product, but so far has decided to brazen it out. Ah well, that’s what happens when the primary aim of developing vital drugs is to bring in profits for the shareholders. People with end-stage HCV related disease will die, but they are only following company policy. It would be nice to think the world could come up with an alternative model.

Lancet 5 Dec 2015 Vol 386
There is enough mega-epidemiology in this week’s printed Lancet to keep you going through a couple of blustery winter afternoons, but it all appeared online three months ago and I don’t intend to comment on it again. I’m pretty up to date with the website too, though there is a new observational childbirth study worth telling you about:
Stillbirth and obesity in Sweden
OL More than 30% of women in Sweden are either overweight (body mass index [BMI] 25 kg/m2 to <30 kg/m2) or obese (BMI ≥30 kg/m2) during early pregnancy. I must confess that I have never visited Sweden, so I get my Swedish stereotypes from Wallander and a tweet from Trish Greenhalgh earlier this year, asking why everybody around her is slim and beautiful. Clearly, we have got this wrong. This whole population study looks at changes in BMI in relation to stillbirth and infant mortality. “Compared with women with a stable BMI (change between −1 kg/m2 and <1 kg/m2) between pregnancies, the adjusted RRs for women who gained at least 4 BMI units between pregnancies were 1.55 (95% CI 1.23–1.96) for stillbirth and 1.29 (1.00–1.67) for infant mortality. Stillbirth risks increased linearly with increased BMI gain.” So here we have a potentially modifiable risk factor, although I’m not aware that anyone has an effective intervention for it. In my stereotype version of Sweden, small portions of fermented fish on crispbread would be delivered by weary members of the police force to every pregnant woman, leading to inappropriate emotional attachments and the uncovering of people smuggling drug dealers who kill babies.

The BMJ 5 Dec 2015 Vol 351
Hip pain doesn’t show on X-rays
Every doctor quickly discovers the fact that the amount of pain in a hip does not bear much relationship to changes on an X-ray. Now the Framingham study has come along and quantified it. Actually, the data come from two separate Framingham cohorts, neither of which is the famous Heart Study. These are called the Framingham Osteoarthritis Study—limited to the town in Massachusetts—and the Framingham Osteoarthritis Initiative, sampling the whole of the United States. In the Study, only 15.6% of hips in patients with frequent hip pain showed radiographic evidence of hip osteoarthritis, and 20.7% of hips with radiographic hip osteoarthritis were frequently painful. In the Initiative, only 9.1% of hips in patients with frequent pain showed radiographic hip osteoarthritis, and 23.8% of hips with radiographic hip osteoarthritis were frequently painful. I guess the main question is to what extent radiographic changes predict resolution of pain by total hip replacement. Somehow or other this surgery usually seems to work.

Nausea and vomiting in palliative care
I learnt “terminal care” as it was known in the early 1980s at the bedside with Robert Twycross in Oxford, who had a burning intellectual passion to make symptom relief as complete and scientifically based as possible. It is a year now since I last treated a dying patient with nausea and vomiting. The drugs and delivery methods have changed a great deal over 30 years, but the principles remain the same, and they are well outlined in this review. The evidence base, alas, remains poor, because drugs are scarcely ever tested in a real palliative care context. Proving something makes a difference in a small placebo-controlled trial really tells one nothing about its usefulness in the general armoury of treatments across the wide range of causes in dying patients. Our mainstay, especially in gastric stasis, used to be metoclopramide, but its use is now discouraged because of the risk of adverse effects. I guess that ondansetron and its analogues are a better choice, though old habits die hard. This is a changing area that every generalist doctor needs to keep up to date with.

Plant of the Week: Erigeron karvinskianus

This is the time of year when all the smaller plants look dead or dormant among drifts of leaves in the mud. Apart from one or two hellebores, which are trying to offer a premature flower, the main exception in our garden is the Mexican daisy.

Although it’s common and easy, I always think of this as a “posh” plant because it used to flourish in the cottage gardens of my well spoken country lady patients, all now long departed. I never asked for a piece, and I was never given one, probably because they assumed that everyone had it. I actually had to go through the indignity of buying this erigeron, from the extremely posh and amiable owner of Kiftsgate House.

I don’t know how long our big clump of daisy flowers will go on through the winter, but they’ve come through the first frosts unscathed and undiminished. I had hoped that little daughter plants would fill our front garden by now, but it’s proving slow to disperse. Some time soon I will give it a bit of help by dividing it. But the real joy will come when it simply appears out of a crack in a stone wall—its best habitat by far, as can be seen in every village in the Cotswolds.