Alex Nowbar reviews the latest research from the top medical journals

RECOVERY: the landmark that was

When I first heard of the RECOVERY trial in March this year I thought that the original list of therapies that were being tested—lopinavir-ritonavir, interferon-beta, and steroids (hydroxychloroquine and tocilizumab arms were added later)—were unlikely to be of any use. By April though, people’s heads were turned by the sheer efficiency of recruitment to this randomised trial of those hospitalised with covid-19. And by June there were results. Dexamethasone was found to reduce mortality, albeit only in advanced disease (those requiring oxygen). We also had answers for hydroxychloroquine and lopinavir-ritonavir—that is, no benefit. I usually dislike the concept of a landmark trial, but, for once, this description is justified. When will society (medical staff included) wake up to the idea that people should only receive a new therapy for any condition (or a therapy without prior randomised evidence) in the context of randomised trials such as this so the therapy gets properly evaluated. Belief in therapies based on anecdotes and mechanistic and observational studies, however earnest, is not good enough. If RECOVERY can do a swift, large scale, randomised trial, why can’t we apply the model to the rest of medicine? The resources, the willing, and the coordination; it’s all there.

N Engl J Med doi:10.1056/NEJMoa2021436

The if and when of renal replacement therapy for acute kidney injury

Intensive care and renal physicians have always given me the impression that we should try to hold off renal replacement therapy. But physicians often wonder if pre-emptive renal replacement could be better for a person physiologically. To tackle the question of the timing of initiating renal replacement therapy, the STARRT-AKI trial randomised over 3000 people in 68 countries to either accelerated or standard delivery, looking at 90-day death rates. Accelerated delivery meant starting it before complications of acute kidney injury such as hyperkalaemia and acidosis developed (within 12 hours of randomisation). In the standard group, clinicians were discouraged from starting renal replacement therapy unless complications developed and there was no obligation to start it. There was no difference in mortality between the two groups, and it was advantageous to be in the standard group, because they were less likely to start renal replacement (61.8% v 96.8% in the accelerated group), become dependent on it, or have adverse events. This is crucial evidence to support everyday clinical decision making in intensive care.

N Engl J Med doi:10.1056/NEJMoa2000741

The age of the aged is coming

Vollset and colleagues modelled fertility, mortality, and migration rates to project population changes across the world by 2100. They anticipate declines in fertility rates due to increased female educational attainment and access to contraception, and slow population growth. They estimate the population will peak in 2064 at 9.73 billion and decline to 8.79 billion in 2100. At first you might think, “What’s the big deal? The population is too big anyway.” The issue is the age distribution of the population. If new people aren’t being born then the population will be dominated by older people. This could be unsustainable economically. But I resent the implication (in the media reaction to this research) that women having fewer children is a problem to be solved. Perhaps this is an argument only people who don’t do modelling make, but surely if an aging population is not sustainable then it just won’t be sustained. We might see life expectancy stop increasing because the economy can’t support older people without enough young people. Lead author, Christopher Murray, says: “We’ll have to reorganise societies.” Yes, probably.

Lancet doi:10.1016/S0140-6736(20)30677-2

Universal covid-19 testing at care homes

Bigelow and colleagues tested 893 residents who hadn’t been tested (universal testing) at 11 long term care facilities in Maryland, USA, that had been using a symptom-based testing strategy and had had positive cases. Unsurprisingly, this detected additional cases. In fact, 40% of these residents tested positive. The researchers had two-week follow-up data for a subgroup of residents at seven of the care homes, and, of those who were found to be positive through the universal testing strategy, 87% were asymptomatic. These data are striking support for a universal testing strategy. However, identification of cases is only the beginning. It then has to be supported with subsequent measures to control spread, such as tracing contacts.

JAMA Intern Med doi:10.1001/jamainternmed.2020.3738

Coronary artery bypass graft choice

I am still in awe that surgeons can plumb a bit of artery or vein to your coronary artery. The preferred pipe is the left internal mammary artery, but the saphenous vein and radial artery are also used. Gaudino and colleagues performed an individual patient data meta-analysis of five randomised trials comparing the saphenous with the radial artery for isolated bypass grafting with long term follow-up. The radial came out on top in this group of 1036 patients with 10 years median follow-up for the composite endpoint of death, myocardial infarction, and repeat revascularisation. The radial was also superior for the composite endpoint of death and myocardial infarction. As described aptly by the authors in the limitations section, this may not reflect current practice because the procedures were performed more than a decade ago, standardised outcome definitions were lacking, and loss to follow-up was much greater in the saphenous vein group which could have introduced bias. Nevertheless, these are convincing results.

JAMA doi:10.1001/jama.2020.8228

Alex Nowbar is a clinical research fellow at Imperial College London

Competing interests: None declared