Alex Nowbar reviews the latest research from the top medical journals

The biggest risk factor for death from covid-19

Tartof et al conducted a retrospective cohort study of almost 7000 people diagnosed with covid-19 treated in the Kaiser Permanente Southern California healthcare system. They confirmed a clear relation between death and body mass index (BMI): risk of death was four times higher in the people with a BMI over 45 compared with those with a normal BMI. Increasing age was also associated with death, and having had an organ transplant, but not cancer. 

The weirdest finding was that hypertension, cardiovascular disease, and black, Asian, and Hispanic ethnicities weren’t associated with death. There are countless papers about these risk factors and mechanisms of harm, and yet this study found that it didn’t matter that much. What is the truth? Was obesity always the biggest risk factor, thus confounding the relationships previously reported? Is this Californian population intrinsically different, or did they receive systematically different healthcare that mitigated the effects of the risk factors?

Ann Intern Med doi:10.7326/M20-3742

Positive trial for children with myopia

The BLINK trial is a US double-blind randomised controlled trial of three contact lens types in 7 to 11 year olds who were followed up for three years. The three types were commercially available high add power multifocal lens, medium add power multifocal lens, and single vision lens (non-multifocal). The idea behind using high add power is that increasing the light to the retina may slow the progression of myopia. The high add power multifocal lens slowed the rate of myopia progression compared with both other groups. The effect size wasn’t large enough to be considered a clinically meaningful slowing of myopia according to an FDA workshop, but the concept of clinically meaningful is disputable. This is a well conducted trial showing a clear advantage with the high add power lens in a dose-response manner. The authors raise the interesting point that “the dose-response result exhibited in this study only examined up to a +2.50 D add power. Speculation remains about whether add powers outside of the standard range may provide better myopia control.”

JAMA doi:10.1001/jama.2020.10834

Alternative masks

Sickbert-Bennett and colleagues assessed the filtration efficiency of 29 masks that aren’t new in-date N95 respirators, including used sterilised ones (with a variety of sterilisation methods), expired ones, and other non-approved masks. They tested the masks on volunteers using a standard US fit testing protocol with sodium chloride. Impressively, the fitted filtration efficiency was greater than 95 for N95s up to 11 years after their expiration date and for used N95s sterilised with ethylene oxide or vaporised hydrogen peroxide. The other non-approved masks did not reach this threshold of filtration efficiency. Interestingly, masks with ties consistently outperformed those with ear loops. This study has immediate practical implications for areas affected by mask shortages but has one major limitation. Testing was performed on only two subjects, one man and one woman. The masks might perform less well on other individuals. 

JAMA Intern Med doi:10.1001/jamainternmed.2020.4221

Viral shedding in covid-19

One of the disturbing features of covid-19 is asymptomatic spread. Lee and colleagues analysed 303 patients isolated in a centre in the Republic of Korea in March 2020. Polymerase chain reaction testing was being used to determine when they could be released from isolation. The cohort were mainly female, in their 20s, and 64% were symptomatic. A fifth of asymptomatic people later developed symptoms. Viral load was similar in symptomatic and asymptomatic people, which supports the idea of searching for and isolating asymptomatic people. Knowing the viral load is a surrogate for knowing how transmissible the virus is, but the threshold isn’t known, and we certainly won’t get any closer to knowing using serial testing of cohorts without their contacts. I hope that the legacy of covid-19 research will be something more than a lesson about research waste.

JAMA Intern Med doi:10.1001/jamainternmed.2020.3862

Trends in lung cancer

In a large US registry, both incidence of and mortality from lung cancer are decreasing, and improvements have sped up in more recent years. This is great news, and, of course, we are interested in how this has been achieved. Lower rates of smoking are probably important. However, the authors’ primary conclusion is that advances in targeted treatment may be responsible for improvements in mortality in non-small cell lung cancer.

Of course, that is possible—that’s what the treatments were designed to do after all. Yet the authors’ overstate the certainty of this explanation, dismissing the idea that mortality improvements could be due to diagnoses being made at an earlier stage with two arguments. First, they say stages have moved from unknown to later stages rather than from later to earlier with a supporting graph in the supplementary appendix. This is the kind of argument that sounds correct but when you examine the graph you realise it doesn’t add up. While the graph shows a decrease in the unknown stages, it also shows an increase in stage IA diagnoses and a decrease in stage II and III.

Second they say that although lung cancer screening has been recommended since 2014, the uptake was low. Again, this argument sounds correct but a significant proportion of diagnoses aren’t made via screening. So the claim that targeted therapies are saving people from cancer deaths should be viewed with caution. To draw conclusions about the impact of these therapies, we should simply look to the randomised trials rather than to these trends.

N Engl J Med doi:10.1056/NEJMoa1916623

Alex Nowbar is a clinical research fellow at Imperial College London

Competing interests: None declared.