Alex Nowbar’s journal reviews—22 May 2020

Alex Nowbar reviews the latest research from the top medical journals.

Annals of Internal Medicine

Caution on negative covid-19 test

If there is an upside to events of recent months, it’s that people seem to have grasped that test results are not absolute. There is an understanding that a negative PCR test does not necessarily equate to not having covid-19. Kucirka et al. study the false negative rate in relation to time of exposure and time of symptom onset using data from 1330 individuals across 7 previously published studies. After exposure but before symptom onset the false negative rate is unacceptably high – 100% on day 1 and 67% on day 4. In other words, a negative test done too early after exposure gives you no information. The probability of a false-negative result on the day of symptom onset (day 5 after exposure) was 38% (confidence interval 18% to 65% – wide!). Three days after symptom onset (day 8 after exposure) the false-negative rate was still pretty high at 20% (confidence interval of 12% to 30% – still pretty wide!). This has very serious implications for controlling spread. Hospitals have been using negative PCR to allow workers to return to work. And infective precautions have been removed from patients who test negative. Frankly, this seems irresponsible and yet it’s the guidance. A second upside to events of recent months may be that people realise that “guidance” does not necessarily equate to a rigorous evidence-based approach. Kucirka et al’s data based on time of exposure would be informative for developing contact tracing and testing protocols.


Kawasaki-like disease with covid-19

For many clinicians, paediatric conditions were merely a footnote in the cacophony of material absorbed at medical school. Kawasaki disease was one of the exotic-sounding conditions that always stood out though because of its multiorgan manifestations and limited treatment options. Verdoni et al identified 10 children with Kawasaki-like presentations at a hospital in Bergamo, Italy between February and April 2020. They compared the features with that of 19 patients with Kawasaki disease in the previous 5 years. For starters, this represented a 30-fold rise in the incidence of such cases during the covid-19 pandemic compared to previously. If that wasn’t alarming enough, 6 of the 10 patients in this case series had cardiac involvement and five had Kawasaki disease shock syndrome (KDSS) i.e. they appeared to be having a severe form of Kawasaki disease. Cardiac involvement and KDSS barely featured in the historical comparator group of 19 patients. More cases of Kawasaki-like disease are to be expected so it is useful to have documentation that there was intravenous immunoglobulin resistance and that all were treated with adjunctive steroids.

Predicting the death toll

Covid-19 has and is going to cause more deaths than would otherwise have happened. Banerjee et al have used a large UK dataset of people over 30 with their ages, gender and underlying conditions to estimate excess deaths. Of the almost 4 million individuals, 20% had a least one risk factor such as age over 70, cardiovascular disease or COPD. They modelled 4 scenarios, full suppression of the virus, partial suppression, “mitigation” and a “do nothing” scenario, each at 3 different relative risks. This produced 12 estimates ranging from 2 excess deaths if there was full suppression and a relative risk of 1.5, to over 500,000 excess deaths in a “do nothing” scenario and a relative risk of 3. Yes, that sounds about as good as asking someone in a pub how many excess deaths there will be. Not that you’d be in a pub lately. However Banerjee et al’s work highlights that tens to hundreds of thousands of lives hang in the balance of how much the virus is suppression through prevention strategy. I hope policymakers make good use of this elegant study and accompanying online calculator.


Organs open for COVID

Which organs does SARS-CoV-2 like to attach to other than the lungs? Puelles et al analysed tissue samples in their autopsy series. They found a high viral load in the lungs, phaynx, kidneys, liver and heart, and to a lesser extent in the brain and the blood. They delved further into the kidneys by looking in all kidney compartments in 6 patients. In 3 of these patients they detected the virus in all kidney compartments but noticed preferential targeting of the glomerular cells. These data cannot be described as game-changing but the best research doesn’t have to be. The best research, in my humble opinion, is research that answers a question using appropriate methodology which is what Puelles et al have done.

Cat-to-cat transmission

We already know that animals can get coronaviruses. First, Halfmann et al infect three fellow cats with SARS-CoV-2. Then they cohouse each cat with an uninfected cat. Within five days all the cats end up with the virus but without visible symptoms. It is worth noting that this study did not include humans. Therefore it is not known how likely it is for the virus to be transmitted from humans to cats or from cats to humans. However, it would be prudent for cat-owners to take precautions. Of course, one doesn’t want one’s cat to get the infection because it could make the cat sick but more importantly, you may need to include your cat in the group of individuals you isolate or socially distance from.

The ACE link

You will have heard, at the very least, whispers about the role of the human enzyme ACE2 in COVID-19. (To sound clued up you can say “coronaviruses engage the ACE2 receptor for cell entry”.) Then, you will recall, there was a frenzy about the safety of ACE-inhibitors especially as people with hypertension were more likely to die from covid-19. Chung et al compared influenza rates between those who were and weren’t prescribed an ACE-inhibitor using a national UK dataset. They found no difference or perhaps even lower rates of influenza in those prescribed an ACE-inhibitor with a hazard ratio of 0.66. No, they weren’t mad talking about a completely different virus; Influenza A has been shown to use the ACE2 receptor to mediate lung damage. This study has all the limitations of observational studies, namely confounding, but it indirectly supports the idea that being on an ACE-inhibitor doesn’t increase the risk of covid-19.

Alex Nowbar is a clinical research fellow at Imperial College London, UK.

Competing interests: None declared