Alex Nowbar’s journal review—8 January 2021

Alex Nowbar reviews the latest research from the top medical journals

Vaccine PR

Vaccination is the phrase on everyone’s lips these days. Szilagyi et al’s analysis of the Understanding America Study is a laudable effort to get a handle on changes in public opinion in the US from April to December 2020 on the likelihood of having a vaccine. Public opinion on taking up the vaccine is critical as high uptake is thought to be the only way out of the pandemic. Thousands of people completed the survey biweekly, but the survey was not asking the same people each time. It asked “How likely are you to get vaccinated for coronavirus once a vaccine is available to the public?” There was a disappointing reduction in the proportion of people self-reporting that they were “somewhat” or “very likely” to get the vaccine from 74% in April to 56% in December. The authors conclude “Educational campaigns to raise the public’s willingness to consider covid-19 vaccination are needed.” Public trust is complex. This is not going to be easy, especially since vaccine claims are necessarily tempered by the short term nature of the evidence.

Inpatient hypertension management

In hospital I think it’s fair to say we overmeasure blood pressure in a lot of people for the sake of detecting the blood pressure changes that matter clinically. In and of itself this isn’t bad (unless you’re the poor person having their arm painfully squeezed for no reason). Outside of situations where it is immediately relevant, like after a stroke, what we do with high blood pressure values in hospital is a subject of great debate. We know in general that treating hypertension improves cardiovascular outcomes. But this is practiced in primary care. In hospital inpatients, the significance of adjusting anti-hypertensives (by what is essentially guesswork, because you don’t have any home blood pressure or ambulatory blood pressure readings) is unknown. Rastogi et al compare outcomes (myocardial injury, stroke and acute kidney injury) in those who had conservative versus intensive management of inpatient hypertension in an observational study of over 20,000 US adults admitted with non-cardiac diagnoses. Before I even look at the results, I think they will be pretty meaningless because of the study design. When I say meaningless, I mean you can’t base treatment decisions on observational data because the correlation does not imply causation. But since we are here, I may as well supply you with the results: intensification of antihypertensive treatment when there was no evidence of end-organ damage was associated with worse outcomes. Interestingly intensification at discharge was not associated with better blood pressure control in the following year. And as I hope you can tell, this neither means you should intensify or not intensify. Only randomised studies would be helpful. But the way we do randomised studies at the moment is too cumbersome to study this question. 

How deadly is covid-19?

I will answer this in my capacity as spokesperson for the confused-yet-opinionated: quite deadly. Deadly enough for the chaos to be warranted. But people want a lot more detail than this. In particular, people are interested in whether covid has become less deadly. People love to speculate that the virus itself is different now, the people it is infecting are different, and/or infected people are receiving better medical care. We really don’t know any of this with any certainty. Asch et al tracked hospital death rates in those admitted with covid-19 in US hospitals based on a large insurance claims database (a broad sample although of course it won’t represent the totality of hospital admissions). Actually they gathered 30-day death rates coupled with rates of hospice referral. This makes rates more comparable across hospitals with different practices concerning timing and thresholds for hospice referral. The mean hospital-level risk-standardised rate of death or hospice referral was 11%. In other words, 1 in 10 people admitted to hospital with covid-19 die or are referred to a hospice. This was more likely in those over 85 and those who had been admitted from a nursing facility and those that had metastatic cancer. The good news is that they found a decrease in death rates between January and June 2020. They found that this decrease correlated with the hospital’s local community case rate. I can’t think of a plausible explanation for that. It is important to note that these data apply to people who got admitted to hospital with covid-19. If you really wanted to know the mortality rate you’d need the denominator to be all people infected with covid, not just those admitted, and not even just those who test positive because there are people with covid who haven’t been tested. 

Another flop for monoclonal antibody for covid

It was the stuff PR dreams are made of. A drug derived from the serum of a covid-19 survivor, Bamlanivimab (also known as LY-CoV555) looked to improve clinical outcomes. I’m sure the name was derived in some clever way too, but to me it sounds like something out of the Flinstones. Anyway, it made no difference to outcomes of people hospitalised with covid-19 based on an ordinal outcomes scale in a randomised placebo-controlled trial.

A little piece of the covid reinfection puzzle

You will often read me denigrating observational research but it does have an important place in epidemiological research. For example it is how we estimate disease prevalence and incidence. Lumley et al’s cohort study of over 12,000 UK healthcare workers addresses an incidence question but also an immunological one. The question was what are the re-infection rates in seropositive healthcare workers compared to healthcare workers without antibodies. In other words, does seropositivity (having positive anti-spike antibodies) provide protection against re-infection? In short, this study suggests it does. The rate of PCR-positive infection was 0.13 per 10,000 days at risk in people who were seropositive (and these infections were asymptomatic) compared to 1.09 infections per 10,000 days at risk in those who were seronegative (and more than half of these infections were symptomatic). The appearance of protection against re-infection is good news but whether this is mediated by the presences of antibodies isn’t clear. There is potential for confounding by behavioural differences between those with and without antibodies. Also, the participants were all working-age adults.

Alex Nowbar is a clinical research fellow at Imperial College London