Ann Robinson’s research reviews—7 May 2020

Ann Robinson reviews the latest research from the top medical journals

Beware the second wave

In late January 2020, China implemented a widescale public health intervention to contain the spread of covid-19. By mid-March, there had only been 120 deaths reported as being covid-19 related in the whole of mainland China outside of the epicentre in Hubei province. I still find that figure hard to believe. This impact assessment attributes the low (reported) death rate to stringent public health measures. The instantaneous reproduction number (Rt), which estimates how many people one case would infect if there were no change in circumstances, decreased sharply and remained below 1, and the confirmed case-fatality risk (cCFR) outside Hubei was only 0.98% (compared with 5.9% in Hubei). The authors say that relaxing the interventions (allowing Rt>1) when the epidemic size was still small would have resulted in an exponential rise in cases and that measures should be relaxed only gradually, keeping Rt below 1. There is a real risk of a second wave “particularly from overseas importation,” says this study from Hong Kong, which urges ongoing monitoring of Rt and cCFR.

Lancet doi:10.1016/S0140-6736(20)30746-7

A tale of five unequal boroughs

Seventeen percent of all US cases of covid-19 have been in the five boroughs of New York City, each of which is demographically and socioeconomically distinct. This study found that the Bronx—with the poorest, least educated population and the most residents from ethnic minorities—had the highest rates of covid-19 related hospitalisations and deaths despite having the most hospital beds per capita. Manhattan, by contrast, had the lowest rates of hospitalisations and deaths despite being even more densely populated than the Bronx. The difference is that people who live in Manhattan are rich and mostly white. Limitations are the study design, limited follow-up, and lack of demographic detail about the patients who died. The number of cases couldn’t be evaluated because of the variability of testing. This stark contrast in death rates between boroughs may be the result of comorbidities, occupational exposure, socioeconomic status, and “race-based structural inequities” say the authors. That the poor suffer more is hardly a surprise.

JAMA doi:10.1001/jama.2020.7197

ACE and ARBs don’t increase the risk of dying of covid-19, smoking does

This US study of 8910 patients confirms that underlying cardiovascular disease is associated with an increased risk of dying among patients hospitalised with covid-19. But it didn’t find any association with angiotensin converting enzyme (ACE) inhibitors or angiotensin receptor blockers (ARBs). Mean hospital stay was 10 days, and 94% of patients survived to go home (though a quarter of those admitted to intensive care died). Being over 65 years old, having cardiovascular disease, having COPD, and being a current smoker were independent variables associated with a doubling of overall mortality to around 10%. There was no increased risk of in hospital deaths associated with use of ACE inhibitors or ARBs and no reason—from this study at least—to support stopping them.

N Engl J Med doi:10.1056/NEJMoa2007621

Close encounters of the covid kind

How transmissible is covid-19 to close contacts (and how close is “close”?) This prospective case-ascertained Taiwan study of 100 laboratory-confirmed cases of covid-19 found that only 0.7% of their 2761 close contacts (who were traced and quarantined for 14 days after contact with the index case) developed the infection themselves. The infection rate was highest (1%) if the contact took place within the first five days of symptoms and declined after that (to 0 cases), which certainly shows that quarantine Taiwan-style is effective. One problem with this study is that the quarantined contacts were tested for covid-19 only if they developed symptoms (fever, cough, or difficulty breathing); asymptomatic infection would have gone undetected. Unsurprisingly, contact with family members resulted in higher transmission rates than contact at health centres or other settings. But I found it interesting that rates of transmission between family members weren’t higher and were similar whether they lived in the same household or not (4.6%, 5.3%). It’s a small study, and Taiwan is not typical of other countries with less draconian quarantining. But the fact that the virus was most infectious before and in the early days of being symptomatic, suggests that finding and isolating symptomatic patients is unlikely to be enough to contain spread and that continuing some measures of physical distancing may be necessary.

JAMA Intern Med doi:10.1001/jamainternmed.2020.2020

Holes in ventilation policies

This study found that over half of 67 US hospitals surveyed didn’t have approved, adequate, or clear ventilator triage policies. Policies differed widely; most used similar clinical criteria to assess need and potential benefit, but important additional criteria (such as age, resource allocation, priority for health workers) and key information (such as who activates the policy and how to minimise potential bias in the triage process) were missing in many centres. All the hospitals were academic centres with bioethics departments; it’s probably fair to assume that smaller district hospitals were even less likely to have policies fit for purpose. You’d think that, after the H1N1 (“swine flu”) epidemic of 2009, contingency planning, including a decent ventilator triage policy, would be embedded in every health institution. Many hospitals are now scrambling to update or develop a policy on the hoof, but without the input from stakeholders and the public that these complex ethical issues really require.

Ann Intern Med doi:10.7326/M20-1738

Ann Robinson is an NHS GP and health writer and broadcaster