Scientists estimate fatality ratio across full spectrum of COVID-19 disease

By Dr. Liji Thomas, MDApr 19 2020

As of 19^th April 2020, there are nearly 2.4 million cases of the novel coronavirus and over  164,000 deaths. Earlier during the spread of the virus, most of the cases were reported from Wuhan, in Hubei province. However, the most significant number of cases of the virus rapidly came to be reported from outside China – in Italy, Spain, France, the UK, Korea, and now the United States.

The symptoms of the virus are similar to that of influenza and viral pneumonia – they include coughing, sneezing, a sore throat, and a feeling of tiredness (called myalgia). The cases have been classified in China in two main categories: the first is severe, those which exhibit more than 30 or more breaths per minute (tachypnoea), and oxygen saturation of 93% or less at rest. The second category is critical, those who require mechanical ventilation or develop septic shock and multi-organ failure.

According to the World Health Organization, 80% of the coronavirus cases in China up to 20^th February 2020 displayed mild-to-moderate symptoms, about 14% had severe disease, and 6% progressed to a critical stage, requiring mechanical ventilation. Those at the highest risk of severe disease were those above 60, along with those who had underlying conditions like hypertension, diabetes, chronic respiratory disease, cardiovascular disease, and cancer.

Study: Estimates of the severity of COVID-19 disease. Image Credit: shutter_o / Shutterstock

How is the severity of COVID-19 assessed?

All potentially lethal diseases have a property called the case fatality ratio. This is, simply put, the ratio of people who die from the disease to the number of people who contract the disease. While many infectious diseases like yellow fever, influenza, and the Ebola virus have a reasonably accurate case fatality ratio (or CFR), the proposed ratios for SARS-CoV-2 vary too substantially to be reliable.

There are many reasons why it’s hard to obtain a reliable CFR estimate for SARS-CoV-2. Firstly, 2-3 weeks may elapse between a case developing symptoms and the clinical conclusion of the case (death or recovery). In a spreading epidemic like COVID-19, says the team, the clinical conclusion for many reported cases is unknown. Therefore, merely dividing the cumulative number of deaths by the number of reported cases will not be accurate: many deaths are unreported, and the calculated case fatality ratio will be too low in the early part. This effect was noticed in past epidemics of respiratory diseases, including SARS and H1N1 influenza.

Another factor that makes it difficult is the fact that during the period of exponential spread of a virus, the length of the period between symptoms and outcome (recovery or death) is often censored. Estimates of this period are often biased. Moreover, during the early stages of an epidemic, surveillance is usually aimed towards the more severe cases, especially when diagnostic capacity is limited (as is the case with COVID-19). Since the measured set consists of the very ill, the fraction of fatalities is higher than it would have been with a more evenly dispersed dataset. This makes the CFR higher. Data from Wuhan has been primarily from hospitals, where the severest cases are – most coronavirus patients recover quietly at home.

Now for the denominator of this fraction: the sum of all those who have contracted the virus, including those with mild symptoms and the asymptomatic. Finding this number with any degree of accuracy requires a robust population survey of infection prevalence. The most reliable way to do this, says the team, is with serological assays: however, very few of these assays are available. To acquire their required dataset, therefore, the team used the data from polymerase chain reaction (PCR) tests done on international residents of Wuhan who were repatriated to China. PCR tests are quite sensitive and thus suitable for this study.

With both of these in mind, a team of researchers from Imperial College London has conducted a study aimed at calculating a reasonably accurate case fatality ratio and infection fatality ratio, which is based on Chinese data as well as international reports. The research is published on the pre-print server MedRxiv.

What did the researchers find?

The team concluded that the average time from onset to death in a SARS-CoV-2 case is 18.8 days, which could vary by about half a day. Given the later decline in the epidemic in China, they later revised this to a mean of about 18 days.

On another dataset of cases outside China, they calculated the onset-to-recovery period to be 22.6 days.

They also found that the highest CFR was in the elderly (those aged above 80): the CFR for this group is about 13.4, in contrast to the 1% in the group aged 50-59 years. In the cases outside China, they estimated a CFR of 2.7%. Putting all the data together, they estimate a current CFR of 1.4% in China.

Underdiagnosis of cases due to varying case definitions, surveillance methods, and health system priorities is a repeated motif both in China and outside. If this is adjusted for, the infection fatality rate is estimated at 0.66%, but this is again different in various age groups. It rises sharply from the age of 50 years and even higher in the age group above 80 years.

What does this mean for public health?

The researchers point out that while their estimated CFR is lower than that of other coronaviruses like SARS or MERS, it is much higher than that of the H1N1 influenza strain in 2009. It is also lower than the crude CFR calculated in China and other countries, primarily because it excludes mild and asymptomatic cases. It will also vary with the average age of the patient population and the frequency of underlying conditions. This, in turn, is likely to vary with the geographical region and the income levels, since the risk of co-morbidities is very different in these different socioeconomic groups.

The age-related increase in CFR is evident, but the very low rates in the below-20 group could be due to intrinsic resistance to infection, or greater resilience, or both. Serologic testing in this population is the only way to understand the true prevalence of infection in this group and its contribution to the spread of the virus.

The researchers suggest the strong probability of large epidemics of COVID-19 within many communities. They conclude: “Our estimates of the underlying IFR of this infection will inform assessments of health impacts likely to be experienced in different countries and thus decisions around appropriate mitigation policies to be adopted.”

Important Notice

medRxiv publishes preliminary scientific reports that are not peer-reviewed and, therefore, not be regarded as conclusive, guide clinical practice/health-related behavior, or treated as established information.