The COVID-19 Pandemic

Analysis and comparison of the Covid-19 genome suggest that it originated in bats and was transmitted to humans through an intermediary host, possibly a pangolin or other non-domestic animal. Bats appear to have an altered innate immune system that allows them to tolerate some viral infections without developing disease and consequently, they are the reservoir for a wide range of infections, some of which infect humans. Multiple different types of coronaviruses (>500) have been identified in bats and the coronavirus that caused SARS in 2002 almost certainly originated in bats and was transferred to humans via an intermediary civet cat. This probably came about because there is a substantial trade in China of “exotic” animals obtained from the wild; these are sold for food and traditional medicines. There is some speculation that the endangered pangolin (or scaly ant-eater) may have been the intermediary host in the current outbreak. Despite this, the actual origin of Covid-19 has not yet been identified. The closest bat coronavirus that has been identified shares about 96% of its genome with Covid-19 while the most recent report suggests that the closest coronavirus found in pangolins only shares 85 to 92% of the genetic material. (https://www.nature.com/articles/d41586-020-00548-w)

COVID-19 is a respiratory virus (like the common cold) which is spread through respiratory droplets, meaning drops of fluid from the nose or mouth that are emitted during coughs, sneezes or even talking. It is possible that some of the viral particles emitted this way end up on surfaces (door handles, subway poles, coins) where they might remain viable. These objects then become “fomites,” inanimate objects that can transfer infection between people. It is also possible that COVID-19 can be transmitted as an aerosol—in other words, through the airborne route, so through direct inhalation of virus suspended in the air—but so far, there is no conclusive evidence of that (although that is challenging to prove one way or the other). The virus has also been identified in stool and less often, in other body fluids (blood, urine) raising the possibility that other routes of transmission—such as fecal-oral—are possible although it is not clear if that has contributed to the outbreak.

 

COVID-19 disease usually begins with mild fever, dry cough, sore throat and malaise. Unlike the coronavirus infections that cause the common cold, it is not usually associated with a runny nose. In the early phase of the disease, illness is usually mild and most often, meaning in about 80% of cases, it remains mild and may not require direct medical intervention. About 14% of people develop severe pneumonia accompanied by hypoxia (poor oxygenation) and 5% are considered critical, meaning they experience respiratory failure requiring mechanical ventilation. Although we know that older people and those with cardiovascular disease or Diabetes mellitus are at especially high risk for severe disease, it is not yet clear why these people experience these outcomes. One theory is that these people have an altered immune response that is not self-regulating. Whereas normal inflammatory responses to pathogens are eventually dampened by other immune responses, this down-regulation may not occur in some people. This could lead to the overproduction of certain immune cells and the massive release of inflammatory cytokines (cytokine storm) that is typical of acute respiratory distress syndrome (ARDS).

A study in the Journal of the American Medical Association from China’s CDC reported that the case fatality rate (CFR) in approximately 45000 patients with confirmed COVID-19 in mainland China was 2.3%. The CFR varies by age and underlying health status, and ranges from 0 in patients 9 years of age and under to 14.8% in those 80 and older. The CFR is high in people with underlying cardiovascular disease (10.5%), diabetes (7.3%), chronic respiratory disease (6.3%) and hypertension (6%). This makes COVID-19 less lethal than the two other coronaviruses that have caused recent outbreaks; in 2002, SARS had a CFR of 9.6% and MERS of 34.4% - but it is nonetheless an order of magnitude more lethal than influenza (between .05-.1% depending on the year). For unclear reasons, the CFR for COVID-19 is lower in other parts of China than Hubei with an estimated mortality rate less than 1%. The mortality rate appears to be higher in men than in women. There has been some speculation that this may reflect the gender difference in smoking in China.

The term incubation period refers to the time from an exposure that results in infection until the occurrence of symptomatic disease in an affected person. It can be measured in people who have a known discrete exposure and who go on to develop disease. Current data suggests that the mean incubation period is 4-5 days but that there is significant variability among people in this parameter – with some developing disease as early as 2 days after exposure and others as long as 14 days later. Recent case reports have raised the possibility that some outliers could have incubation periods of 24 days or even longer. It is not yet clear how the incubation period relates to the infectious period (defined as the number of days in which a person can transmit an infection). Some transmission events have occurred from asymptomatic people prior to the development of clinical disease, suggesting that people can transmit before they are aware that they are ill, but it is not known how often asymptomatic transmission occurs or how long people remain infectious after they have been diagnosed.

The transmissibility of any infectious agent depends on several things: the probability of an infection event given a contact between a susceptible person and an infectious person; the duration of infectiousness – or number of days that a person can transmit - and the number of contacts that an infectious person has per unit time. This means that the transmissibility can vary in different settings and will depend on things like crowding which increases the number of contacts. Based on a summary of multiple studies, it seems that each infectious person with Covid-19 is expected to infect between 2 and 3 people on average. But this term – on average – obscures the substantial variability observed in different people. Some people are much more infectious than others and other people don’t transmit at all. Epidemiologists refer to this as dispersion around a basic reproductive number (defined below). High levels of dispersion or variance affect the likelihood that an introduced case will cause an outbreak – so if 80% of people do not transmit and 20% infect 10 people each, the average will be 2 but the probability that a single introduction will lead to an outbreak is only 20% whereas if everyone infects exactly two people, there would be 100% probability of an outbreak. It appears that the infectiousness of Covid-19 is quite widely dispersed but more data is needed.

Another way to estimate the transmissibility of an infectious agent is to measure the secondary attack rate (SAR), or the proportion of people who develop disease after a discrete exposure. A Lancet study that used data on secondary transmission associated with specific discrete social events reported 48 secondary infections that occurred among 137 attendees of these events, for an SAR among close contacts of 35%. https://www.thelancet.com/journals/lancet/article/PIIS0140-6736(20)30462-1/fulltext

The term “basic reproductive number” or R0 refers to the number of people who would be infected by a single infectious case if it were (hypothetically) introduced into an entirely susceptible population. If the parameters listed above (per contact infectivity, duration of infectiousness and number of contacts per unit time) are known, one can estimate the basic reproductive number simply by multiplying them. But we usually do not know these parameters, and some are very hard to measure, in particular the number of contacts over time. More often the R0 is estimated during the early phase of an epidemic by estimating the growth rate (as represented by the slope of the epidemic curve). Mathematically, the R0 = 1 + growth rate X the serial interval where the serial interval is defined as the time between one infection and the next in a transmission chain. Some possible issues with estimating R0 is that not all cases are reported and so they are not included in the epidemic curve. There are now more than 20 different publications/preprints estimating R0 using slightly different data sets and methods and the results range from 1.5 to 6 but most fall within the 2.3-3.2 range.

A basic rule of thumb is that the proportion of people who will not be infected by an epidemic pathogen (once the infectious disease has come to an equilibrium) is approximately the reciprocal of the R0. This is where the estimate that 40-70% of the world’s population could be infected by COVID-19 comes from. Of course, this depends on the efficacy of interventions, the mobility of infectious people and as noted above, the dispersion of the reproductive number.

There are several different approaches to restricting movements in order to control epidemic disease. One can isolate people with the disease to try to prevent them from infecting others but this will only be completely effective if they are diagnosed with the disease at or before the time that they become infectious. If people are infectious before they have symptoms or if some infectious people never develop symptoms at all, transmission can take place before a person is diagnosed. For diseases with significant asymptomatic spread, quarantine is used to separate and restrict the movements of people without signs of illness who may have been exposed to an infectious case so that they do not infect others during that period. A less extreme measure is social distancing – asking people to avoid congregate settings such as schools, work-places, or large gatherings. For example, the Wuhan “lockdown” is an example of fairly rigorous social distancing.

All of these methods can have specific downsides. Patients that are isolated within health care facilities may receive suboptimal care if isolation measures make it more difficult for health care workers to attend to them. Quarantine can result in the housing of uninfected people with asymptomatic infectious people and can lead to much higher rates of spread within the quarantine facilities. If social distancing measures involve loss of employment, education or routine medical care, they too can have serious negative effects both on individuals’ physical and mental health as well as on the economy. In the case of COVID-19, it is unclear whether school closures would be of benefit since few children develop the disease, although we do not yet know if they are asymptomatic carriers of the infection.

The question of the efficacy of quarantine, isolation and social distancing depends on when in the course of the infection most transmission is taking place. If most transmission occurs during the asymptomatic period—as it does, say, for HIV—isolation of patients with disease will have little impact. If, on the other hand, most transmission takes place when people have identified themselves as ill (as it did for SARS in 2002), isolation can be a very effective way to reduce spread. The benefits of quarantine—restricting the movements of people who are known to be in contact with an infectious case—depend on how effectively one can identify all contacts and prevent them from mixing with the general public. For obvious reasons, this can be very challenging and can have unintended consequences if quarantined people are housed together and become infected in that setting. Social distancing cannot prevent all transmission but could have a substantial impact on delaying transmission since contact rates are often much higher in congregate settings such as schools, prisons and other residential facilities. None of these measures is likely to lead to complete control of an epidemic since transmission is expected to resume once these are discontinued. But they may delay spread and give health systems time to develop better responses to the disease, whether those are new drugs, vaccines or simply improved efficiency of supportive care.

There are several different types of masks available to prevent infection. Surgical masks are used to prevent surgeons from contaminating a surgical site with respiratory droplets and are designed to protect others, but not necessarily the wearer. N95 masks are much more heavy-duty and fit tightly around the nose and mouth, blocking most transmission of even small airborne particles. These are worn either by patients themselves or by health care workers who come into close contact with known cases. They are quite uncomfortable and very expensive but they probably do reduce transmission of infections transmitted through the respiratory route. WHO and the CDC are urging the general public not to buy N95 masks as they are needed by health-care workers and patients and are in short supply.

COVID-19 is currently diagnosed by RT-PCR (real time polymerase chain reaction) or sequencing of respiratory and/or blood samples using “primers” based on the COVID-19 RNA sequence. This detects ongoing infection with live virus. The recent experience in China and elsewhere suggests that only 30-60% of cases are correctly diagnosed by RT-PCR during the initial presentation although it is still unclear if this low sensitivity is due to issues with sample collection, transportation or faulty kits. In the US, testing by RT-PCR has been slow to take off in part because kits initially provided by the CDC were found to give a higher than expected number of indeterminate test results. Current reports indicate that the problem with the tests was due to a faulty reagent and that this issue has now been resolved. The use of “home brew” or “laboratory developed tests” developed by hospital laboratories was not endorsed by the FDA in the case of this coronavirus although hospital labs are usually allowed to develop and use their own tests. (Apparently, this rule has been relaxed as of February 28.) Some reports recommend Chest CT scans as a more sensitive diagnostic tool than the PCR tests with one study suggesting that CT was 98% sensitive compared to 70% sensitivity of PCR.

While PCR tests indicate whether are not a person is currently infected, they do not indicate recent infection that has been resolved. For this, antibody tests are necessary. A team at Duke-NUS Medical School in Singapore reports having developed such a test which it used to trace a cluster of cases that have already cleared infection. Several other companies and teams have also reported developing this type of test. The antibody test will likely be more useful for surveillance and for retrospective studies of transmission clusters than for routine diagnostic testing of patients.

There are no FDA approved treatments yet available for COVID-19, although multiple clinical trials of antiviral drugs have been initiated. Remdesivir is an investigational “broad-spectrum” antiviral which is currently being studied in China as well as in the US—for the latter, specifically among people from the Diamond Princess now quarantined in Nebraska. Other agents that target the virus include favipravir, ribavirin, and galidesivir, nucleoside analogues that target RNA-dependent RNA polymerase. Protease inhibitors that might be effective include disulfiram, lopinavir and ritonavir; these agents were reported to be active against SARS and MERS. Another approach to treatment is to target host responses; pegylated interferon has been proposed as has the immune modulator, chloroquine. In the absence of specific therapy, most treatment of critically ill patients has included mechanical ventilation, treatment of sepsis and other types of supportive ICU care. I have been told that as many as 400 clinical trials are now underway in China.

Some respiratory viruses (influenza, RSV, the coronaviruses that cause the common cold) are seasonal, meaning that they tend to peak during winter months and decline in summer. This seasonal pattern is due to multiple factors. In temperate climates, schools tend to be in session in winter and people tend to congregate in warm buildings in cold weather; these behavioral factors mean that the contact rate is often higher in winter than in summer. Humidity is known to play a role in the transmission of influenza with higher rates of transmission during periods when the air is drier—which tends to be the case in winter in many areas. Some evidence exists that there are seasonal differences in host immune response. This is often attributed to vitamin D levels—which are higher in summer because they reflect exposure to UV light–and this theory is supported by the results of a recent meta-analysis which showed that vitamin D supplementation modestly reduced the occurrence of acute respiratory infections. One study of two other novel coronaviruses (SARS and MERS) found that these persisted on inanimate surfaces for longer periods of time in colder and drier conditions. (https://www.journalofhospitalinfection.com/article/S0195-6701(20)30046-3/fulltext)

In contrast, multiple observers note that COVID-19 has already circulated widely in Singapore where temperatures are in the 80s F. Several studies have compared the epidemic growth rates in different areas in China with differing levels of absolute humidity and found that changes in weather alone would be unlikely to reduce COVID-19 incidence without the implementation of public health interventions. (https://www.medrxiv.org/content/10.1101/2020.02.12.20022467v1)

Even if COVID-19 transmission declines with increasing temperatures in the Northern hemisphere, the virus has already been detected in the Southern hemisphere and transmission in those regions could intensify as the weather there cools down.

Vaccine development has proceeded at an unprecedented pace. A number of companies and research teams already have candidate vaccines that are either ready or close to ready to trial in humans. ModernaTx has submitted its mRNA-1273 vaccine to the NIH—a Phase 1 clinical trial to measure safety and immunogenicity is scheduled to begin in April in Seattle. Other companies that report having vaccines include Innovio, Janssen, Sanofi, Curevac and Clover Biopharmaceuticals. The speed with which these are being developed is partly due to the fact that a great deal of work was done on a SARS vaccine after the 2002 epidemic and some of that can be applied to this organism. Much of this research was funded through the Norway-based organization CEPI (Coalition for Epidemic Preparedness Innovations) which is funded by the Wellcome Trust, The Gates Foundation, and the World Economic Forum as well as some governments.

Although it is impossible to know exactly what proportion of cases of COVID-19 have been detected, it is highly likely that most cases have remained undetected. Several research groups have tried to estimate the proportion of cases that go undetected by looking at data on new cases occurring in countries that received air travelers from high burden areas. The basic idea is that if we know that—say—2% of people in Wuhan were infected at some period, and we know how many air travelers arrived from Wuhan in other countries, and we know the R0 or the SAR, we should be able to estimate how many cases should have occurred as a result of those introductions. The difference between the actual number of reported cases and the estimated number is meant to be a proxy for undetected cases. Although there are some flaws to this reasoning – for example, the prevalence in Wuhan might have been much higher than reported, these biases would tend to underestimate the proportion of undetected cases(https://www.medrxiv.org/content/10.1101/2020.02.04.20020495v2). Using these methods, research teams have estimated that between 60-75% of cases are undetected (https://www.imperial.ac.uk/media/imperial-college/medicine/sph/ide/gida-fellowships/Imperial-College---COVID-19---Relative-Sensitivity-International-Cases.pdf). Other groups have used a sort of reverse method to estimate the number of undetected cases in Iran–that is, going back from the number of exported cases that have been detected in people traveling from Iran. In one study, the group estimated that 18,300 COVID-19 infections had occurred in Iran although to date, only 388 (2% of the estimated number) have been reported. (https://www.medrxiv.org/content/10.1101/2020.02.24.20027375v1)

On February 23, the first case of locally transmitted COVID-19 was reported in the US in a woman with no known travel history and no ill contact. She came from Solano in Northern California and was transferred to the UC Davis Medical Center on February 19. At the time of her transfer, she had already been diagnosed with a severe pneumonia and was intubated and on mechanical ventilation. With her medical team suspecting that she might have COVID-19, she was kept in isolation at UC Davis but it is unclear from news reports how long she had been ill prior to her transfer, for how many days she had been hospitalized, and whether she had been in isolation prior to arriving at UCD. The UC Davis team was unable to obtain CDC permission to get COVID-19 testing until February 23 when the test came back positive. Notably, Solano is also in close proximity to the Travis Air Force Base where American evacuees from China were quarantined. A whistle-blower has complained that federal workers at the AFB were in contact with quarantined evacuees without having received proper training and without protective gear. These events strongly suggest that community transmission of COVID-19 has occurred in this region and given the estimated proportion of cases that go undetected and the lack of easy access to testing in the US, that other infectious cases may currently be present in the community.

Over the past couple of days, a number of new US cases have been reported that had no known travel or exposure history. One of these was a person in Snohomish county in Washington State whose COVID-19 genome was sequenced and found to be very closely related to a travel-related case reported in the same county on January 20. This suggests that the earlier case is likely to have transmitted to others in this region and that the second detected case was likely to have been infected by an undetected person who was part of a transmission chain that stems back to the first imported case. By examining the differences in the two COVID-19 genomes and knowing how much time has elapsed between the two infections, the investigator (Dr. Trevor Bedford) is able to estimate of the number of people who could have been infected through this transmission chain. His best estimate is several hundred although given the limited number of genomes analyzed, the number is quite uncertain and could range from less than 100 to over 1000. This is an unpublished report available on Twitter but multiple experts have concurred with these conclusions. The most important point here is that there has almost certainly been undetected transmission of COVID-19 in the US for at least 5-6 weeks and the true scope of this problem will not be fully realized until testing is much more widespread.

Most experts believe that it is inevitable that COVID-19 will spread in the US. It may be possible to slow transmission through the adoption of the kinds of interventions listed above but it is unlikely that a vaccine will be available in the near future. There are several factors that may make control of an epidemic especially difficult in the US. We are in the midst of a particularly bad influenza season and it will be difficult to know if one has seasonal flu or COVID-19. This may increase the number of people who need to be tested markedly, making this both more expensive and more logistically challenging. Many patients may be reluctant to present for diagnosis of what they consider a mild illness because the uninsured and under-insured may need to pay out of pocket for testing or for a doctor’s visit. Sick individuals may also be reluctant to stay home from work if they do not have accumulated sick leave or receive any paid time off from their employer.

One of the consequences of the Wuhan epidemic is that people who needed medical care for other conditions have not been able to obtain that care because hospitals and medical staff are at full capacity dealing with the virus. News reports describe people in the region who have been unable to get dialysis or chemotherapy for the past month. So far, we have been unable to obtain data on whether general mortality has increased in Wuhan as a result of this lack of access to care. It is worth noting that during the Ebola epidemic, rates of maternal mortality increased because of the effect of Ebola on the availability, uptake and outcomes of maternal and newborn health services in Sierra Leone.

There are lots of very practical suggestions available from the CDC or WHO on measures one can take to protect oneself from infection and to prepare for the possibility of an epidemic. Some of these are obvious:

• Wash your hands frequently.
• Try not to touch your face.
• Avoid people who are coughing or obviously ill.
• Avoid large crowds if possible.
• Don’t go to work if you are sick. Send your sick workers home.
• If you need to seek medical care for a flu-like illness, call in advance and ask for instructions on where to go.
• If you are sick, don’t go and visit your elderly or immuno-compromised friends and neighbors.

Less obvious

• Consider having a plan for what you might do if social distancing measures are put into effect or if you were quarantined.
• Consider forgoing unnecessary travel (and possibly even necessary travel if it is to high risk places).
• Think through how you and your teams can work from home – what are the best options for conference calls, etc.
• Get whatever books you might need out of the library now.
• Make sure you have a reasonable supply of any prescription drugs you need.
• Have some emergency provisions but don’t go crazy buying up the grocery stores’ entire supply of canned goods.
• Consider using a humidifier.

Beyond the efficacy of new and repurposed drugs and vaccine candidates and the development and validation of rapid, reliable diagnostics, some of the most urgent questions that need to be resolved include the following:

1. What role do children play in the transmission of COVID-19? Are they getting infected but just not getting disease and if so, are they infectious in this asymptomatic state? Or are they somehow protected from being infected as well? This is important since one social distancing approach is school closures, which of course would be fruitless if children are not involved in transmission. In principle, this question could be easily resolved though wider testing.
2. Are people who have had COVID-19 infection immune from re-infection and if so, for how long? One of the things that slows down disease spread is when enough people in a population are immune so that “herd immunity” kicks in - either through people having acquired immunity through previous infection or through vaccination. If natural or vaccine-induced immunity is not robust or long-lasting, this means that the “brakes” that usually end epidemics may not work. Evidence from other coronaviruses suggests that some immunity to this class of organisms is acquired but it is not clear how long it lasts.  Several reports from Asia indicate that some people who have had COVID-19 have cleared the virus, only to have it re-appear later. It is not clear if this is due to re-infection or to a stuttering or relapsing course in which virus levels were undetected but not fully cleared.