Twenty Years of Research and Discovery

Prof Kelvin To Kai-wang
Chairperson and Clinical Professor
Department of Microbiology

The COVID-19 pandemic illustrated the immense health and socioeconomic impact of emerging infections. Not too long ago, however, emerging infections were largely ignored because it was believed that the improvement of hygiene and healthcare could keep them under control. But the threat has increased substantially due to rapid increase in international travel and human-animal contact.

In 1997, Hong Kong faced the first emerging infection with a major public-health impact in the modern era. Eighteen patients were infected with the novel avian influenza A(H5N1), resulting in six deaths. The outbreak was only stopped after territory-wide culling of chickens.

In 2002-03, the bat coronavirus jumped species barrier and caused the severe acute respiratory syndrome coronavirus (SARS-CoV-1) pandemic. The 2003 SARS-CoV-1 pandemic was characterized by a high rate of transmission within hospitals, and with explosive community transmission under favourable environmental conditions. The 2003 SARS-CoV-1 not only led to loss of lives, but has severely affected the economy. The 2003 SARS-CoV-1 was a wake-up call for the society that an emerging infection can paralyze a community.

After the 2003 SARS-CoV-1 epidemic, we have encountered several notable emerging infections in Hong Kong, including the 2009 influenza A(H1N1) pandemic and the 2013 avian influenza A(H7N9). Furthermore, the seasonal influenza surge continues to put tremendous pressure on the healthcare system. In the past 20 years, we have equipped ourselves with the knowledge and skills to better combat emerging infections.

Viral load kinetics

Understanding viral load changes can help us understand the pathogenesis of severe infections and the risk of infection during different stages of infection, and guide us on the timing of intervention strategies. During the 2003 SARS-CoV-1 epidemic, it was found that the viral load peaked 10 days after symptom onset. Since most patients were admitted during the first few days of illness, the spread of SARS-CoV-1 mostly occurred within hospitals. Community spread was uncommon, except under special circumstances.

During the 2009 influenza A(H1N1) pandemic, we compared the viral load changes over the course of illness between mild and severe cases. We found that although mild cases have high viral load initially, the viral load declined very rapidly. By contrast, severe cases were unable to clear the virus despite the use of antivirals. Thus, our currently available antivirals only had modest effectiveness for patients with influenza virus infection.

When SARS-CoV-2 first emerged in Hong Kong in early 2020, we traced the viral load changes in our patients. Unlike the 2003 SARS-CoV-1 patients, COVID-19 patients had very high viral load upon hospital admission. This explained the rapid spread of SARS-CoV-2 in the community, and helped us predict that it would be more difficult to control COVID-19 than the 2003 SARS-CoV-1.

Furthermore, the viral load data suggests that we need an antiviral to lower the viral load as quickly as possible. Another notable finding is that the viral RNA can be detected by reverse transcriptase-polymerase chain reaction for a long period even when the virus is no longer infectious. Therefore, it became apparent that it is unnecessary and not practical to isolate all COVID-19 patients until the virus is completely cleared.

Using saliva for diagnostic testing and viral load monitoring

Traditionally, doctors and nurses are taught to collect nasopharyngeal specimens for testing of respiratory viruses. But using this method poses several problems, especially when a novel respiratory virus is emerging. Specimens must be collected by healthcare workers, putting them at risk of infection. Personnel are trained to perform the procedure inside negative-pressure rooms, wearing full personal protective clothing. These infection-control requirements have led to a huge delay in specimen collection, which in turn results in a prolonged unnecessary stay for some patients. Furthermore, collecting a nasopharyngeal specimen is an invasive procedure that creates substantial discomfort for the person being tested.

To tackle this problem, we needed to find a specimen type that does not require collection by healthcare workers or any invasive procedure. We turned to saliva, which can be self-collected. In 2017, we published our first study demonstrating that saliva has a high sensitivity in the detection of influenza virus and other respiratory viruses. In 2019, we published another study demonstrating that we could detect respiratory viruses in saliva specimens using a rapid PCR testing machine widely available in Hospital Authority hospitals.

When SARS-CoV-2 first arrived in Hong Kong in early 2020, we quickly found that saliva has high sensitivity in the detection of SARS-CoV-2. This study led to the use of saliva specimen collection for community testing, allowing early detection and early isolation of patients and quarantine for contacts which prevented further spreading of the virus. Saliva testing also allows screening of asymptomatic healthcare workers to prevent the spread of infection within a hospital. Saliva testing also allows us to have a non-invasive and safe procedure to collect serial specimens from patients to monitor viral load and determine when the patient can be deisolated. Since its successful implementation in Hong Kong, it has become a popular method worldwide for screening specimens, including regular testing of professional athletes.

Saliva

Vending machine for distributing saliva collection packs

Antibody testing and the burden of infection and susceptibility to infection

Antibodies against a specific virus can be found in the blood about two weeks after an infection, and these antibodies can last for a long period of time. Antibody testing is important for studying novel emerging infections. We can use antibody testing to estimate the true burden of infection. Unlike PCR testing which can only identify patients with active infection, antibody testing can also detect patients with past infection. Since the level of antibodies correlate with protection, we can also use antibody testing to predict the severity of an upcoming outbreak. And we can use it to identify novel variants which escape from immunity elicited from prior infection or vaccination.

During the 2009 influenza pandemic, we conducted a seroprevalence study and found that the highest risk of death occurred for individuals aged 50 or over. This suggests that we should recommend the influenza vaccine for those older than 50 rather than 65.

During the H7N9 outbreak in 2013, we wondered whether the public-health measures were sufficient to protect the workers in live-poultry markets from infection due to H7N9 virus. Using antibody testing, we did not find any evidence of H7N9 infection. We did determine that about 60 percent of the live-poultry market workers had evidence of infection due to influenza A(H5).

This study alerted the government to the risk of avian influenza among high-risk groups in Hong Kong. Although there have not been any human cases of avian influenza virus infection in the city for a long time, we should not be complacent. Recently, there have been H5N1 outbreaks among mammals, including mink and seals, which suggest that the current H5N1 strains can transmit efficiently among these animals. In February 2023, a fatal human case of H5N1 was confirmed in Cambodia.

We have also monitored the seroprevalence of seasonal influenza viruses. In doing so, we successfully predicted the seasonal influenza virus subtype that affected the 2018-19 winter influenza season.

When SARS-CoV-2 emerged, we quickly determined the seroprevalence among our population. We found that most blood specimens collected from individuals before 2020 did not carry antibodies against SARS-CoV-2, suggesting that we were dealing with a novel virus instead of one that had been circulating in the community for a long time.

In March 2020, we found that about four percent of individuals returning from Hubei province (mainly from Wuhan) carried antibodies against SARS-CoV-2, which indicated that they were infected. This study helped us to understand the burden of infection in Wuhan, which helped us make a risk assessment for Hong Kong. Our figure was very similar to those obtained from other studies conducted in Wuhan during the same period. After the fifth wave of the virus arrived in Hong Kong, we found that about 90 percent of elderly had antibodies against SARS-CoV-2, suggesting that most of these senior citizens had at least partial protection and therefore the risk of another big outbreak would be unlikely. These efforts helped us to make evidence-based risk assessments and provide realistic recommendations to the government. SARS-CoV-2 variants, such as Omicron, can escape antibodies produced from prior infection or vaccination. We have used antibody testing to trace the immune escape of these novel variants. By testing population immunity against these novel variants, we can estimate the risk of a variant in causing a severe outbreak.

Building up genomic capacity

Understanding the viral genome is important for tracing the dynamics of the spread of infection. Over the past 20 years, there was a big leap in sequencing technology. We are equipped with the latest tools so that we are prepared to fight novel infections. In 2013, we used Sanger sequencing to determine the viral genome of the avian influenza A(H7N9) strains in Hong Kong and found that it was most closely related to a virus strain detected in Guangdong Province.

Next-generation sequencing (NGS) technology has become much cheaper and easier to use. Using whole viral genome sequencing, we have designed novel-influenza primers based on the PB2 gene. This novel PB2 PCR was found to have a lower limit of detection than the standard M gene PCR recommended by the World Health Organization (WHO) and was able to detect both seasonal, pandemic and avian influenza viruses.

During COVID-19, we performed whole viral genome sequencing and have deposited over 2,000 strains on the public viral genome database. Our sequencing has revealed how new strains were introduced into Hong Kong during each wave and how SARS-CoV-2 evolves. Whole viral genome sequencing is also important for outbreak investigations. Using it, we confirmed the crucial fact that SARS-CoV-2 can spread over a long distance via the airborne route within hospitals.