3

Pathogen Amplification and Dissemination

PATHOGEN DYNAMICS AND SPILLOVER

Vincent Munster, National Institutes of Health, discussed the patient zero paradigm and approaches to detect new infectious diseases. Based on his work on reservoirs, drivers, and the dynamics of emerging viruses, he said that the ability to detect a pathogen is directly related to the severity of the symptoms it causes (Munster et al., 2020). For example, pathogens causing viral hemorrhagic fever (VHF) are relatively easy to detect owing

to the visible symptoms of these diseases, which helps with measures to control and contain the outbreaks. Ebola virus stands out in its ability to transmit relatively efficiently from human to human. Origins of Ebola disease in human populations can be complex; outbreaks may come from a single spillover event from a reservoir species such as the fruit bat, but multiple spillover events can also occur in a relatively small time frame within a region. There may also be epizootic spillovers, he said, where the pathogen is introduced in nonhuman animal species. In a later phase of outbreak response, when the disease transmission has been controlled, there may still be low-level circulation of the pathogen in surviving patients, which can spark new cases later through the human-to-human transmission of latent cases. Munster said that, considering how spillover events might occur quite frequently in some settings, it is important to understand what factors can lead to transmission from the point of spillover onward to more people.

Munster noted that there is a lag between the spillover event and the detection of the first patient. This is important, as the longer it takes to detect the first patient, the harder it will be to control an outbreak as additional transmission events may occur during this time. Access to Ebola treatment units and decentralized diagnostics are essential to control outbreaks, Munster said, but these tools are meaningless if there is no actionable response. Munster shared an example of work on the West African Ebola virus outbreak in Liberia, where he and his team reanalyzed patient samples and found that the viral load in these samples decreased over time (Jeremiah Matson et al., 2022). That sparked discussion among virologists about whether there was viral adaptation and how that might affect the potential for more human-to-human transmission. However, this trend actually reflected the success of the Ministry of Health’s intervention, as people were accessing the clinic and receiving diagnostics within fewer days of symptom onset (i.e., when there were lower levels of replicating virus in the patient at the time of diagnosis), and this helped control the outbreak.

Munster then discussed ongoing work in predicting the dynamics of Marburg virus outbreaks in people. When considering a specific fruit bat species as a reservoir for Marburg virus, Munster said, an assessment of the bats’ population distribution can show the likelihood of having Marburg virus in certain places and thus potential for an outbreak (Amman et al., 2023). Ebola virus, in contrast, has no clear reservoir species, and larger areas are considered at risk for outbreaks, which makes prediction of infection and disease dynamics in humans more challenging. Therefore, more data are needed on host–pathogen interactions to allow for the specific prediction of disease emergence, he said.

In a study Munster conducted with a local health clinic in the Republic of Congo, a population of local indigenous hunter gatherers were analyzed for seropositivity for mpox, which increased over time and suggested that

people were regularly exposed to the virus. Further, a study showed that people with occupational risks, such as bat hunters, have a different serological profile than people living in cities, indicating exposure to different pathogens. Munster and team set out to assist in laboratory response efforts by the local ministry of health to mpox. They used next-generation sequencing (NGS) to determine which variants of mpox were causing infections and to elucidate potential transmission routes, whether human-to human transmission or spillover from natural reservoirs (Yinda et al., 2024). However, he pointed out, diagnostics are critical in identifying cases and providing epidemiologically important data such as the NGS findings, but they do not affect outbreaks if the findings cannot lead to actionable responses. Mpox, for example, has been on the rise for decades, and while there are approved vaccines, only recently have attempts been made to introduce these vaccines in countries where mpox has been circulating for years. Diagnostics should be used to inform robust public health response to get ahead of outbreaks.

AMPLIFICATION IN SECONDARY HOSTS

Richard Webby, St. Jude Children’s Hospital, discussed the amplification of disease in secondary hosts. He explained that wild waterfowl are natural hosts for influenza, but the threat of influenza virus transmission to humans comes primarily from an intermediate host, like swine, chickens, or cattle. However, intermediate hosts are not all the same in terms of risk for a spillover event into humans. Selective pressures within an intermediary host can drive changes in the virus, said Webby. These evolutionary adaptations are driven by the type of host, tropism of infection, densities of host species populations, and the local conditions where the virus must survive. Further, there may be chances of interaction with related pathogens within a host species, said Webby, which can also drive evolution of the virus. All these factors affect the risk of transmission to humans.

There are many other variables to consider for human exposure, Webby said, including how many people are exposed to the pathogen in a particular setting, how often they are exposed to the pathogen, who is exposed (e.g., different ages, existing comorbidities or other susceptibilities), and the transmission route of exposure (e.g., air, fluids through permeable entry points, ingestion). One question with the spread of highly pathogenic avian influenza (H5N1) among dairy cattle in the United States is why there has not been more severe disease observed in dairy workers. This could be attributable to it affecting a relatively healthy population with fewer underlying health conditions or susceptibilities given the physical nature of the work, or the exposure route, which is dependent on the intermediate host.

Webby then discussed how different selective pressures in intermediate hosts affect the risk of disease in humans. If a virus that is not well adapted to humans gets into a new host, the virus may evolve in that new host and adapt in a way that allows it to transmit to and infect humans (Shinya et al., 2006). For instance, avian influenza virus binds to the sialic acid receptor to gain entry to host cells. Avian hosts express the α-2,3-linked sialic acid receptors, and humans express the α-2,6-linked sialic acid receptor, but swine express both versions of this receptor. If avian influenza virus infects swine, selective pressure within the swine host could drive evolution in the virus to adapt to binding the α-2,6-linked receptor and enable the virus to more efficiently infect humans. However, whether this happens depends on many factors, he explained. Despite the ongoing H5N1 outbreak in dairy cattle, the virus has not developed further markers indicative of adaptation to mammalian hosts. This might be caused by where the virus is replicating in the cows, Webby said, since the virus is present in the udders and there is no selective pressure for it to replicate in the milk. He also noted that if transmission to humans is occurring through the milking apparatus and procedures, it may lower selection pressure for the virus to evolve in order to transmit.

Another means of viral evolution, Webby noted, is through interaction with other influenza viruses present in a host. Recombination assortment, in which genetic material from different viruses is exchanged, allows viruses to adapt to different hosts and environments. Reassortment has been a driver of past influenza pandemics, he said.

In summary, the secondary or intermediate host can be an important conduit to human infection for many pathogens. In the search for patient zero, Webby concluded, understanding which secondary hosts might drive evolution of the most concerning phenotypes leading to human infections can help prioritize research. Gaining a better understanding of the risk to humans will require a better understanding of the selective pressures present in secondary hosts, their microbiota, and how management of animal intermediate hosts changes these pressures.

AMPLIFICATION IN HEALTH CARE SETTINGS

Mary Choi, Centers for Disease Control and Prevention, discussed the amplification of VHF in health care settings. Current known VHFs are caused by zoonotic viruses that spread from person to person and frequently cause illness in humans with severe morbidity or mortality.¹ Currently, there are no or limited medical countermeasures against human infection with VHF. Choi explained that the clinical presentation varies for each VHF, but typically begins with nonspecific symptoms like fever,

___________________

¹A zoonotic pathogen is one that can spread between animals to humans.

malaise, fatigue, and muscle or joint pain. This is followed by vomiting and diarrhea, and hemorrhage occurs in some patients, which can lead to multisystem organ failure, shock, and eventually death. Amplification in health care settings is common in VHF outbreaks, she said, and a cluster of illness among health care workers is often the first indication of a VHF outbreak.

Choi explained that several factors contribute to VHF amplification in health care settings. The first is that patients exhibit high care-seeking behavior, as they have an illness that cannot be adequately treated by their usual sources of care. Choi’s research in the Democratic Republic of Congo revealed that most patients seek care from two or more facilities before diagnosis is reached. Another factor in the amplification of VHF is nonadherence to infection prevention and control (IPC) practices. In the past, poor IPC practices such as reuse of needles in a clinic have propagated VHF outbreaks. Choi noted that there is generally insufficient quantities of personal protective equipment (PPE) in countries where VHF are endemic, limiting the appropriate use or resulting in reuse of PPE. A third factor is that VHFs are hard to diagnose, since the diseases are rare and occur in countries where diseases with similar symptoms are endemic (e.g., malaria, typhoid). Further, Choi noted that no single sign or symptom is specific to VHF, as less than 50 percent of patients experience hemorrhage. Therefore, a key factor to diagnosis is eliciting risk factors (e.g., travel to area of outbreak, contact with patients), but patients may not recognize or acknowledge VHF risk factors.

Additionally, these diseases are rapidly fatal without treatment. Therefore, Choi said, by the time a physician realizes the patient is not responding to treatment for initial diagnosis, the patient is likely very sick, and the virus has often already spread in the facility. Finally, there is limited VHF testing capacity, and most testing is performed at centralized regional laboratories. Infections are rare, reminded Choi, so some countries may not have any testing capacity. Logistical issues in transporting samples to conduct diagnostic testing can impede the rapid detection of outbreaks.

Choi provided an example comparing two patients who contracted Marburg virus in Equatorial Guinea. A 44-year-old male patient was admitted at hospital A based on early symptoms, discharged, and then admitted to hospital B a few days later. At request of the family, the patient was transferred to hospital C. He had to be resuscitated and intubated in the emergency room but ultimately succumbed to his illness. A sample later tested positive for Marburg virus, and this triggered a notification from the Ministry of Health to the three hospitals, with a request for contact tracing. There was delay in allowing contact tracing teams into hospital C, but the teams identified two ill health care workers who both tested positive for the virus. This triggered the request for IPC personnel to strengthen practices and engagement to prepare for future cases.

A second patient, a 9-year-old female, was first admitted to hospital B and was transferred two days later to hospital C, where she was resuscitated and intubated, but also succumbed to illness. A health care worker suspected Marburg virus and reported the case to the Ministry of Health. Sample testing later confirmed Marburg virus infection. Unlike for the first case, no secondary cases of Marburg virus infection were detected. Choi suggested the reason for the difference was the level of awareness about the VHF outbreak among health care workers. Following the first patient infection, health care workers were not made aware of the outbreak, but by the time of the second patient infection, health care workers knew there was an ongoing Marburg outbreak in their city and were able to take precautions to prevent the spread of disease.

Choi concluded that disease amplification in health care settings is common and, owing to multiple contributing factors, identification of patient zero in these settings is challenging. Awareness about ongoing outbreaks such as with VHF plays a key role in the early recognition of cases, but strict adherence to standard precautions is critical when caring for any patient, she said.

OVERCOMING BLIND SPOTS IN PATHOGEN DETECTION

Erik Karlsson, Institut Pasteur du Cambodge, discussed his work on early warning disease surveillance in Cambodia. He highlighted the need for disease surveillance to get ahead of endemic and emerging diseases at high-risk interfaces. Karlsson expressed the need to develop active disease surveillance tools and strategies that are better, cheaper, faster, and safer to implement.

Karlsson explained that environmental samples provide a viable option to improve early warning for avian influenza at high-risk interfaces, as samples can be collected quickly and are conducive to studying multiple pathogens. Environmental samples are a safer option for places that may be difficult to reach in terms of biosafety considerations, are better for animal welfare, and enable the use of less expensive test assays.

Karlsson explained that his team has performed disease surveillance in many key hotspots in Cambodia. The team has focused performing disease surveillance in markets by taking air samples. For the collection, a small portable aerosol collector can be used, and these samplers can be strapped to remote-controlled drones or small cars to obtain samples, so researchers do not have to enter high-risk areas. In addition to markets, this aerosol collector can be attached to drones and used in farms, caves, and other areas with wild and captive animals to collect samples while reducing risk to researchers and increasing efficiency, Karlsson explained.

Karlsson showed an example where his team collected air samples at various places within a live bird market. Avian influenza could be detected by quantitative reverse transcription polymerase chain reaction (qRT-PCR) in paired samples collected directly from poultry and from the air. Viral load was shown to be present with a gradient-like pattern in the air samples, increasing from areas of slaughter to storage and to food sales. Based on this, Karlsson and team hypothesized that this might be attributable to the air-handling systems within these markets, and they investigated the airflow in the market to inform potential interventions. Further, they provided personal air samplers to the sellers who worked at the market, which revealed that there was a 100 percent chance of being exposed to avian influenza viruses in just 30 minutes when there is high circulation of avian influenza in the area (Horwood et al., 2023). This example demonstrates the value of these technologies in defining risk.

Karlsson and team are also working on developing rapid diagnostics that can be coupled with environmental surveillance. He shared an example in which nanopore sequencing—a real-time technique for DNA or RNA sequencing—was integrated with whole genome sequencing to reduce sample processing and sequencing time from 33–48 hours to 5–7 hours. This approach allowed his team to process a sample of influenza virus and post its full genome sequence on the Global Initiative on Sharing All Influenza Data in the same day. Karlsson described how his laboratory has applied nanopore sequencing in the field, allowing them to analyze avian influenza cases in humans and quickly determine whether poultry-to-human transmission is the source of infection. They were also able to quickly recognize the emergence of a new lineage of avian influenza virus, which is also critical for informing outbreak response. Further, Karlsson said, his laboratory is combining environmental disease surveillance with metagenomics in order to survey for pathogens more broadly and receive information more rapidly.

Validation studies found comparable results between the next-generation sequencing methods and standard PCR methods. The researchers also tested whether environmental samples could effectively replicate the sampling of live poultry in the market and found that the viral taxa in avian samples could be detected in the environment at each visit; furthermore, while around 70 viruses were detectable from poultry samples, the same viruses as well as an additional 300 pathogens were detected from environmental samples. These results suggest that another advantage of environmental sampling in that it allows testing for the overall presence of viruses without having to test every bird, surface, or equipment at the market, Karlsson noted.

Karlsson added that their environmental testing approach has also been applied to bat caves, flying fox roosts, and bat farms. His team found that

all virus genera detected in feces, urine, and organs of bats can be accurately detected in air samples taken at the same time, except for sexually transmitted viruses that are not released in the air.

Karlsson stated that standards developed by the international community are needed, especially for the use of next-generation sequencing in both clinical and surveillance settings. The adoption of international standards requires a paradigm shift, he said, as environmental testing can be more effective than animal sampling to determine the location of risk hotspots. Once hotspots are located, then animals in specific places can be sampled. He said that currently, his team is scaling up and is expanding this type of sampling to multiple places in Asia and Africa, as well as developing data and sharing methods to inform decision making. Further, Karlsson said that he is engaged in discussions on how to make these data actionable, which requires collaboration with social science experts to determine how to most effectively present their data to politicians and decision makers.

DISCUSSION

High-Risk Areas for Amplification

Keith Klugman, Gates Foundation, began the discussion by asking Karlsson whether closing food markets is being considered as a potential strategy to mitigate the risk of disease spread. Karlsson replied that widespread closure of live animal markets is not practical, as these places are important to society and culture in many parts of Asia and are critical for people to access food. A better solution, he suggested, is to make them safer by applying interventions that can reduce risks, such as taking rest days and not keeping animals in the markets overnight, he said. Environmental disease surveillance can help monitor the effectiveness of preventative measures, and these data can help inform authorities and decision makers.

Risk Factors for Amplification

Emily Gurley, Johns Hopkins University, asked about the amplification of infections in hospitals, and whether it is known how characteristics such as age and sex affect heterogeneity of the number of secondary cases in each disease outbreak. Choi replied that many factors influence how many people one initial case may go on to infect. For instance, cases in children and people who need more physical assistance in daily life may result in more secondary cases. Another factor is occupation. Infections in health care workers tend to result in more secondary cases owing to the nature of their work and interactions with multiple people. In response to this, another attendee asked whether staffing patterns could be altered to

decrease the number of people who routinely contact a critically ill patient to reduce the risk of amplification.

Choi replied that altering staffing patterns depends on the initial suspicion and recognition of a disease. For instance, as soon as VHF is suspected, the recommendation of Centers for Disease Control and Prevention is to isolate the patient and limit the number of health care workers who are in contact, but that does not help all the patients and providers that they were in contact with before the diagnosis. Patients often seek care in multiple facilities before diagnosis, and because some infections are very rare, clinicians and patients may not immediately consider VHFs as a possible diagnosis. Another participant noted that health care worker perception of risk in encountering a patient zero is low, resulting in lax adherence to standard PPE protocols. Choi agreed that in a clinical setting it is difficult on a daily basis to consider if different PPE protocols are necessary for each patient that arrive at the facility. She noted that awareness of the interconnectedness of the world is important for keeping this real risk in mind, but that adherence to PPE protocols is ultimately a matter of personal responsibility.

This page intentionally left blank.