Increased severity of influenza-associated hospitalizations in resource-limited settings: Results from the Global Influenza Hospital Surveillance Network (GIHSN) (preprint)

Background Influenza disease data remain scarce in middle and lower-income countries. We used data from the Global Influenza Hospital Surveillance Network (GIHSN), a prospective multi-country surveillance system from 2012-2019, to assess differences in the epidemiology and severity of influenza hospitalizations by country income level. Methods We compiled individual-level data on acute respiratory hospitalizations, with standardized clinical reporting and testing for influenza. Adjusted odds ratios (aORs) for influenza-associated intensive care unit (ICU) admission and in-hospital death were estimated with multivariable logistic regression that included country income group (World Bank designation: high-income countries: HIC; upper middle-income countries: UMIC; lower middle-income countries: LMIC), age, sex, number of comorbidities, influenza subtype and lineage, and season as covariates. Findings From 73,121 patients hospitalized with respiratory illness in 22 countries, 15,660 were laboratory-confirmed for influenza. After adjustment for patient-level covariates, there was a two-fold increased risk of ICU admission for patients in UMIC (aOR 2.31; 95% confidence interval (CI) 1.85-2.88, p < 0.001), and a 5-fold increase in LMIC (aOR 5.35; 95% CI 3.98-7.17, p < 0.001), compared to HIC. The risk of in-hospital death in HIC and UMIC was comparable (UMIC: aOR 1.14; 95% 0.87-1.50; p > 0.05), though substantially lower than that in LMIC (aOR 5.05; 95% 3.61-7.03; p < 0.001 relative to HIC). A similar severity increase linked to country income was found in influenza-negative patients. Interpretation We found significant disparities in influenza severity among hospitalized patients in countries with limited resources, supporting global efforts to implement public health interventions. Funding The GIHSN is partially funded by the Foundation for Influenza Epidemiology (France). This analysis was funded by Ready2Respond under Wellcome Trust grant 224690/Z/21/Z. Research in Context Evidence before this study In the past 35 years, fewer than 10% of peer-reviewed articles on influenza burden of disease have reported analyses from lower middle- or lower-income settings. Whereas the impact of influenza in upper middle- and high-income countries – regions where influenza seasonality is well-defined and where high numbers of influenza-related clinic visits, hospital admissions, and deaths are well-documented – has been clearly quantified, data scarcity has challenged our ability to ascertain influenza burden in resource-limited settings. As a result, policy decisions on vaccine use in lower-income countries have been made with limited data, slowing the development of influenza vaccine recommendations in these settings. In this study, we have conducted prospective influenza surveillance in the hospital setting in multiple countries to assess potential geographic differences in the severity of influenza admissions and have shown that influenza is a global concern, and report poorer clinical outcomes among patients admitted to hospitals in resource-limited settings. In these settings, it is especially important to consider the role of preventive measures, such as vaccines, in providing protection against severe disease. Added value of this study Since 2012, in collaboration with over 100 clinical sites worldwide, the Global Influenza Hospital Surveillance Network (GIHSN) has provided patient-level data on severe influenza-like illnesses based on a core protocol and consistent case definitions. To our knowledge, this is the first study to analyze multiple years of global, patient-level data generated by prospective, hospital-based surveillance across a large number of countries to investigate geographic differences in both influenza morbidity and mortality. Our study provides information on influenza burden in under-researched populations, particularly those in lower middle-income countries, and highlights the need for continued global collaboration and unified protocols to better understand the relationships between socio-economic development, healthcare, access to care, and influenza morbidity and mortality. After adjustment for differences in the characteristics of individual patients admitted to the hospital for influenza, we find an increased severity of disease in lower-income settings. In particular, the risk of ICU admissions increases two- and five-fold in upper middle- and lower-middle income countries, compared to high-income countries. The risk of in-hospital death is five-fold higher in lower-middle income countries, compared to more affluent countries. Implications of all the available evidence We find evidence of increased severity in influenza admissions in lower-income countries, which could point at structural differences in access to care between countries (patients arriving at the hospital later in the disease process) and/or differences in care once in the hospital. Understanding the mechanisms responsible for these disparities will be important to improve management of influenza, optimize vaccine allocation, and mitigate global disease burden. The Global Influenza Hospital Surveillance Network serves as an example of a collaborative platform that can be expanded and leveraged to address geographic differences in the epidemiology and severity of influenza, especially in lower and upper middle-income countries.

Sampling strategies and pre-pandemic surveillance gaps for bat coronaviruses (preprint)

The emergence of SARS-CoV-2, and the challenge of pinpointing its ecological and evolutionary context, has highlighted the importance of evidence-based strategies for monitoring viral dynamics in bat reservoir hosts. Here, we compiled the results of 93,877 samples collected from bats across 111 studies between 1996 and 2018, and used these to develop an unprecedented open database, with over 2,400 estimates of coronavirus infection prevalence or seroprevalence at the finest methodological, spatiotemporal, and phylogenetic level of detail possible from public records. These data revealed a high degree of heterogeneity in viral prevalence, reflecting both real spatiotemporal variation in viral dynamics and the effect of variation in sampling design. Phylogenetically controlled meta-analysis revealed that the most significant determinant of successful viral detection was repeat sampling (i.e., returning to the same site multiple times); however, fewer than one in five studies longitudinally collected and reported data. Viral detection was also more successful in some seasons and from certain tissues, but was not improved by the use of euthanasia, indicating that viral detection may not be improved by terminal sampling. Finally, we found that prior to the pandemic, sampling effort was highly concentrated in ways that reflected concerns about zoonotic risk, leaving several broad geographic regions (e.g., South Asia, Latin America and the Caribbean, and most of Sub-Saharan Africa) and bat subfamilies (e.g., Stenodermatinae and Pteropodinae) measurably undersampled. These gaps constitute a notable vulnerability for global health security and will likely be a future barrier to contextualizing the origin of novel zoonotic coronaviruses.

Zoonotic Risk Technology Enters the Viral Emergence Toolkit (preprint)

In light of the urgency raised by the COVID-19 pandemic, global investment in wildlife virology is likely to increase, and new surveillance programs will identify hundreds of novel viruses that might someday pose a threat to humans. Our capacity to identify which viruses are capable of zoonotic emergence depends on the existence of a technology—a machine learning model or other informatic system—that leverages available data on known zoonoses to identify which animal pathogens could someday pose a threat to global health. We synthesize the findings of an interdisciplinary workshop on zoonotic risk technologies to answer the following questions: What are the prerequisites, in terms of open data, equity, and interdisciplinary collaboration, to the development and application of those tools? What effect could the technology have on global health? Who would control that technology, who would have access to it, and who would benefit from it? Would it improve pandemic prevention? Could it create new challenges?