The SARS-CoV-2 pandemic: remaining uncertainties in our understanding of the epidemiology and transmission dynamics of the virus, and challenges to be overcome

Abstract

Great progress has been made over the past 18 months in scientific understanding of the biology, epidemiology and pathogenesis of SARS-CoV-2. Extraordinary advances have been made in vaccine development and the execution of clinical trials of possible therapies. However, uncertainties remain, and this review assesses these in the context of virus transmission, epidemiology, control by social distancing measures and mass vaccination and the effect on all of these on emerging variants. We briefly review the current state of the global pandemic, focussing on what is, and what is not, well understood about the parameters that control viral transmission and make up the constituent parts of the basic reproductive number R ₀. Major areas of uncertainty include factors predisposing to asymptomatic infection, the population fraction that is asymptomatic, the infectiousness of asymptomatic compared to symptomatic individuals, the contribution of viral transmission of such individuals and what variables influence this. The duration of immunity post infection and post vaccination is also currently unknown, as is the phenotypic consequences of continual viral evolution and the emergence of many viral variants not just in one location, but globally, given the high connectivity between populations in the modern world. The pattern of spread of new variants is also examined. We review what can be learnt from contact tracing, household studies and whole-genome sequencing, regarding where people acquire infection, and how households are seeded with infection since they constitute a major location for viral transmission. We conclude by discussing the challenges to attaining herd immunity, given the uncertainty in the duration of vaccine-mediated immunity, the threat of continued evolution of the virus as demonstrated by the emergence and rapid spread of the Delta variant, and the logistics of vaccine manufacturing and delivery to achieve universal coverage worldwide. Significantly more support from higher income countries (HIC) is required in low- and middle-income countries over the coming year to ensure the creation of community-wide protection by mass vaccination is a global target, not one just for HIC. Unvaccinated populations create opportunities for viral evolution since the net rate of evolution is directly proportional to the number of cases occurring per unit of time. The unit for assessing success in achieving herd immunity is not any individual country, but the world.

Keywords: COVID-19; SARS-CoV-2; epidemiology; herd immunity; mass vaccination; transmission.

Figure 1.

Radial plot (‘Florence Nightingale’ plot) of excess deaths in England and Wales by year and week number (1–53). Data from the Office for National Statistics (ONS) during the pre-pandemic years 2012–2019 and during the pandemic in England & Wales [15]. Data for 2021 were only recorded up to week 34 at the time of plotting the Figure. Therefore, the deaths in weeks 35–52 in 2021 are zero.

Figure 2.

Patterns in reported SARS-CoV-2 cases per week in Thailand, United Kingdom, India, Brazil, US and China, up to 20 August 2021 illustrating the great heterogeneity in the pattern of the epidemic in different countries (source: Johns Hopkins University, USA [2]).

Figure 3.

The percentage of cases reported in which sequence information of the virus is acquired by country. Only the 20 countries with the highest frequencies are shown (data source: Johns Hopkins University and Medicine Coronavirus Resource Center [30].

Figure 4.

The flow chart of a simple epidemic model for COVID-19 of individual states and pathways representing rates of transfer between states [25]. The top pathway is for the fraction (1 − p) of those infected who are asymptomatic, with no symptoms or non-specific, unrecognized symptoms, who stay in the community and eventually recover—but contribute to transmission. The bottom pathway is for the fraction (p) who have identifiable symptoms who either self-isolate (minimal level of behaviour change, reducing the infection rate by (1 − b)) or get admitted to hospital (mandatory self-isolation/hospitalization, which is assumed to be far more effective, reducing the infection rate by (1 − m)). I₁ is the pre-symptomatic stage, during which individuals are infectious before the p fraction develop symptoms. Two phases are added to produce an Erlang distribution for time to symptoms. No social distancing or lockdown is represented in the diagram. Below the flow chart is the R_t equation that arises from this flow chart. It is defined as an effective reproduction number because self-imposed or mandatory isolation is assumed to take place. In this equation 1/α is the average number of days it takes from symptom onset to isolation. λ is the force of infection term: the rate at which susceptible individuals become infected, which is dependent on the infection rates (expected amount of people an infected person infects per day) for each infectious state in the model. The β term is the infection rate for the pre-symptomatic infectious state, while q₂ and q₃ are the relative increases in infection rate for symptomatic individuals in phases 2 and 3 of infection, respectively. The γ terms define rates of leaving a given state (1/γ is the average duration of stay). See electronic supplementary material, Information for the expression of λ and derivation of R_t.

Figure 5.

Data from Ke et al. [52] describing viraemia in hospitalized patients over time since infection, employing data from Germany in the early stages of the pandemic [42]. The red dots are viral load in the URT and the blue crosses are viral load in the LRT. The dotted horizontal line is the limit of detection. The red and blue lines are fits of the model of viral dynamics within patients. The model assumes target cell limitation and spatial spread of virions within the host. The model was fitted to data points by minimizing the residual sum of squares. For details see Ke et al. [52]. The filled dots represent values at the limit of detection.

Figure 6.

Exposure settings as documented in contact tracing data compiled by Public Health England, August–December 2020. All data show categories where exposure to infected contacts occurred in England by week and do not infer transmission. (a) Contacts by exposure setting. These are the settings where a person who tested positive for SARS-CoV-2 reported that they met with and potentially exposed their contacts (forward contact tracing). Work is ongoing to link contacts to future cases and to determine where transmission occurs. (b) Locations reported by people who tested positive for SARS-CoV-2 as possible exposure settings, defined as locations visited in the 3–7 days prior to symptom onset, or test date if asymptomatic (backward contact tracing). (c) Common locations reported by people who tested positive for SARS-CoV-2. Two or more individuals who tested positive reported the same location, defined by the same postcode, as a possible exposure setting in the 3–7 days prior to symptom onset, or test date if asymptomatic (backward contact tracing). Information on this type of event and the location are recorded but not information on contacts. (d) As (c) but with relative frequency of reported settings on the y-axis.

Figure 7.

Percentages of populations vaccinated against SARS-CoV-2 infection in a selection of countries. The data are only available for countries that report the breakdown of doses administered by first and second doses in absolute numbers (source: Our World in Data. Data as of 19 August 2021 [110]).

Figure 8.

Case reports of SARS-CoV-2 infection, hospitalizations resulting from infection, deaths reported due to COVID-19 and vaccinations per day in the United Kingdom from the start of the epidemic to 8 September 2021 (excepting hospitalizations where data only began being reported in late March 2020; there was only targeted testing for COVID-19 cases up to late May 2020 in the UK, before wider testing gradually became available) (source: Our World Our Data [112]). The data illustrate how the relationship between the three indicators of impact (cases, hospitalizations and deaths) becomes uncoupled following mass vaccination (0.6% of UK population fully COVID-19 vaccinated 10 January 2021; 64% fully vaccinated 5 September 2021 [112]). Note: y-axis for hospital patients, new cases and new deaths are per million. Vaccination numbers are actual number received.