key: cord-0943848-567whmi7
authors: Wang, Jingxuan; Chen, Xiao; Guo, Zihao; Zhao, Shi; Huang, Ziyue; Zhuang, Zian; Lai-yi Wong, Eliza; Chung-Ying Zee, Benny; Ka Chun Chong, Marc; Haitian Wang, Maggie; Kiong Yeoh, Eng
title: Superspreading and heterogeneity in transmission of SARS, MERS, and COVID-19: a systematic review
date: 2021-09-01
journal: Comput Struct Biotechnol J
DOI: 10.1016/j.csbj.2021.08.045
sha: 2be445111d9e88fe108635f7efa9bc3112cb39b1
doc_id: 943848
cord_uid: 567whmi7

BACKGROUND: Severe acute respiratory syndrome (SARS), Middle East respiratory syndrome (MERS), and coronavirus disease 2019 (COVID-19) have caused substantial public health burdens and global health threats. Understanding the superspreading potential of a virus is important for characterizing transmission patterns and informing strategic decision-making in disease control. This systematic review aimed to summarize the existing evidence on superspreading features and to compare the heterogeneity in transmission within and among various coronavirus epidemics of SARS, MERS and COVID-19. METHODS: PubMed, Medline, and Embase databases were extensively searched for original studies on the transmission heterogeneity of SARS, MERS, and COVID-19 published in English between January 1, 2003, and February 10, 2021. After screening the articles, we extracted data pertaining to the estimated dispersion parameter (k) which has been a commonly-used measurement for transmission heterogeneity and superspreading potential. Findings We included a total of 60 estimates of transmission heterogeneity from 26 studies on outbreaks in 22 regions. The majority (90%) of the k estimates were small, with values less than 1, indicating an over-dispersed transmission pattern. The point estimates of k for SARS and MERS ranged from 0.12 to 0.20 and from 0.06 to 2.94, respectively. Among 45 estimates of individual-level transmission heterogeneity for COVID-19 from 17 articles, 91% were derived from Asian regions. The point estimates of k for COVID-19 ranged between 0.1 and 5.0. CONCLUSIONS: We detected a substantial over-dispersed transmission pattern in all three coronaviruses, while the k estimates varied by differences in study design and public health capacity. Our findings suggested that even with a reduced R value, the epidemic still has a high resurgence potential due to transmission heterogeneity.

We detected a substantial over-dispersed transmission pattern in all three coronaviruses, while the k 60 estimates varied by differences in study design and public health capacity. Our findings suggested 61 that even with a reduced R value, the epidemic still has a high resurgence potential due to 62 transmission heterogeneity. In general, a small k value indicates higher heterogeneity in transmission. If k is <1, the NB 96 distribution has an exponential tail [20, 21] , which indicates that the transmission pattern is 97 substantially over-dispersed. Different from the typical phenomenon, an over-dispersed transmission 98 manifest the concept that a small proportion of people generate a large proportion of transmission. 99 This phenomenon has also been described as the '20/80' rule, which stipulates that 20% of the most 158 We consistently used k to represent the superspreading potential. If an article did not explicitly report 159 k, but reported R and the transmission distribution profiles in the form of the '20/80' rule, we 160 generated an estimation of k by using the framework proposed by Endo et al. [14] , which has also 161 been adopted in other studies [15] . For given values of R and the '20/80' rule, the overdispersion 162 parameter k is given by

Here, P is the expected proportion of the most infectious individuals responsible for Q of all 165 transmissions.

represents the NB distribution for secondary cases, with mean R and ( • ) 

The estimates of k for SARS and MERS are shown in Fig. 2 had a lower bound of less than 1 (Fig. 2) . In total, 40% of the k estimates (18/45) were derived from subsets of data that focused on specific 223 age ranges, transmission patterns, and generations (see Supplementary Table S3 for details) . 224 Additionally, all of these subgroup estimates were derived from studies conducted in China and the 225 US. To avoid duplications and overweighting of these two countries, only the 27 estimates using data 226 from the entire study population are shown in Fig. 3 The type of data also influenced the estimate of transmission heterogeneity. The data type used for 259 calculating k mainly fell into two categories: 'cluster size data' and 'transmission pair data'. The 260 former comprises only information on the total cluster size, whereas the latter contains 261 primary-secondary case pairs constructed by contact tracing. In general, we found that studies using 262 cluster size data [14,45] tended to estimate a higher heterogeneity than those using transmission pair 263 data [19, 36, 37, 41, 46, 49] . This tendency was also noted in [15] , in which the authors separately 264 estimated k using both cluster size and transmission pair data collected during the same period. This study has some limitations. First, as the superspreading potential is context-dependent in nature, 317 meta-analysis of k was not performed to avoid misinterpretation. Previous studies also discussed the 318 strong influence of differences in the study period, data type, and statistical methods [46, 49] . Second, 319 although this review provided a comprehensive picture of the superspreading potential of the three 320 coronaviruses, the quality of evidence varied across diseases. For instance, compared to that for In conclusion, this systematic review provided a comprehensive overview of the superspreading 331 potential and transmission heterogeneity of SARS, MERS, and COVID-19. We found that while the 332 k estimates varied across studies due to the differences that arose in public health capacity and 333 aspects of study design, the small k values among majority of the studies demonstrated a substantial 334 over-dispersed transmission pattern in all three coronaviruses. Our findings suggested that even with 335 a reduced R value, the epidemic of these three coronavirus diseases still has a high resurgence 336 potential due to transmission heterogeneity. 

WHO (2019) WHO MERS global summary and assessment of risk

A brief history of R 0 and a recipe for its calculation

Estimating the COVID-19 R number: a bargain with the devil? The Lancet 372

Preliminary case report on the SARS-CoV-2 cluster in the UK

Estimating the asymptomatic proportion 376 of coronavirus disease 2019 (COVID-19) cases on board the Diamond Princess cruise ship

The Korean clusters: how coronavirus cases exploded 379 in South Korean churches and hospitals, in Reauters

Role of disease-associated 381 tolerance in infectious superspreaders

The characteristics of Middle Eastern 384 respiratory syndrome coronavirus transmission dynamics in South Korea

Dimensions of superspreading

Transmission dynamics of the etiological agent of SARS in Hong Kong: 389 impact of public health interventions

Estimating and modelling the transmissibility of Middle East Respiratory Syndrome CoronaVirus 392 during the 2015 outbreak in the Republic of Korea. Influenza and other respiratory viruses 11

Estimating the overdispersion in COVID-19 transmission using outbreak sizes 395 outside China

Clustering and superspreading potential of SARS-CoV-2 infections in Hong Kong

Transmission 400 characteristics of MERS and SARS in the healthcare setting: a comparative study

Exploration of superspreading events in 2015 mers-cov 403 outbreak in korea by branching process models

Superspreading and the effect of 406 individual variation on disease emergence

Inference of 408 person-to-person transmission of COVID-19 reveals hidden super-spreading events during the early 409 outbreak phase

Modelling the large-scale yellow fever outbreak in 411

Angola, and the impact of vaccination

Evidence that coronavirus superspreading is fat-tailed

Heterogeneities in the transmission of infectious agents: implications for the 416 design of control programs

COVID-19 superspreading suggests mitigation by 418 social network modulation

Ebola superspreading

Chopping the tail: How preventing 421 superspreading can help to maintain COVID-19 control

Characterizing superspreading of SARS-CoV-2: from mechanism 423 to measurement

Preferred reporting items for systematic review and meta-analysis protocols (PRISMA-P) 426 2015 statement

Development of a critical appraisal 428 tool to assess the quality of cross-sectional studies (AXIS)

Associations between social jetlag and mental 430 health in young people: A systematic review

Assessing the risk of observing multiple 432 generations of Middle East respiratory syndrome (MERS) cases given an imported case

Clinical and epidemiologic characteristics of spreaders of Middle East respiratory 436 syndrome coronavirus during the 2015 outbreak in Korea

Risk factors for transmission of Middle East respiratory syndrome coronavirus infection 440 during the 2015 outbreak in South Korea

The role of superspreading in Middle East respiratory 442 syndrome coronavirus (MERS-CoV) transmission

A dynamic compartmental model for the Middle East 444 respiratory syndrome outbreak in the Republic of Korea: a retrospective analysis on control 445 interventions and superspreading events

Epidemiologic parameters of the Middle East 447 respiratory syndrome outbreak in Korea

Transmission heterogeneities, kinetics, and controllability of SARS-CoV-2

Low 451 dispersion in the infectiousness of COVID-19 cases implies difficulty in control

Superspreading in early transmissions of COVID-19 in Indonesia

Pattern of early human-to-human transmission of Wuhan

A single holiday was the turning point of the 458

COVID-19 policy of Israel

Descriptive epidemiology of SARS-CoV-2 infection in Karnataka state, South India: 461 Transmission dynamics of symptomatic vs. asymptomatic infections

Inferring super-spreading from transmission clusters of COVID-19 in Hong Kong

Characterizing 466 superspreading events and age-specific infectiousness of SARS-CoV-2 transmission in Georgia

USA

Full genome viral sequences inform patterns of SARS-CoV-2 spread into and 470 within Israel

Real-time monitoring 472 the transmission potential of COVID-19 in Singapore

Inferencing superspreading potential using zero-truncated negative binomial model: exemplification 475 with COVID-19

Evaluating transmission heterogeneity and 477 super-spreading event of COVID-19 in a metropolis of China

Epidemiology and transmission of COVID-19 in 391 cases and 1286 of their close contacts in 481

China: a retrospective cohort study

Effective control of SARS-CoV-2 transmission in Wanzhou, China

Epidemiology and clinical 485 characteristics of 43 COVID-19 patients in Weifang, China. Annals of palliative medicine 486 51. Lloyd-Smith JO (2007) Maximum likelihood estimation of the negative binomial dispersion 487 parameter for highly overdispersed data, with applications to infectious diseases

Why many countries failed at COVID contact-tracing-but some got it right. 489 Nature

The importance of contact tracing in Singapore and the role technology plays

World 493 Health Organization, 494 55. Endo A (2020) Implication of backward contact tracing in the presence of overdispersed 495 transmission in COVID-19 outbreaks

Superspreading of SARS-CoV-2 in the USA

Variability of individual infectiousness derived from

Identifying and interrupting superspreading events-implications for 501 control of severe acute respiratory syndrome coronavirus 2

Superspreading sars 504 events

Influenza 506 transmission in households during the 1918 pandemic

Social contacts, vaccination decisions and influenza in Japan

SARS-CoV-2 (COVID-19) superspreader events

Asymptomatic Cases, the Hidden Challenge in Predicting 514 COVID-19 Caseload Increases

Clusters of coronavirus disease in communities

Association of higher MERS-CoV virus load with severe disease and 520 death, Saudi Arabia

SARS-CoV-2 viral load is associated with increased disease 523 severity and mortality

Heterogeneity in 525 transmissibility and shedding SARS-CoV-2 via droplets and aerosols

Temporal dynamics in viral shedding and transmissibility of COVID-19

Evidence of 530 airborne transmission of the severe acute respiratory syndrome virus