Characterizing transmission dynamics and severity of 2009 H1N1 pandemic influenza in Hong Kong

Abstract

Background: The first influenza pandemic in the 21st century, the past 2009 influenza pandemic (pdmH1N1), was caused by a novel H1N1 influenza virus. The virus was first described in April 2009 and is now believed to emerge from re-assortment of bird, pig and human flu viruses. Although this pandemic was relatively mild compared to the past pandemics, better knowledge about its characteristics in transmission dynamics and severity is still of public health interest in order to better prepare for future pandemics. Data: Clinical surveillance data were obtained from eFlu database maintained by Hong Kong Hospital Authority. Information was extracted from all pdmH1N1 virologically confirmed infections (which were all symptomatic) about their dates of symptom onset, and, if applicable, dates of hospitalization, ICU admission and death. Serological data were obtained from various sources: 1) community cross-sectional serological survey; 2) convalescent serological data (from symptomatic and virologically confirmed infections); and 3) serological response kinetics data (from symptomatic and virologically confirmed infections). These serological data combined described serological responses against pdmH1N1 infections in the Hong Kong population from different aspects. Methods: I constructed an age-structured natural history model to mimic the pdmH1N1 transmission dynamics in Hong Kong. The transmission model was linked to hospitalization and serology in order to match the observed data. Based on all the data comprehensively, characteristic transmission parameters (basic reproductive number R0, mean generation time E(Tg), attack rates etc.) in the model were estimated using likelihood-based statistical inferences by Bayesian inference with Markov chain Monte Carlo (MCMC). Results: I estimated that R0 is 1.37 and E(Tg) is 2.16 days, which are both comparable to seasonal flu. Younger age groups <20 years were found to be more susceptible (2.72 times compared to 20-29 age group) to pdmH1N1 infection but older age groups 30-59 years were less susceptible (0.55 times). School closure reduced 0-12 year olds’ within-age-group transmission effectively during the reactive kindergarten and primary school closure from Jun 10 to Jul 9 by 93%. Summer holidays from Jun 10 to Aug 31 also reduced within-group transmission by 65% and 13% for 0-12 and 13-19 years olds respectively. Estimates of infection hospitalization probabilities ranged from 0.2% to 0.9% across age groups. I found that not all infected individuals would have serological response strong enough to be positive in serological test but younger age groups were more likely to have stronger serological response after infection. Conclusions: Clinical surveillance data have been used to estimate the transmission dynamics of pdmH1N1 in 2009. Here, I combined hospitalization surveillance data with serological data collected throughout the first pandemic wave (April to December 2009) from different sources, which could better characterize the transmission dynamics and severity of pdmH1N1 in Hong Kong. Although further validation is needed, serological surveillance should be considered as a supplementary alternative to clinical surveillance in influenza surveillance.