A comparison of epidemiology and clinical outcomes between influenza A H1N1pdm09 and H3N2 based on multicenter surveillance from 2014 to 2018 in South Korea

Abstract Background After pandemic, A(H1N1)pdm09 is generally known to be associated with younger adults' infection and greater severity than seasonal A(H3N2) but some inconsistences between recent studies exist. Objectives We aimed to compare the epidemiology and clinical outcomes of A(H1N1)pdm09 and A(H3N2) to verify and consolidate about the knowledge of known differences of subtypes. Methods Data were retrospectively collected from the hospital‐based influenza morbidity and mortality surveillance in South Korea in nine tertiary care hospitals, from August 31, 2014, to August 25, 2018. Patients with H1N1pdm09 or H3N2 infection admitted in the emergency room or ward were recruited. Results A total of 1747 patients had influenza A and were divided into two groups those with A(H1N1)pdm09 (n = 240) and those with A(H3N2) (n = 1507). A(H1N1)pdm09 group had younger age (mean age ± standard deviation 50.0 ± 18.8 in H1N1 vs 53.4 ± 21.1 in H3N2, P = .030), lower influenza vaccination (27.9% vs 43.9%, P < .001) and pneumococcal vaccination rates (41.0% vs 51.9%, P < .001), and fewer underlying diseases (67.5% vs 74.0%, P = .035) than the A(H3N2) group. Influenza A subtypes were not associated with pneumonia risk (adjusted odds ratios [AOR] of A(H1N1)pdm09: 0.7 [95% confidence interval [CI]: 0.4‐1.2, P = .172]) and in‐hospital mortality (hazard ratio (HR) of A(H1N1)pdm09: 1.0 (95% CI: 0.3‐3.1, P = .983)). Influenza vaccination reduced in‐hospital mortality in hospitalized patients (HR: 0.3 (95% CI: 0.1‐0.7), P = .005). Conclusions A(H1N1)pdm09 infection was more common in younger patients without significant difference in pneumonia risk and in‐hospital mortality between subtypes. Influenza vaccination was associated with reduced in‐hospital mortality.

The difference in clinical manifestation and outcome between A(H1N1)pdm09 and A(H3N2) has not been clearly defined. Before 2009, several observational data analysis described higher influenza-associated hospitalization rate and mortality in A(H3N2) predominant seasons, but they are generally regarded as insignificant findings. 3,4 After the H1N1 pandemic in 2009, A(H1N1)pdm09 became a major seasonal strain which was reported to be affected more younger groups, occasionally leading to more serious outcomes than previous strains. [5][6][7] This result suggests that individuals aged below 30 years lacked the cross-reactive antibody against A(H1N1) pdm09. 8 Data in the post-pandemic era generally showed that A(H1N1)pdm09 was associated with younger patients and greater severity compared to A(H3N2). [9][10][11][12][13] However, some inconsistencies exist due to heterogeneous setting of each study such as different participated countries, healthcare resources, vaccination rates, and other potential confounders. Therefore, we pursue to reveal more Influenza vaccination reduced in-hospital mortality in hospitalized patients (HR: 0.3 (95% CI: 0.1-0.7), P = .005).
Conclusions: A(H1N1)pdm09 infection was more common in younger patients without significant difference in pneumonia risk and in-hospital mortality between subtypes. Influenza vaccination was associated with reduced in-hospital mortality.
(with or without admission) or admitted to the hospital via outpatient clinic were enrolled when they were agreed to participate in the study.
Written informed consent was obtained before the specimen collection. After enrollment, two viral samples were collected via nasopharyngeal swab. One sample was immediately underwent rapid antigen tests (RAT) using BD Veritor System to diagnose influenza infection.

| Data management and statistical analysis
All categorical clinical data were calculated and analyzed using chisquare and Fisher's exact test. Mann-Whitney's U test was used to compare the mean age and length of hospital stay between two groups, and the results were expressed as median (interquartile range). A P value of <.05 was considered to significant. Potential variables that might be associated to the development of pneumonia during influenza A infection were analyzed by multivariable linear model logistic regression as adjusted odds ratios (ORs). Variables which were significantly associated with pneumonia in the univariable analysis (P < .05) and generally known risk factors of pneumonia (eg, age and underlying diseases) were selected and included in multivariable regression analysis. The association between A(H1N1) pdm09 and A(H3N2) in relation to the overall mortality of hospitalized patients was investigated using a time-dependent Cox regression analysis to obtain the hazard ratios (HRs).

| Clinical outcomes
Clinical outcomes of subtype-confirmed influenza A infection are displayed in Table 3 (Figures 2 and 3). The high influenza immunization rate in senior groups might also be associated with a high pneumococcal immunization rate, which was also introduced as a national immunization program in South Korea.

Subtypes of influenza
Total morbidity and mortality of influenza are generally known to be related to the influenza-bacterial co-infection, superinfection However, the incidence of serious influenza infection requiring hospitalization and the mortality rate were lower in younger patients. 24 All the above factors including age and vaccination history had complex relationships with each other which may confound the study results. We used a logistic regression analysis to compare these factors, and no significant difference was observed in the pneumonia incidence between the A(H1N1)pdm09 and A(H3N2) subtypes. A recent systematic literature review assessed 47 studies and did not observe any significant difference in secondary bacterial pneumonia, ICU admission, and death between the subtypes of influenza A and B. 9 Statistically insignificant results about subtype difference might be associated with sparse number of participants in the study.
However, up to now, results in our study and related researches showed that A(H1N1)pdm09 generally does not seem to be more virulent than A(H3N2).
In the Cox regression analysis, hospitalized patients who received influenza vaccination had low risk of in-hospital mortality, and the age group, sex, subtypes, and pneumococcal vaccination did not show any significant effect on in-hospital mortality. High hazards ratio in older age groups (≥65 years) without statistical significance might be associated with insufficient number of participants and deaths in the analysis. The effectiveness of influenza vaccine is well known. However, there is a controversy regarding whether the vaccine can certainly reduce mortality. 25 Recently, an analysis of high-risk individuals with chronic obstructive pulmonary disease or heart failure showed a benefit of all-cause mortality in the influenza vaccinated group. 26,27 This finding is comparable to our result because all hospitalized participants in our analysis had more than one underlying illness. Moreover, South Korea has a relatively well-organized immunization program with a high vaccination rate; the impact of vaccination decreasing mortality might be more