Lessons from the Past: Comparison of the Disease Burden of the Inuenza A (H1N1) Pandemic 2009-10 and Seasonal Inuenza 2010-2019 in the United States

Abstract


Results
With regard to seasonal in uenza, rates of illnesses and medical visits were highest in infants aged 0-4 years, followed by adults aged 50-64 years. Rates of hospitalizations and deaths evinced a starkly different pattern, both dominated by elderly adults aged 65 and over. Youths aged 0 to 17 years were especially adversely affected by the H1N1 pandemic relative to hospitalizations and mortality compared to seasonal in uenza; but curiously the opposite pattern was observed in elderly adults (aged 65 and older).

Conclusions
The disease burden of the 2009-10 in uenza A pandemic was strikingly unlike that observed in the subsequent in uenza seasons 2010 to 2019, in the United States: the past did not predict the future.

Background
A reasonable approach to control or mitigate the adverse impact of seasonal in uenza and pandemics in the United States would be to focus resources on those subsets of the population most heavily affected by the disease. To that end, we summarize the burden of disease as re ected by illnesses, hospitalizations and mortality associated with the 2009-10 in uenza A (H1N1) pandemic, compared to the subsequent in uenza seasons 2010 to 2019, as compiled by the Centers for Disease Control. Our goal is to identify speci c at-risk populations, for whom public health resources should be marshaled appropriately and equitably.

Methods
We utilized tables prepared by the Centers for Disease Control (CDC) relating to the estimated in uenza disease burden in the United States, from the 2010-11 through the 2019-2020 in uenza seasons. 1 Each seasonal table contains estimates of the numbers of symptomatic illnesses, medical visits, hospitalizations, and deaths attributable to in uenza, by age groups 0-4 years, 5-17 years, 18-64 years, and 65 + years, with associated estimated ranges 2,3 . In addition, estimated rates of these in uenza disease outcomes per 100,000 population in each of these age groups are also provided.
We obtained data relating to the impact of the in uenza A (H1N1) pandemic in the United States during the period April 2009 through April 2010 from CDC investigators. 4 These data consist of estimates of illnesses, hospitalizations, and deaths attributable to in uenza A, along with ranges, by age groups 0-17 years, 18-64 years, and 65 + years. We converted these counts into rates (per 100000) using 2010 population gures from the US Census Bureau 5 . We remark that the CDC has implemented rigorous protocols 1-4 for data tabulation and presentation, furthering credence in their reports.
To ensure comparability between the in uenza A (H1N1) data from 2009-10 and observations in subsequent years, we pooled the age groups 0-4 and 5-17 years in the CDC annual tabulations, and used a random effects beta binomial model 2,6 with linear time trend 7,8 to summarize the weighted average (consensus) disease burden over the period 2010-2020 relating to rates of illnesses, hospitalizations, and deaths for the age groups 0-17 years, 18-64 years, and 65 + years. Calculations were performed in Stata v.14 (StataCorp, College Station, Texas, 2015), as outlined by Guimaraes 9 .

Results
In Fig. 1, we depict the rates of illnesses, medical visits, hospitalizations, and mortality across the age groups 0-4, 5-17, 18-64, and 65 + attributable to in uenza, from the CDC compilations. Rates of illnesses and medical visits are highest in infants age 0-4 years, followed by adults age 50-64 years.
Rates of hospitalizations and deaths evince a starkly different pattern, both dominated by elderly adults age 65 and over.
We show the age-speci c rates of illnesses, hospitalizations, and deaths in Fig. 2, along with summary measures of the rates over 2010-19, from a beta binomial random effects model. We also depict the corresponding rates from the 2009-2010 in uenza A (H1N1) pandemic 4 . Estimated ranges are also shown for the yearly age-speci c rates, and 99% con dence intervals are given for the summary consensus rates.
Lastly, we abstracted in Fig. 3 the rates and ranges from Fig. 2 corresponding to the 2009-10 in uenza A pandemic, the consensus values for 2010-19, and the individual rates from 2011-12 ( a "low impact" u season, from Fig. 1) and 2017-18 (a "high impact" u season, from Fig. 1), to facilitate comparisons across different age groups.

Discussion
Patterns in the annual rates of illnesses, medical visits, hospitalizations, and deaths attributable to seasonal in uenza according to age are revealing (Fig. 1). Clearly, the overall health impact varies by year, but relative differences between age groups are fairly stable. Rates of illnesses and medical visits are highest in infants age 0-4 years, followed by adults age 50-64 years. One might conjecture that infants naive to novel in uenza viruses would be highly susceptible to illness, but recovery would typically be expected. Rates of hospitalizations and deaths evince a starkly different pattern, both dominated by elderly adults age 65 and over. Vulnerability in this age group is probably exacerbated by pre-existing health conditions, so avoidance of initial infection would be a prudent strategy.
How does the in uenza A (H1N1) pandemic of 2009-2010 compare to the subsequent "normal" in uenza seasons? Rates of illnesses during the H1N1 pandemic (2009-10) were about 2.4 times greater than the consensus rates for seasonal in uenza (2010-19) across all age groups. For youths aged 0 to 17, the H1N1 hospitalization rate was about 2.6 times greater than the consensus rate for seasonal in uenza, and the mortality rate was 2.9 times higher. One might thereby infer that in uenza A (H1N1) is intrinsically more severe among youths than seasonal in uenza. Among adults aged 18 to 64, hospitalization rates and mortality rates from H1N1 were 1.3 and 1.4 times greater respectively than the corresponding consensus rates for seasonal in uenza. On the other hand, the hospitalization and mortality rates for elderly adults (aged 65 and older) during the H1N1 pandemic appear anomalous at a mere one-tenth the corresponding consensus rates for seasonal in uenza, and their corresponding ranges seem disproportionately small. In light of subsequent ndings, we conjecture that hospitalizations and deaths (and their spread) in this age cohort during the H1N1 pandemic were underestimated; in particular, one might expect that the hospitalization and mortality rates would be higher than in the younger adults aged 18 to 64. It is therefore doubtful that the in uenza A pandemic of 2009-10 is a suitable model for subsequent in uenza epidemics relative to the experience of elderly adults.
An important limitation of this retrospective analysis is the absence of information on other potential population characteristics and risk factors, such as gender and comorbidities, which likely affect disease morbidity and mortality. And, as noted above, possible under-detection of morbidity and mortality during the H1N1 pandemic in elderly adults might lead to erroneous inferences in highlighting speci c at-risk populations or focusing preventive measures toward them.

Conclusions
The disease burden of the 2009-10 in uenza A pandemic was strikingly unlike that observed in the subsequent in uenza seasons 2010 to 2019, in the United States, though questions remain concerning the validity of some of the 2009-10 pandemic ndings.

Declarations
Availability of data and materials The data relating to the in uenza burden of disease for the years 2010-2019 in the United States are freely available on the Centers for Disease Control website, https://www.cdc.gov/ u/about/burden/index.html .
The data relating to the 2009-10 in uenza A pandemic burden of disease uis also freely available, and were taken directly from reference #4.
Ethics approval and consent to participate Not Applicable Consent for publication Not Applicable

Competing interests
The author declares no competing interests. Funding