Change in patterns of hospitalization for influenza during COVID‐19 surges

Abstract Background Hospitalization due to influenza has been stable in recent years. In March 2020, New York was an epicenter for coronavirus disease 2019 (COVID‐19). Because influenza and COVID‐19 present similarly, there were serious concerns that coinfection of these viruses would burden the healthcare system. We compared incidence and outcomes of patients hospitalized with influenza before and during COVID‐19 (seasons 2017–2021). Methods We conducted a retrospective study evaluating hospitalized patients with influenza. Four influenza seasons were evaluated, 2017–2021, pre‐ and during COVID‐19 pandemic. We compared incidence of influenza and clinical outcomes across the seasons. Results We found 412 patients hospitalized due to influenza in the study period; 394 had influenza, and 18 had both influenza and COVID‐19 infections. Demographics across the four influenza seasons were comparable; the cohort was predominantly female (61%) and had an average age of 60 years old. Comorbid conditions were common. No outcome differences were found for patients with influenza when comparing influenza seasons prior to and during the COVID‐19 pandemic. The mortality for the entire cohort was 6.5%. During the COVID‐19 pandemic, there were 18 (4.4%) influenza patients coinfected with COVID‐19 and 32 (7.8%) patients with bacterial super infection. Predictors of mortality in patients with influenza included presence of shock, heart failure, bacterial pneumonia, and use of mechanical ventilation. Coinfection with COVID‐19 did not increase mortality. Conclusion We observed a significant decrease in the incidence of hospitalization due to influenza during the COVID‐19 pandemic. Clinical presentations and outcomes for patients with influenza remain stable. Being aware of possible increased mortality for patients with both influenza and bacterial pneumonia is important. Although coinfection with COVID‐19 did not increase mortality in influenza patients, identifying the specific virus responsible for infections has major therapeutic implications.

We compared outcomes across the four influenza seasons. For the last two seasons, which coincided with the COVID-19 pandemic, we compared outcomes for patients with only influenza and for patients with influenza and COVID-19 infections. COVID-19 infections were identified using the Roche Cobas SARS-CoV-2 test.
The primary outcome was incidence of influenza in the four time periods. Secondary outcomes were needed for critical care admission, length of hospital stay, development of shock, use of mechanical ventilation, and mortality from all causes during hospital admission.
Ethics approval: Our study protocol was approved by the Institutional Review Board (approval number 004082105). We followed the amended Declaration of Helsinki.

| Data abstraction
All data were retrospectively extracted from medical records, including demographic, clinical, and laboratory information. Radiology reports of chest radiographs (CXR) were used to collect radiological findings. We completed a multivariate regression analysis with mortality as the predicted variable and the parameters as the explanatory vari-  Table 3 shows pertinent laboratory and radiological investigations. We detected no significant differences between the influenza seasons for various biomarkers including D-dimers, serum creatinine, platelets, and pro-BNP. We did not observe significant differences in ejection fraction between the seasons. Finally, we did not observe sig-   length of hospital stay in the 2020-2021 season; these data need to be interpreted with caution as the total number of influenza cases was low and some of these patients were also infected with COVID-19.

| Statistical analysis
Predictors of mortality in patients with influenza are shown in Table 5. Admission to ICU, shock, need for mechanical ventilation, length of hospital stay, low ejection fraction, presence of secondary bacterial pneumonia, and unilateral infiltrates on chest radiograph    were predictive of mortality. Normal radiograph was associated with lower mortality.

| DISCUSSION
Our study is consistent with data reporting a decrease in influenza Age; comorbid illnesses such as atrial fibrillation, acute heart failure, pneumonia, malnutrition, and admission to intensive care; shock; and respiratory failure are predictors of mortality for patients with influenza. [8][9][10][11] Additionally, certain laboratory findings, including a change in serum D-dimer values, can serve as a predictor of mortality. 12 An association between low platelets and mortality has been reported in avian flu. 13 Our study reveals no association between comorbid conditions and mortality; however, our data support prior studies showing an increased influenza mortality in patients requiring ICU care, presence of shock, and need for mechanical ventilation. 8 The presence of COVID-19 pandemic and overlap of COVID-19 in the two last seasons did not change the association of mortality with ICU admission, shock, and need for mechanical ventilation.
Comparison of clinical outcomes in critically ill patients hospitalized with COVID-19 or influenza has been published. 14 Stowe et al. reported that co-circulation of influenza and COVID-19 viruses could have a significant impact on morbidity, mortality, and health service demand. 15 Our study contributes new and unique data to the field because we evaluated the last four seasons of influenza in our innercity hospital including an overlap of two COVID-19 waves. Furthermore, we evaluated COVID-19 as a secondary infection and confounders of mortality. The presence of COVID-19 coinfection does not seem to impact mortality in our patient population; we hypothesize that this could be due to the few cases of coinfections in our cohort.
Influenza coinfections have been reported with bacterial infections; rates range from 2% to 65% with Streptococcus pneumonia been the most common bacterial pathogen. 16 Influenza infection is associated with increased hospital morbidity and mortality in patients with heart failure. 22,23 Our study revealed a clear association between ejection fraction and mortality in patients with heart failure.
The presence of chest radiological abnormality is predictive of worse outcomes in patients with influenza; pulmonary infiltrate, consolidation, and effusion are associated with the worst outcomes. 24 Aviram et al. reported that initial chest radiography may have significance in helping predict clinical outcome, but normal initial radiographs cannot exclude adverse outcome. 25 28 This demographic also has lower rates of influenza vaccination. 29 To the best of our knowledge, limited research has been done examining the outcomes of coinfections between influenza and COVID-19. Limitations of the present study include the retrospective nature of the study as well as that it was performed in a single center. Thus, further research is necessary to support widespread changes in practice.

| CONCLUSION
Influenza infections have shown a downward trend after the emergence of COVID-19. Predictors of poor outcomes and mortality in hospitalized patients with influenza remain basically unchanged throughout seasons in our inner-city community. It seems that the risk of coinfection with another virus like COVID-19 is small and does not increase mortality.
We anticipate that COVID-19 will remain a significant pathogen in the upcoming influenza seasons. Early identification of viral etiology has major therapeutic implications; antiviral treatments are available for influenza patients, and selected patients with COVID-19 could benefit from steroids.
Patients with dual infection will present a management challenge for the clinicians. However, with the limited available data on coinfections, it is too early to suggest specific treatments. Prevention of infection will continue playing a significant role in our high-risk population. As we approach the coldest months of the year, vaccinating for COVID-19 and influenza as well as continuing with some nonpharmacological measures is of the utmost importance.

CONFLICT OF INTEREST
There are no conflicts of interest declared from all the authors. The authors have no disclosure.

PEER REVIEW
The peer review history for this article is available at https://publons. com/publon/10.1111/irv.12900.

DATA AVAILABILITY STATEMENT
The data that support the findings of this study are available from the corresponding author upon reasonable request and pending additional ethical approval.