Association of antiviral use for influenza among non-severe cases with subsequent hospitalization and mortality

Abstract

Background

Evidence regarding a preventive impact of antiviral medication for influenza on hospitalization or mortality among non-severe cases is scarce. The objective of this study was to assess the association between the use of antiviral medications in non-hospitalized patients with influenza and their subsequent hospitalization and mortality.

Methods

This multi-center retrospective cohort study used TriNetX, a global database to assess the clinical impact of antiviral medications (oseltamivir, laninamivir, zanamivir, peramivir, and baloxavir) for influenza among non-hospitalized patients. The outcomes included hospitalization, emergency department visit, intensive care unit (ICU) admission, and mortality between 3 and 30 days following influenza diagnosis with the 2-day window period between diagnosis and antiviral administration. Propensity score matching with a 1:1 ratio was conducted to estimate the risk ratio (RR).

Results

A total of 1,016,581 patients with influenza were identified. After matching, 426,275 patients from each cohort were compared. The antiviral-treated group was associated with lower risks of hospitalization (RR 0.91; 95% confidence interval 0.85–0.95; p < 0.001); emergency department visits (RR 0.80; 95% CI 0.79–0.82; p < 0.001), ICU admission (RR 0.84; 95% CI 0.73–0.97; p = 0.016); and mortality (RR 0.60; 95% CI 0.47–0.77; p < 0.001). The sensitivity analysis revealed that the statistical significance in hospitalization was maintained by replacing the 2-day window period with 0 days (RR 0.82; 95% CI 0.79–0.84; p < 0.001). The sensitivity analysis limited to those aged 2 to 64 years without any high-risk comorbidities in the antiviral-treated group compared to the antiviral-untreated group showed fewer outcome events for hospitalization (RR 0.76; 95% CI 0.68–0.85; p < 0.001); ED visit (RR 0.68; 95% CI 0.65–0.71; p < 0.001); and ICU admission (RR 0.48; 95% CI 0.30–0.78; p = 0.002).

Conclusions

Antiviral treatment was negatively associated with hospitalization, ICU admission, and mortality. The results need to be interpreted with caution, given limitations.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12916-026-04841-9.

Background

Influenza, a common respiratory infectious disease, causes significant morbidity and mortality [1–3]. Clinical presentations of influenza are variable, ranging from mild symptoms associated with upper respiratory tract infection to pneumonia and secondary complications, including encephalitis, encephalopathy, and cardiovascular events, some of which are fatal [4, 5].

Antiviral medications, including neuraminidase inhibitors such as oseltamivir, laninamivir, zanamivir, and peramivir, and endonuclease inhibitors such as baloxavir, are available for the treatment of influenza [6]. Multiple studies have shown that the early use of antiviral medications for influenza is associated with a shorter duration of symptoms [7–9]. However, the impact of antiviral medications on influenza-related emergency department (ED) visits, hospitalization, intensive care unit (ICU) admission, or deaths remains unclear. A systematic review and meta-analysis showed that the use of antiviral medications was associated with a more rapid improvement in symptoms; however, few pieces of evidence were found on reduced hospitalization or death [10]. Another recent systematic review and meta-analysis concluded that all seven antiviral medications investigated for treating non-severe influenza had little or no impact on mortality, with high certainty of evidence [11]. However, there was no mortality or only a single mortality event in each arm across the seven comparisons (i.e., antiviral treatment vs. control groups) in this systematic review, which could raise concerns about the lack of statistical power to detect a true impact on mortality reduction, if it exists. Similarly, this systematic review reported that antiviral medications used for treating non-severe influenza, except for baloxavir, had little to no impact on hospitalization.

Recently, real-world data have been used in clinical research, especially for relatively rare outcomes [12–14]. Given the significant global disease burden of influenza, evaluating whether antiviral medications can reduce ED visits, hospitalization, and mortality is crucial. The objective of this study was to assess the association between the use of antiviral medications in non-hospitalized patients with influenza and their subsequent hospitalization and mortality using a global database.

Methods

This multi-center retrospective cohort study used TriNetX to evaluate the clinical impact of antiviral medications on non-severe patients with influenza. TriNetX is a multi-center global network with aggregated de-identified electronic health records, both inpatient and outpatient, diagnostic codes, laboratory data, and medications of more than 177 million patients from 150 healthcare organizations as of September 2025. The TriNetX database allows for cohort generation and matching, as well as analysis of outcome events, while maintaining the anonymity of aggregate data. While the TriNetX platform contains individual data and calculates effect sizes using propensity score matching (PSM) within the platform, researchers do not have access to the individual data but command the platform on statistical implementations and obtain the aggregated results.

The data extraction period was between August 1, 2014 (the year of data inception in TriNetX) and July 31, 2025. A search run using the TriNetX global database was performed on January 19, 2026. We included patients with a confirmed diagnosis of influenza who did not require hospitalization within two days of diagnosis. In this study, a confirmed influenza diagnosis was defined as a laboratory confirmation (i.e., positive results for nucleic acid or antigen tests) and the presence of influenza-related diagnoses as defined by the International Classification of Diseases 10th Revision [ICD-10] codes. The exclusion criteria were the presence of any outcome event or secondary complications of influenza (i.e., pneumonia, encephalitis/encephalopathy, cardiovascular events, or sepsis) from 7 days before to 2 days after the diagnosis. Two cohorts were generated in this study. The study cohort represented non-hospitalized patients without these influenza-associated complications around the peri-diagnosis period. The peri-diagnosis period was set to mitigate the potential influence of these pre-existed events and complications on the outcome evaluation. The antiviral-treated group comprised patients with a confirmed influenza diagnosis who received antiviral treatment within 2 days of diagnosis, whereas the antiviral-untreated group comprised patients with a confirmed influenza diagnosis who did not receive any antiviral treatment within 2 days of diagnosis. Antivirals used included five medications (i.e., oseltamivir, laninamivir, zanamivir, peramivir, and baloxavir). Other antivirals (e.g., umifenovir, favipiravir, amantadine, and rimantadine) were excluded in this study given their extremely low numbers of administrations for influenza [15]. The details of the search run are provided in Additional file 1: Tables S1 and S2.

The primary outcome in this study was all-cause hospitalization (ward), and the secondary outcomes included all-cause emergency department (ED) visit, all-cause ICU admission, and all-cause mortality (Additional file 1: Table S3). The outcome follow-up period was between 3 and 30 days after the influenza diagnosis. Thus, antiviral exposure was defined as antiviral prescription within 2 days after diagnosis, while outcome events are excluded through 2 days after diagnosis. Consequently, this study evaluated outcomes occurring from 3 days after diagnosis onward among individuals who remained event-free in the peri-diagnosis period. More than one type of outcome events (e.g., hospitalization and ICU admission) which occurred within the follow-up period in a single individual were counted for each outcome.

Given that only aggregate de-identified data were used in this study, written informed consent was not required. The Institutional Review Board at the Nara Prefecture General Medical Center reviewed the study and approved the study to be submitted to a peer-reviewed journal (Approval No. 1003).

Statistical analysis

Propensity score matching (PSM) was conducted using a 1:1 greedy nearest-neighbor method with a caliper width of 0.1 pooled standard deviations with the difference between propensity scores below 0.1. According to reported risk factors for severe influenza, variables in PSM included age, sex, ethnicity, the presence of high risk of socioeconomic condition, receipt of influenza vaccination, and medical comorbidities as presented in Table 1 [16]. History of these medical comorbidities was extracted by ICD-10 codes. Following PSM, the risk ratio (RR) with a 95% confidence interval (CI) was calculated for each outcome using χ² tests. The absolute risk reduction (ARR) was estimated by subtracting the risk in the antiviral-treated group from that in the antiviral-untreated group. Statistical significance was set at a two-sided p-value of <0.05. Statistical analyses were performed using the TriNetX platform and Microsoft Excel 2019.

Table 1.

Patient characteristics before and after propensity score matching

	Before propensity matching			After propensity matching
	Antiviral-treated group (N = 517,785)	Antiviral-untreated group (N = 466,350)	Standardized mean difference	Antiviral-treated group (N = 426,275)	Antiviral-untreated group (N = 426,275)	Standardized mean difference
Age group
< 2 years	39,963 (7.7%)	29,156 (6.3%)	0.058	29,112 (6.8%)	29,005 (6.8%)	0.001
2–4 years	52,546 (10.1%)	53,965 (11.6%)	0.046	45,903 (10.8%)	46,557 (10.9%)	0.005
5–17 years	166,382 (32.1%)	168,616 (36.2%)	0.085	145,915 (34.2%)	146,535 (34.4%)	0.003
18–49 years	162,962 (31.5%)	144,773 (31.0%)	0.009	135,856 (31.9%)	135,702 (31.8%)	0.001
50–64 years	54,875 (10.6%)	40,835 (8.8%)	0.062	40,419 (9.5%)	39,887 (9.4%)	0.004
65–79 years	32,511 (6.3%)	22,923 (4.9%)	0.059	22,973 (5.4%)	22,596 (5.3%)	0.004
≥ 80 years	8546 (1.7%)	6082 (1.3%)	0.029	6097 (1.4%)	5993 (1.4%)	0.002
Gender
Females	273,308 (52.8%)	249,916 (53.6%)	0.016	228,512 (53.6%)	226,710 (53.2%)	0.008
Race/ethnicity
White	264,370 (51.1%)	256,492 (55.0%)	0.079	233,984 (54.9%)	231,481 (54.3%)	0.012
Black or African American	90,705 (17.5%)	94,490 (20.3%)	0.070	81,897 (19.2%)	82,166 (19.3%)	0.002
Asian	38,403 (7.4%)	17,751 (3.8%)	0.157	18,064 (4.2%)	17,747 (4.2%)	0.004
Others	29,375 (5.7%)	27,755 (6.0%)	0.012	25,021 (5.9%)	25,556 (6.0%)	0.005
Socioeconomic risk factor a	17,532 (3.4%)	23,954 (5.1%)	0.087	16,828 (3.9%)	16,838 (4.0%)	< 0.001
Receipt of influenza vaccination	147,398 (28.5%)	109,663 (23.5%)	0.113	108,432 (25.4%)	106,471 (25.0%)	0.011
Comorbidity
Endocrinal, nutritional and metabolic diseases	161,822 (31.3%)	133,221 (28.6%)	0.059	124,351 (29.2%)	124,466 (29.2%)	0.001
Mental, behavioral and neurodevelopmental disorders	133,937 (25.9%)	136,022 (29.2%)	0.074	119,351 (28.0%)	117,089 (27.5%)	0.012
Diseases of the nervous system	118,989 (23.0%)	111,330 (23.9%)	0.021	100,592 (23.6%)	98,564 (23.1%)	0.011
Pregnancy, childbirth, and puerperium	26,321 (5.1%)	27,858 (6.0%)	0.039	23,754 (5.6%)	23,548 (5.5%)	0.002
Diseases of the circulatory system	104,915 (20.3%)	84,996 (18.2%)	0.052	81,379 (19.1%)	79,691 (18.7%)	0.010
Chronic lower respiratory tract diseases	108,904 (21.0%)	86,797 (18.6%)	0.061	82,525 (19.4%)	82,499 (19.4%)	< 0.001
Diseases of liver	19,288 (3.7%)	15,871 (3.4%)	0.017	14,669 (3.4%)	14,799 (3.5%)	0.002
Acute kidney failure and chronic kidney disease	16,472 (3.2%)	12,753 (2.7%)	0.026	12,308 (2.9%)	12,120 (2.8%)	0.003
Neoplasms	57,570 (11.1%)	47,667 (10.2%)	0.029	45,639 (10.7%)	44,312 (10.4%)	0.010
Human immunodeficiency virus disease	1291 (0.2%)	0.850 (0.2%)	0.014	0.847 (0.2%)	0.835 (0.2%)	0.001
Long term use of steroids	10,344 (2.0%)	10,020 (2.1%)	0.011	8818 (2.1%)	8845 (2.1%)	< 0.001
Long term use of aspirin	12,202 (2.4%)	10,726 (2.3%)	0.004	10,165 (2.4%)	9877 (2.3%)	0.004
Immunocompromising conditions	6335 (1.2%)	4258 (0.9%)	0.030	4246 (1.0%)	4163 (1.0%)	0.002

^aDefined by ICD-10 code Z55-65: Persons with potential health hazards related to socioeconomic and psychosocial circumstances

Subgroup analyses were conducted based on age group (children vs. adults); region (North America, Latin America, Europe/Middle East/Africa, and Asia–Pacific); influenza type (A vs. B); antiviral type (such as neuraminidase inhibitors and baloxavir); and study period (2014–2020 vs 2020–2025).

Sensitivity analyses were conducted by replacing the 2-day window period from confirming influenza diagnosis to initiating antiviral treatment, excluding those with pre-existing outcome events or complications and initiating outcome follow-up (since next day) (1) with 0 days (the same day) to investigate the potential impact of the time difference in the index event between the antiviral-treated and untreated groups and (2) with 5 days to explore whether the clinical impact of antivirals, if any, was maintained by including cases initiating antiviral treatment at a later stage of illness. In the sensitivity analysis with the 0-day window period, further analyses were implemented by limiting participants with documented symptoms associated with respiratory infection (i.e., fever, headache, malaise or fatigue, throat pain, cough, nasal congestion, or dyspnea) within 2 days before diagnosis. This analysis was also supplemented by the presence of dyspnea (suggesting more severe condition compared with other symptoms) to investigate whether the antiviral impact was affected by severity of condition. An additional sensitivity analysis was conducted by limiting patients to those aged 2–64 years (non-high-risk age group) without any of the high-risk comorbidities described above. Propensity score matching was conducted at each subgroup and sensitivity analysis to ensure < 0.1 standardized mean difference for all confounding factors in all analyses.

Results

In total, 1,016,581 patients with influenza were identified, of whom 32,340 and 116 patients were excluded due to meeting severe criteria in the peri-diagnosis period and the use of antiviral medication not investigated in this study (Fig. 1). Of the remaining patients, 517,785 received antiviral treatment for influenza within 2 days of diagnosis, and 466,350 did not receive any antiviral treatment for influenza within 2 days of diagnosis. PSM was used to compare 426,275 patients from each cohort.

Fig. 1.

Cohort selection flowchart

Before PSM, children (< 18 years old) accounted for 50.0% and 54.0% of the antiviral-treated and antiviral-untreated groups, respectively, and older adults comprised 7.9% and 6.2% of each group, respectively (Table 1). The prevalence of endocrinal, nutritional, and metabolic diseases, diseases of the circulatory system, chronic lower respiratory diseases, liver diseases, kidney diseases, neoplasms, human immunodeficiency virus disease, and immunocompromising conditions was higher in the antiviral-treated group compared to the antiviral-untreated group. The outcome events, effect sizes, and overlap cases for each outcome before PSM were presented in Additional file 1: Tables S4 and Table S5.

After PSM, the standardized differences of all defined covariates between the two groups were < 0.1. After PSM, hospitalization was observed in 6579 (1.5%) and 7249 (1.7%) of patients in the antiviral-treated and -untreated groups, respectively (RR 0.91; 95% CI 0.88–0.94; p < 0.001 and ARR/1000 cases 1.56; 95% CI 1.05–2.07) (Table 2). All secondary outcomes were fewer in the antiviral-treated compared to the antiviral-untreated group, and the results were as follows: RR 0.80; 95% CI 0.79–0.82; p < 0.001 and ARR 8.69; 95% CI 7.94–9.44 for ED visits, RR of 0.84; 95% CI 0.73–0.97; p = 0.016 and ARR 0.16; 95% CI 0.03–0.27 for ICU admission, and RR 0.60; 95% CI 0.47–0.77; p < 0.001 and ARR 0.15; 95% CI 0.09–0.20 for mortality.

Table 2.

Primary and secondary outcomes of antiviral treatments in all included patients

	Antiviral-treated (N = 426,275)	Antiviral-untreated (N = 426,275)	RR [95% CI]	p-value	ARR [95% CI] (/1000 cases)
Primary outcome
Hospitalization	6579 (1.54%)	7249 (1.70%)	0.91 [0.88, 0.94]	< 0.001	1.56 [1.05, 2.07]
Secondary outcome
ED visit	15,087 (3.54%)	18,799 (4.41%)	0.80 [0.79, 0.82]	< 0.001	8.69 [7.94, 9.44]
ICU admission	364 (0.09%)	432 (0.10%)	0.84 [0.73, 0.97]	0.016	0.16 [0.03, 0.27]
Mortality	98 (0.02%)	163 (0.04%)	0.60 [0.47, 0.77]	< 0.001	0.15 [0.09, 0.20]

ARR absolute risk reduction, CI confidence interval, ED emergency department, ICU intensive care unit, RR risk ratio

The results of subgroup analyses are presented in Fig. 2 and in Additional file 1: Tables S6–S10. Fewer hospitalizations were observed in the antiviral-treated group for both adults and children (RR 0.91; 95% CI 0.87–0.94; p < 0.001 for adults and RR 0.79; 95% CI 0.74–0.85; p < 0.001 for children). Regarding secondary outcomes, the antiviral-treated group had significantly fewer ED visits (RR 0.81; 95% CI 0.79–0.83; p < 0.001), ICU admissions (RR 0.80; 95% CI 0.68–0.94; p = 0.001), and mortality (RR 0.67; 95% CI 0.51–0.89; p = 0.005) in adults, as well as significantly fewer ED visits (RR 0.77; 95% CI 0.75–0.80; p < 0.001) in children.

Fig. 2.

Risk ratios of hospitalization in antiviral-treated compared to antiviral-untreated group. ICU: intensive care unit

In the subgroup analysis based on region, a significant reduction in primary outcome events was maintained in North America (RR 0.88: 95% CI 0.85–0.91; p < 0.001), and in the Asia–Pacific region (RR 0.46; 95% CI 0.35–0.60; p < 0.001). Sufficient sample size could not be obtained in Europe, the Middle East and Africa as well as Latin America. The statistical significance of the primary outcome was maintained in the type A influenza group (RR 0.89; 95% CI 0.86–0.93; p < 0.001), whereas no statistical significance was observed in the type B influenza group (RR 0.94 [CI 0.87–1.01]; p = 0.098). Significantly fewer primary outcome events were not observed in 2014–2020 (RR 0.93; [CI 0.86–1.00]; p = 0.054), but in 2020–2025 (RR 0.90; 95% CI 0.86–0.93; p < 0.001). Subgroup analysis based on each antiviral medication revealed that oseltamivir (RR 0.91; 95% CI 0.88–0.94; p < 0.001) and baloxavir (RR 0.28; 95% CI 0.18–0.44; p < 0.001) were significantly associated with fewer primary outcome events. Oseltamivir was also associated with fewer events in all secondary outcomes. On the other hand, no significant reduction in ICU admissions or mortality was observed with baloxavir. No significant reduction for any primary or secondary outcomes was observed in zanamivir, laninamivir, or peramivir. Compared with the neuraminidase inhibitors group, the baloxavir group had fewer primary outcome events (RR 0.47; 95% CI 0.29–0.76; p = 0.001).

Sensitivity analysis (Additional file 1: Tables S11 and Table S12) revealed that the statistical significance in the primary outcomes was maintained by replacing the 2-day window period with 0 days (RR 0.82; 95% CI 0.79–0.84; p < 0.001) and 5 days (RR 0.96; 95% CI 0.92–0.99; p = 0.022). With the 0-day window period, RR of the primary outcome based on the symptoms’ documentation within 2 days of diagnosis was as follows: RR 0.84; 95% CI 0.81–0.88; p < 0.001 by limiting to participants with any documented symptoms associated with influenza, RR 0.98; 95% CI 0.88–1.09; p = 0.695 by limiting to documented dyspnea, and RR 0.83; 95% CI 0.79–0.87; p < 0.001 by limiting to documented symptoms associated with influenza except for dyspnea, respectively. Sensitivity analysis limited to those aged 2 to 64 years without any high-risk comorbidities in the antiviral-treated group compared to the antiviral-untreated group showed fewer outcome events for hospitalization (RR 0.76; 95% CI 0.68–0.85; p < 0.001); ED visit (RR 0.68; 95% CI 0.65–0.71; p < 0.001); and ICU admission (RR 0.48; 95% CI 0.30–0.78; p = 0.002).

Discussion

This multi-center study using real-world data revealed that antiviral treatment was associated with fewer hospitalizations, ED visits, ICU admissions, and mortality in cases with non-severe influenza. One of the advantages of using multi-center real-world data is the large sample size, which helps detect rare events, such as mortality in patients with influenza who did not require hospitalization at their initial presentation. Using multi-center global data, we observed that antiviral use for non-severe influenza cases had prevented ED visits, hospitalizations, and deaths.

The association of antiviral treatment for influenza on fewer ED visits, hospitalization, and ICU admission was maintained among patients without any high-risk factors in our study. The impact on shortening the symptoms associated with influenza among healthy children and adults has been reported [7, 10, 17]. However, evidence supporting the protective effect of antiviral use in reducing hospitalization and ICU admission among healthy populations with non-severe influenza is limited. Our findings revealed fewer hospitalizations and mortality with the use of antivirals among healthy children and adults, which might have important implications on indicating antiviral treatments for influenza in this population. Other important factors guiding treatment recommendations include the adverse effects of antivirals and the cost-effectiveness of treating healthy non-severe cases of influenza.

A previous systematic review and meta-analysis showed a trend of reduction in hospitalization (but not in mortality) when using baloxavir in treating non-severe influenza cases, although the meta-analysis could not reveal a statistical significance due to an insufficient sample size [11]. The widespread use of baloxavir has raised concerns about the potential emergence of antiviral resistance. Some studies have reported variants of the influenza virus induced by baloxavir treatment [18, 19]. The acquisition of baloxavir resistance might be associated with the diminishing clinical impact of baloxavir [20]. The results based on the type of influenza medication, with the exception of those for oseltamivir, should be interpreted with caution due to the small sample size in each medication group. In addition, differences in the routes of administration of each antiviral might have affected our results. For example, peramivir is administered intravenously, and patients requiring intravenous administration are more likely to have had severe conditions (e.g., dehydration), even if they did not require hospitalization at the time of diagnosis or antiviral administration. This potential difference in disease severity at the time of antiviral administration could not be adjusted for using the propensity matching method employed in this study.

In our sensitivity analysis based on influenza type (type A vs. B), antiviral use was associated with a significant reduction in all primary and secondary outcomes observed in the influenza type A group. However, no significant association, except for a reduction in ED visits, was observed in the influenza type B group. Some studies reported that oseltamivir, the most frequently used antiviral for influenza, was less effective for symptom improvement in type B compared to type A influenza cases, whereas others reported similar improvement in symptoms for both types of influenza [21, 22].

Because documentation of symptoms and vaccination was inadequate in the database using electronic health records, additional cautions are required when interpreting the results. Inadequate documentation of vaccinations is common in many healthcare organizations worldwide, particularly when vaccinations are administered across various healthcare organizations [23, 24]. The development of a global system to accurately determine vaccination history is crucial for evaluating the future impact of vaccination.

This study had some limitations. First, we could not assess the potential effects of unmeasured confounders in the PSM. Although we adjusted for known factors associated with severe influenza outcomes, there may have been other unmeasured factors that could not be adjusted for. Second, information on symptom onset was insufficiently documented in the TriNetX database. Therefore, we could not evaluate the effect of the timing of antiviral initiation on symptom onset. Third, we could not evaluate causality between influenza and outcome event. Other unmeasured confounders could have influenced the observed results. Future research may include the impact evaluation of antiviral medications on influenza-attributable severe outcomes. Fourth, because this study evaluated the outcome events from 3 days after diagnosis (not symptom onset) onward by excluding outcomes which occurred within 2 days after diagnosis, the potential introduction of selection bias, including immortal time bias, needs to be recognized [25]. The sensitivity analysis by removing the 2-day period provided similar results to the base-case analysis, which indicated that the impact of immortal time bias could be small. However, the complete removal of the potential immortal time bias in the base-case analysis would be more ideal. Fifth, data regarding dosage, frequency, and duration of antiviral treatment were not available, which could affect the observed study results. Sixth, while we did not exclude participants with missing covariates in this study, misclassification or under-ascertainment of medical comorbidities should be recognized given that these comorbidities were captured by ICD-10 coding [26]. Seventh, although we provided regional results, sample sizes in Europe, Middle East, Africa, and Latin America were small. In addition, we could not evaluate sample size of each country contributing to the data in Asia and Pacific. These limit the generalizability of the study. Eighth, the TriNetX built-in platform only allows for the uniform statistical analysis. For example, McNemar’s test and conditional logistic regression, which are often used for matched data analysis, were not available within the TriNetX platform. Thus, the observed associations may be influenced by immortal time bias and unmeasured confounding, and they should not be over-interpreted as causal effects

Conclusions

Using real-world global data, this study showed that the use of antiviral medications for influenza was negatively associated with hospitalization, ED visits, ICU admissions, and mortality among patients with non-severe influenza. Although the observed results were relatively robust, irrespective of the presence of high-risk factors age group (i.e., children and adults) or presence of documented symptoms, the results of this study need to be interpreted with caution, given the number of limitations, including the potential impact of immortal time bias and unmeasured confounders.

Abbreviations

CI

Confidence interval

ED

Emergency department

ICU

Intensive care unit

NNT

Number needed to treat

RR

Risk ratio

Authors’ contributions

T.K., S.T., and S.Y contributed to the conceptualization, study design, data interpretation and visualization. T.K. was involved with data extraction, data analysis, data curation, and original draft writing. S.T., and S.Y. were involved with visualization and critical revision of the manuscript for important intellectual content. All authors read and approved the final manuscript.

Funding

The study was funded by JSPS KAKENHI (Grant number JP23K27865).

Data availability

The data that support the findings of this study are available from the TriNetX Network. (https://trinetx.com). Codes to build cohorts, define outcomes, and run propensity score matching in the TriNetX Network is provided in the Additional file 1. Individual-level data are not available due to the regulation of TriNetX. Additional analyses are available upon reasonable request to the corresponding author. Data access to the TriNetX Network is restricted to protect patient health information, but researchers requesting the access to the data network may apply through the TriNetX Network. The approval process may take several weeks.

Declarations

Ethics approval and consent to participate

Given that only aggregate de-identified data were used in this study, written informed consent was not required. The institutional review board at the Nara Prefecture General Medical Center reviewed the study and approved the study to be submitted to a peer-reviewed journal (Approval No. 1003).

Consent for publication

Not applicable.

Competing interests

Dr Kitano received grants from Japan Foundation for Pediatric Research and bioMérieux. Dr Tsuzuki received grants from Japan Society for the Promotion of Science, Japan Science and Technology Agency, Ministry of Health, Labour and Welfare, and Japan Institute for Health Security and honoraria from Gilead Sciences. Dr Yoshida received a grant from Tauns Laboratories and Mizuho Medy.