Seasonal influenza infections cause significant morbidity and mortality annually in the United States with an estimated 290,000 hospitalizations and 36,000 deaths on average each year.1,2 The exact rates of influenza-associated hospitalizations and death vary widely from season to season, depending on characteristics of circulating virus strains, availability and uptake of vaccine, and relatedness between vaccine and circulating strains.3 In particular, rates of hospitalization and death are higher during seasons when influenza A (H3N2) viruses predominate.1,2,4,5 The 2014–2015 season, in which H3N2 comprised 99.6% of influenza A specimens, was moderately severe among all age groups and was especially severe among individuals 65 years or older with an influenza-related hospitalization rate of 65.5 per 100,000 individuals of all age groups and 322.8 per 100,000 individuals 65 years or older.4 This is considerably higher than the last H3N2 predominant season, 2012–2013 (44.3 per 100,000 individuals of all age groups and 183.2 per 100,000 individuals ≥65 years old).4,6 Furthermore, the rate of influenza-related hospitalizations among individuals 65 years or older was significantly higher compared with the previous 4 years, where the incidence range was 30.2 to 183.2 per 100,000 individuals. The 2014–2015 season had the highest rate of influenza-related hospitalizations in this age group since this type of annual surveillance began in 2005.4 The especially high burden of disease during the 2014–2015 season was, in part, due to the circulation of an antigenically drifted influenza A (H3N2) viral strain, leading to a suboptimal match with the vaccine.4–6
Annual vaccination is the single most effective strategy for preventing influenza infections.7 Influenza vaccination of healthy adults during seasons with good vaccine match resulted in lower rates of influenza illness, missed work days, and physician visits.8 Vaccination against influenza is less effective at preventing infection in older individuals and those with underlying medical comorbidities.9–12 It remains unclear whether vaccination affords protection against influenza-associated complications and serious outcomes in instances where infection is not prevented. There are several observational studies that suggest vaccination attenuates illness severity and may therefore be associated with less morbidity and mortality in various populations.13–18 However, all of these studies examined seasons with adequate vaccine match, and less is known about whether vaccination may reduce illness severity when the circulating strain is antigenically drifted from the vaccine. The 2014–2015 season presents an opportunity to evaluate this, as the majority of the circulating H3N2 strains did not match the vaccine, contributing to low vaccine effectiveness.4,19,20
We seek to describe the characteristics of adults hospitalized with influenza A infection in a large tertiary care center in the United States during the 2014–2015 season and evaluate the hypothesis that vaccination is associated with lower morbidity and mortality.
We performed a retrospective observational study of all adults with laboratory-confirmed influenza A hospitalized at the University of Michigan Health Center, a large tertiary care center, during the 2014–2015 season. The University of Michigan Health Center is a 900-bed hospital with more than 100,000 emergency department visits per year. All inpatients 18 years or older with an influenza A–positive reverse transcriptase–polymerase chain reaction (ProFlu+ Assay by Prodesse [Hologic, San Diego, Calif]) nasal swab, oropharyngeal swab, tracheal aspirate, or broncheoalveolar lavage were included in the review. Influenza polymerase chain reaction testing occurs in the emergency department prior to admission or soon after admission on the inpatient floor, and reports, during the influenza season, are available within 2 to 4 hours. Study participants were considered “vaccinated” if they had received a 2014–2015 season Centers for Disease Control and Prevention–approved vaccine at least 2 weeks prior to influenza symptom onset. Otherwise, patients were designated as “unvaccinated.” Vaccination status, including product when available, was obtained from the chart (Epic-MiChart) on all patients. We assumed that, in line with national and statewide data, greater than 99% of influenza A infections at our institution were subtype H3N2.4,19,20 Additional data on patient demographics, medical history, and clinical outcomes were abstracted from the medical record by study personnel blinded to vaccination status. The University of Michigan Health Center guidelines recommend that all patients who are admitted with influenza should be treated with oseltamivir. This study was approved by the institutional review board at the University of Michigan Medical School (no. HUM00075270).
The primary outcomes included need for intensive care unit (ICU) admission, need for mechanical ventilation, and death during hospitalization or within 30 days of hospital discharge. Because of relatively low event rate of the primary outcomes, a composite end point was created by combining death or ICU admission, labeled “serious outcome.” Secondary outcomes included meeting 2 of 4 systemic inflammatory response syndrome (SIRS) criteria, diagnosis of pneumonia, positive bacterial respiratory culture (from sputum sample, tracheal aspirate, or bronchoalveolar lavage), bacteremia, and hospital length of stay longer than 1 week.
We used descriptive analyses to summarize patients' demographic and clinical characteristics. Comparisons of demographic and clinical features by vaccination status, patient outcome, and other variables were conducted using a 2-sided χ2 test (or Fisher exact test where appropriate).
We performed multivariable logistic regression analyses to examine the relationship between vaccination status and the primary and secondary outcomes while adjusting for demographic and clinical variables. We adjusted for all collected demographic variables. Clinical variables were included in the final model if there was an association with vaccination status based on a relaxed P < 0.20.
Several variables were prespecified for assessment of effect modification, including age 65 years or older, current smoking status, immunosuppressed status, presence of chronic lung disease, and whether antiviral medications were administered. Based on existing evidence suggesting lowered flu vaccine effectiveness in elderly and chronically ill patients,9–12 we hypothesized that these characteristics would attenuate the effect of vaccination on illness severity. Effect modification was tested by adding multiplicative interaction terms into each model. All comparisons were considered statistically significant at P < 0.05. All data analyses were conducted using SAS University Edition (Cary, NC).
During the 2014–2015 influenza season, 217 adults were hospitalized with influenza A. The majority of patients were vaccinated (54% [n = 118]; Table 1). The median age was 64 years (range, 19–99 years), and 48% were 65 years or older (n = 105); 54% were female (n = 118); 78% were white (n = 173). The median length of hospitalization was 3 days (range, 1–90 days). Chronic medical conditions were common, with 95% of patients having at least 1 high-risk condition (n = 207). Chronic lung disease and cardiovascular disease were the most prevalent conditions at 43% (n = 94) and 47% (n = 102), respectively. Serious clinical outcomes were frequent: 14 patients (7%) died, 31 (14%) required ICU admission, and 27 (12%) required mechanical ventilation. Complications from flu were also common, with 79 patients developing pneumonia (36%), 6 developing ARDS (3%), 7 developing rhabdomyolysis (3%), 116 meeting more than 2 of 4 SIRS criteria (53%), 28 having a positive bacterial respiratory culture (13%), and 8 having a positive blood culture (4%).
Univariable analyses revealed vaccinated patients with influenza A had lower odds of serious outcomes compared with unvaccinated patients (Table 2). Mechanical ventilation (odds ratio [OR], 0.37; 95% confidence interval [CI], 0.16–0.87) and severe outcome (defined as ICU admission or death) (OR, 0.42; 95% CI, 0.20–0.87) were negatively associated with vaccination status. Rates of death (OR, 0.31; 95% CI, 0.09–1.03) and ICU admission (OR, 0.48; 95% CI, 0.22–1.04) were also lower but did not reach statistical significance. Additionally, vaccinated patients had lower odds of pneumonia (OR, 0.53; 95% CI, 0.30–0.92) and meeting 2 or more of 4 SIRS criteria (OR, 0.54; 95% CI, 0.32–0.94). Rates of prolonged hospital length of stay (≥7 days) (OR, 0.57; 95% CI, 0.29–1.11), ARDS (OR, 0.16; 95% CI, 0.02–1.40), rhabdomyolysis (OR, 0.13; 95% CI, 0.02–1.12), positive bacterial respiratory culture (OR, 0.59; 95% CI, 0.26–1.31), and bacteremia (OR, 0.27; 95% CI, 0.05–1.36) were also lower in vaccinated patients; however, these did not reach statistical significance.
After adjusting for age, race, sex, presence of chronic lung and cardiovascular disease, and smoking status (despite meeting criteria for inclusion, morbid obesity could not be analyzed because of its low event rate and overall small sample size), vaccination was associated with a decreased odds of death (OR, 0.29; 95% CI, 0.09–0.99), mechanical ventilation (OR, 0.35; 95% CI, 0.14–0.83), and combined severe outcome (ICU admission or death) (OR, 0.41; 95% CI, 0.19–0.87) (Table 3). While rate of ICU admission (OR, 0.46; 95% CI, 0.21–1.05) was decreased among vaccinated patients, this was not statistically significant.
The effect of vaccination on several secondary outcomes, including hospital length of stay, presence of 2 or more of 4 SIRS criteria, pneumonia, and positive bacterial respiratory culture, was also examined. Adjusting for the same covariables as the primary outcome models, multivariable logistic regression was used to examine these relationships (Table 4). Vaccination was associated with a decreased odds of a hospital stay of 7 days or longer (OR, 0.50; 95% CI, 0.25–0.99), meeting 2 or more of 4 SIRS criteria (OR, 0.56; 95% CI, 0.31–0.95), and developing pneumonia (OR, 0.43; 95% CI, 0.24–0.79). The rate of positive bacterial respiratory culture (OR, 0.46; 95% CI, 0.19–1.09) was decreased among vaccinated patients; however, this did not reach statistical significance. As mentioned previously, occurrences of ARDS, rhabdomyolysis, and bacteremia were decreased in vaccinated patients; however, the small number of events of these outcomes precluded multivariable analysis.
Age 65 years or older, current smoking status, immunosuppressed status, presence of chronic lung disease, and whether antiviral medications were provided were examined for effect modification of vaccination status. None of the interaction terms reached statistical significance and thus were not included in the final models.
Based on our data from hospitalized patients with influenza A infection during the 2014–2015 season, vaccination may afford protection against serious clinical outcomes in cases where infection was not prevented. We demonstrate a decreased odds of death, mechanical ventilation, and combined severe outcome (ICU admission or death) in vaccinated patients. Moreover, vaccination was associated with reduced rates of most of our secondary outcomes—including rates of hospital stay 7 days or longer, meeting 2 or more of 4 SIRS criteria, developing pneumonia, and developing rhabdomyolysis, lending further support for our primary end-point results. These results are particularly salient because the vaccine was poorly matched to the predominant circulating strain (>99%) of influenza A (H3N2) that season.4,19,20
Several studies have suggested an illness severity benefit with influenza vaccination.13–18 Baxter et al13 and Kwong et al14 both demonstrated an association between influenza vaccine and a reduced risk of hospitalizations for influenza and pneumonia in patients 50 years or older. VanWormer et al17 studied symptom severity in an outpatient population with confirmed influenza infection and demonstrated that symptom severity was significantly lower (31%) in adults 65 years or older who were vaccinated compared with those who were not. Deiss et al16 found that symptom severity was significantly decreased in vaccinated versus unvaccinated outpatients who were infected with influenza A/H3N2 but not in those infected with influenza A/H1N1. Heidemann et al15 found that influenza A–infected patients hospitalized during the 2012–2013 flu season were significantly less likely to meet 2 of 4 criteria for SIRS or develop ARDS if they had previously been vaccinated. Again, this study was done during a season in which influenza A/H3N2 predominated. Most recently, Arriola et al18 demonstrated that vaccination in adults hospitalized with influenza during the 2013–2014 season was associated with a reduced in-hospital mortality and ICU admission rates as well as shorter ICU and hospital length of stay. Notably, the 2013–2014 season's vaccine was a good match for the circulating viruses.21 Our study expands on these results to demonstrate reduced illness severity and death among hospitalized patients with influenza specifically during a season where the circulating virus was antigenically drifted from the vaccine strain.
The mechanism of vaccine-induced illness severity benefit is not clear. Some individuals who mount an antibody response following vaccination may still become infected, and it is possible that these antibodies could mediate severe disease, even if it did not prevent infection. The role of antibody specificity in mediating the severity of vaccine failures has yet to be fully investigated. It cannot be ignored that vaccination behavior is often correlated from year to year within individuals, and so the vaccinated individuals in our study may have had more frequent previous annual vaccinations. Contemporary exposures to influenza infection or vaccination have been shown to prompt broad antibody responses to previously encountered antigens in a phenomenon known as the “back-boost.”22 It is possible that the vaccinated individuals in our study had a sufficiently broad antibody repertoire in response to the vaccine hemagglutinin to prevent severe disease, but lacked the specificity needed to prevent all infection. One relatively unexplored component of the vaccine effect in 2014–2015 is the role of antibodies to neuraminidase,23 which have been suggested to protect against symptomatic or severe disease, more so than infection itself.24,25
There are a number of limitations in our study. We captured data during only one season, and the effect of influenza vaccination on illness severity may vary from year to year with changes in circulating viral strains and vaccine matching. We only captured data from hospitalized patients in a single large tertiary care center, and our findings do not speak to the association between vaccination and rates of hospitalization, which itself is an important marker of illness severity. Moreover, our results may not be generalizable across other hospital settings or to community-level infections in healthy people. Hospitalized patients generally have a higher level of chronic disease and obesity, and this is even truer of our tertiary care center, which is a referral center in the Midwest. Another possible limitation is that we did not have influenza A subtype data on many of our observations. However, given that the overwhelming majority (>99%) of circulating strains, both nationally and statewide, were H3N2, it is reasonable to expect similar rates in our own population.4,19,20 Also, we were not able to get the specific form of vaccine administered in all patients. Vaccine data are collected at admission into the medical record, and recall of specific product information was unreliable. Lastly, given the few occurrences of death, caution must be used when generalizing our findings on the relationship between vaccination and mortality to other populations. It is worth noting that we found an association between vaccination and reduced illness severity consistently across most of our outcomes, supporting our conclusion that even during seasons with circulation of antigenically drifted strains vaccination may afford protection against serious clinical outcomes in patients hospitalized with influenza.
1. Thompson WW, Shay DK, Weintraub E, et al. Influenza
-associated hospitalizations in the United States. JAMA
2. Centers for Disease Control and Prevention (CDC). Estimates of deaths associated with seasonal influenza
—United States, 1976–2007. MMWR Morb Mortal Wkly Rep
3. Russell CA, Jones TC, Barr IG, et al. Influenza
vaccine strain selection and recent studies on the global migration of seasonal influenza
. 2008;26(suppl 4):D31–D34.
4. Appiah GD, Blanton L, D'Mello T, et al. Centers for Disease Control and Prevention (CDC). Influenza
activity—United States, 2014–15 season and composition of the 2015–16 influenza
vaccine. MMWR Morb Mortal Wkly Rep
5. Centers for Disease Control and Prevention. CDC Health Advisory Regarding the Potential for Circulation of Drifted Influenza
A (H3N2) Viruses 2014. Available at: https://emergency.cdc.gov/han/han00374.asp
. Accessed May 10, 2016.
6. Centers for Disease Control and Prevention (CDC). Influenza
activity—United States, 2012–13 season and composition of the 2013–14 influenza
vaccine. MMWR Morb Mortal Wkly Rep
7. Grohskopf LA, Sokolow LZ, Broder KR, et al. Prevention and control of seasonal influenza
with vaccines: recommendations of the Advisory Committee on Immunization Practices—United States, 2017–18 Influenza
Season. MMWR Recomm Rep
8. Bridges CB, Thompson WW, Meltzer MI, et al. Effectiveness and cost-benefit of influenza vaccination
of healthy working adults
: a randomized controlled trial. JAMA
9. Herrera GA, Iwane MK, Cortese M, et al. Influenza
vaccine effectiveness among 50–64-year-old persons during a season of poor antigenic match between vaccine and circulating influenza
virus strains: Colorado, United States, 2003–2004. Vaccine
10. Hak E, Nordin J, Wei F, et al. Influence of high-risk medical conditions on the effectiveness of influenza vaccination
among elderly members of 3 large managed-care organizations. Clin Infect Dis
11. Nichol KL, Nordin JD, Nelson DB, et al. Effectiveness of influenza
vaccine in the community-dwelling elderly. N Engl J Med
12. Jefferson T, Rivetti D, Rivetti A, et al. Efficacy and effectiveness of influenza
vaccines in elderly people: a systematic review. Lancet
13. Baxter R, Ray GT, Fireman BH. Effect of influenza vaccination
on hospitalizations in persons aged 50 years and older. Vaccine
14. Kwong JC, Campitelli MA, Gubbay JB, et al. Vaccine effectiveness against laboratory-confirmed influenza
hospitalizations among elderly adults
during the 2010–2011 season. Clin Infect Dis
15. Heidemann LA, Reagan-Steiner S, Lopez KE, et al. Impact of vaccination
on hospitalized adults
A, 2012–2013. Infect Dis Clin Pract
16. Deiss RG, Arnold JC, Chen W-J, et al. Vaccine-associated reduction in symptom severity among patients with influenza
A/H3N2 disease. Vaccine
17. VanWormer JJ, Sundaram ME, Meece JK, et al. A cross-sectional analysis of symptom severity in adults
and other acute respiratory illness in the outpatient setting. BMC Infect Dis
18. Arriola C, Garg S, Anderson EJ, et al. Influenza vaccination
modifies disease severity among community-dwelling adults
hospitalized with influenza
. Clin Infect Dis
19. Chambers BS, Parkhouse K, Ross TM, et al. Identification of hemagglutinin residues responsible for H3N2 antigenic drift during the 2014–2015 influenza
season. Cell Rep
20. Flannery B, Zimmerman RK, Gubareva LV, et al. Enhanced genetic characterization of influenza
A(H3N2) viruses and vaccine effectiveness by genetic group, 2014–2015. J Infect Dis
21. Centers for Disease Control and Prevention (CDC). Flu Activity During the 2013–2014 Season 2014. Available at: https://www.cdc.gov/flu/pastseasons/1314season.htm
. Accessed July 20, 2004.
22. Fonville JM, Wilks SH, James SL, et al. Antibody landscapes after influenza
virus infection or vaccination
23. Monto AS, Petrie JG, Cross RT, et al. Antibody to influenza
virus neuraminidase: an independent correlate of protection. J Infect Dis
24. Johansson BE, Grajower B, Kilbourne ED. Infection-permissive immunization with influenza
virus neuraminidase prevents weight loss in infected mice. Vaccine
25. Marcelin G, Sandbulte MR, Webby RJ. Contribution of antibody production against neuraminidase to the protection afforded by influenza
vaccines. Rev Med Virol