Mortality Associated With Influenza and Respiratory Syncytial Virus in the United States

Context Influenza and respiratory syncytial virus (RSV) cause substantial morbidity and mortality. Statistical methods used to estimate deaths in the United States attributable to influenza have not accounted for RSV circulation.

Objective To develop a statistical model using national mortality and viral surveillance data to estimate annual influenza- and RSV-associated deaths in the United States, by age group, virus, and influenza type and subtype.

Design, Setting, and Population Age-specific Poisson regression models using national viral surveillance data for the 1976-1977 through 1998-1999 seasons were used to estimate influenza-associated deaths. Influenza- and RSV-associated deaths were simultaneously estimated for the 1990-1991 through 1998-1999 seasons.

Main Outcome Measures Attributable deaths for 3 categories: underlying pneumonia and influenza, underlying respiratory and circulatory, and all causes.

Results Annual estimates of influenza-associated deaths increased significantly between the 1976-1977 and 1998-1999 seasons for all 3 death categories (P<.001 for each category). For the 1990-1991 through 1998-1999 seasons, the greatest mean numbers of deaths were associated with influenza A(H3N2) viruses, followed by RSV, influenza B, and influenza A(H1N1). Influenza viruses and RSV, respectively, were associated with annual means (SD) of 8097 (3084) and 2707 (196) underlying pneumonia and influenza deaths, 36 155 (11 055) and 11 321 (668) underlying respiratory and circulatory deaths, and 51 203 (15 081) and 17 358 (1086) all-cause deaths. For underlying respiratory and circulatory deaths, 90% of influenza- and 78% of RSV-associated deaths occurred among persons aged 65 years or older. Influenza was associated with more deaths than RSV in all age groups except for children younger than 1 year. On average, influenza was associated with 3 times as many deaths as RSV.

Conclusions Mortality associated with both influenza and RSV circulation disproportionately affects elderly persons. Influenza deaths have increased substantially in the last 2 decades, in part because of aging of the population, underscoring the need for better prevention measures, including more effective vaccines and vaccination programs for elderly persons.

Influenza infections result in substantial morbidity and mortality nearly every year^1,2 and estimates of this burden have played a pivotal role in formulating influenza vaccination policy in the United States.³ However, numbers of deaths attributable to influenza are difficult to estimate directly because influenza infections typically are not confirmed virologically or specified on hospital discharge forms or death certificates. In addition, many influenza-associated deaths occur from secondary complications when influenza viruses are no longer detectable.^4,5 Nonetheless, wintertime influenza epidemics have been shown to be associated with increased hospitalizations and mortality for many diagnoses, including congestive heart failure, chronic obstructive pulmonary disease, pneumonia, and bacterial superinfections.^6-9

Respiratory syncytial virus (RSV) epidemics often overlap with influenza epidemics,^8,10 and RSV infections have been associated with substantial morbidity and mortality in young children and more recently in older adults.^10-14 Like influenza, RSV infections can precipitate both cardiac and pulmonary complications.^15-17 Respiratory syncytial virus infections are rarely diagnosed in adults, in part because available rapid antigen-detection tests are insensitive in adults and few tests for RSV are requested for this age group by medical practitioners.^16,18 It is likely that some deaths previously attributed to influenza are actually associated with RSV infection.^13,14,19

In this study, we provide age-specific estimates of deaths attributable to influenza, by virus type and subtype, and to RSV using Poisson regression models that incorporates national respiratory viral surveillance data. Recent deliberations of the Advisory Committee on Immunization Practices (ACIP) regarding influenza vaccination recommendations³ guided our choice of age groups for these analyses.

Influenza and RSV typically circulate during winter months and across calendar years. Therefore, we defined each annual respiratory season as the period from July 1 through June 30 of the following year.

In the United States, laboratory-based surveillance for influenza viruses is conducted from October through mid-May (calendar week 40 through week 20). For the influenza virus surveillance periods from the 1976-1977 through 1998-1999 seasons, we obtained numbers of influenza virus isolates reported weekly by 50 to 75 World Health Organization collaborating virology laboratories in the United States to the Influenza Branch of the Centers for Disease Control and Prevention (CDC). The collaborating laboratories provided weekly numbers of total respiratory specimens tested for influenza and positive-influenza isolates by virus type and subtype²⁰ (Table 1).

Weekly RSV data were obtained from the National Respiratory and Enteric Virus Surveillance System for the 1990-1991 through 1998-1999 seasons. During this period, 63 to 72 clinical and public health laboratories in 44 states reported to CDC weekly numbers of specimens tested for RSV by antigen-detection and viral-isolation methods and numbers of positive results.²¹ We used the results of both antigen-detection and isolation tests to determine the circulation pattern of RSV. The weekly percentages of specimens that tested positive for both influenza and RSV were used in estimating the association of virus circulation with weekly deaths in the United States (Table 1).

National mortality data were obtained from the National Center for Health Statistics (NCHS).²² Deaths were categorized using the International Classification of Diseases, Ninth Revision (ICD-9) codes for NCHS mortality data obtained from 1976 through 1998.²³ The International Classification of Diseases, 10th Revision (ICD-10) was used for classifying NCHS mortality data obtained for 1999.²⁴ Analyses were based on the underlying cause-of-death because it represents the disease or injury that initiated the chain of morbid events that led directly to the death.²⁵ The change from ICD-9 to ICD-10 in January 1999 resulted in a 30% decrease in the number of coded underlying pneumonia deaths.²⁶ Therefore, all analyses of underlying pneumonia and influenza deaths were carried out using data collected through the 1997-1998 season, which limited the analyses to ICD-9 coded deaths.

The 3 death categories modeled were underlying pneumonia and influenza deaths (ICD-9 codes 480-487), underlying respiratory and circulatory deaths (ICD-9 codes 390-519 and ICD-10 codes I00-I99, J00-J99), and all-cause deaths (all ICD codes). Underlying pneumonia and influenza deaths exclude some deaths, such as those related to exacerbations of underlying cardiac and pulmonary conditions, which are associated with both influenza and RSV infections.^6,18 Influenza-associated all-cause death estimates have been previously used to represent the full spectrum of deaths associated with influenza infections.^1,2 However, these estimates include deaths such as those caused by fires and motor vehicle crashes, which are not directly associated with respiratory viral infections. Therefore, we also modeled a third category of deaths, underlying respiratory and circulatory deaths (which includes pneumonia and influenza deaths), to provide an estimate of deaths that was more directly associated with viral respiratory infections. These estimates would be expected to be more sensitive than estimates using underlying pneumonia and influenza deaths and more specific than estimates using all-cause deaths.

For the influenza model, we developed an age-specific Poisson regression model that used weekly influenza circulation data. Deaths were stratified into the following 5 age groups: younger than 1 year, 1 to 4 years, 5 to 49 years, 50 to 64 years, and 65 years or older. Influenza-associated deaths were estimated for influenza A(H1N1), A(H3N2), and B viruses. The viral circulation terms represented percentages of specimens testing positive for each of the 3 influenza virus types and subtypes during a particular week. Estimates of the weekly age-specific population size were used to account for changes in population trends over time. The US population estimates by age group were obtained from the US Census Bureau.²⁷

For the influenza and RSV model, we used a model identical to the influenza model except it included an additional coefficient for RSV viral circulation. The full model was written as follows:
Y = α exp{β₀ + β₁[t] + β₂[t²] + β₃[sin(2π/52)] + β₄[cos(2π/52)] + β₅[A(H1N1)] + β₆[A(H3N2)] + β₇[B] + β₈[RSV]}
where Y represents the number of deaths during a particular week for a specific age group, α is the offset term and is equal to the log of the age-specific population size, β₀ represents the intercept, β₁ accounts for the linear time trend, β₂ accounts for nonlinear time trends, β₃ and β₄ account for seasonal changes in deaths, and β₅ through β₈ represent coefficients associated with the percentage of specimens testing positive for a given week.

We fit the influenza model to national influenza surveillance data available from the 1976-1977 through 1998-1999 seasons. The influenza and RSV model was fit to data available from the 1990-1991 through 1998-1999 seasons, when both weekly influenza and RSV data were available (PROC GENMOD, SAS, version 8.2; SAS Institute Inc, Cary, NC).

National influenza and RSV surveillance data are summarized in Table 1. Influenza isolate data were available for the 1976-1977 through 1998-1999 seasons. A mean of 27 360 specimens (range, 13 275-52 505) was tested for influenza viruses during each of the influenza surveillance periods (October through May). During weeks that testing for influenza occurred, an average of 12% of specimens tested positive for influenza. Influenza A(H1N1), A(H3N2), and B viruses, respectively, comprised 15%, 60%, and 25% of the positive influenza isolates. From the 1990-1991 through 1997-1998 season, the annual mean number of specimens tested for RSV was 107 711 (range, 67 374-133 648) with an average of 17 220 specimens (16%) testing positive each season for RSV.

From the 1976-1977 through 1998-1999 seasons, there was an annual mean of 69 140 (range, 47 133-90 895) underlying pneumonia and influenza deaths, 1 135 724 (range, 1 069 560-1 203 728) underlying respiratory and circulatory deaths, and 2 126 740 (range, 1 879 039-2 407 494) all-cause deaths. The numbers of deaths in each of these categories increased linearly during this period. From the 1976-1977 through 1997-1998 seasons, underlying pneumonia and influenza deaths increased by 83%, substantially more than underlying respiratory and circulatory deaths or all-cause deaths (11% and 28%, respectively). From the 1990-1991 through 1998-1999 seasons, there was an annual mean of 82 239 (range, 74 872-90 895) underlying pneumonia and influenza deaths, 1 158964 (range, 1098 086-1203 728) underlying respiratory and circulatory deaths, and 2 277 268 (range, 2 135 976-2 407 494) all-cause deaths.

The mean annual estimates of underlying pneumonia and influenza deaths, underlying respiratory and circulatory deaths, and all-cause deaths associated with influenza were 5977, 25 420, and 34 470 (Table 2). Each of these 3 estimates increased significantly during the study period (P<.001 for trend for all 3 death categories).

For underlying pneumonia and influenza deaths, we estimated an annual mean of 8097 (SD, 3084; range, 3515-13 033) influenza-associated deaths, representing 9.8% (8097/82 239) of these deaths (Table 3). Influenza A(H1N1), A(H3N2), and B viruses were associated with annual means of 381 (SD, 617; range, 0-1742), 6613 (SD, 3928; range, 944-12 941), and 1103 (SD, 1030; range, 53-2619) deaths, respectively. Respiratory syncytial virus was associated with an annual mean of 2707 (SD, 196; range 2336-2880) deaths or 3.3% (2707/82 239) of all such deaths. The year-to-year variation in influenza-associated deaths was higher than the year-to-year variation in RSV-associated deaths.

For underlying respiratory and circulatory deaths, we estimated an annual mean of 36 155 (SD, 11 055; range, 17 056-51 296) influenza-associated deaths, representing 3.1% (36 155/1 158 964) of these deaths. Influenza A(H1N1), A(H3N2), and B viruses were associated with annual means of 1960 (SD, 3372; range, 0-10 080), 28 940 (SD, 14 848; range, 4435-50 855), and 5255 (SD, 4513; range, 253-12 067) deaths, respectively. Respiratory syncytial virus was associated with an annual mean of 11 321 (SD, 668; range, 10 047-12 385) deaths or 1.0% (11 321/1 158 964) of all such deaths.

For all-cause deaths, we estimated an annual mean of 51 203 (SD, 15 081; range, 25 570-71 416) influenza-associated deaths, representing 2.2% (51 203/2 277 268) of these deaths. Influenza A(H1N1), A(H3N2), and B viruses were associated with annual means of 2836 (SD, 4909; range, 0-14 727), 40 017 (SD, 20 656; range, 6033-70 701), and 8349 (SD, 7105; range, 404-19 030) deaths, respectively. Respiratory syncytial virus was associated with an annual mean of 17 358 (SD, 1086; range, 15 464-19 262) deaths or 0.8% (17 358/2 277 268) of all such deaths. Influenza-associated deaths again showed higher year-to-year variability than did RSV-associated deaths.

In children younger than 1 year, RSV was associated with annual means of 124 underlying pneumonia and influenza deaths, 211 underlying respiratory and circulatory deaths, and 214 all-cause deaths (Table 4). In this age group, influenza viruses were associated with annual means of 13 underlying pneumonia and influenza deaths, 26 underlying respiratory and circulatory deaths, and 88 all-cause deaths. There were more influenza-associated deaths relative to RSV-associated deaths among children aged 1 to 4 years for all 3 death categories.

Among underlying pneumonia and influenza deaths, 90% (7326/8097) of influenza-associated deaths and 88% (2388/2707) of RSV-associated deaths occurred among persons aged 65 years or older. For underlying respiratory and circulatory deaths, 90% (32 651/36 155) of influenza-associated deaths and 78% (8811/11 321) of RSV-associated deaths occurred among persons aged 65 years or older. For all-cause deaths, 43 979 and 9812 all-cause deaths were attributable to influenza and RSV, respectively.

Annual mean influenza-associated mortality rates for underlying pneumonia and influenza deaths, underlying respiratory and circulatory deaths, and all-cause deaths were 3.1, 13.8, and 19.6 per 100 000 person-years, respectively (Table 5). Similarly, annual mean RSV-associated mortality rates for these death categories were 1.0, 4.3, and 6.6 per 100 000 person-years, respectively. The relative risks (RRs) and 95% confidence intervals (CIs) comparing influenza mortality rates with RSV mortality rates for the 3 death categories were 3.0 (95% CI, 2.9-3.1), 3.2 (95% CI, 3.1-3.3), and 2.9 (95% CI, 2.9-3.0), respectively.

Annual mean influenza-associated mortality rates for underlying pneumonia and influenza deaths in persons younger than 1 year, 1 to 4 years, 5 to 49 years, 50 to 64 years, and 65 years or older were 0.3, 0.2, 0.2, 1.3, and 22.1 deaths per 100 000 person-years, respectively. The RSV-associated mortality rates for underlying pneumonia and influenza deaths in persons younger than 1 year, 1 to 4 years, 5 to 49 years, 50 to 64 years, and 65 years or older were 3.1, 0.1, <0.1, 0.5, and 7.2 deaths per 100 000 person-years, respectively. For children younger than 1 year, the RR for RSV vs influenza mortality rates was 9.5 (95% CI, 5.4-16.9) for underlying pneumonia and influenza deaths and 8.1 (95% CI, 5.4-12.2) for underlying respiratory and circulatory deaths. For all-cause deaths, the RR among this age group was substantially lower (RR, 2.4; 95% CI, 1.9-3.1).

Periseason influenza rate-difference models¹⁰ were fit to the 1976-1977 through the 1998-1999 seasons and revealed substantial differences in relative influenza-attributable mortality rates among elderly persons. Persons aged 85 years or older were 32 times more likely to die of an influenza-associated underlying pneumonia and influenza death compared with persons aged 65 to 69 years (RR, 32.1; 95% CI, 31.3-32.9). Persons aged 85 years or older were 16 times more likely to die of an influenza-associated all-cause death compared with persons aged 65 to 69 years (RR, 14.8; 95% CI, 14.6-14.9). However, there were no statistically significant increases in any of the 5-year age-specific mortality rates from the 1976-1977 through the 1998-1999 seasons (P>.05 for all).

The number of persons aged 65 years or older increased substantially between the 1976-1977 and 1998-1999 seasons.²⁷ During the 1990s, the growth rate for the number of persons aged 50 to 64 years also increased substantially relative to the period from 1976 through 1990.²⁷

Morbidity and mortality associated with seasonal epidemics of influenza in the United States have provided the impetus for public health policies and strategies to control influenza infections, particularly among specific target groups.³ Mortality associated with influenza can vary dramatically by season and models developed to assess influenza-associated mortality date back to 1847.²⁸ These approaches have been feasible because well-defined peaks in deaths occur in association with influenza outbreaks in temperate countries. In the recent past, the CDC has used a linear regression model, applied to either complete national mortality data or more immediately available mortality surveillance data from 122 cities, to estimate annual deaths associated with influenza.^1,2 The influenza and RSV model presented in this study will be used to provide future estimates of influenza-associated mortality in the United States, because the model permits estimates of influenza subtype-specific mortality and also simultaneously estimates RSV-associated mortality.

Our results indicate that US influenza-associated deaths have increased substantially from the 1976-1977 through 1998-1999 seasons. We believe this is explained in part by the aging of the US population. Between 1976 and 1999, the number of persons aged 85 years or older doubled in the United States.²⁹ We found that persons in this age group were 16 times more likely to die of an influenza-associated all-cause death than persons aged 65 to 69 years during a period in which all-cause age-specific death rates have remained stable. Other studies have also shown that influenza-attributable mortality rates increased rapidly with age among persons aged 65 years or older.^30-32 For example, Nordin et al³² found that persons aged 75 years or older were 3 to 9 times more likely to die from influenza infections than persons aged 65 to 74 years. Another important factor contributing to the increase in influenza-associated deaths during the 1990s was the predominance of influenza A(H3N2) viruses, the most virulent of the recently circulating influenza viruses. Influenza A(H3N2) viruses were 1 of the predominant strains in 8 of 9 seasons we analyzed during the 1990s.

The influenza and RSV model confirmed that influenza A(H3N2) viruses were associated with the highest attributable mortality rates, followed by RSV, influenza B, and influenza A(H1N1) viruses. The annual effect of RSV on mortality was relatively stable, although the numbers of deaths associated with influenza viruses varied substantially, depending on the predominant circulating virus type or subtype. In this study, RSV was the most common viral cause of death in children younger than 5 years, particularly in children younger than 1 year. However, RSV-associated mortality rates were higher in elderly persons and substantially more RSV-associated deaths occurred among elderly persons than among young children.

Determining the most appropriate death category for characterizing the burden of influenza on mortality is difficult. Pneumonia and influenza deaths are highly correlated with the circulation of influenza, and these estimates are useful for monitoring year-to-year trends and variability in the severity of influenza seasons. However, this death category underestimates the total burden of influenza because many deaths are caused by other secondary complications of influenza (eg, congestive heart failure).⁷ Traditionally, all-cause deaths have been used to estimate the total burden of influenza on mortality.^1,2 However, this death category is also not ideal because it includes deaths that are not causally linked with respiratory viral infections. Therefore, we analyzed underlying respiratory and circulatory deaths to provide a more specific estimate of the total burden of influenza and RSV on mortality. Our estimate of annual mean influenza-associated underlying respiratory and circulatory deaths was 36 155 (29% lower than the annual mean all-cause estimate).

The Poisson regression models used to estimate influenza- and RSV-associated deaths were more complex compared with previous influenza models, but the models also provided more specific estimates, including independent estimates of deaths associated with influenza and RSV. The model we used in this study is capable of incorporating additional factors that could not be included in the previous CDC model, such as temperature, that might independently influence winter season mortality. Therefore, we believe this new model represents a step forward in current efforts to better understand the burden of viral respiratory infections on mortality.

We applied the new model to age groups relevant to policy deliberations by ACIP regarding influenza-vaccination recommendations for persons younger than 5 years and persons 50 to 64 years.³ Currently, ACIP recommends annual influenza vaccination for all persons aged 65 years or older. Because most influenza-associated deaths occur in persons aged 65 years or older, understanding age-specific effects within this age group is also of considerable interest. As a first step, we fit simple periseason rate difference models among those aged 65 years or older by 5-year age intervals and demonstrated increases with age in influenza-associated mortality. Future research will focus attention on ACIP discussions regarding age-specific vaccination policy.

We believe the results of this study have important policy implications. Deaths associated with viral respiratory infections have increased substantially during the past decade and it appears that they will continue to increase as the population continues to age. Increased numbers of patients with serious respiratory infections may further stress hospital systems that are already struggling to cope with wintertime surges in patient visits during influenza seasons. For example, during the 1997-1998 season, a severe influenza outbreak in Los Angeles County resulted in a dramatic increase in hospitalizations and the need to divert patients to other facilities. This problem may have been less severe if more bed capacity had been available.³³

Vaccinating elderly persons will continue to be the primary strategy for preventing influenza-associated deaths. Studies directly comparing outcomes in vaccinated vs unvaccinated groups have shown that the currently available trivalent inactivated influenza vaccine is approximately 68% effective in preventing deaths from complications of influenza infections.^34-36 However, the effectiveness of influenza vaccines in preventing deaths among elderly persons with associated chronic conditions is significantly lower,^7,35 underscoring the need for influenza vaccines that are more immunogenic in elderly persons.^37,38 Recent studies have also raised the question of whether vaccinating young children against influenza may decrease transmission rates and thus decrease influenza-associated morbidity and mortality among elderly persons,^39,40 but the effectiveness of this approach remains uncertain.

Although the importance of RSV among young children is well recognized,^41,42 we found that more than 78% of RSV-associated underlying respiratory and circulatory deaths occured among persons aged 65 years or older. This finding highlights the need for an effective RSV vaccine in both young children and elderly persons.^12,16,18,43 A number of candidate RSV vaccines are being developed including vaccines based on cold-adapted live attenuated RSV strains^44,45 and subunit vaccines intended for use in RSV nonnaive populations.⁴⁶ Effective and safe vaccines for use among persons aged 65 years or older are needed to decrease deaths associated with RSV infection.