More Related Content
What's hot (20)
Similar to NEJMp0903906 (18)
NEJMp0903906
- 1. Perspective The NEW ENGLAND JOURNAL of MEDICINE n engl j med 10.1056/nejmp0903906 1 insights for current and future planning. Having conducted ar- cheo-epidemiologic research, we can clarify certain signature fea- tures of three previous influenza pandemics A/H1N1 from 1918 through 1919, A/H2N2 from 1957 through 1963, and A/H3N2 from 1968 through 1970 that should inform both national plans for pandemic preparedness and re- quired international collabora- tions. Past pandemics were charac- terized by a shift in the virus sub- type, shifts of the highest death rates to younger populations, suc- cessive pandemic waves, higher transmissibility than that of sea- sonal influenza, and differences in impact in different geographic regions. Although influenza pan- demics are classically defined by the first of these features, the other four characteristics are fre- quently not considered in re- sponse plans. Yet the second feature, the shift in mortality toward younger age groups, was the most strik- ing characteristic of the 20th-cen- tury pandemics.1,2 Exposure to influenza A/H1 subtypes before 1873 may have offered some pro- tection to adults over 45 years of age during the pandemic of 1918 and 1919. A similar mechanism of antigen recycling might explain the partial protection against in- fluenza-related death that was observed among people over 77 years of age during the 19681970 pandemic a possibility sup- ported by the prepandemic pres- ence of antibodies to H3, which were isolated in people born before 1892.1 Another possible mechanism is immune potentia- tion, leading to an increased like- lihood of lethal outcomes after influenza infection in specific age groups. Still other hypotheses in- clude the possibility of bacterial superinfection due to asymmetric carriage rates, given that higher rates were found among young people in 1918 and 1919.1,2 Al- though the elderly frequently have the highest death rates during seasonal epidemics, their relative sparing during pandemics has not been generally appreciated. Advance knowledge of which sub- populations are most likely to be at increased risk for death can shape the optimization of control strategies. The third feature, a pattern of multiple waves, characterized all The Signature Features of Influenza Pandemics Implications for Policy Mark A. Miller, M.D., Cecile Viboud, Ph.D., Marta Balinska, Ph.D., and Lone Simonsen, Ph.D. Vast amounts of time and resources are being invested in planning for the next influenza pandemic, and one may indeed have already be- gun. Data from past pandemics can provide useful Copyright 息 2009 Massachusetts Medical Society. All rights reserved. Downloaded from www.nejm.org on May 18, 2009 . For personal use only. No other uses without permission.
- 2. PERSPECTIVE n engl j med 10.1056/nejmp09039062 three 20th-century pandemics, each of which caused increased mortality for 2 to 5 years (see chart).1 The lethal wave in the autumn of 1918 was preceded by a first wave in the summer that led to substantial morbidity but relatively low mortality in both the United States and Eu- rope. Recent studies suggest that these early mild outbreaks partially immunized the popula- tion, decreasing the mortality impact of the main pandemic wave in the fall of 1918.2 In the United States, the 1957 influen- za A/H2 pandemic had three waves in the United States, with notable excess mortality in the nonsuccessive winter seasons of 1959 and 1962 the latter be- ing 5 years after the initial emergence of the pandemic strain.1 From 1968 through 1970, Eurasia had a mild first influenza season, with the full effects on morbidity and mortal- ity occurring in the second sea- son of pandemic-virus circula- tion. The reasons for multiple waves of varying impact are not precisely understood, but they probably include adaptation of the virus to its new host, demo- graphic or geographic variation, seasonality, and the overall im- munity of the population.1,2 The occurrence of multiple waves po- tentially provides time for health authorities to implement control strategies for successive waves. Increased transmissibility of influenza because of high sus- ceptibility of the population, the fourth feature, has also been doc- umented for all the past pandem- ics, although estimates of repro- ductive numbers a measure of the average number of secondary infections caused by each indi- vidual case vary considerably among studies and pandemics.2,3 Recent studies suggest that dur- ing the early mild wave of the 19181919 pandemic, the repro- ductive number (i.e., the number of new cases attributable to a The Signature Features of Influenza Pandemics Implications for Policy 3x col A B 19181919, Copenhagen18891892, London C D 19681970, England and Wales19571963, United States AUTHOR: FIGURE JOB: ISSUE: 4-C H/T RETAKE 1st 2nd SIZE ICM CASE EMail Line H/T Combo Revised AUTHOR, PLEASE NOTE: Figure has been redrawn and type has been reset. Please check carefully. REG F 3rd Enon ARTIST: Miller 1 of 1 07-02-09 ts 36101 45% 45% 10% January1890M arch 1890June 1890 Septem ber1890 D ecem ber1890M arch 1891June 1891 Septem ber1891 D ecem ber1891M arch 1892 60% 35% 5% January1918 M arch 1918 July1918 N ovem ber1918 M ay1918 Septem ber1918 January1919 M arch 1919 July1919 M ay1919 Septem ber1919 28% 29% 43% April1958 O ctober1957 April1959 O ctober1958 O ctober1959April1960 O ctober1960April1961 O ctober1961April1962 O ctober1962April1963 85% 15% January1968M arch 1968M ay1968 July1968 Septem ber1968 N ovem ber1968 January1969M arch 1969M ay1969 PercentageofTotalMortality perPeriod PercentageofTotalMortality perPeriod PercentageofTotalMortality perPeriod PercentageofTotalMortality perPeriod Mortality Distributions and Timing of Waves of Previous Influenza Pandemics. Proportion of the total influenza-associated mortality burden in each wave for each of four previous pandemics is shown above the blue bars. Mortality waves indicate the timing of the deaths during each pandemic. The 1918 pandemic (Panel B) had a mild first wave during the summer, followed by two severe waves the following winter. The 1957 pandemic (Panel C) had three winter waves during the first 5 years. The 1968 pan- demic (Panel D) had a mild first wave in Britain, followed by a severe second wave the following winter. The shaded columns indicate normal seasonal patterns of influenza. Copyright 息 2009 Massachusetts Medical Society. All rights reserved. Downloaded from www.nejm.org on May 18, 2009 . For personal use only. No other uses without permission.
- 3. n engl j med 10.1056/nejmp0903906 PERSPECTIVE 3 single established case) may have ranged between 2 and 5,2,3 as compared with the average of 1.3 for seasonal influenza. Since models of containment and pan- demic control assumed lower re- productive numbers for the cur- rent epidemic than those that have been historically observed, they are likely to be overly optimistic regarding the success of contain- ment strategies. Great heterogeneity among re- gions in terms of incidence and mortality is also a characteristic of pandemics. This variability is probably explained by the complex heterogeneity in the degree of im- munity in local populations to the circulating influenza strains, as well as by transmission factors such as geographic conditions, social mixing, degree of viral in- fectiousness, and seasonal forc- ing (small seasonal changes in the effective transmission rate).4 The benefits of sharing data on all these variables provide major incentives for international col- laboration. Although the A/H5N1 influ- enza subtype has spread to avian populations in more than 30 countries and infected nearly 400 persons, with a case fatality rate above 50%, scientists disagree about its pandemic potential. Such a highly pathogenic virus does not usually adapt well to its host, since it tends to kill fast- er than it can be transmitted. Other avian subtypes are also con- sidered to be pandemic threats. Although avian viruses have a different tropism for respiratory cellular receptors in birds than for those in humans, gradual vi- ral mutations or gene-segment reassortments in a mammal mix- ing vessel could result in a novel viral clade or subtype that spreads rapidly in a population that has largely not previously been ex- posed to it. Such changes may have occurred in the current swine H1N1 circulating strain. The death toll of a future pan- demic depends not only on the virulence of the virus in question but also on the rapidity with which we are able to introduce effective preventive and therapeu- tic measures. Although A/H5N1 has been associated with a cy- tokine-storm phenomenon rem- iniscent of that observed in 1918 and 1919, new methods for the timely manufacture and admin- istration of antiviral agents and influenza and pneumococcal vac- cines could mitigate the effects of a pandemic. The evidence of multiple waves in the 20th-century pan- demics underlines the impor- tance of active real-time viral surveillance on a global scale. Transnational collaborations are crucial for the effective exchange of genomic, clinical, and epide- miologic data that will make possible the development of vac- cines and treatment protocols and the identification of the best population-based strategies. Al- though our ability to produce a vaccine in sufficient quantities to cover people who are exposed in a first pandemic wave is very limited with todays technology, an interwave period would pro- vide time to increase the produc- tion of biomedical tools and to vaccinate populations, thereby mitigating the morbidity and mortality associated with suc- cessive and potentially more le- thal waves. This possibility, too, is a powerful incentive for inter- national collaboration, since all would potentially share the ben- efits. If an effective vaccine had been available and used even a year after the emergence of the A/H3N2 viruses in 1968, most of the deaths in Europe and Asia could probably have been pre- vented. The signature pandemic fea- ture of shifts in age-specific mor- tality patterns should influence vaccination priorities.5 Given that the supplies of vaccine and anti- viral agents are likely to be con- strained, the efficient allocation necessary for reducing mortality will require consideration of lo- cal demography, expected shifts in age-specific incidence, direct and herd effects of vaccination in various age groups, and ethical issues regarding life expectancies and potential years of life saved in various groups. The indirect ef- fects of reducing transmission also warrant further consider- ation. The role of preexisting an- tibodies in the elderly, their re- duced immune response because of immune senescence, and great- er transmission among children should prompt the targeting of younger age groups as the sound- est policy in a 1918-like scenario. However, these attributes do not necessarily apply to other pan- demics to the same extent.5 Nonmedical interventions primarily social distancing could be useful in staving off transmission. Simulation models suggest that such interventions would considerably decrease the incidence of infection only if the basic reproductive number was less than 2, a rate that is lower than that observed in past pan- demics.3 Though the rapidity of trans- mission of influenza virus during pandemics necessitates immedi- ate action, it can be hoped that close collaborations and lessons drawn from previous pandemics will contribute to reducing na- tional and global mortality. The The Signature Features of Influenza Pandemics Implications for Policy Copyright 息 2009 Massachusetts Medical Society. All rights reserved. Downloaded from www.nejm.org on May 18, 2009 . For personal use only. No other uses without permission.
- 4. PERSPECTIVE n engl j med 10.1056/nejmp09039064 documented relevant signature features can help health authori- ties prioritize national strategies and aid international collabora- tors in addressing the initial and successive waves of illnesses and deaths. Dr. Miller reports being named on a pending patent held by the National Insti- tutes of Health on a novel influenza vaccine; and Dr. Simonsen, receiving consulting fees from Merck and research support from Wy- eth. No other potential conflict of interest relevant to this article was reported. Dr. Miller is the associate director for re- search, Dr. Viboud a staff scientist, and Dr. Balinska a research associate at the Fogar足 ty International Center of the National Insti- tutes of Health, Bethesda, MD. Dr. Simon- sen is an adjunct professor and research director of the Department of Global Health, George Washington University School of Public Health and Health Services, Wash- ington, DC. This article (10.1056/NEJMp0903906) was published at NEJM.org on May 7, 2009. Simonsen L, Olson DR, Viboud C, et al.1. Pandemic influenza and mortality: past evi- dence and projections for the future. In: Kno- bler SL, Mack A, Mahmoud A, Lemon SM, eds. The threat of pandemic influenza: are we ready? Workshop summary. Washington, DC: National Academies Press, 2005:89-114. Andreasen V, Viboud C, Simonsen L. Epi-2. demiologic characterization of the 1918 in- fluenza pandemic summer wave in Copen- hagen: implications for pandemic control strategies. J Infect Dis 2008;197:270-8. Viboud C, Tam T, Fleming D, Handel A,3. Miller MA, Simonsen L. Transmissibility and mortality impact of epidemic and pan- demic influenza, with emphasis on the un- usually deadly 1951 epidemic. Vaccine 2006; 24:6701-7. Richard SA, Sugaya N, Simonsen L, Mill-4. er MA, Viboud C. A comparative study of the 1918-1920 influenza pandemic in Japan, USA and UK: mortality impact and implica- tions for pandemic planning. Epidemiol In- fect 2009 February 12:1-11 (Epub ahead of print). Miller MA, Viboud C, Olson DR, Grais5. RF, Rabaa MA, Simonsen L. Prioritization of influenza pandemic vaccination to minimize years of life lost. J Infect Dis 2008;198: 305-11. Copyright 息 2009 Massachusetts Medical Society. The Signature Features of Influenza Pandemics Implications for Policy Copyright 息 2009 Massachusetts Medical Society. All rights reserved. Downloaded from www.nejm.org on May 18, 2009 . For personal use only. No other uses without permission.