Effectiveness of influenza vaccination in preventing influenza-like illness among community-dwelling elderly: Population-based cohort study in Japan

Vaccine 24 (2006) 5546–5551

Effectiveness of influenza vaccination in preventing influenza-like illness among community-dwelling elderly: Population-based cohort study in Japan Megumi Hara ∗ , Tatsuhiko Sakamoto 1 , Keitaro Tanaka 1 Department of Preventive Medicine, Faculty of Medicine, Saga University, 5-1-1 Nabeshima, Saga 849-8501, Japan Received 18 February 2006; received in revised form 28 March 2006; accepted 21 April 2006 Available online 5 May 2006

Abstract A population-based cohort study was conducted during the 2003–2004 season to examine the effectiveness of influenza vaccine among community-dwelling elderly. The subjects consisted of 4787 elderly, ranging from 65 to 79 years. We either interviewed the elderly directly or their families regarding acute febrile illness, hospital visits, hospitalization and death by telephone every month. The vaccination status and physician-diagnosed clinical influenza (hereinafter referred as clinical influenza) were determined based on data obtained from the city office and hospitals, respectively. Influenza-like illness (ILI) was defined as an acute febrile illness (≥38.5 ◦ C) during the epidemic period. After adjusting for confounders, vaccination decreased ILI significantly (odds ratio [OR], 0.38; 95% confidence interval [CI], 0.17–0.85), but not clinical influenza (OR, 0.76; 95% CI, 0.28–2.06). The results were inconclusive for preventing hospitalization for influenza or pneumonia (OR, 0.37; 95% CI, 0.09–1.47) and death (OR, 3.68; 95% CI, 0.75–18.12), due to the inadequate sample size. In conclusion, the influenza vaccination was thus found to be associated with a decreased ILI during the epidemic period in community-dwelling elderly. © 2006 Elsevier Ltd. All rights reserved. Keywords: Influenza vaccine; Effectiveness; Community-dwelling elderly

1. Introduction According to the Advisory Committee on Immunization Practices (ACIP) in the United States, inactivated influenza vaccination is 30–70% effective in preventing hospitalization for pneumonia and influenza among elderly persons not living in nursing homes or similar chronic-care facilities. Annual vaccinations are recommended for these groups, as well as the residents of nursing homes and for groups with high-risk medical conditions [1]. Most observational studies about the effectiveness of influenza vaccine in communitydwelling elderly have been investigated by linkage to large scale databases, such as health maintenance organization, Medicare, Medicaid, national health insurance, data base of ∗ 1

Corresponding author. Tel.: +81 952 34 2289; fax: +81 952 34 2065. E-mail address: [email protected] (M. Hara). Tel.: +81 952 34 2289; fax: +81 952 34 2065.

0264-410X/$ – see front matter © 2006 Elsevier Ltd. All rights reserved. doi:10.1016/j.vaccine.2006.04.027

general practitioner, population registry, and mortality registry [2–11], while few population-based studies have been made [12–15]. Linkage studies, which examine the effectiveness in preventing death and hospitalization among senior citizens, are important for evaluate the vaccinating program in public health. However, a high fever by influenza is an important health hazard in the elderly in a community and evaluations of the effectiveness of vaccination regarding the onset of ILI including high fever is thus required. In a linkage study, information about influenzal infection and ILI are limited to patients who had visited hospitals, otherwise they remained unidentified. All the subjects should be followed equally through influenza season to examine the effect of influenza vaccination for ILI and high fever, whereas such studies have so far mostly been limited to elderly residents in nursing home [16–19]. In general, little is known about the vaccine effectiveness for ILI in community-dwelling elderly [20,21].

M. Hara et al. / Vaccine 24 (2006) 5546–5551

5547

2. Materials and methods

were self-reported and 10 subjects were known with verification) inoculated vaccine from October 1, 2003 to December 31, 2003, and the vaccination rate was 67.7%. According to reports of Infectious Disease Information in Saga Prefecture as recorded by the National Epidemiological Surveillance of Infectious Disease, an influenza epidemic was experienced in Saga between 1 January 2004 and 31 March 2004, and the influenza epidemic was mild in comparison with the previous 10 seasons. The predominant influenza strain circulating all over Japan was A (H3N2), and only 10% of the Influenza A (H3N2) isolates were similar to the vaccine strain while about 90% of the isolates were similar to the drift variant, A/Fujian/411/2002 (H3N2) [22].

2.1. Subjects

2.3. Follow-up and outcomes

We selected 10,000 community-dwelling elderly, ranging from 65 to 79 years old at 1 January 2003, randomly from a population registry, and sent them a letter about the explanation for the study and request for participation in it at 1 December 2003. The eligibility criteria to participate to study were as follows; not being hospitalized, not being institutionalized, not being having any long-term absence, not living alone, and possible to contact by telephone at least once a month. We asked them to answer the self-administered questionnaire about baseline characteristics that might act as potential confounders: including a history of influenza vaccination (this season, a pre-season, the season before last), diagnosis of influenza (a pre-season, the season before last), comorbid conditions at baseline period, such as hypertension, cardiovascular disease (heart failure, angina pectoris, arrhythmia), respiratory tract disease (pulmonary emphysema, chronic obstructive pulmonary disease, and asthmatic bronchitis), diabetes mellitus, cerebrovascular disease (cerebral infarction, cerebral hemorrhage) and the other, frequency of hospitalizations for pneumonia, health status, smoking habit, exercise habit, going out to a crowd, day care use, handwashing and gargling habit, and family constitution. Subjects who reported at least 1 of listed above comorbid conditions were classified as the high-risk group. Among 10,000 elderly citizens, 7357 responded and 4787 agreed to participate and also matched our eligibility criteria. We obtained written informed consent from all subjects. This study was approved by the institutional review board associated with Saga University.

The survey period was from 1 December 2003 to 31 March 2004. If the participants had fever equal to or more than 37 ◦ C degrees during the survey period, they were asked to record their signs or symptoms prospectively onto the provided record sheet, including a check list of symptoms, such as cough, sore throat, nasal congestion, muscle ache and arthralgia, hospital visits, and medication. Research nurses performed telephone interviews to them or their families about the recorded symptoms, hospital visit, hospitalization and death at least one time per a month. If the participants had visited hospital with fever equal to or more than 37 ◦ C as the chief complaint, we referred to their doctor for their clinical diagnosis. ILIs were defined as acute febrile illness with symptoms described above, which occurred during the epidemic period of influenza in study area. The following outcome measures were considered: ILI, clinical diagnosed influenza, hospitalization for all causes, hospitalization for influenza or pneumonia, and total death.

Individual medical conditions are well-known risk factors for serious influenza associated complications. However, the benefits of vaccination among elderly according to the risk category have not yet been well described, even for influenzarelated mortality and hospitalization [3,5,7,9–11]. To clarify the effectiveness of influenza vaccination against various outcomes, including ILI, according to riskconditions among elderly citizen, we conducted a populationbased cohort study during the 2003/2004 influenza season in Saga city, Japan.

2.2. Influenza vaccination and epidemic in the study area The vaccine contained the following antigens, which were predicted as possible epidemic strains by World Health Organization (WHO): 30 ␮g/ml each of A/New Caledonia/20/99 (H1N1), A/Panama/2007/99 (H3N2), and B/Shandong/7/97. The vaccination status of the study subjects was identified by self-reporting verification and a list of recipients of partially funded vaccination. After all 3240 subjects (3230 subjects

2.4. Data analysis A univariate logistic regression model was used to assess the association between the vaccination uptake and baseline characteristics, as well as the association between each outcome and the vaccination status or baseline characteristics. To evaluate the vaccine effectiveness independent of confounding factors, such as significantly associated with the vaccine uptake, the adjusted odds ratio (OR) and its 95% confidence intervals (95% CI) were calculated by a multivariate analyses. Vaccine effectiveness (VE) was calculated as (1 − OR) × 100%.

3. Results A total of 4787 community-dwelling elderly were followed during the 2003/2004 influenza season, 60 subjects were lost to the follow-up while 18 subjects died, among them only two deaths were due to influenza or pneumonia. Among the remaining 4709 subjects, we observed 115 cases

5548

M. Hara et al. / Vaccine 24 (2006) 5546–5551

Table 1 Baseline characteristics of the study subjects and the odds ratios of the influenza vaccine uptake Number (%)

OR

95% CI

Vaccinated (n = 3240)

Non-vaccinated (n = 1547)

Sex Male Female

1808 (55.8) 1432 (44.2)

1007 (65.1) 540 (34.9)

1.00 1.48

1.30–1.67

Vaccination history 2002/2003 season 2001/2002 season

1917 (59.2) 1285 (39.7)

73 (4.7) 59 (3.8)

29.39 17.06

23.00–37.56 13.03–22.34

Clinical diagnosed Influenza 2002/2003 season High-riska Hypertension Cardiovascular disease Respiratory tract disease Diabetes mellitus Cerebrovascular disease Other comorbid conditions Hospitalization for pneumonia

139 (4.3)

42 (2.7)

1.61

1.13–2.38

2411 (76.1) 1282 (39.6) 576 (17.8) 251 (7.7) 392 (12.1) 269 (8.3) 1240 (38.3)

830 (53.9) 424 (27.4) 198 (12.8) 58 (3.7) 162 (10.5) 69 (4.5) 399 (25.8)

2.72 1.74 1.48 2.17 1.18 1.95 1.79

2.39–3.10 1.52–1.98 1.24–1.76 1.62–2.90 0.97–1.44 1.49–2.56 1.57–2.05

1.48

1.15–1.90

256 (7.9)

85 (5.5)

Health status Good Fair Poor

511 (15.8) 2318 (71.5) 409 (12.6)

330 (21.3) 1065 (68.8) 152 (9.8)

1.00 1.40 1.20–1.64 1.73 1.37–2.18 (P for trend P < 0.01)

Smoking Never Past Current

1667 (51.5) 1213 (37.4) 356 (11.0)

679 (43.9) 569 (36.8) 298 (19.3)

1.00 0.87 0.76–0.99 0.49 0.41–0.58 (P for trend P < 0.01)

Exercise habit None 1–2 day/week 3–4 day/week Almost everyday

1405 (43.4) 651 (20.1) 460 (14.2) 720 (22.2)

737 (47.6) 219 (14.2) 208 (13.4) 383 (24.8)

1.00 1.56 1.30–1.86 1.16 0.96–1.39 0.98 0.84–1.15 (P for trend P = 0.95)

Going out to a crowd None 1–2 day/week 3–4 day/week Almost everyday

913 (28.2) 1203 (37.1) 699 (21.6) 422 (13.0)

454 (29.3) 547 (35.4) 336 (21.7) 209 (13.5)

1.00 1.09 0.94–1.27 1.03 0.87–1.23 1.00 0.82–1.23 (P for trend P = 0.98)

Day care use None Less than 1 day/week 1 day/week 2 day/week Equal to more than 3 day/week

3045 (94.0) 23 (0.7) 47 (1.5) 83 (2.6) 37 (1.1)

1499 (96.9) 7 (0.5) 9 (0.6) 23 (1.5) 8 (0.5)

Washing hand when they return Rare Sometimes Every time

462 (14.3) 1387 (42.8) 1387 (42.8)

295 (19.1) 648 (41.9) 604 (39.0)

1.00 1.36 1.14–1.61 1.45 1.22–1.73 (P for trend P < 0.01)

Gargling when they return Rare Sometimes Every time

1038 (32.0) 139 (4.3) 803 (24.8)

621 (40.1) 625 (40.4) 301 (19.5)

1.00 1.33 1.16–1.52 1.59 1.35–1.87 (P for trend P < 0.01)

390 (12.0)

154 (10.0)

Living with infant or children

OR: odds ratio; CI: confidence interval. a High-risk was defined as having one or more disease, as listed below, at baseline.

1.00 1.99

1.24

1.44 –2.76

1.02–1.51

M. Hara et al. / Vaccine 24 (2006) 5546–5551

5549

Table 2 Crude odds ratios of influenza-like illness with fever after influenza vaccination Fever threshold

Number (%)

≥37.8 ◦ C ≥38.0 ◦ C ≥38.5 ◦ C ≥39.0 ◦ C

Crude

Vaccinated (n = 3169)

Non-vaccinated (n = 1540)

OR

95% CI

70 (2.21) 54 (1.70) 20 (0.63) 9 (0.28)

45 (2.92) 40 (2.60) 22 (1.43) 11 (0.71)

0.75 0.65 0.44 0.37

0.51–1.10 0.43–0.98 0.24–0.81 0.16–0.96

OR: odds ratio; CI: confidence interval.

of ILI with a fever of more than 37.8 ◦ C, 28 cases with clinical influenza, 137 hospitalizations for all causes and 17 hospitalizations for Influenza or pneumonia (IP). A comparison of the baseline variables between the vaccinated group and the non-vaccinated group is shown in Table 1. In contrast to the non-vaccinated subjects, the vaccinated subjects were significantly more likely to be females, to have received the influenza vaccine in the previous season and two seasons before, to be diagnosed with influenza in the previous season, to have baseline risk conditions such as underlining hypertension, heart disease, respiratory tract disease, cerebrovascular disease and other diseases, to have been previously hospitalized for pneumonia, to have a poor health status, to regularly exercise, to use day care, to regularly wash their hands and gargle, and to live together with their grandchildren, who were still infants. Non-vaccinated subjects were more likely to smoke. The crude ORs of the ILI with fever differed according to fever degree (Table 2). The risk reduction for ILI with fever was greater as fever was higher. The higher threshold of

fever was, the greater risk reduction of ILI was. A statistically significant risk reduction was found when the fever was more than 38 ◦ C. The ORs of ILI with fever more than 38, 38.5, and 39 ◦ C were 0.65 (95% CI, 0.43–0.98), 0.44 (95% CI, 0.24–0.81), and 0.37 (95% CI, 0.16–0.96), respectively. The OR of ILI in more than 38.5 ◦ C fever was considered to be around 0.4, thus we defined this threshold as ILI with highfever. Table 3 gives the ORs and vaccine effectiveness of influenza vaccine in reducing the outcomes according to risk condition. The vaccination was associated with fewer ILI with a high fever, and the effectiveness of the vaccination was higher among the low-risk group. The vaccine effectiveness for ILI was 62% (95% CI, 15–83%) in all subjects, 52% (95% CI, −13–80%) in the high-risk group, and 87% (95% CI, 13–98%) in the low-risk group. Vaccination tended to reduce the risk of clinical influenza, hospitalization for all causes and hospitalization for Influenza or pneumonia in the highrisk group, although the ORs were not statistically significant. Vaccination was not associated with a reduction in death.

Table 3 Crude and adjusted odds ratios of the vaccination effectiveness among elderly according to the risk condition Outcome

Partial modela

Complete modelb

95% CI

OR

95% CI

OR

95% CI

0.44 0.87 0.99 0.89 3.89

0.24–0.81 0.40–1.90 0.69–1.43 0.33–2.41 0.89–16.93

0.46 0.85 0.98 0.87 3.88

0.25–0.85 0.39–1.87 0.68–1.41 0.32–2.38 0.89–17.00

0.38 0.76 0.72 0.37 3.68

0.17–0.85 0.28–2.06 0.46–1.13 0.09–1.47 0.75–18.12

15 7 33 5 2

0.39 0.59 0.76 0.62 2.24

0.19–0.78 0.23–1.50 0.50–1.16 0.21–1.85 0.50–9.94

0.39 0.56 0.76 0.62 2.31

0.19–0.79 0.22–1.45 0.50–1.15 0.21–1.88 0.52–10.32

0.48 0.53 0.66 0.29 1.90

0.20–1.13 0.16–1.76 0.39–1.09 0.07–1.28 0.37–9.85

7 3 12 1 0

0.40 1.88 1.42 1.88 NA

0.10–1.55 0.47–7.55 0.68–2.96 0.17–20.73

0.46 1.91 1.42 1.91 NA

0.12–1.82 0.47–7.64 0.67–2.99 0.17–21.26

0.13 1.66 0.81 3.45 NA

0.02–0.87 0.31–9.01 0.30–2.18 0.10–117.95

Vaccinated (n = 3169)

Non-vaccinated (n = 1540)

Crude OR

Total subjects ILI with high fever Clinical Influenza Hospitalization for all Hospitalization for IP Death from all causes

20 18 92 11 16

22 10 45 6 2

High-risk condition ILI with high fever Clinical Influenza Hospitalization for all Hospitalization for IP Death from all causes

17 12 74 9 13

Low-risk condition ILI with high fever Clinical Influenza Hospitalization for all Hospitalization for IP Death from all causes

3 6 18 2 3

OR: odds ratio; CI: confidence interval; ILI: Influenza-like illness; IP: Influenza or pneumonia; NA: not applicable due to absence number in the outcome category. a The odds ratios for the partial models were adjusted for age and sex. b The odds ratios for the complete models were adjusted for age and sex plus the potential confounders at baseline, which were significantly associated with the vaccine uptake.

5550

M. Hara et al. / Vaccine 24 (2006) 5546–5551

4. Discussion In this population-based prospective cohort study among community-dwelling elderly in Japan, influenza vaccination reduced the risk of ILI with fever more than 38 ◦ C. The protective effect was greater among low-risk group than among high-risk group. The study of influenza vaccine effectiveness in persons aged 65 years and over living in the community was limited [12–15], especially the effectiveness for ILI [20,21]. Vu et al. performed a meta-analysis for ILI, and estimated 35% as the reduction, which was similar to the estimates in the study of healthy adults [21]. A placebo-controlled trial from the Netherlands confirmed that vaccination reduced the risk of clinical and serologic influenza among the elderly by 58%, however, vaccine effectiveness was less pronounced for self-reported influenza [20,21]. In that study, all participants were sent questionnaires regarding possible influenza episodes and symptoms twice, namely at 10 and 23 weeks after vaccination. Since it was a retrospective investigation, some information bias might have influenced the results. Accuracy of self-information is critical to evaluate the vaccine effectiveness for ILI, and a complete follow-up and an adequate case definition is essential. The advantage of our study is that we observed each individual study subject prospectively with an equal intensity throughout the epidemic period. To measure the outcome, we asked the study subjects to note their body temperature and symptoms on the provided record sheet (symptoms diary) when they had fever (
37.0 ◦ C) and we interviewed them for ILI by telephone at least once a month. Therefore, we could confirm ILI, as well hospitalization and death. In addition, this prospective design could minimize selection bias and recall bias, in comparison to a retrospective design. One population-based follow-up study with a telephone among elderly citizens in Taiwan reported the vaccine effectiveness for reducing hospitalization and death, however, it could not evaluate the vaccine effectiveness for ILI, due to a large number of lost subjects during the follow-up (13.8%) [15]. If the untraceable subjects included many ILI and serious influenza cases, then the results would be negatively affected. Devising the criteria for the study participation enabled us to achieve an excellent follow-up rate (98.7%). Observational cohort studies among community-dwelling elderly mostly address the benefits of influenza vaccination for hospitalization and death [2–11]. Although several studies have demonstrated the vaccine effectiveness for clinically diagnosed influenza, such reductions were inconsistent, ranging 15–50 % [10,23]. Vaccinations tended to reduce the number of clinically diagnosed influenza cases in the present study, however, protective effect was not statistically significant and it also differed based on the risk conditions. Because the seriousness of influenzal symptoms and hospital visits are not necessarily related in Japan, a misclassification of clinical diagnosed influenza may occur due to the fact that not every influenza patient visits a hospital. In comparison to the lowrisk group, the high-risk group tended to visit the hospital

more frequently and was also more likely to be diagnosed to have influenza, so that a reduction of clinical influenza might thus be observed. Our results indicated that self-reported ILI may thus be a suitable method for showing the protective effect of vaccination among community-dwelling elderly individuals, as long as they are reported completely, because they are not affected by hospital visits. In general, non-differential misclassification, which means that ILI including non-influenzal disease and non-ILI including influenza, results in an underestimation of vaccine effectiveness. We observed that the ORs decreased as the fever threshold increased. When fever was equal to and higher than 38.5 ◦ C, the OR was almost 0.4. Two possibilities to explain this may thus be proposed. First, the misclassification of “non-influenza” was avoided by setting a clear limit for fever (namely, equal to or higher than 38.5 ◦ C) thereby allowing us to better identify the true effect of influenza vaccination. Second, vaccination may also prevent higher fever itself. This study provided an additional possibility that the target outcome of vaccination differed according to the subjects’ medical conditions, even among community-dwelling elderly. The protective effect for ILI was greater among the low-risk group, on the other hand, that for influenza hospitalization was only seen among the high-risk group. Several studies investigated the vaccine effectiveness according to the risk-conditions [3,5,7,9–11], however, the results are inconsistent. A previous cohort study using a computerized database in the Netherlands investigated the effectiveness of an influenza vaccine according to the risk conditions in the community and also reported mortality to be significantly reduced in elderly with comorbidity (OR, 0.67; 95% CI, 0.48–0.94), although vaccination failed to show any protective effect for pneumonia [10]. Whereas the sample size of our study was sufficient to evaluate the effect for ILI by means of a telephone follow-up, it was not large enough to detect any statistically significant effect for mortality and hospitalization. However, our findings suggest that vaccination to elderly in community is beneficial for both high-risk and lowrisk groups. In the absence of randomization, potential problems of this study might be related information bias and confounding by indications. Information bias for vaccination was unlikely, since we verified all self-reported information with the records of the city vaccination program. The information biases for self-reported symptoms were minimized, due to prospective, frequent and complete follow-up. On the other hand, some confounding factors regarding the indications were likely associated because some chronic diseases, as well as a history of hospitalization might be independent risk factors for an increased risk of influenzal disease, as was a history of influenza vaccination. Even when these factors were adjusted by a multivariate analysis, no all residual confounding factors could be excluded. However, such confounding factors may have resulted in an underestimation of vaccine effectiveness. Hence, our estimates are conserva-

M. Hara et al. / Vaccine 24 (2006) 5546–5551

tive, and the real effects may actually be higher than herein reported. In conclusion, our results indicated that, influenza vaccination was thus found to be associated with a decreased ILI during the epidemic period in community-dwelling elderly. The above risk reduction was greater under low-risk conditions than that under high-risk conditions. The results were inconclusive for preventing hospitalization and death, due to an inadequate sample size. However, our findings support the finding that all elderly individuals substantially benefit from vaccination even in a season of mild influenza activity and also when the antigenic match between the vaccine strains and the circulating strains are not closely matched. Acknowledgements This study was supported by a research grant for emerging and re-emerging infectious diseases from the Ministry of Health, Labor and Welfare of Japan. The authors would like to thank the staff of the Saga-city government office for their help in confirming the vaccination status of the study subjects. We also thank the doctors of the Saga-city medical association for their cooperation in confirming the clinical diagnosis of influenza. References [1] Harper SA, Fukuda K, Uyeki TM, Cox NJ, Bridges CB. Prevention and control of influenza. Recommendations of the Advisory Committee on Immunization Practices (ACIP). MMWR Recomm Rep 2005;54(RR-8):1–40. [2] Barker WH, Mullooly JP. Influenza vaccination of elderly persons. Reduction in pneumonia and influenza hospitalizations and deaths. JAMA 1980;244(22):2547–9. [3] Mullooly JP, Bennett MD, Hornbrook MC, Barker WH, Williams WW, Patriarca PA, et al. Influenza vaccination programs for elderly persons: cost-effectiveness in a health maintenance organization. Ann Intern Med 1994;121(12):947–52. [4] Nichol KL, Margolis KL, Wuorenma J, Von Sternberg T. The efficacy and cost effectiveness of vaccination against influenza among elderly persons living in the community. N Engl J Med 1994;331(12):778–84. [5] Nichol KL, Wuorenma J, von Sternberg T. Benefits of influenza vaccination for low-, intermediate-, and high-risk senior citizens. Arch Intern Med 1998;158(16):1769–76. [6] Nordin J, Mullooly J, Poblete S, Strikas R, Petrucci R, Wei F, et al. Influenza vaccine effectiveness in preventing hospitalizations and deaths in persons 65 years or older in Minnesota, New York, and Oregon: data from three health plans. J Infect Dis 2001;184(6):665–70. [7] Davis JW, Lee E, Taira DA, Chung RS. Influenza vaccination, hospitalizations, and costs among members of a Medicare managed care plan. Med Care 2001;39(12):1273–80.

5551

[8] Nichol KL, Nordin J, Mullooly J, Lask R, Fillbrandt K, Iwane M. Influenza vaccination and reduction in hospitalizations for cardiac disease and stroke among the elderly. N Engl J Med 2003;348(14):1322–32. [9] Hak E, Nordin J, Wei F, Mullooly J, Poblete S, Strikas R, et al. Influence of high-risk medical conditions on the effectiveness of influenza vaccination among elderly members of three large managed-care organizations. Clin Infect Dis 2002;35(4):370–7. [10] Voordouw BC, van der Linden PD, Simonian S, van der Lei J, Sturkenboom MC, Stricker BH. Influenza vaccination in communitydwelling elderly: impact on mortality and influenza-associated morbidity. Arch Intern Med 2003;163(9):1089–94. [11] Wang ST, Lee LT, Chen LS, Chen TH. Economic evaluation of vaccination against influenza in the elderly: an experience from a population-based influenza vaccination program in Taiwan. Vaccine 2005;23(16):1973–80. [12] Foster DA, Talsma A, Furumoto-Dawson A, Ohmit SE, Margulies JR, Arden NH, et al. Influenza vaccine effectiveness in preventing hospitalization for pneumonia in the elderly. Am J Epidemiol 1992;136(3):296–307. [13] Nicholson KG, Kent J, Hammersley V. Influenza A among community-dwelling elderly persons in Leicestershire during winter 1993–1994; cigarette smoking as a risk factor and the efficacy of influenza vaccination. Epidemiol Infect 1999;123(1):103– 8. [14] Crocetti E, Arniani S, Bordoni F, Maciocco G, Zappa M, Buiatti E. Effectiveness of influenza vaccination in the elderly in a community in Italy. Eur J Epidemiol 2001;17(2):163–8. [15] Wang CS, Wang ST, Chou P. Efficacy and cost-effectiveness of influenza vaccination of the elderly in a densely populated and unvaccinated community. Vaccine 2002;20(19–20):2494–9. [16] Gross PA, Quinnan GV, Rodstein M, LaMontagne JR, Kaslow RA, Saah AJ, et al. Association of influenza immunization with reduction in mortality in an elderly population. A prospective study. Arch Intern Med 1988;148(3):562–5. [17] Potter J, Stott DJ, Roberts MA, Elder AG, O’Donnell B, Knight PV, et al. Influenza vaccination of health care workers in long-termcare hospitals reduces the mortality of elderly patients. J Infect Dis 1997;175(1):1–6. [18] Loeb M, McGeer A, McArthur M, Walter S, Simor AE. Risk factors for pneumonia and other lower respiratory tract infections in elderly residents of long-term care facilities. Arch Intern Med 1999;159(17):2058–64. [19] Monto AS, Hornbuckle K, Ohmit SE. Influenza vaccine effectiveness among elderly nursing home residents: a cohort study. Am J Epidemiol 2001;154(2):155–60. [20] Govaert TM, Thijs CT, Masurel N, Sprenger MJ, Dinant GJ, Knottnerus JA. The efficacy of influenza vaccination in elderly individuals. A randomized double-blind placebo-controlled trial. JAMA 1994;272(21):1661–5. [21] Vu T, Farish S, Jenkins M, Kelly H. A meta-analysis of effectiveness of influenza vaccine in persons aged 65 years and over living in the community. Vaccine 2002;20(1314):1831–6. [22] Infectious Agents Surveillance Report, National Institute of Infectious Disease and Tuberculosis and Infectious Disease Control Division, Ministry of Health and Welfare, 2004;25:278– 279. [23] Nichol KL, Margolis KL, Wouremna J, von Sternberg T. Effectiveness of influenza vaccine in the elderly. Gerontology 1996;42(5):274–9.