Effectiveness of inactivated trivalent influenza vaccine in long-term care institutions, Toronto, 2003-2004

Introduction

A traditional public health role is the support of influenza immunization programs targeted at high-risk individuals such as residents and staff of long-term care facilities. In Toronto, vaccination of both staff and residents of long-term care facilities is encouraged through yearly promotional campaigns that coincide with the vaccine's release. The inactivated trivalent influenza vaccine is publicly funded for all residents of Ontario.

When there is a good match between the circulating and vaccine influenza strains, immunization is recognized as one of the most effective strategies for preventing and reducing the risk of influenza infection and its secondary complications, such as pneumonia and death, especially in high-risk populations^(1,2). This is particularly important in long-term care facilities, which house elderly people who often have co-morbidities that make them vulnerable to the complications of influenza. Studies have found that adults aged >= 65 account for 90% or more of influenza-related deaths^(3,4). The National Advisory Committee on Immunization (NACI) states that the vaccination of health care workers (HCWs) is an essential component of influenza prevention, and during a confirmed influenza outbreak it supports exclusion of unim- munized HCWs who are not receiving antiviral prophylaxis⁽¹⁾. As HCWs are often the source of influenza for residents of long-term care facilities, this NACI recommendation has been adopted by Toronto Public Health.

In February each year, NACI recommends the influenza strains to be included in that year's vaccine. This decision is based on global influenza surveillance reports predicting the strains of the impending influenza season. The 2003-2004 vaccine for Ontario (Fluviral®, Shire Biologics, and Vaxigrip®, Aventis Pasteur) was composed of three antigens: A/New Caledonia/20/99 (H1N1)-like, A/Panama 2007/99 (H3N2) A/Moscow/10/99 (H3N2)-like, and B/Hong Kong/330/2001-like (Fluviral®) or B/Shangdong/7/97-like (Vaxigrip®)⁽⁵⁾.

The Situation in Toronto

As of the end of January 2004, all influenza A specimens from Toronto⁽⁶⁾ subtyped at the Ontario Ministry of Health and Long-Term Care's Central Public Health Laboratory and 97.5% of the isolates submitted from across Ontario⁽⁷⁾ had been anti-genically characterized as a new variant strain, A/Fujian/411/2002 (H3N2)-like. Similar results have been noted in the United States as well as around the globe⁽⁸⁾. The 2003-2004 vaccine did not contain this new strain, but given that the A/Fujian and the vaccine strains shared neuraminidase proteins it was expected that the 2003-2004 vaccine composition would provide some protection against the Fujian strain⁽¹⁾. Early serologic studies had demonstrated that adults and elderly individuals immunized with vaccines containing the A/Panama/411/2002-like viruses developed antibody against A/Fujian-like viruses⁽⁹⁾.

Given the mismatch between the circulating influenza strain and the composition of the vaccine, a retrospective study was undertaken to estimate vaccine effectiveness in preventing influenza-like illness (ILI), pneumonia, and associated death in residents of Toronto long-term care facilities reporting outbreaks of influenza A during the 2003-2004 season.

Under the provisions of the Health Protection and Promotion Act, all Ontario hospitals and long-term care facilities are required to report suspected respiratory outbreaks to their local Medical Officer of Health (MOH). An outbreak is declared when one laboratory-confirmed case of influenza or at least two cases of ILI within a 48-hour period are detected in a specific unit or facility. Outbreaks are either declared within part of the institution such as a ward or floor, or for the entire institution. Upon being notified of an outbreak, Toronto Public Health works with institutions to implement outbreak infection control practices, such as enhanced hand washing, isolation of staff and residents, visitor restriction, and enhanced environmental cleaning. An outbreak is declared over only after an 8-day period without a new resident case or a 3-day period without a new staff case (whichever is longer).

Methods

Study design

This study included all institutional outbreaks that met certain criteria: they involved at least one resident or staff member found on laboratory testing to be positive for influenza A, they were reported to Toronto Public Health on or after 26 October, 2003, and they were declared over by 27 January, 2004. The data were extracted from outbreak investigation reports prepared by public health investigators as part of their outbreak management team duties. The cases counted for evaluating vaccine effectiveness included those that occurred during the period from the onset of symptoms in the index case up to 2 days after the outbreak had been reported. This was a period that ended before outbreak control measures (including antiviral drugs) were likely to have been effective and was chosen to reduce their effect as confounders.

Laboratory confirmation

Once a respiratory outbreak had been detected, HCWs collected three to six nasopharyngeal swabs from the most recent cases with ILI in the facility, and these were sent to the Central Public Health Laboratory. All submitted specimens from an outbreak were cultured for respiratory viruses. Up to three specimens were tested using an enzyme-linked immunofluorescence assay for influenza A (BD Directigen Flu A Test Kit, Franklin Lakes, NJ) and, for those specimens submitted after 31 December (when oseltamivir became available), for influenza B (BD Directigen Flu A & B Test Kit, Franklin Lakes, NJ). If influenza virus was detected by either method the outbreak was confirmed as an influenza outbreak. All individuals with ILI symptoms within the same institution during the outbreak were considered to be cases of influenza. If the outbreak had a prolonged or unusual course, up to three additional nasopharyngeal swabs were resubmitted for testing to detect co-infections.

Prophylaxis

Once an influenza A outbreak had been confirmed, chemoprophylaxis was recommended, using either amantadine or oseltamivir (Tamiflu®), for all healthy residents regardless of vaccination status. Unimmunized staff were required to take antiviral prophylaxis or they were excluded from work. Treatment of ill residents with oseltamivir was recommended if it could be started within 48 hours of symptom onset. Symptomatic staff were restricted from working in the institution until 5 days after their symptoms began or until they were asymptomatic. The vaccination status of residents and staff was obtained from the institution. A 2-week interval after vaccination was required in order for them to be considered as immunized.

Health outcome measures

Health outcome measures were as follows: ILI, as defined by the Ontario Ministry of Health and Long-term Care's respiratory outbreak guidelines⁽¹⁰⁾ , radiographically confirmed pneumonia, and deaths attributed to pneumonia that occurred in the facility during the study period. Vaccine status was based on records shared by long-term care facility staff. Data were entered and analyzed in Windows Excel and SPSS, and Epi Info 2000. Vaccine effectiveness was calculated as follows: (1 - attack rate of those vaccinated/attack rate of those unvaccinated) x 100%⁽¹¹⁾. An epidemiologist with Toronto Public Health and a biostatistician were consulted.

Results

The first influenza A outbreak of the 2003-2004 season was reported to Toronto Public Health on 26 October, 2003. A total of 61 respiratory institutional outbreaks were reported from that time to 27 January, 2004. Of these, 41 (67.2%) were due to influenza A (38 influenza A alone, 2 influenza A and respiratory syncytial virus, 1 influenza A and parainfluenza type 1), two to Chlamydia pneumoniae, and one to parainfluenza type 2; in 17 outbreaks the infectious agent was not identified. All samples from the influenza A outbreaks that were subtyped at the Central Public Health Laboratory were A/Fujian/411/2002 (H3N2)-like⁽⁷⁾. This report summarizes 40 of the confirmed influenza A outbreaks. One outbreak was excluded because of incomplete data at the time of this review.

Twenty-nine outbreaks (72%) occurred in long-term care facilities, six (15%) in chronic care hospitals, and five (12.5%) in retirement residences. The median number of days between the occurrence of symptoms in the index case and reporting to Public Health was 5 days (range 0 to 22 days). The median outbreak duration was 20 days (range 11- 44 days). There were 685 resident cases (median 16, range 2-45 per institution) and 223 staff cases (median 3, range 0-31 per institution) reported across all 40 outbreaks.

The average institutional vaccination rate among residents at the onset of the outbreak was 90.4% (range 73.3%-100%), and one outbreak occurred before residents had been vaccinated. The average institutional vaccination rate among staff was 64.5% (range 29.2%-100%). Of those who developed ILI during the entire outbreak period, 89.5% of resident and 71.8% of staff cases were reported as having been vaccinated.

Vaccine effectiveness in residents

Of the 5711 residents at risk (in wards, on floors or in the whole institution), 620 (10.9%) developed ILI symptoms within 2 days of the outbreak being reported; 558 (90%) of these residents were reported as vaccinated. The relative risk of illness in vaccinated residents was not significant (risk ratio [RR] = 0.92, 95% confidence interval [CI] 0.72-1.18). The corresponding vaccine effectiveness was 7.9% (p = 0.513) (Table 1). There were 49 (7.9%) resident cases with pneumonia confirmed by chest radiography, of whom 41 (83.7%) had been immunized. The relative risk of pneumonia in those who were vaccinated was not significant (RR = 0.525, 95% CI 0.25-1.11). There were 14 deaths attributed to pneumonia, of which 12 (85.7%) involved residents who had been immunized (RR = 0.61, 95%, CI 0.14-2.47).

Vaccine effectiveness in staff

Of the 6569 staff members who worked in an outbreak area, 175 (2.7%) developed ILI within 2 days of the outbreak being reported; 122 (69.7%) of these were reported as having been vaccinated. Vaccinated staff were at greater risk of illness (RR = 1.52, 95% CI 1.11- 2.09) than those unvaccinated. This corresponded to a vaccine efficacy of -52.2% (p = 0.011) (Table 1).

Table 1. Vaccine effectiveness in study period from onset of symptoms in the index case up to 2 days after reporting of the influenza outbreak

	Residents		Staff
	Vaccinated	Unvaccinated	Vaccinated	Unvaccinated
ILI* cases	558	62	122	53
Well	4623	468	3832	2562
Total persons	5181	530	3954	2615
ILI* attack rate	10.8%	11.7%	3.1%	2.0%
Risk ratio	0.92 (0.72-1.18)		1.52 (1.11-2.09)
Vaccine effectiveness	7.9%		-52.2%
*ILI = influenza-like illness

Use of influenza antiviral prophylaxis

Thirty-five (87.5%) of the institutions used oseltamivir for treatment and prophylaxis in the outbreak, two used a combination of oseltamivir and amantadine, and three did not use any antiviral drugs. In these three institutions influenza A was isolated in culture and reported after the outbreak had been declared over.

The total number of cases of ILI reported >= 2 calendar days after antiviral chemoprophylaxis had been started was 64 residents (median 0, range 0-15 per institution) and 48 staff (median 0, range 0-10 per institution).

Discussion

Studies have shown that the trivalent influenza vaccine is cost-effective in reducing influenza-related morbidity and mortality if the vaccine strain is well matched to the circulating strain^(12-14). Among the elderly in nursing homes, the vaccine has been shown to be 50% to 60% effective in preventing hospitalization or pneumonia and 80% in preventing death; its effectiveness in preventing influenza infection has been estimated to range from 30% to 40%^(12,13). In a meta-analysis of case-control studies of influenza vaccine efficacy among the elderly, effectiveness ranged from 32% to 45% in preventing hospitalization for pneumonia, 31% to 65% in preventing deaths from pneumonia and influenza, and 43% to 50% in preventing deaths from respiratory conditions⁽¹⁴⁾.

The influenza vaccine for the 2003-2004 season was expected to provide some cross-protection because of the antigenic similarities between the vaccine strains and the A/Fujian strain. However, the results of this study fail to demonstrate any statistically significant effectiveness of the vaccine in preventing ILI in residents of long-term care facilities in Toronto. There was a trend towards a reduction in the number of pneumonia cases and deaths, but the small numbers of patients who had pneumonia or who died limited the power of this study to find statistical significance. For staff working in long-term care facilities there was a significant increase in the risk of ILI among those who had been immunized. These findings concur with another study of influenza efficacy in a children's hospital in Denver⁽¹⁵⁾, where the preliminary data suggested that the 2003-2004 vaccine had no or low efficacy against ILI.

Even if the true efficacy of the vaccine among staff were very low or negligible, factors that preferentially increase the risk of exposure or of acquiring influenza in immunized staff could contribute to a negative efficacy. It is possible that staff with greater direct patient care, such as nurses or physiotherapists, are more likely to be vaccinated than other staff who are at lower risk of exposure (e.g. clerical staff) and illness. In light of their perceived increased risk of exposure and infection, it is conceivable that unimmunized staff may be more vigilant in exercising personal infection control procedures such as hand washing.

Despite annual promotion of the vaccine to the staff of long-term care facilities, the institutional vaccination rate among staff ranged from 29.5% to 100% with a median of 64.5% before the onset of these outbreaks. As vaccination status was obtained from the institution, we could not verify the individual vaccination status or the accuracy of the reported data. It is possible that the reported vaccination rates may not reflect the true immune status of the staff, and staff may have been misclassified. As unimmunized staff who refused to take antiviral chemprophylaxis were excluded from work, there may be a disincentive to disclose an unvaccinated status.

Individuals who had a diagnosis of ILI in an institution with laboratory-confirmed influenza were assumed to be influenza cases in the course of the outbreak. Other viruses may have been responsible for the ILI, especially among staff, who had broader exposure to various circulating viruses than residents. However, the predominant respiratory virus circulating and isolated at the Central Public Health Laboratory during this period was influenza A, and only three outbreaks were found to involve co-infections later in the outbreak.

There were 64 cases (median 0, range 0-15) of ILI in residents and 48 (median 0, range 0-10) in staff after chemoprophylaxis had been instituted in those outbreaks. It appeared that most of the cases occurred in a few institutions. It is not known how many of the 48 staff cases had been immunized and therefore were not required to take antiviral prophylaxis during the outbreak. In most institutions, unimmunized staff were required to obtain antiviral prophylaxis independently, often at their own expense, and we were unable to verify the start date or the compliance with these drugs.

This study is limited by the small number of cases occurring in each outbreak, which prevents analysis of each outbreak separately and misses the possibility of heterogeneity in the vaccine effectiveness between outbreaks. The small numbers limit the power of the study to show statistically reliable results for the residents, especially among the few reports of unimmunized residents. Lack of the data needed to perform subanalysis or adjustments based on age, sex, comorbidity or pneumococcal vaccination status also posed limits. These results may be generalizable to other long-term care facilities and to HCWs, but there is likely limited generalizability of the vaccine effectiveness to the general community.

When the annual influenza vaccine is well matched to the circulating influenza strains, vaccination remains an effective prevention strategy in reducing the morbidity and mortality associated with influenza. Although serologic and laboratory investigations may suggest that the 2003-2004 trivalent influenza vaccine was expected to give some degree of cross-protection to the predominant strain, these results do not support this clinically in the arena of long-term care facilities. With antigenic shifts and drifts of influenza virus, it is not always possible to accurately predict the recombination that will emerge in the impending influenza season. The overall year-to-year benefits of vaccination in vulnerable populations outweigh the years of poor efficacy.

However, it remains important to assess the clinical efficacy of influenza vaccine early in the influenza season. Once there is evidence that vaccine efficacy in a given year is zero or very low, it is important to reassess outbreak control policies and reinforce infection control practices for all staff and not just those who are unimmunized. It is also important to consider influenza antiviral prophylaxis for all staff regardless of immunization status.

Conclusion

This study did not find any statistical evidence that the 2003-2004 trivalent inactivated influenza vaccine was effective in reducing ILI, pneumonia or deaths in residents during influenza A outbreaks in long-term care facilities in Toronto. However, staff members who were reported to be immunized were at increased risk of acquiring ILI. Further studies involving larger numbers are required to assess the effectiveness of the influenza vaccine in reducing ILI, pneumonia, and death related to pneumonia in long-term care facilities. It is important to assess vaccine effectiveness annually and early in the influenza season to allow a timely review of influenza outbreak management policies and maximize the effectiveness of infection control measures.

Acknowlegements

The authors thank the following for their assistance: F Goettler, S Korman, E Kefalas, and the infection control investigators and managers, Toronto Public Health, Toronto, Ontario.

References

1. National Advisory Committee on Immunization. Statement on influenza vaccination for the 2003-2004 season. CCDR 2003; 29(ACS-4):1-20.

2. Centers for Disease Prevention and Control. Prevention and control of influenza: recommendations of the Advisory Committee on Immunization Practices [ACIP]. MMWR 2003;52(RR08):1-36.

3. Simonsen L, Clarke MJ, Schonberger LB et al. Pandemic versus epidemic influenza mortality: a pattern of changing age distribution. J Infect Dis 1998;178:53-60.

4. Thompson WW, Shay DK, Weintraub E et al. Mortality associated with influenza and respiratory syncytial virus in the United States. JAMA 2003;289:179-86.

5. US Food and Drug Administration, Center for Biologics Evaluation and Research. Influenza virus vaccine 2003-2004. URL: <http://www.fda.gov/cber/flu/flu2003.htm>. Date accessed: 15 January, 2004.

6. Toronto Public Health. TPH weekly influenza news. 30 January, 2004.

7. Ministry of Health and Long-Term Care. Ontario influenza bulletin. 24 January, 2004.

8. Public Health Agency of Canada. FluWatch. 11-17 January, 2004.

9. World Health Organization, Communicable Disease Surveillance and Response. Influenza vaccine for the northern hemisphere 2003-2004: additional Information. URL: <http://www.who.int/csr/disease/influenza/vaccine2003/en/index.html>.
   Date of Access: 15 January, 2004.

10. Ontario Ministry of Health, Public Health Branch. A guide to the control of respiratory disease outbreaks in long-term care facilities. Toronto, 1997.

11. Halloran ME, Struchiner CJ, Longini IM. Study designs for evaluating different efficacy and effectiveness aspects of vaccines. Am J Epidemiol 1997;146:789-803.

12. Patriarca PA, Weber JA, Parker RA et al. Efficacy of influenza vaccine in nursing homes: reduction in illness and complications during an influenza A (H3N2) epidemic. JAMA 1985; 253:1136-39.

13. Monto AS, Hornbuckle K, Ohmit SE. Influenza vaccine effectiveness among elderly nursing home residents: a cohort study. Am J Epidemiol 2001;154:155-60.

14. Gross PA, Hermogenes AW, Sacks HS et al. The efficacy of influenza vaccine in elderly persons: a meta-analysis and review of the literature. Ann Intern Med 1995:123(7);518-27.

15. Centers for Disease Prevention and Control. Preliminary assessment of the 2003-04 inactivated influenza vaccine - Colorado, December 2003. MMWR 2004;53(01);8-11.

Source: A Banerji, MD, Communicable Diseases Control, Toronto Public Health; I Tyler, MD, Community Medicine resident, University of Toronto; MS Finkelstein, MD, E Gournis, MSc, MF Pritchard, MHSc, CPHI(C), F Kolbe, MSc, CPHI(C), and R Shahin, MDCM, Communicable Diseases Control, Toronto Public Health.

[Table of Contents]