Methods

VaccinesThe investigational HD vaccine assessed in this trial is a trivalent, inactivated, split-virus influenza vaccine prepared from influenza viruses propagated in embryonated chicken eggs. Three lots of HD vaccine (lot 1 [UD08635], lot 2 [UD08636], and lot 3 [UD08637]) were prepared, each containing 60 μg of HA per 0.5-mL dose from the following influenza strains: A/New Caledonia/20/99 (H1N1), A/Wisconsin/67/2005 (H3N2), and B/Malaysia/2506/04. The control product was licensed trivalent influenza vaccine (Fluzone [lot U2165C]), which contained 15 μg of HA per 0.5-mL dose from the same 3 influenza strains

Study designThe study was conducted as a multicenter, randomized, double-blind, controlled trial during the fall of 2006. Subjects were assigned by block randomization in a 2:1 ratio (block size, 18) to receive either HD vaccine or SD vaccine. Subjects in the HD vaccine group were further randomized to receive 1 of the 3 vaccine lots

ParticipantsSubjects who were ⩾65 years of age, living in the community, and medically stable were recruited. Volunteers were excluded for the following reasons: allergy to eggs, history of Guillain-Barré syndrome, immunodeficiency or receipt of immunosuppressive therapy, and active neoplastic disease. Written, informed consent was obtained prior to enrollment, and all procedures were reviewed and approved by local and central institutional review boards. The study was conducted in the United States in accordance with the Investigational New Drug (IND) regulations

Clinical proceduresAfter a baseline blood sample was collected, vaccine—identifiable only by code—was administered as a single intramuscular injection in the deltoid muscle by use of a 22-gauge, 2.54-cm needle, and subjects were observed for 30 min. Participants measured their oral temperature daily and recorded injection site and systemic symptoms in a diary for 7 days after vaccination. At day 28 after vaccination, participants’ medical histories were reviewed and second blood samples were obtained

Approximately 6 months after vaccination, subjects were contacted by phone to document all concurrent serious adverse events (SAEs) and use of health care. Injection site reactions—pain, erythema, and swelling—were categorized as mild, moderate, and severe. Pain was categorized as none (0), mild (1, easily tolerated), moderate (2, discomfort interferes with daily activity), and severe (3, incapacitating pain). Erythema and swelling were assessed in terms of the diameter of the maximum reaction size, and classified as follows: mild (<2.5 cm), moderate (⩾2.5–<5 cm), and severe (⩾5 cm). Systemic symptoms were graded as follows. Fever (oral temperature) was categorized as mild (⩾37.5°C–⩽38°C), moderate (>38°C–⩽39°C), and severe (>39°C). The symptoms of headache, malaise, and myalgia were considered mild if they were noticeable but did not interfere with daily activities, moderate if they interfered with daily activity, and severe if they prevented daily activities

Laboratory assaysSerum samples were assessed for antibody to each of the 3 components of the vaccine by hemagglutination inhibition (HAI) testing performed in accordance with standard methods [15]. In brief, serial 2-fold dilutions of neuraminidase-treated serum samples were incubated with 4 HA units of influenza antigen and a chicken red blood cell suspension. After incubation, assay plates were evaluated for the ability of the serum antibodies to inhibit hemagglutination. The serum HAI titer was defined as the dilution factor of the highest serum dilution that completely inhibited hemagglutination. This assay was validated and performed using Good Laboratory Practice standards [16]

Definition of end pointsThe lot consistency of HD vaccine was assessed for each virus strain by use of the pairwise ratios of the HAI geometric mean titers (GMTs) for recipients of the 3 lots of vaccine. Acceptable lot consistency was defined as a ratio for which the limits of the 95% confidence interval (CI) fell between 0.67 and 1.5 for all pairwise ratios for each vaccine strain. Results from the 3 lots of HD vaccine were pooled for further analyses after lot consistency had been demonstrated

The immunogenicity of HD vaccine was assessed in terms of rates of seroconversion and ratio of GMTs for each virus strain, relative to the values obtained for the SD vaccine. Seroconversion was defined as either an increase in HAI titer from <1:10 to ⩾1:40 after vaccination or a ⩾4-fold increase in HAI titer after vaccination from a prevaccination titer⩾1:10. Superiority was demonstrated if the lower limit of the 95% confidence interval for the difference in seroconversion rates (i.e., HD vaccine minus SD vaccine) was >10%, and noninferiority was shown if the lower limit of the 95% confidence interval was >−10%. The ratio of the HAI GMTs for HD vaccine recipients and SD vaccine recipients was assessed for all vaccine strains; superiority was demonstrated if the lower limit of the 95% confidence interval for the ratio was >1.5, and noninferiority was defined as an HAI GMT ratio value >0.67. For HD vaccine to be considered superior to SD vaccine overall, for each measure it was required to demonstrate superiority for at least 2 of the 3 vaccine strains without demonstrating inferiority for any strain

As secondary objectives, we also assessed the rate of seroprotection, defined as an HAI titer ⩾1:40, and the safety profile of HD vaccine compared to SD vaccine. The systemic safety of HD vaccine was considered noninferior to that of SD vaccine if the upper limit of the 2-sided 95% confidence interval for the relative risk was <3. Post hoc analyses were conducted to determine whether age, sex, or the presence of underlying cardiopulmonary disease affected the antibody response to HD vaccine, compared with the response to SD vaccine. We also compared the proportions of vaccinees who had HAI titers of at least 1:80 and 1:160

Statistical methodsThe sample size for this study was selected to achieve the primary objectives of demonstrating lot consistency and superiority of HD vaccine. A total sample size of 3896 was planned (2598 HD vaccine and 1298 SD vaccine) to provide >80% overall power to detect the primary immunologic objectives and allow for a 5% drop-out rate. Although this study was not powered to make safety determinations, it had >99% power to detect a 3-fold increase in the number of subjects experiencing solicited systemic reactions

The full analysis set (FAS) was defined as the set of subjects who received a study vaccine and provided data for at least 1 postvaccination assessment. The immunogenicity analysis was performed on the FAS according to the vaccine the subjects were randomized to receive, whereas the safety analysis was performed on the FAS according to the vaccine subjects actually received. For analysis of lot consistency, at 1 month after vaccination we analyzed HAI GMT ratios for the pairwise comparisons of the 3 HD vaccine lots and their associated 95% confidence intervals for each virus strain. To evaluate the superiority of the HD vaccine, 95% confidence intervals were calculated for the HAI GMT ratios and for the difference in seroconversion rates for the pooled HD vaccine lots and compared to the values for SD vaccine. All the 95% confidence intervals were calculated using normal approximation. Demographic characteristics, safety data, and other secondary endpoints for HAI titers were analyzed using descriptive statistics. The statistical analysis was performed using SAS (version 8.2; SAS Institute)

Results

Between 9 October and 21 December 21, 2006, a total of 3876 subjects were randomized at 30 centers throughout the United States. Twenty-five subjects were excluded from the immunogenicity analysis, as shown in figure 1A; this analysis was performed using data for the 2576 subjects randomized to receive 1 of 3 lots of HD vaccine and the 1275 participants randomized to receive SD vaccine. The safety analysis was performed using actual vaccination data, but 18 subjects were excluded from this analysis because researchers could not verify the vaccine formulation they received. Overall, 46 subjects who received HD vaccine and 32 subjects who received SD vaccine withdrew prior to the end of the study (figure 1B). There were 33 participants who withdrew because of SAEs, but none of their SAEs were considered to be related to vaccination. The mean (± SD) age of study participants was 73 ± 6 years (range, 65–97 years), and there were no statistically significant differences between the groups that received different lots of HD vaccine or between the HD group and the SD group with respect to age, race, sex, or the presence of underlying diseases (table 1)

Immunogenicity analysisBaseline and postvaccination HAI GMTs for subjects who received 1 of the 3 HD vaccine lots were comparable (data not shown). For all vaccine strains, the ratios of HAI GMT for each lot-to-lot comparison of lots 1, 2, and 3 were between 0.94 and 1.04; thus, acceptable manufacturing consistency was demonstrated in accordance with the predefined criteria, and all subsequent analyses were performed on the pooled responses of HD vaccine recipients

Prevaccination HAI GMTs were similar in the HD and SD vaccine groups (table 2). In contrast, HAI GMTs at 28 days after vaccination were higher in the HD vaccine group than in the SD vaccine group for all 3 strains. The HAI GMT ratios were 1.7 (95% CI, 1.6–1.8) for A/H1N1, 1.8 (95% CI, 1.7–2.0) for A/H3N2, and 1.3 (95% CI, 1.2–1.4) for strain B. The absolute difference between the HD group and the SD group with respect to the percentage of subjects who experienced seroconversion was 25.4% for A/H1N1, 18.4% for A/H3N2, and 11.8% for B. Thus, in terms of GMT ratios and rates of seroconversion, HD vaccine met superiority criteria for the 2 A strains and showed noninferiority for the B strain; it demonstrated overall superiority in accordance with predefined criteria

Seroprotective titers for all 3 virus strains were also achieved in a significantly greater proportion of subjects who received HD vaccine than subjects who received SD vaccine (table 2). A significantly higher percentage of subjects in the HD group had postvaccination HAI titers for A/H1N1, A/H3N2, and B of at least 1:80 (1857 [73%] of 2543 vs. 642 [51%] of 1252; 2461 [97%] of 2544 vs. 1116 [89%] of 1252; 1322 [52%] of 2542 vs. 485 [39%] 1252; all P values <.001) and 1:160 (1155 [45%] 2543 vs. 323 [26%] 1252; 2324 [91%] 2544 vs. 982 [78%] 1252; 569 [22%] 2542 vs. 200 [16%] 1252; all P values <.001), compared to subjects in the SD vaccine group

Post hoc analyses were conducted to determine whether sex, age, or the presence of cardiopulmonary disease influenced the immune response. Female subjects experienced a greater response to both vaccines than did males, but both males and females experienced a greater response to HD vaccine than to SD vaccine (table 3). There were no statistically significant differences in postvaccination HAI GMTs in subsets of subjects ⩾75 years of age and those with history of cardiopulmonary disease, compared with younger subjects and those without cardiopulmonary disease. However, the improved immunogenicity of HD vaccine, relative to SD vaccine, was maintained in the subsets of subjects ⩾75 years of age and subjects with a history of cardiopulmonary disease. The HD vaccine also elicited greater responses in subjects with prevaccination titers <1:10. In this subgroup, day 28 HAI GMTs for recipients of HD vaccine were 83 vs. 45 for A/H1N1, 283 vs. 124 for A/H3N2, and 42 vs. 27 for B strain (all P values <.001)

Vaccine safetyRecipients of HD vaccine reported higher rates of local reactions than did SD vaccine recipients during the first 7 days after vaccination (figure 2). For both groups, the most commonly reported reaction was pain; pain of any intensity was reported by 915 (36%) of 2572 HD vaccine recipients and 306 (24%) of 1260 SD vaccine recipients. For the majority of subjects, the pain was of mild intensity and resolved within 3 days; by day 4, only 65 (3%) of 2569 subjects in the HD group and 25 (2%) of 1258 in the SD group reported pain. Erythema of any intensity was reported by 384 (15%) of 2572 in the HD group, compared with 136 (11%) of 1260 in the SD group; moderate and severe reactions were more common in the HD group. Swelling was the least common local reaction reported, and severe reactions were more common in the HD vaccine group (165 [6%] of 2572 vs. 45 [4%] of 1260). Most episodes of erythema and all episodes of swelling resolved within 3 days. Although local reactions were significantly more common in the HD group, the actual difference in the mean maximal zone size of erythema or swelling was ⩽5 mm, and the difference in mean maximum pain scores was very modest (1.12 vs. 1.08)

The rate for any solicited systemic reaction (i.e., fever, headache, malaise, or myalgia) during the first 7 days after vaccination was not significantly different when HD and SD vaccine recipients were compared (34% vs. 29%) (table 4). Most systemic reactions were mild and resolved within 3 days. The predefined criteria indicated that HD vaccine was noninferior to SD vaccine with respect to headache, malaise, myalgia (any or moderate to severe), and any fever, but inferior to SD vaccine with respect to moderate to severe fever, with a relative risk of 3.6 (95% CI, 1.25–10.08) (table 4). However, the number of subjects with moderate or severe fever was relatively low—29 (1.1%) of 2569 subjects in the HD vaccine group and 4 (0.3%) of 1258 in the SD vaccine group—and most fever reactions were moderate, with only 1 report of severe fever in each vaccine group. Also, the mean and median temperatures in the 2 groups during days 0–7 were the same (36.2°C and 36.3°C, respectively)

The rate of unsolicited adverse events within 28 days after vaccination was comparable for the 2 vaccine groups; such events were reported by 559 (22%) of 2563 subjects in the HD group and 276 (22%) of 1252 subjects in the SD group. Of these, 129 (23%) of the events in the HD group and 58 (21%) of the events in the SD group were considered to be related to vaccination. During the 6-month follow-up period, SAEs were reported less commonly by HD vaccine recipients (159 [6%] of 2541) than by SD vaccine recipients (93 [7%] of 1240). Two SAEs were deemed related to vaccination; a 72-year-old woman with a history of Crohn disease experienced an exacerbation requiring hospitalization 2 days after receiving HD vaccine, and an 83-year-old man was diagnosed with myasthenia gravis ∼1 month after receiving SD vaccine. A total of 23 deaths occurred, all of which were considered unrelated to vaccination; the deaths were distributed equally among the 2 vaccine groups (16 [0.6%] of 2573 who received HD vaccine and 7 [0.6%] of 1260 who received SD vaccine died). There was no statistically significant difference between the HD group and the SD group with respect to healthcare use during the 6-month period after vaccination

Discussion

Despite increasing rates of vaccination among older and high-risk adults, the rates of influenza-associated morbidity and mortality continue to increase, sparking debate regarding the efficacy of currently licensed vaccines [1, 17]. Although the precise efficacy of influenza vaccine remains controversial, it appears that efficacy is diminished in older age groups [18]. The presence of chronic diseases, medication use, poor nutrition, and immunosenescence likely all play a role in decreased vaccine responsiveness. A recent review of antibody response in 31 influenza vaccine studies conducted from 1986 to 2002 concluded that elderly adults were 2–4 times less likely to seroconvert or achieve protective HAI titers after vaccination, compared with young adults [6]. Approaches to improving immune response, including live, cold-adapted influenza vaccines, new adjuvants, and virosomal vaccines, have met with mixed success [19–22]. In the present study of older adults, HD vaccine induced significantly higher levels of serum antibody response than did SD influenza vaccine, confirming and extending the results of preliminary studies [11, 14]. Superior immune responses to HD vaccine were demonstrated by both the higher rates of seroconversion and the GMT ratios for the 2 influenza A virus strains included in the HD vaccine. Improved immunogenicity was demonstrated for subjects ⩾75 years old and those with cardiopulmonary disease; these groups are considered to be at the highest risk of influenza-related complications, and they have had suboptimal responses to standard influenza vaccine [12, 23]

Given the higher dose of antigen, the increased rates of local reactogenicity observed in the HD vaccine group were not unexpected. However, injection site symptoms were generally not severe and resolved quickly. The overall rates of systemic complaints were judged to be comparable to those of the SD vaccine. Thus, we believe the increased reactogenicity of HD vaccine is clinically acceptable and outweighed by its superior immunogenicity

The primary limitation of the current study is the lack of data demonstrating clinical efficacy against influenza infection and illness. The collection of meaningful efficacy data was not possible in a study of this size, given the relatively low incidence of influenza infection in vaccinated populations. However, there is substantial evidence that HAI antibody titers represent a good correlate of protection from severe illness in young adults [24]. A consistent inverse quantitative relationship between HAI titer and risk of influenza infection has been observed in both experimental challenge and natural exposure studies [7–9]. In addition, HAI titers correlate inversely with severity of illness and duration of viral shedding, with HAI titers of ⩾1:40 conferring ∼50% protection against infection and higher levels associated with greater degrees of protection [24, 25]. Although HAI titers correlate with protection, the predictive value of these measurements in older adults is imperfect. In this study, antibody responses to SD vaccine in subjects <75 years of age were similar to responses in those ⩾75 years of age, a group in which vaccine efficacy has been shown to be diminished. However, it is possible that the higher postvaccination HAI titers induced by HD vaccine may translate into better protection either by increasing the proportion of subjects who attain a protective titer or by providing protective titers longer into the influenza season

Given the rapidly expanding population of elderly individuals and the immediate medical need, high priority should be given to bringing new vaccines into clinical practice to reduce the substantial morbidity and mortality associated with annual epidemics of influenza. The HD vaccine evaluated in this study represents a straightforward approach and an important step forward in efforts to improve the clinical benefits of influenza vaccination programs for elderly individuals