Results

Safety in children. With the exception of age, the demographic characteristics of children ⩾60 months of age who participated in the efficacy field trial were similar to the demographic characteristics of all study children. Of children ⩾60 months of age, 87% of CAIV-T recipients and 84% of placebo recipients were white. Sex distribution in this group was balanced (56% of CAIV-T recipients and 50% of placebo recipients were female).

Comparisons of the reactogenicity events that occurred in the 10 days after administration of dose 1 of CAIV-T or placebo spray to participants in the pediatric efficacy study are shown in tables 1 and 2. Previously, it was reported that the occurrence of low-grade fever (temperature, >37.8°C) after administration of dose 1 of CAIV-T occurred significantly more often among CAIV-T recipients than among placebo recipients [1]. Analysis of the data for children ⩾60 months of age revealed no significant increase in the frequency of temperatures >37.8°C (table 2). Likewise, it was previously noted that, for children 15–71 months of age, runny nose and/or nasal congestion, vomiting, and muscle aches were significantly more common among vaccine recipients than among placebo recipients after administration of dose 1; however, no significant increase in these events was observed among children ⩾60 months of age, although the size of the sample in this subset was smaller than the overall sample size.

In the large safety trial that evaluated medically attended events that occurred in the 42 days after administration of vaccine [4], the frequencies of the 4 prespecified group diagnoses of acute respiratory events, acute gastrointestinal events, systemic bacterial infection, or rare events potentially associated with wild-type influenza were not significantly increased in any of the prespecified analyses of children 1–8 years, 9–17 years, 1–17 years, 12–17 months, or 18–35 months of age. Likewise, for the subjects in the indicated age group (i.e., children 5–17 years of age), none of the frequencies of these 4 prespecified group diagnoses were significantly increased (table 3). In fact, acute respiratory events were significantly decreased among vaccine recipients (upper bound of the 90% CI, <1.00).

In this large safety trial, an increased risk for asthma and/or reactive airway disease (RAD) within 42 days of administration of dose 1 of CAIV-T was observed in one prespecified analysis of children 18–35 months of age, but it was not observed in the other prespecified analyses that included children 12–17 months or 1–8 years of age; however, it should be noted that power was limited in the study of children who were 12–17 months of age [4]. No temporal clustering of medical utilization for asthma and/or RAD was seen for the children 18–35 months of age, and no children were hospitalized for this diagnosis during the 42-day observation window. In a post hoc analysis, the risk of asthma and/or RAD among children ⩾12 months of age was assessed by adding groups of older children, with the age of the older children increasing by 6 months for each group added. In this post hoc analysis, an increased risk for asthma and/or RAD was observed, with a risk ratio of 3.5 (90% CI, 1.09–15.54) for the vaccine recipients among children 12–59 months of age (table 4). It should be noted that, although the statistically significant association persisted in the analysis of children 12–59 months of age, no statistically significant increase in risk was observed after vaccine administration, when children 36–59 months of age were analyzed separately. Furthermore, among children 5–17 years of age, the relative risk (RR) was not increased after administration of dose 1 (RR, 0.74; 90% CI, 0.42–1.33) or dose 2 (RR, 0.33; 90% CI, 0.10–0.96) (table 4). Of note, after administration of dose 2, a significant decrease in the RR for asthma and/or RAD was observed among children 5–17 years of age, both in the clinic setting and in all 3 settings combined (hospital, emergency department, and clinic). An analysis of both doses combined also showed a significant reduction in the RR for asthma and/or RAD events occurring in the clinic setting among children 5–17 years of age in this study.

Safety in adults. In the effectiveness trial involving healthy working adults, the demographic characteristics of adults 18–49 years of age were similar (with the exception of age) to the demographic characteristics of the entire study cohort. Eighty-four percent of the vaccine recipients versus 83% of the placebo recipients were white, and 55% of each group of recipients were women.

For the subset of healthy adults 18–49 years of age, as well as for the study population as a whole, vaccine recipients were significantly more likely than placebo recipients to experience some of the same reactogenicity events (tables 5 and 6). Statistically significant increases, with ⩾10% differences between treatment groups, were observed for runny nose and sore throat. Fever defined as a temperature >37.8°C occurred at a similar rate among both the CAIV-T recipients and the placebo recipients, in the study population as a whole and in the population of 18–49 year olds.

Efficacy in children. In year 1 of the pediatric efficacy study, the influenza strains that circulated in the United States were antigenically similar to the strains represented in the vaccine, and, among children 15–71 months of age, efficacy against culture-confirmed influenza was 93.4% (95% CI, 87.5%–96.5%) for the primary end point (2 doses received) [1]. Nineteen percent of the children (312 of 1602 children) who were participating in the study in year 1 were ⩾60 months of age, and a subset of 238 (15% overall) of these 312 children received 2 doses of CAIV-T or placebo (the primary end point). Among this subset of children, the efficacy of CAIV-T was 87% (95% CI, 59.4%–97.9%), for 2 doses received, and 90.6% (95% CI, 70.3%–97.1%), for 1 or 2 doses received. In year 2 of the trial, after a single annual revaccination, the overall efficacy for the population as a whole was 100% against antigenically matched strains and 87% (95% CI, 78%–93%) against any culture-confirmed influenza [2, 3]. Likewise, in children ⩾60 months of age (age range, 60–83 months), efficacy was 100%, for protection against antigenically matched strains, and 86.9% (95% CI, 70.8%–94.1%), for protection against all culture-confirmed influenza, regardless of antigenic match. Of importance, efficacy against the predominantly circulating influenza type A/Sydney/5/97 (H3N2) strain that was antigenically drifted from the recommended H3N2 vaccine strain was, for the overall study population, 85.9% (95% CI, 75.3%–91.9%), and, for children ⩾60 months of age, efficacy was 86.3% (95% CI, 69.4%–93.9%). When efficacy was analyzed by age in 1-year increments, the vaccine had significant and high efficacy in all age groups for both study years, including the group of subjects ⩾60 months of age (table 7).

Effectiveness in children. Table 8 presents effectiveness data for each study year in the pediatric efficacy study, for the overall population of children 15–71 months of age, and for the subset of children ⩾60 months of age. For the overall study population, in year 1 of the study, when the circulating wild-type influenza strains were well matched to the strains represented in the vaccine, the effectiveness end points restricted to culture-positive illness were all highly significant (P < .01): effectiveness against febrile illness with antibiotic use was 97.1%, effectiveness against febrile otitis media with antibiotic use was 100%, effectiveness against days of day care/preschool/school missed in association with illness (hereafter known as “illness-associated days of day care/preschool/school missed”) was 94.4%, effectiveness against illness-associated days of work missed was 97.7%, and effectiveness against illness-associated health care provider visits was 93.9% [2, 6]. Likewise, these culture-positive effectiveness end points for children ⩾60 months of age in year 1 of the trial were all significantly reduced (P ⩽ .0085), and the reductions ranged from 93.7% to 100% (table 8). For the overall study population, regardless of the influenza culture result, the reductions for these effectiveness end points were 9.4%–35%; those that were significantly reduced included febrile illness with antibiotic use (reduction, 31%), febrile otitis media with antibiotic use (reduction, 35%), and illness-associated health care provider visits (reduction, 13.4%). For the subset of children ⩾60 months of age, these reductions were from -10.2% to 53.9%, and 1 effectiveness end point was significantly reduced (reduction for illness-associated days of work missed, 53.9% [P = .005]).

In year 2 of the pediatric efficacy study, when the wild-type influenza strain that predominately circulated was the drifted A/Sydney/5/97 (H3N2) strain, the effectiveness against the culture-positive end points of illness was 87.8%–95% for the overall study population: each culture-confirmed end point of effectiveness was statistically significant (P ⩽ .01). In year 2, for the overall study population, the end points of effectiveness that were significantly reduced, regardless of influenza culture result, were febrile otitis media with antibiotic use (reduction, 20.9%; P = .04) and illness-associated days of day care/preschool/school missed (reduction, 16.6%; P = .01). For the subset of children ⩾60 months of age, illness-associated days of day care/preschool/school missed was the end point that was statistically significant (reduction, 10.8%; P = .04).

Effectiveness in adults. The effectiveness of CAIV-T was evaluated in healthy working adults 18–64 years of age, during the 1997–98 influenza season, when the drifted strain A/Sydney/5/97 (H3N2) predominantly circulated (table 9). Evaluation revealed that the 9.7% reduction in the proportion of participants with 1 or more febrile illnesses defined as AFI (the primary end point) was not significant. However, there were significant reductions in the occurrence of the other 2 prespecified defined illnesses (SFI and FURI), as well as in the occurrence of 2 additional defined illnesses (CDC-ILI and DOD-ILI) [5, 7]. Likewise, the effectiveness analyses involving healthy working adults 18–49 years of age revealed no significant protection against the occurrence of AFI (reduction, 10.9%; P = .16), but significant reductions were observed regarding the occurrence of SFI (reduction, 19.5%; P = .02), FURI (reduction, 23.7%; P = .007), CDC-ILI (reduction, 24.4%; P < .002), and DOD-ILI (reduction, 21.1%; P = .0089) (table 9). Across the 5 illness definitions for the 3 illness-associated outcome measures (occurrence of illness, the number of episodes of illness, and the number of days of illness), all 15 comparisons were consistent for the study population as a whole and for the subset of individuals 18–49 years of age. Specifically, 13 of 15 comparisons were significantly reduced in both age strata, among vaccine recipients; the occurrence of AFI (reduction, 9.7% in 18–64 year olds vs. 10.9% in 18–49 year olds) and the number of AFI episodes (reduction, 10.0% in 18–64 year olds vs. 11.2% in 18–49 year olds) were not significantly reduced. In contrast to AFI, for the other 4 illness definitions (SFI, FURI, CDC-ILI, and DOD-ILI), occurrence of illness, the number of episodes of illness, and the number of days of illness were significantly reduced for both age strata, among the vaccine recipients compared with the placebo recipients.

Across the 5 illness definitions and the 4 illness-associated outcome measures (illness-associated days of work missed, illness-associated health care provider visits, antibiotic use, and use of over-the-counter medication), 17 of the 20 comparisons were significantly reduced in the overall study population and in the subset of 18–49 year olds (table 9). However, in the subset of 18–49 year olds, illness-associated days of work missed, for SFI (reduction, 12% [P = .1326] vs. 17.9% [P = .012] for 18–64 year olds); illness-associated days of work missed, for CDC-ILI (reduction, 12.9% [P = .1033] vs. 19.9% [P = .0043] for 18–64 year olds); and illness-associated health care provider visits, for CDC-ILI (reduction, 11.1% [P = .1991] vs. 20.0% [P = .0079] for 18–64 year olds) were not significantly reduced, compared with data for the overall study population. These differences reflect that inclusion of the data for the 641 (14%) of 4561 adults 50–64 years of age provided significant reductions in these 3 outcome measures for the overall population.

In an additional analysis that assessed whether there was any evidence of an interaction between age and vaccine effectiveness, we compared results for persons 18–49 years of age (n = 3920) with results for persons ⩾50 years of age (n = 641). There were no statistically significant differences, according to age group, in estimates of vaccine effectiveness for the number of episodes of illness or the number of days of illness. Thus, CAIV-T was effective both in adults <50 years of age and in adults ⩾50 years of age. Of note, there were numerically greater reductions in the number of illness-associated days of work missed and the number of illness-associated health care provider visits among persons ⩾50 years of age, compared with persons 18–49 years of age (data not shown). The older adults may have more-severe disease and may take time off work to seek medical attention and to recover.