Published ahead of print on December 4, 2003, doi:10.1164/rccm.200309-1251OC
© 2004 American Thoracic Society Influenza Vaccination in Children with AsthmaRandomized Double-Blind Placebo-controlled TrialDepartment of General Practice; Department of Pediatrics, Subdivision of Pediatric Respiratory Medicine; Department of Virology; and Institute for Medical Technology Assessment, Erasmus MC-University Medical Center Rotterdam, Rotterdam, The Netherlands Correspondence and requests for reprints should be addressed to Johannes C. van der Wouden, Ph.D., Department of General Practice, Room Ff305, Erasmus MC-University Medical Center Rotterdam, P.O. Box 1738, 3000 DR Rotterdam, the Netherlands. E-mail: j.vanderwouden{at}erasmusmc.nl
There is little evidence that influenza vaccination reduces asthma exacerbations. We determined whether influenza vaccination is more effective than placebo in 6–18-year-old children with asthma. We performed a randomized, double-blind, placebo-controlled trial. Parenteral vaccination with inactivated influenza vaccine or placebo took place approximately November 1, and children were followed until April 1 of the next year. Airway symptoms were reported in a diary. When symptom scores reached a predefined level, a pharyngeal swab was taken. Primary outcome was the number of asthma exacerbations associated with virologically proven influenza infection. Three hundred forty-nine children were assigned placebo, and 347 were assigned vaccine. Pharyngeal swabs positive for influenza were related to 42 asthma exacerbations, 24 in the vaccine group and 18 in the placebo group, a difference of 33% favoring placebo (31% after adjustment for confounders; 95% confidence interval, -34% to 161%). Influenza-related asthma exacerbations were of similar severity in both groups; they lasted 3.1 days shorter in the vaccine group (95% confidence interval, -6.2 to 0.002 days, p = 0.06). We conclude that influenza vaccination did not result in a significant reduction of the number, severity, or duration of asthma exacerbations caused by influenza. Additional studies are warranted to justify routine influenza vaccination of children with asthma.
Key Words: influenza vaccination • asthma • children Infections of viral origin, including influenza, cause exacerbations in patients with asthma (1, 2). Most guidelines support the opinion that influenza vaccination should be given to patients with asthma, including children (3). Over the years, the necessity of vaccinating patients with asthma has been discussed (4–6). A Cochrane systematic review on influenza vaccination in people with asthma concludes that because of the lack of randomized trials, the evidence to assess the benefits and risks of influenza vaccination is insufficient (7). Inhalation medication for relief and maintenance therapy has made it easier to control asthma, and exacerbations can be prevented or suppressed; thus, patients can lead normal lives. The proportion of children with asthma that is vaccinated varies widely and is lower than in the older population (8). This is partly due to doubts among patients and doctors about the benefit of the vaccination and the lack of evidence for the protective effect of influenza vaccination against asthma exacerbations (9, 10). An answer to the question of whether influenza vaccination can prevent exacerbations in children with asthma is necessary to either validate or modify current guidelines and may enhance evidence-based practice. We conducted a study to investigate whether influenza vaccination in children with asthma prevents asthma exacerbations provoked by influenza infection. Some of the results of this study have been previously published in the form of abstracts (11–13).
Participants We performed a randomized, double-blind, placebo-controlled parallel trial in children 6–18 years of age with asthma who were recruited through family physicians in the city of Rotterdam and its surroundings during the winter seasons of 1999–2000 and 2000–2001. Patients could only participate for one season. Exclusion criteria were other chronic diseases, allergy to chicken protein, and insufficient understanding of the Dutch language (see online supplement for further details). We informed parents and children orally and in written text about the purpose and contents of the study and obtained informed consent for those willing to participate. The Medical Ethical Committee of Erasmus MC-University Medical Center Rotterdam approved the study.
Interventions
Objectives
Outcomes
Respiratory Tract Symptoms and Episodes
Virologic and Serologic Assessments Throat swab material was analyzed by culture, immunofluorescence, and reverse transcription-polymerase chain reaction. A full description of the methods used is provided in the online supplement. We took blood samples (4 cc) before vaccination (sample 1), 14–21 days afterward (sample 2), and at the end of the season, around the 1st of April (sample 3). The presence of influenza virus–specific antibodies was performed by a hemagglutination inhibition assay (details are given in the online supplement). A fourfold increase in postseason serum antibody titer (comparing sample 2 and 3) was considered an indication for an influenza virus infection.
Sample Size The online supplement provides additional information on sample size calculation, randomization procedure, blinding, assessment of adverse effects, and statistical analyses.
Altogether, 144 family practices (200 physicians) participated. They selected 3,220 children and informed these children and their parents by letter about the objectives of our study. In total, 1,365 children were willing to participate. After inclusion and informed consent, 697 children were randomized (first season, 297; second season, 400) (Figure 2) . One participant, after having received the study intervention, was mistakenly vaccinated by her family doctor as well and excluded because of protocol violation. Of the remaining participants, 347 were vaccinated with influenza vaccine and 349 with placebo. In each group, 344 participants provided diary data for primary outcomes. The proportion of diary days for which no data were available was 4.7%. Because the number of missing diaries was very small, we decided not to impute data but to exclude these children from the analysis for these outcomes.
Baseline Characteristics The groups were similar for baseline characteristics except for the history of allergy (Table 1) . Almost 90% of children had used maintenance medication for asthma during the previous 12 months; 55% of the children had been vaccinated against influenza before.
Influenza Activity In our study, in the season 1999–2000, the first influenza positive swab was taken on December 17, 1999, and the last one on January 20, 2000. All were influenza A virus, and the four cultures from these swabs all yielded H3N2. Dutch primary care sentinel stations detected influenza A H3N2 from week 45 (November 8) in 1999 until week 8 (February 21) in 2000 (14). In the season 2000–2001 in our study, the first influenza A virus was detected on November 21, 2000, and the last on March 23, 2001. All cultures yielded H1N1. Three influenza B positive swabs were obtained March 7–13, 2001. This season sentinel stations reported influenza A H1N1 from week 50 (December 11) in 2000 until week 8 (February 19) in 2001. Influenza B was spotted in week 8 (February 19–26, 2001) (15).
Primary Outcome
The mean length of an influenza-related asthma exacerbation was 9.2 (vaccine group) and 11.2 (placebo group) and after adjustment for confounders 3.1 days longer in the placebo group (p = 0.06). There were no differences in severity (Table 3) (p = 0.11).
Respiratory Tract Episodes Subgroup analysis by vaccination history did not reveal any differences between children that had been vaccinated before and those that were vaccinated for the first time.
Medication, Health Care Use, and Absenteeism One of the outcomes not reported in this article is quality of life during the asthma exacerbations. An article describing the translation and validation of the questionnaire used to assess quality of life has been submitted. Another article studying various secondary outcomes in more detail is in preparation.
Serologic Results
Adverse Reactions
This study shows that influenza vaccination was not more effective than placebo in reducing the number of asthma exacerbations caused by influenza infections in children. The duration of influenza-related asthma exacerbations was 3 days shorter in the vaccine group, but this was not statistically significant. There was no difference in severity. Regarding the other secondary endpoints, vaccinated children had shorter exacerbations (irrespective of their cause) and fewer seroconversions than the placebo group but reported more side effects. There was no difference in proportion of days with asthma symptoms. For both seasons, the vaccine strain matched well with epidemic virus strains (14, 15). For children, the Dutch influenza bulletins reported a low number of influenza positive cases in 1999–2000 (16) and a possible epidemic increase in 2000–2001 (17), both of which are in accordance with our findings. The second season was unusual because of the relatively extensive circulation of subtype H1N1 and the low activity of subtype H3N2 (15). We found no difference in the number of influenza-related asthma exacerbations between both groups, although the vaccine group had a significantly higher protective level of IgG. Is there a possible explanation for this lack of effect? Studies in healthy subjects have shown that parenteral inactivated influenza vaccine does not induce as good an IgA response in the mucous membranes of the respiratory tract (being the first line of defense) (18) as does priming by natural infection or intranasally administered influenza vaccine (19–23). Hence, we speculate that parenteral vaccination may be relatively ineffective in preventing mucosal infection with influenza virus and subsequent asthma exacerbations. After the 1986 study of Stenius-Aarniala and colleagues (24), this is the first randomized trial investigating parenteral influenza vaccination that takes influenza-related asthma exacerbations as an outcome (7). As the study of Stenius-Aarniala and colleagues was unsuccessful because of the extremely low influenza activity, we cannot compare our findings. Recent nonrandomized studies in children with asthma showed conflicting results (25–29). Was our study sample large enough? We were surprised by the low number of influenza-positive swabs, especially in the first season. Our choice to use pharyngeal swabs may have played a role but compared with nasopharyngeal aspirates sensitivity is only approximately 20% lower (30). By recruiting more children than originally planned in the second season, we hoped to compensate for the low incidence in the first season. The final answer to the power question is given by our results, especially the 95% confidence interval for the primary endpoint (31). The interval ranges from 34% reduction to 161% increase. This firmly excludes a 50% reduction of influenza-related asthma exacerbations by vaccination, which we considered to be the threshold for clinical relevance. Even a 35% reduction by vaccination is excluded by the confidence interval. Our finding of no difference is further corroborated by the lack of differences on almost all of our other clinical outcome measures, as shown in Table 3. In previous studies, seroconversion throughout the influenza season was often the most important outcome measure to assess effectiveness of influenza vaccines. However, this outcome measure has serious shortcomings as a proxy measure for influenza infection (32). The most important problem is that seroconversion is more difficult or even impossible to attain in subjects with elevated titer levels at the start of the season (33). A closer look at the difference in mean duration of all asthma exacerbations revealed that extremely long episodes (more than 3 SDs above mean) were found more often in the placebo group than in the vaccine group (n = 12 vs. n = 4). Most of these episodes began before influenza activity started in the Netherlands, which makes it likely that the difference in duration was a chance finding, not caused by the difference in treatment assignment. Regarding adverse reactions, we found differences between groups for local symptoms (red or stiff arm) and myalgia, as reported in previous studies (33, 34). There has been some debate about exacerbations as a direct result of vaccination (34, 35). In this study, we found no differences in airway symptoms during the week after vaccination, making vaccination-induced airway symptoms unlikely. We found relatively few influenza-related asthma exacerbations (9.1% of all calls, 1.8% of all asthma exacerbations). Given the proportion of children that showed a fourfold increase of antibodies during the season, a considerable number of influenza infections apparently did not give symptoms or gave symptoms that were not severe enough to pass our predefined threshold. Inevitably, our episode definition, although used before (2), has arbitrary elements. Other cutoff points for the number of days between two episodes would give different results. For our primary endpoint, four children had two episodes related to the same throat swab that could be counted as one episode by stretching this criterion from 2 to 3 days. Three of these children were in the vaccine group. However, the reduction in number of episodes in the vaccine group as a consequence of changing this criterion (still negative, and the confidence interval still excluding a 50% reduction) would be balanced by an increase in length of the episodes. In conclusion, our randomized placebo-controlled study showed no relevant effect of influenza vaccination on the number and severity of influenza-related asthma exacerbations in children with asthma. Vaccinated children tended to have shorter exacerbations throughout the season but reported more (mild) adverse effects after vaccination than children receiving placebo. Both the limited effectiveness of influenza vaccination found in this study and the low incidence of influenza, observed in the first season, warrant additional studies to justify routine influenza vaccination of children with asthma in general practice.
The authors thank Philip Rothbarth for challenging us to design the study. They also thank the children, parents, and general practitioners who took part in the study; the research nurses and Lya Euser, who coordinated fieldwork; Frank Pistoor for his assistance on virologic and serologic matters; Kris Sieradzan and Roel Verkooyen for helping with data entry; and AstraZeneca BV, Zoetermeer, the Netherlands, for providing local anesthetic plasters.
Supported by Health Research and Development Council (ZON-MW), The Hague, the Netherlands. This article has an online supplement, which is accessible from this issue's table of contents online at www.atsjournals.org Conflict of Interest Statement: H.J.B. has no declared conflict of interest; R.M.D.B. has no declared conflict of interest; J.C.D. has no declared conflict of interest; L.W.A.V. has no declared conflict of interest; G.F.R. has no declared conflict of interest; A.D.M.E.O. has no declared conflict of interest, but Solvay contributes to a foundation that sponsors research on respiratory issues although it has no direct influence on the research conducted on behalf of the foundation at the Dept. of Virology of the Erasmus Medical Center, Rotterdam; M.P.M.H.R-V. has no declared conflict of interest; S.T. has no declared conflict of interest; J.C.V. has no declared conflict of interest. Received in original form September 9, 2003; accepted in final form November 27, 2003
This article has been cited by other articles:
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
TOP
ABSTRACT
METHODS
= 0.05 and 80% power. Assuming a 30% influenza incidence in the placebo group and 50% of these infected children reacting with an exacerbation, we aimed at a total number of 600 children, that is, 300 children per season. 
