Abstract
OBJECTIVES. We conducted this study to test the hypothesis that intradermal influenza vaccination at one fifth of a standard dose elicits comparable immunogenicity to full-dose intramuscular vaccination in children.
PATIENTS AND METHODS. We conducted a randomized, open-label study in 112 healthy children aged 3 to <18 years to compare the immunogenicity and safety of intradermal vaccination at one fifth of a dose with standard intramuscular vaccination. Analyses of hemagglutination inhibition antibody titers to each antigen in each group included geometric mean titers before and 21 days after vaccination, fold increase in geometric mean titers after vaccination, seroprotection rate, and seroconversion rate.
RESULTS. The mean age of the subjects was 10.11 ± 4.04 years in the intradermal vaccination group and 10.57 ± 3.91 years in the intramuscular group. Intradermal vaccination was safe. Induration and mild erythema at the injection site were reported at 25% and 57%, respectively, in the intradermal group. Fold increase of geometric mean titers against influenza A/Caledonia was robust in both groups (11.1-fold and 12.9-fold increase in the intramuscular and intradermal groups, respectively), whereas that for B/Shandong was more modest (4.3–4.4). Both approaches elicited very high geometric mean titers against influenza A/Panama: 1360.5 and 893.9 for the intramuscular and intradermal groups, respectively, but because the prevaccination antibody titers were high, the fold increase of geometric mean titers was only 4.5 and 2.6, respectively.
CONCLUSION. The immunogenicity of one fifth of a dose of influenza vaccine delivered by the intradermal route is comparable to the standard-dose intramuscular vaccination in children as young as 3 years of age.
Seasonal interpandemic influenza infection has serious health impacts worldwide, and vaccination is an important public health intervention.1–5 Vaccination is also a key intervention in containing an influenza pandemic.6, 7 Even with well-established vaccine production for interpandemic influenza, shortage of supply sometimes occurs. Therefore, there is justified concern that when large populations need to be vaccinated within a short period of time, demand may exceed production capacity.
Intradermal vaccination exploits the abundance of antigen-presenting cells (macrophages and dendritic cells) that allow a robust immune response to be elicited with a small dose of antigen delivered directly to the skin, which may be a solution to vaccine shortage. Intradermal vaccination is not new. Bacille Calmette-Guerin vaccination, which is widely practiced in many parts of the world, is routinely performed by the intradermal route. Studies exploring intradermal vaccination were conducted with hepatitis B, rabies, and in influenza vaccines as early as the 1940s.10–16 However, these early studies either had no control group or had control groups using the subcutaneous method of administration, were conducted with monovalent or bivalent vaccines with antigen dosages not directly comparable to those available now, lack a clear definition of response in at least 1 study, or used assays that were crude when compared with what is available today. Two recent articles reported on the effectiveness of intradermal influenza vaccination at a reduced dose in adults.17–18 It is important to investigate the possibility of influenza vaccination using the intradermal approach in children.
METHODS
Study Design
This was an open-label, randomized study to compare the safety and immune responses of intramuscular administration of an inactivated influenza vaccine with intradermal administration of the same vaccine at one fifth of a dose in healthy children 3 to <18 years of age. The hypothesis was that intradermal vaccination with one fifth of a standard dose of influenza vaccine elicited comparable immunogenicity as full dose of intramuscular vaccination of the same vaccine. The study protocol was approved by the joint institutional review board of the University of Hong Kong and Queen Mary Hospital (Hong Kong) and was in accordance with the Helsinki Declaration of 1975, as revised in 1983. Subjects were patients who were admitted to the hospital for an acute illness or followed in our outpatient clinic. Written informed consent was obtained from parents and the older children before enrollment. Subjects were considered ineligible if they had an underlying condition that rendered them at risk for influenza complications or a need for regular medication. For children between 3 and <9 years of age, only those who had previous influenza vaccination were recruited to avoid the need for a second dose of immunization. Subjects were matched by age, and a computer-generated randomization list with a block size of 4 was used to assign study subjects to receive an intramuscular dose of 0.5 mL of inactivated trivalent influenza vaccine, or an intradermal dose of 0.1 mL of the same vaccine. Randomization was performed by study personnel immediately before vaccination. The randomization assignment was blinded to the laboratory investigators. All subjects had height and weight measured and the history of influenza vaccination in the previous year elicited. The children were vaccinated in October and November 2005, before the influenza season in Hong Kong that usually peaked in January or February.
Vaccination
The influenza vaccine (Fluarix; GlaxoSmithKline Biologicals, Rixensart, Belgium) we used contained purified surface antigen equivalent to the influenza types and subtypes recommended by the World Health Organization for the 2005–2006 season: A/New Caledonia/20/99 (H1N1)-like strain; A/California/7/2004 (H3N2)-like strain, and B/Shanghai/361/2002-like strain. Each 0.5-mL dose contains at least 15 μg of hemagglutinin antigen per recommended strain.
Intramuscular injection was performed in the deltoid area according to standard procedure. Intradermal injection was performed using a 1-mL syringe calibrated in hundredths (Becton Dickinson, Franklin Lakes, NJ). One research nurse performed all the injections at each subject's deltoid area (Fig 1). The needle was inserted at a 15° angle to the skin according to standard intradermal injection techniques.18 The resulting wheal was graded by the research nurse as 0 (no wheal), 1 (wheal size of 1–2 mm in diameter), 2 (3–5 mm in diameter), or 3 (>5 mm in diameter).
A, Intradermal vaccination using the 1-hand technique with the help of a caretaker in a 3-year-old boy during the study. B, Intradermal vaccination using the 1-hand technique.
Safety
Subjects were observed for 15 to 30 minutes after vaccination for acute reactions. Caretakers were given a diary card for recording of reactions for 3 days. Solicited reactions were fever, malaise, shivering, erythema, induration, and bruising of the injection site of >5 mm in diameter. Adverse effects were graded according to severity. A mild adverse event was one in which that the symptoms were easily tolerated, a moderate reaction caused interference with usual activities, and a serious one resulted in inability to perform usual activities. Research personnel retrieved the adverse events information on day 4 by telephone. Parents were asked to bring back the diary card for cross-checking when the subjects returned for the postvaccination blood draw.
Serologic Studies and Outcome Measures
Serum samples were drawn on day 0 (before vaccination) and day 21 from each subject. The paired samples were tested together by hemagglutination inhibition (HAI) using reference antigens provided as part of the World Health Organization kit provided by the World Health Organization Influenza Collaborating Centre, Centers for Disease Control and Prevention (Atlanta, GA). The sera were treated with receptor-destroying enzyme (RDE) (1:3) at 37°C overnight to remove nonspecific inhibitors, and residual RDE was destroyed by heat inactivation at 56°C for 30 minutes. Serial twofold dilutions of RDE-treated serum (1 in 10) were titrated in a 96-well microtiter plate against 4 hemagglutinin units of reference antigens (H1N1, H3N1, and B/Shanghai/361/2002-like) using 0.25% turkey erythrocytes.19
The following analyses were conducted on the antibody titers to each antigen obtained in the 2 groups of patients: geometric mean titers (GMTs) before and at 21 days after vaccination; fold increase in GMT (calculated as the mean of the ratio of titer before and 21 days after vaccination); the seroprotection rate (defined as the percentage of patients with antibody titers ≥40); and the seroconversion rate (calculated as the percentage of subjects with a prevaccination antibody titer of <40 who developed a fourfold rise or a titer of ≥40 after vaccination).20 Those with a fourfold increase in antibody titer were determined for the intramuscular and intradermal vaccine groups overall, with subset analyses for those whose prevaccine antibody titer was <40 and for those ≥40.
Statistical Analysis
The 2 methods of vaccine administration are considered equivalent if the 95% confidence interval (CI) for the ratio of geometric mean titers (as a percentage) is sufficiently narrow and lies within a range of 80% to 120%. A sample size of 50 subjects per group would have 80% statistical power to identify a difference between the 2 methods using a 2-sided test and a type 1 error rate of 5%, to detect a significant difference of a minimum difference of a ratio of 2.05 in the geometric mean titer and a ratio of 2.33 in the fold increase between the 2 groups. Fifty-six subjects were recruited for each group to take into account an attrition rate of 12%. For the intention-to-treat analysis, missing values of convalescent serology would be assigned the mean value of the group. Pearson correlation was used to detect correlation between individual titers in the intradermal group with age, induration size and BMI. Adverse events for the 2 groups were compared by means of the χ2 test.
RESULTS
Fifty-six subjects each were assigned randomly to receive either intramuscular or intradermal influenza vaccination. Two children in the intramuscular group failed to return for convalescent blood taking. The demographics of the subjects are shown in Table 1. They were comparable in age, male to female ratio, BMI, and history of influenza vaccine in the previous year. When using the 1-hand injection technique, the intradermal vaccination did not pose significant problems, even in young children. Median score of the intradermal lesions was 2 (3- to 5-mm wheal), with a range of 1 to 3.
Demographics of Children Randomly Assigned to Receive Either 0.1 mL of Influenza Vaccine by Intradermal Injection or 0.5 mL of Influenza Vaccine by Intramuscular Injection
Reactogenicity
Complete reactogenicity data were available from all subjects. Both methods of influenza vaccination were well tolerated (Table 2). Not surprisingly, induration was reported more in the intradermal group, but disappeared by the next day in the majority. Mild erythema was reported in 57.1% of the intradermal group, significantly more frequent than in the intramuscular group (3.6%). There was no other significant difference of either local or systemic adverse effects in both groups.
Adverse Effects in Vaccinated Children During the First 3 Days After Vaccination
Immunogenicity
Of the 112 subjects recruited and vaccinated, 110 returned for convalescent blood draw. Two subjects in the intramuscular group, aged 9.4 and 10.1 years, failed to return for convalescent serology and were assigned the mean serology titers of the group. Convalescent blood was drawn at a mean of 23.7 ± 3.1 day and 24.4 ± 5.9 days after vaccination in the intradermal and intramuscular group, respectively (P = not significant). All blood was drawn before the onset of influenza season in January 2006.
There was no statistically significant difference in seroconversion rate and seroprotection rate after vaccination between subjects in the 2 groups against any of the 3 antigens (Table 3). High postvaccination GMT were elicited in both groups against all 3 antigens, especially for the 2 influenza A antigens. There was no correlation of all individual antibody titers with age, BMI, or history of influenza vaccine in the previous year. There was also no correlation between antibody titers and induration size in the intradermal group (data not shown). There was no significant difference between the intramuscular and intradermal groups in the fold increase in GMT antibody titer for H1N1, at 11.1- and 12.9-fold increase respective, or B/Shandong, at 4.4- and 4.3-fold increase, respectively. However, there was a significant difference in fold increase of GMT against H3N2, with a fold rise of 4.5 in the intramuscular group compared with that of 2.6 in the intradermal group (P = .005). Despite this difference, both approaches elicited very high GMT: 1360.5 and 893.9 for the intramuscular and intradermal groups, respectively. For all those with prevaccination antibody titer of <40 to any of the 3 antigens (n = 68), there was no difference in fold-rise in GMT for all antigens combined: 12.5-fold in the intramuscular group and 17.9-fold in the intradermal group, suggesting that the response in seronegative subjects using either method of vaccine administration was equally robust.
Strain-Specific HAI Result for the 3 Antigenic Components of the Trivalent Influenza Vaccine
A significant proportion of subjects in both groups had at least a fourfold increase in titers against the 3 antigens. There was, again, a significant difference in the proportion of children with at least a fourfold rise in GMT against H3N2 between the intramuscular group (66%) and the intradermal group (43%). Those with a lower prevaccination titer were more likely to respond with at least a fourfold rise of titer. A higher percentage of subjects in either group with prevaccination titers of <40 against any of the 3 antigens had at least fourfold increase in GMT when compared with those who had prevaccination titers of ≥40 (Table 3). More subjects with high prevaccination titers had a modest or no increase in titer after vaccination in the intradermal group. Twenty-seven subjects (55%) in the intradermal group with prevaccination titers against H3N2 ≥1: 160 had less than a fourfold rise in titer, as compared with 16 such subjects (33.3%) in the intramuscular group (P = .04).
To validate the power calculation of our sample size, we calculated, with the fold increase defined relative to day 0, the standard deviations of GMT and the mean log10 ratio of GMT to be 0.66 and 0.764, respectively, with little variation according to strain. The sample size of 56 in each group, therefore, had 80% statistical power to identify a difference between the 2 methods using a 2-sided test and a type 1 error rate of 5%, for a ratio of 2.07 in GMT and a ratio of 2.52 in the fold increase.
DISCUSSION
Intradermal administration of influenza vaccine in children was safe and immunogenic. Local reaction was more common in subjects who received the intradermal injection. This finding is similar to that reported in a study of intradermal rabies vaccine in children.15 We did not find any difference in systemic reactions between the 2 groups.
Very high HAI GMTs against all 3 antigens of the influenza vaccine were elicited by the intradermal route by using one fifth of the standard dose. These titers were much higher than those reported in similar adult studies.17, 18 Without direct comparison, it is not known whether intradermal influenza vaccination in children is in fact more immunogenic than in adults. However, available data show that young skin is superior to old skin in resting Langerhans cell numbers and migration response after intradermal injection.21 Because H3N2 has been in circulation worldwide since 1968 and H1N1 since 1977, and young children with previous influenza vaccination were recruited by design, it is not surprising that the majority of children with a mean age of 10 years in this study already had HAI titers of ≥1:40 against both antigens. The protective and high prevaccination GMT supports the notion that children in Hong Kong are heavily influenza experienced. It was suggested that in a partially seropositive population like that in the current study, fold increase in titers tends to underestimate vaccine immunogenicity.22 This, rather than the route of administration, may partly explain the statistical significant difference in GMT fold increase against H3N2 observed. This hypothesis is supported by (1) the comparable fold rises against H1N1 and influenza B/Shandong, against which subjects from both groups had a lower prevaccination GMT and were less likely to have an antibody level ≥1:40, (2) a higher percentage of at least fourfold rise in titer in both groups in those with seronegative prevaccination titers, and (3) a robust and equivalent fold increase for seronegative prevaccination titers against all 3 antigens in both groups. In rabies vaccination of children, there has been concern that antibodies elicited by intradermal vaccination were of lower titer, and may not be as durable.15 With very high postvaccination GMT, it is unlikely that the difference in postvaccination GMT against H3N2 between the 2 groups would be clinically significant. Moreover, with influenza vaccine being given annually, the long-term durability of antibodies is less relevant.
The skin has been the target of immunization for many years, and there is recent renewal of interest in exploiting this immune organ. Studies performed in animals and humans have had promising results.23–25 Epidermal vaccination that includes transcutaneous immunization or needle-free jet injectors targets the epidermis that is abundant in Langerhans cells, and epidermal and dermal vaccination that uses gene gun or electroporation technology targets both layers of the skin that contains Langerhans cells as well as dendritic cells.26 These methods of vaccination require special preparations and devices. Intradermal vaccination targets the dermis that is abundant in dendritic cells and can be performed by a simple needle and syringe using existing vaccine preparation. However, in general, intradermal injection requires more skill than intramuscular injection. Performing intradermal vaccination in struggling young children is more challenging than in obliging small infants. The current study has demonstrated that intradermal injection can be performed safely and effectively even in children as young as 3 years of age using the 1-hand injection technique, which can be easily mastered using standard tuberculin syringes and can also be performed effectively in overweight children with more subcutaneous fat. This has important implication for use in resource poor countries, because sophisticated devices are likely to mean increased costs.
There are limitations in our study. Our subjects were heavily influenza experienced. Although this situation reflects reality, where a proportion of the population is likely to be seropositive to the strains in the vaccine, this will not hold true regarding the pandemic strain. All the intradermal injections were administered by 1 experienced person. A larger variation of results may be expected when many persons give the vaccine. We designed the experiment to obtain the convalescent blood samples 21 days after vaccination. Ideally, another sampling on day 42 postvaccination would have provided additional information on antibody kinetics. However, repeated blood sampling is generally not well received by parents of healthy young children. Future studies should involve children younger than 3 years of age, children seronegative to the vaccine strains, children with various immunocompromised states, comparing 1 intradermal dose with the intramuscular 2-dose regimen in young children, multiple vaccine administrators, longer follow-up, and studies to include efficacy to ascertain that antibodies elicited by different routes of administration are comparable functionally.
This study has provided encouraging data that intradermal injection can elicit comparable immune response against influenza at a much reduced dose in healthy children as young as 3 years of age and is a useful strategy against vaccine shortage. Pandemic vaccines with avian hemagglutinins, especially the nonadjuvanted H5N1 vaccine, are known to be poor immunogens.27, 28 It is not known whether the intradermal route of vaccination will offer any advantages. A recent report of intradermal administration of a subvirion H5N1 vaccine was not immunogenic up to 3 doses in healthy adults.29 We are currently investigating the cellular response of our subjects to better define the role of intradermal influenza vaccination.
Acknowledgments
This work was supported by Research Grants Council of Hong Kong grant HKU 7396/03M, Ellison Medical Foundation grant 1D-1A-0036-02, and the Vice Chancellors Development Fund (University of Hong Kong).
We thank the children and parents who participated in the study, our dedicated team of research staff, and Winnie Lau, who single-handedly performed all the intradermal injections.
Footnotes
- Accepted January 30, 2007.
- Address correspondence to Susan S. Chiu, MD, FAAP, Department of Pediatrics and Adolescent Medicine, Queen Mary Hospital, Pokfulam, Hong Kong Special Administrative Region, China. E-mail: ssschiu{at}hkucc.hku.hk
Financial Disclosure: Dr Lau has conducted clinical trials for Wyeth, GlaxoSmithKline, Medimmune, and Merck Sharp and Dohme (Asia) Limited and is a member of the Steering Committee for Prevention and Control of Infectious Diseases in Asia for GlaxoSmithKline. The other authors have indicated they have no financial relationships relevant to this article to disclose.
REFERENCES
- Copyright © 2007 by the American Academy of Pediatrics
In this issue
Jump to section
Related Articles
Cited By...
- Immunogenicity of the Trivalent Inactivated Influenza Vaccine in Young Children Less than 4 Years of Age, with a Focus on Age and Baseline Antibodies
- Empirical Evaluation of Age Groups and Age-Subgroup Analyses in Pediatric Randomized Trials and Pediatric Meta-analyses
- Intradermal Vaccination with Influenza Virus-Like Particles by Using Microneedles Induces Protection Superior to That with Intramuscular Immunization