Introduction

Epidemiological studies carried out during numerous influenza epidemics have demonstrated that the attack rates of influenza are highest in infants and children [14, 28]. The reported rates of infection during annual epidemics range typically between 20 and 30% of children each year [18, 28, 34] and attack rates of up to 50% have been observed in children attending day care [20]. Thus, children are regarded as the main transmitters of influenza virus in the community. Concentrations of influenza virus in the nasopharynx are higher and duration of viral shedding is longer in children than in adults [11, 16]. In addition, immunity after primary infection lasts in children for much shorter time than in adults and subclinical re-infections are needed to establish natural immunity [9].

The high influenza morbidity during childhood has been reported to result in higher frequencies of outpatients’ visits for acute respiratory disease during influenza epidemics in children up to the age of 9 years [13, 33]. Influenza is associated with a wide spectrum of complications such as croup, tracheobronchitis, bronchiolitis, pneumonia [35], acute otitis media [43], sinusitis [52], acute exacerbations in asthmatic children [33], febrile seizures [6], myositis [25], myocarditis [15], encephalitis [27] and Reye’s syndrome [4]. Several studies have demonstrated that influenza disease in children has considerable socioeconomic impact on the children themselves and concerning families [18, 36]. Especially young children are hospitalized frequently for influenza-related illness [19, 21, 37, 51]. Although the overall influenza-associated mortality is low in childhood population, life-threatening diseases and deaths from influenza and its complications do occur [5, 48]. Furthermore, infants and children with chronic medical conditions are at increased risk of influenza-associated mortality [33].

In spite of the high burden of influenza in children, in most countries, annual influenza vaccination is not recommended for healthy children, but rather only for those with underlying conditions which place them at increased risk for complications from influenza [42]. The Advisory Committee on Immunization Practice (ACIP) of the Center of Disease Control and Prevention (CDC) in the United States and the National Advisory Committee on Immunization (NACI) in Canada recommended universal influenza vaccination for all children aged 6 months to 18 years and 6–23 months, respectively, due to their higher risk of hospitalization and complications from influenza [2, 31, 47]. Recently, a European vaccination expert group has published a similar recommendation [17]. However, more research is needed to warrant universal influenza vaccination during childhood.

The objective of the present study was to provide seroepidemiological information on the prevalence of influenza A and B antibodies in children and adolescents compared to adult controls in Germany. Thus, virus-specific immunoglobulin (Ig) G class antibodies were determined using commercial enzyme-linked immunosorbent assays (ELISA) in a group of 1,111 children and adolescents as well as 809 healthy blood donors serving as controls.

Materials and methods

Patients and serum samples

The study group contained 1,111 sera from children and adolescents aged 0–18 years, living in a central part of Germany (Erfurt, capital of the German federal state Thuringia, and its rural surroundings) (Tables 2, 3). During the first year of life, a total number of 206 infants, between 8 (3.9%) and 39 (18.9%) infants per month of age, were included (Table 1). About 845 (76.1%) subjects were healthy and 266 (23.9%) children and adolescents suffered from adiposity or growth disorders. Sera were randomly and anonymously collected from remaining serum samples which were submitted originally to the laboratory in a time frame of 1999–2006. Of the age group 0–1 year, 130 sera were taken in 1999–2000 and 76 sera were collected between 2001 and 2006. Among the children aged >1–3 years, 108 sera were from 1999 to 2000 and 42 sera were from 2001 to 2006. From the children >3–6 years of age, 145 sera were collected in 1999–2000 and 10 sera were taken between 2001 and 2006. All sera from children and adolescents aged between >6 and 18 years were collected in 1999–2000. Samples were taken for routine serological diagnostics, e.g. medical check-up of school children or determination of immune status. In all subjects included, there were no clinical and laboratory data indicating infectious diseases. Data about the immunization history against influenza could not be obtained. Sera samples were included in this study on the basis of a statement of the Central Ethical Committee of Germany about the use of human samples for research studies [53]. In addition, the Ethical Committee of the University of Jena (Germany) agreed to this study.

Table 1 Prevalence of IgG antibodies against influenza A and B virus in children of the first year of life
Full size table

About 809 sera from healthy voluntary blood donors aged 18–30 years served as controls (Tables 2, 3). These control subjects also lived in the region of Erfurt (Thuringia, Germany). Sera were taken randomly and anonymously during 1999 and 2000. There were no data on the immunization history against influenza. Subjects’ consent was obtained prior to processing the samples.

Table 2 Prevalence of IgG antibodies against influenza A virus in children and adolescents as well as adult controls
Full size table
Table 3 Prevalence of IgG antibodies against influenza B virus in children and adolescents as well as adult controls
Full size table

Sera were stored in aliquots at −20°C without interruption until testing for antibodies to influenza A and B viruses. All sera were allowed to attain room temperature immediately before investigation. Antibody testing was carried out blindly in groups of 20–30 serum samples. Sera were tested twice in two different test procedures and only sera with concurrent results were included in this study.

Serological testing

The influenza virus A IgG ELISA (no. INFG0290, lot INFAG-023 and INFAG-27) and influenza virus B IgG ELISA (no. INFG0300, lot INFBG-024 and INFBG-29) manufactured by NovaTec Immundiagnostica (Dietzenbach, Germany) were used for qualitative and semi-quantitative antibody testing. The micro-titration strip wells were pre-coated with influenza virus type-specific antigens consisting of matrix proteins and nucleoproteins of influenza A and B virus, respectively. According to information from the manufacturer, influenza A virus antigen was prepared from the strain A/Texas/1/77 (H3N2) and the influenza B virus antigen was obtained from the strain B/HongKong/5/72 after viral propagation in embryonated chicken eggs. Testing of sera was carried out according to the manufacturer’s instructions. In short, after 1 in 100 dilution of test samples, coated viral antigen was incubated with the sample dilution, washed by using the microplate washer aw1 (Anthos Labtec Instruments, Wals, Austria) and incubated again with the horseradish peroxidase-labeled anti-human IgG conjugate which binds to the captured influenza virus type-specific antibodies. Washing of plates was followed by the incubation with the tetramethylbenzidine substrate resulting in a blue reaction product. Sulfuric acid was added to stop the reaction producing a yellow endpoint color. Plates were read photometrically using an Anthos reader 2010 (Anthos Labtec Instruments) at 450 nm and a reference wavelength of 620 nm against the blank value. Test results were considered valid on the basis of the absorbance values of substrate blank (lower than 0.100), negative control (lower than 0.200), cut-off control (between 0.250 and 0.900) and positive control (greater than cut-off control). Samples were considered positive if the absorbance value was higher than 10% over the cut-off. Negative samples had an absorbance value lower than 10% below the cut-off. Samples with an absorbance value up to 10% above or below the cut-off were regarded as borderline.

The ELISAs also offered the opportunity for semi-quantitative evaluation of the results. By using the arbitrary units “NovaTec-Units” (NTU), a relative quantitation within the test was possible. The NTUs were calculated with the following formula:

$$ \frac {{\text{Subject's mean absorbance}} \times 10} {\text{Cut-off}} = (\text{NovaTec-Units} = \text{NTU})$$

For refined semi-quantitative determination of antibodies, sera which showed absorbance values equal to or greater than 2.400 were re-tested on the basis of linear regression analysis after dilution of 1:2, 1:5 and 1:10, respectively, in sample dilution buffer. The NTUs were calculated by multiplication with the dilution factor.

According to the manufacturer’s information, the performance of the influenza A and the influenza B virus ELISAs was assessed using the VIR-ELISA Anti-Influenza A IgG and VIR-ELISA Anti-Influenza B IgG (Viro-Immun Labor-Diagnostika, Oberusel, Germany), respectively. The influenza virus A IgG ELISA had 93% specificity and 96% sensitivity, whereas the influenza virus B IgG ELISA had 87% specificity and 96% sensitivity.

Statistical analysis

A sample size of about 150 subjects per pre-specified age group was planned to assure that a single two-sided 95% confidence interval (CI 95%) for the prevalence would extend at most 8% from the observed value for a prevalence range of 10–90%. Since analysis of monthly prevalence during the first year after birth was not planned in advance less precise estimates result from small sample sizes.

Antibody prevalence was calculated within the study and control group as the number of seropositive cases divided by the number of all subjects enrolled. Assuming binominal distribution, the two-sided exact CI 95% was calculated. The Cochran–Armitage test for trend was used to examine the increase of antibody prevalence by age. Differences between the prevalence of antibodies of male and female subjects adjusted to age were investigated by the Mantel–Haenszel test. Using logistic regression odds ratios comparing antibody detection in the different age groups of children and adolescents with the entire control group as reference were evaluated by the Wald statistics. Age and gender-specific antibody concentrations were described by mean and standard deviation (SD). The influence of age and gender on the concentration of antibodies was analyzed using linear regression. Antibody concentrations of the different age groups of children and adolescents were compared with the correspondent data of the control group by the Dunnett test. For both analyses, antibody concentration was transformed by the common logarithm. The level of significance was 0.05 (two-sided). SAS V9.1 software was used for statistical analyses.

Results

Influenza A virus

Data on prevalence of IgG antibodies against influenza A virus are shown in Tables 1, 2, and 3 as well as in the Fig. 1. In the population of children and adolescents, the total prevalence of antibodies (CI 95%) against influenza A virus was calculated as 82.0% (79.6–82.2%) (Table 2). The frequency of positive samples was between 76.2% (52.8–91.8%) and 87.2% (72.6–95.7%) during the first 4 months of life (Table 1). Afterwards, the antibody prevalence declined to 33.3% (7.5–70.1%) by the end of the seventh month of the first year of life. Autochthonous production of antibodies against influenza A virus started during the second half of the first year and increased from 53.3% (55.6–69.2%) among the >1–3 years old to 74.8% (67.2–81.5%) among the >3–6 years old and to 93.3% (88.1–96.8%) among the >6–9 years old (Table 2; Fig. 1). In the age group >9–12 years, 98.0% (94.3–99.6%) of the children had been infected followed by prevalence rates of 99.3% (96.3–100%) among >12–15 years old and 100.0% (97.6–100%) among >15- to 18-year-old adolescents. Statistical analyses demonstrated that the antibody prevalence against influenza A virus increased significantly with an increasing age during childhood (trend test P < 0.001). Adjusted for age, there were no significant differences between the prevalence of antibodies among boys and girls (P = 0.169). In the population of adult blood donor controls, the total prevalence of antibodies against influenza A virus was estimated as 99.4% (98.6–99.8%) (Table 2). There was no influence of age (trend test P = 0.570) and gender (P = 0.145) on the frequency of positive samples (Fig. 1). Children and adolescents up to the age group of >6–9 years had a significantly lower prevalence of antibodies than the adult controls (P < 0.001).

Fig. 1
figure 1

Age-specific distribution of qualitative IgG antibodies (percent) against influenza A virus (IAV) in the study group of children and adolescents as well as in the control group of adults

Full size image

In the study group of children and adolescents, the mean concentration of antibodies against influenza A virus was calculated as 31.7 NTU (SD 37.8 NTU). After the mean antibody concentration declined in children of the first year of life (19.1 NTU, SD 22.5 NTU) to children aged >1–3 years (17.1 NTU, SD 21.4 NTU), the mean concentration increased significantly with increasing age (>3–6 years: 33.5 NTU, SD 50.5 NTU; >6–9 years: 38.3 NTU, SD 49.3 NTU; >9–12 years: 33.9 NTU, SD 24.0 NTU; >12–15 years: 47.1 NTU, SD 46.1 NTU; >15–18 years: 37.6 NTU, SD 30.7 NTU; trend test P < 0.001; Fig. 2). In all age groups, significant differences could not be found between boys and girls (P = 0.667). The adult controls had a mean antibody concentration of 33.2 NTU (SD 23.8 NTU). Significant differences between the age groups selected (Fig. 2, P = 0.214) as well as between men and women could not be detected (P = 0.307). Children up to the age group of >3–6 years had significantly lower concentrations of IgG antibodies against influenza A virus than the control group of adults (P < 0.001), whereas the concentrations were significantly higher among the >12- to 15-year-old adolescents (P < 0.001, Fig. 2).

Fig. 2
figure 2

Age-specific distribution of IgG antibody concentrations [NovaTec-Units (NTU)] against influenza A virus (IAV) in children, adolescents and adult controls. Boxplot shows median, 25 and 75th percentile, minimum, maximum and/or outliers. Open circles values >1.5 inter-quartile, Asterisks values >3 inter-quartile range from the end of the box

Full size image

Influenza B virus

The prevalence of antibodies against influenza B virus among the subjects of the study and control group is shown in Tables 1 and 3 as well as in the Fig. 3. In children and adolescents, the total prevalence of antibodies against influenza B virus was 9.6% (8.0–11.5%) (Table 3). There was a decline of antibody-positives from 45.2% (27.3–64.0%) to 0% (0.0–16.1%) during the first 2 months of life (Table 1). Within the second half of the first year of life, low antibody prevalence up to 12.5% (0.3–52.7%) against influenza B virus was measured. In the age groups of children >1–3 years up to >9–12 years, the frequency of positive samples was between 1.3% (0.2–4.7%) and 3.3% (1.1–7.6%) (Table 3; Fig. 3). Afterwards, the prevalence of antibodies increased to 24.0% (17.7–31.6%) among the >12- to 15-year-old adolescents and to 24.7% (18.0–32.4%) among the >15–18 years old. The prevalence of antibodies against influenza B virus rose significantly with increasing age during childhood (trend test P < 0.001, Fig. 3). There were no significant differences between the prevalence of antibodies as a function of gender in all age groups of children and adolescents (P = 0.427). In the control group of adults, the total prevalence of antibodies against influenza B virus was calculated as 56.7% (53.2–60.2%) (Table 3). There was also a significant increase of the prevalence with an increasing age (tend test P < 0.001, Fig. 3) and adjusted to the age, a significantly lower prevalence in women than in men was detected (P = 0.014). In all age groups of children and adolescents, significantly lower prevalence of antibodies was measured in comparison to the adult controls (P < 0.001, Fig. 3).

Fig. 3
figure 3

Age-specific distribution of qualitative IgG antibodies (percent) against influenza B virus (IBV) in the study group of children and adolescents as well as in the control group of adults

Full size image

The mean antibody concentration against influenza B virus of the study group of children and adolescents was estimated as 5.1 NTU (SD 4.3). After decline of mean antibody concentration in children of the first year of life (5.8 NTU, SD 5.1 NTU) to children aged >1–3 years (2.6 NTU, SD 2.7 NTU), there was a significant increase of mean antibody concentration with an increasing age (>3–6 years: 2.9 NTU, SD 2.6 NTU; >6–9 years: 3.7 NTU, SD 2.0 NTU; >9–12 years: 5.4 NTU, SD 4.3 NTU; >12–15 years: 8.2 NTU, SD 6.6 NTU; >15–18 years: 7.5 NTU, SD 3.8 NTU; trend test p 0.002; Fig. 4). The concentration of antibodies was not dependent on the gender of children and adolescents (P = 0.101). The mean antibody concentration against influenza B virus of adult controls was 11.7 NTU (SD 4.0 NTU). There were no significant differences between the different age groups of adults (P = 0.194, Fig. 4) as well as between men and women (P = 0.536). For all age groups of children and adolescents, significantly lower antibody concentrations were measured in comparison to the control group of adults (P < 0.001, Fig. 4).

Fig. 4
figure 4

Age-specific distribution of IgG antibody concentrations [NovaTec-Units (NTU)] against influenza B virus (IBV) in children, adolescents and adult controls. Boxplot shows median, 25 and 75th percentile, minimum, maximum and/or outliers. Open circles values >1.5 inter-quartile, Asterisks values >3 inter-quartile range from the end of the box

Full size image

Discussion

The findings of this study are based on the determination of antibodies against the highly conserved influenza A and B virus matrix and nucleoprotein antigens [26, 49, 50]. Similar to the indirect immunofluorescence test [1], ELISAs used in this study allow the detection of antibodies induced by influenza A and B viruses currently circulating without subtype-specific diagnosis. The procedures do not require the annual adjustment of the viral antigen preparations and, most likely, they do not interfere with antibodies induced by haemagglutinin antigens of split influenza vaccines. The study group contained 1,111 sera from children and adolescents, from whom three quarters were healthy and one quarter suffered from adiposity or growth disorders. According to the present knowledge, there are no data indicating that adiposity or growth disorders influence the immunologic defense against viral infections. Infections such as the acquired immunodeficiency syndrome (AIDS) [46] or rare congenital viral infections [22, 23], which may be associated with lipometabolic disturbances or growth disorders, were precluded from this study on the basis of patients’ laboratory data and history.

The prevalence of both anti-influenza A and B antibodies was significantly lower in children and adolescents than in the group of adults. In principle, the prevalence of antibodies against influenza A virus in the study population (82.0%) and in the control group (99.4%) is in line with the epidemiological situation in Germany in the years preceding serum sampling: between 1991 and 2006 influenza A has been the predominant influenza virus type in most years, i.e. during the lifetime of the most study subjects. In 1999 and 2000 when 88% of sera of the study group and all sera of the control group were obtained, there was a moderate activity of influenza A viruses of the subtype H3N2 in Germany [3, 45]. After reduction of maternal antibodies over a period of 2–7 months of life, infections with influenza A virus started already at an early stage during childhood. Nearly all children >9 years of age have had contact with influenza A virus suggesting high infection coverage with this virus. These data reflect serological findings from others showing seropositivity rates in the haemagglutinin inhibition test for A/H3N2 of 78% in previously unvaccinated 6–9 years old in the United States [44] and the high attack rates ranging from 15 to nearly 50% of influenza disease in pre-school and school children [12, 34]. However, more than one-third of all infants up to 6 years of age do not have serological evidence of previous contact with influenza A virus and thus, they can be regarded according to their antibody status as immunologically naïve [8]. Presumably, these children cannot mount a disease mitigating humoral immune response as efficiently as those immunologically primed and therefore, they are at high risk of severe course of infection [30, 38]. Several studies have shown consistently that children younger than 5 years of age and particularly those under 12 months of age are at substantially higher risk of hospitalization for influenza than older children [21, 32]. The lower concentrations of antibodies in children up to 6 years of age than in adults is most likely due to the lower number of boosting influenza infections in their life time. By contrast, from the significantly higher concentrations of antibodies in children and adolescents aged >12–15 years, it can be concluded that this age group is affected most frequently with infections caused by influenza A virus or has a higher immune response. This can reflect a considerably higher number of re-infections in the population of older children and young adolescents.

The lower prevalence of influenza B-specific antibodies in children and adolescents as well as in adults in comparison with the prevalence of antibodies against influenza A virus is consistent with the epidemiological situation reported in Germany in the years preceding serum sampling (see above) with generally low level circulation of influenza B virus. During the season 1999/2000, one-third of all isolated influenza viruses could be classified as influenza B viruses and few influenza B virus isolates were obtained in the season 2000/2001 [39–41]. During 2001–2006, when only a small portion of children and adolescents were enrolled, a low incidence of influenza B compared with influenza A was reported with the exception of the moderate influenza seasons 2001/2002 (~53% influenza B isolates) and 2005/2006 (~70% influenza B isolates) [1, 3, 45]. According to the presented data, below the age of 12 years nearly all children were influenza B virus antibody-negative and therefore, they are most likely susceptible to severe courses of influenza B virus infections. These data differ from serological findings from other studies showing seropositivity rates for influenza B of 37.1% in 6–9 years old and 66% in 10–13 years old [44]. The higher seropositivity rates reported in that study might be due to the fact that all children participating in that study had experienced the influenza seasons with pronounced influenza B virus circulation in 1999/2000 and 2001/2002. The present findings suggest that only a low number of influenza infections during childhood and adolescence were caused by influenza B virus. In addition, the incidence of influenza B virus primoinfections was considerably lower in children and adolescents than in adults during the study period. Among the adults, the prevalence of antibodies against influenza B virus was lower in women than in men, even though this difference was not highly significant (P = 0.014). Since child care is undertaken generally by women, the lower antibody prevalence may be associated with the low incidence of influenza B virus infections during childhood.

On the basis of single statistical inquires, it can be assumed that the current annual influenza vaccination covers maximally 10–20% of all children and adolescents in Germany [24, 29]. Considering this low vaccination rate and the requirement of annual revaccination for continuous protection from influenza, it is unlikely that previous vaccination with trivalent inactivated influenza vaccines had a major impact on the incidence of influenza infection in the study population and the serological results of this study. Moreover, as mentioned above, vaccine-induced antibodies most likely do not interfere with the serological testing procedure used. Even though antibodies to influenza virus matrix and nucleoprotein antigens fail to contribute to protection, they indicate the presence of a certain subtype-independent T cell-mediated protection [8]. In view of a possible new pandemic by influenza virus A/H5N1 [7, 10], the present results may be informative for the development and application of a new live attenuated influenza vaccine inducing both humoral and cellular immunity.

Taken together, this study reveals that a considerable proportion of children and adolescents in Germany do not display any serologic markers of previous influenza virus infections. This concerns more than one-third of all young children by the sixth year of age for influenza A, as well as, nearly all children by the 12th year of age and more than three quarters of adolescents >12–18 years of age for influenza B. These children must be regarded as susceptible with regard to influenza virus A and B and therefore, they are at particular risk for severe courses of influenza. The present findings suggest that annual influenza vaccination in children and adolescents may improve considerably the protection against severe influenza virus infections occurring during epidemics.