|Home | About | Journals | Submit | Contact Us | Français|
Annual influenza vaccination is routinely recommended for pediatric solid organ transplant recipients. However, there are limited data defining the immune response to the inactivated vaccine in this population.
This prospective study compared the humoral and cell-mediated immune responses to the trivalent subvirion influenza vaccine in pediatric liver transplant recipients with those in their healthy siblings. All subjects received inactivated influenza vaccine. Hemagglutination inhibition and interferon-γ (IFN-γ) enzyme-linked immunosorbent spot assays for New Caledonia and Shanghai strains were performed at baseline, after each vaccine dose, and 3 months after the series. Seroconversion was defined as a 4-fold increase in antibody titers; seroprotection was defined as an antibody titer ≥1:40. An increase in the number of T cells secreting IFN-γ was considered to be a positive enzyme-linked immunosorbent spot response.
After 1 dose of vaccine, transplant recipients achieved rates of antibody seroprotection and seroconversion that were similar to those achieved by their healthy siblings. However, for both influenza strains, IFN-γ responses by enzyme-linked immunosorbent spot were significantly attenuated in transplant recipients after 2 doses of vaccine. No cases of influenza or vaccine-related serious adverse events were documented in the study.
The diminished cell-mediated immune response to influenza vaccination that was observed in pediatric liver transplant recipients suggests that the current vaccine strategy may not provide optimal protection. Because of concerns regarding potential emergence of more virulent influenza strains, further studies are warranted to determine if IFN-γ responses are predictive of efficacy and to identify the optimal vaccination strategy to protect populations with a high risk of infection.
Recent outbreaks of severe acute respiratory syndrome and avian influenza demand optimized strategies to protect society from pandemic respiratory illnesses [1, 2]. Of particular concern is the population of solid organ transplant recipients. These individuals are at high risk for morbidity and mortality secondary to influenza, and their immune response to vaccination is poorly understood. Pediatric transplant recipients are, perhaps, the most vulnerable, because even immunocompetent children are susceptible to serious influenza disease [3, 4]. At the time of transplantation, infants and young children often lack prior exposure to influenza and its subsequent protective immunologic priming .
Pediatric solid organ transplant recipients are at risk for influenza-related complications, including pneumonia, sepsis, CNS disease, acute graft rejection, and death [6–8]. In a retrospective review involving 42 pediatric solid organ transplant recipients with influenza or parainfluenza, 3 patients with influenza died and 4 developed concurrent infections with cytomegalovirus (CMV) and bacteremia .
Liver transplant recipients represent a growing proportion of the population of pediatric transplant recipients. How these children respond to vaccinations, including the inactivated trivalent subvirion influenza vaccine, is poorly understood. Evaluations of the inactivated influenza vaccine in other immunocompromised populations, such as children with HIV infection and cancer, suggest a diminished response to the vaccine that is relative to immunocompetent children [9, 10]. However, the data regarding the immune response in pediatric solid organ transplant recipients are less substantial. Prior studies have yielded conflicting results, with some authors suggesting a diminished antibody response to vaccination [11–13].
The generation of serum antibodies following influenza vaccination is crucial for protection from illness and is an important correlate for vaccine efficacy [5, 14]. However, clearance of influenza and prevention of influenza-associated complications also require a vigorous cell-mediated immune response. Influenza-specific CD8+ T cells mediate the killing of infected host cells and up-regulate proinflammatory cytokines in animal models [15, 16]. Adults with baseline cytotoxic T cell immunity against influenza clear virus more effectively than those with no pre-existing cell-mediated immunity, and cytotoxic T cells may demonstrate cross-reactivity when responding to new influenza A virus subtypes . Influenza in infants stimulates cytotoxic T lymphocyte proliferation, although the degree of this response may not correlate with serum hemagglutination inhibition (HI) antibody levels . Following vaccination, secondary antibody response requires the expansion of memory CD8+ T cells and CD4+ T cell assistance [19–21]. Although small studies involving both pediatric hematopoietic stem cell transplant recipients and adult solid organ transplant recipients suggest a diminished T cell response to inactivated influenza vaccine [22, 23], to our knowledge, no prospective studies have evaluated this response in pediatric solid organ transplant recipients. We present the results of a single center, prospective, comparative evaluation of humoral and cell-mediated immune responses to inactivated influenza vaccine in pediatric liver transplant recipients and their healthy siblings.
Inactivated trivalent subvirion influenza vaccine (Fluzone; Sanofi Pasteur) was provided for the 2004–2005 and 2005–2006 influenza seasons. The H3N2 viral strains included in the vaccine differed between seasons. The H1N1 and B viral strains (A/New Caledonia/20/99[H1N1]–like virus and B/Shanghai/361/2002–like virus) remained consistent and were used to assess immune response.
Written informed consent was obtained from the parents or guardians of all subjects, and subject assent was obtained when applicable. The study protocol was approved by the National Institutes of Health Division of Microbiology and Infectious Diseases and by the Mount Sinai School of Medicine Institutional Review Board. The trial consisted of the 2 following study arms: pediatric liver transplant recipients (n = 41) and their healthy siblings (n = 19). Healthy siblings were selected as a control group, because the burden of influenza exposure and disease is similar among members of a household. Subjects were eligible for enrollment if they were either a liver transplant recipient or a healthy sibling of a liver transplant recipient, if they were 6 months to 18 years of age, and if 3 months had elapsed since transplantation. Exclusion criteria were as follows: (1) previous vaccination with an influenza vaccine for the 2004–2005 influenza season, (2) known hypersensitivity reaction to inactivated influenza vaccine or vaccine component (including eggs or gelatin), (3) history of Guillain-Barré syndrome, (4) receipt of an immunoglobulin product (including cytomegalovirus hyperimmunoglobulin or varicella-zoster immunoglobulin) within 3 months of vaccination, and (5) receipt of any live viral vaccine within 4 weeks or any inactivated viral vaccine within 2 weeks of enrollment.
Study subjects were vaccinated according to the recommendations of the Centers for Disease Control and Prevention's Advisory Committee on Immunization Practices . All liver transplant recipients received 2 intramuscular doses of vaccine, separated by 4 weeks. Healthy siblings received 2 doses of vaccine if they were ≤9 years of age and had no history of influenza vaccination. Siblings who received influenza vaccination in previous seasons or who were >9 years of age received 1 dose of vaccine. Subjects were monitored for immediate adverse events for 30 min following vaccination. Parents or guardians received a symptom diary and recorded vaccine reactions for 10 days following vaccination, including pain or redness at the injection site, fever, malaise, myalgia, headache, and nausea. Subjects were observed for 6 months following vaccination and were assessed for development of influenza disease, acute allograft rejection, Epstein-Barr virus disease and/or viremia, CMV disease and/or viremia, or change in immunosuppressive medications.
Blood samples for HI antibody and IFN-γ enzyme-linked immunosorbent spot (ELISPOT) assays were obtained at baseline (prior to vaccination) and at 4 weeks after each dose of vaccine. PBMCs and serum samples were collected by ficoll purification and frozen at −80°C until the time of analysis. Cell viability was assessed with phytohemagglutinin stimulation, which was conducted in parallel with ELISPOT assays. One subject's cells were found to be nonviable; this subject was excluded from ELISPOT analysis.
Humoral immune response was evaluated by performing HI antibody assays for the A/New Caledonia/20/29(H1N1) and B/Shanghai/361/2002–like strain components of the vaccine by microtiter technique . Seroprotective antibody response was defined as an HI antibody titer ≥1:40, and seroconversion was defined as a 4-fold increase in antibody titers. Cell-mediated immune response was measured by IFN-γ ELISPOT assay (Pierce Biotechnology), with a positive response represented by an increase in the number of T cells secreting IFN-γ in response to the New Caledonia or Shanghai strains [26–29].
Subjects enrolled in the 2 trial arms were compared with regard to age, race, and sex by Fisher's exact test. Subjects were compared by Fisher's exact test for HI antibody response at baseline, after 1 dose of vaccine, and after 2 doses of vaccine. ELISPOT results for each viral strain were analyzed by log transformation to adjust for non-Gaussian distribution and nonequivalent variances. The mean ELISPOT values for both transplant recipients and their siblings were then compared by 2 sample-independent Student's t tests at baseline, after 1 dose of vaccine, and after 2 doses of vaccine. For transplant recipients, Student's t test was used to compare the mean ELISPOT value at baseline with that after the first dose of vaccine, as well as the mean ELISPOT value after the first dose with that after the second dose of vaccine. A parallel analysis was performed for healthy siblings. Spearman rank correlation analysis was performed to determine if HI antibody responses or ELISPOT results correlated with age, time from transplantation, T cell counts, or immunosuppressive medication dose.
Forty-one pediatric liver transplant recipients and 19 siblings were enrolled in the study. Thirty transplant recipients and 13 siblings completed the study and were included in the final analysis. Demographic data are presented in table 1. Liver transplant recipients were similar to their healthy siblings with regard to age, CD4+ T cell count, and total T cell count. There were significantly more female participants in the group of transplant recipients than in the group of healthy siblings (P = .02). For transplant recipients, the mean duration from transplantation to study enrollment was 52.3 months (range, 3–170 months). Transplant recipients received low-dose immunosuppressive treatment and had a mean tacrolimus level ± SD of 5.53 ± 2.49 ng/mL for the duration of the study. Twenty-six of the 30 transplant recipients received tacrolimus therapy, 5 received mycophenolate mofetil therapy, 2 received cyclosporine therapy, 1 received azathioprine therapy, and 19 received prednisone therapy (with a mean prednisone dosage ± SD of 0.12 ± 0.09 mg/kg/day). Fifteen subjects underwent transplantation secondary to extrahepatic biliary atresia, 5 underwent transplantation secondary to fulminant hepatic failure, 2 underwent transplantation secondary to autosomal recessive polycystic kidney disease, and 8 underwent transplantation secondary to other causes.
Baseline antibody titers for both New Caledonia and Shanghai strains were significantly lower in transplant recipients than in their healthy siblings (P = .003 and P = .02, respectively) (figure 1A and 1B). Only 50% of transplant recipients demonstrated seroprotective antibody titers (≥1:40) for the New Caledonia strain prior to the first vaccine dose (figure 1A); nearly 60% began the study with seroprotective titers for the Shanghai strain (figure 1B). Almost 100% of healthy siblings began the study with seroprotective titers for both viral strains. This discrepancy resolved following receipt of the first dose of vaccine, with >75% of transplant recipients achieving seroprotective titers for both strains. No statistically significant difference was noted in rates of seroprotection between transplant recipients and their healthy siblings following receipt of 1 dose of vaccine (P = .3, for the New Caledonia strain; P = .3, for the Shanghai strain). Receipt of a second dose of vaccine resulted in no statistically significant increase in rates of seroprotective antibody levels in transplant recipients or healthy siblings.
Nearly 75% of healthy siblings achieved seroconversion, defined as a 4-fold increase in antibody titers, following receipt of 1 dose of vaccine. Fewer than 50% of transplant recipients achieved seroconversion to the New Caledonia or Shanghai strain following receipt of the first dose of vaccine. This trend was not statistically significant when the rate of seroconversion among transplant recipients was compared with that among healthy siblings (P = .1, for the New Caledonia strain; P = .97, for the Shanghai strain). Receipt of a second dose of vaccine did not increase the rate of seroconversion in either group.
IFN-γ responses, as measured by ELISPOT, were attenuated for both viral strains in transplant recipients (figures 2A and 2B). Prior to vaccination, transplant recipients and siblings had similar ELISPOT responses of <10 spot-forming units per 105 PBMCs (P = .36, for the New Caledonia strain; P = .9, for the Shanghai strain). ELISPOT responses in siblings increased to nearly 30 spot-forming units per 105 PBMCs after receipt of 1 dose of vaccine. Transplant recipients exhibited a significantly diminished response, with an increase to a mean value of only 20 spot-forming units per 105 PBMCs (P = .01, for the New Caledonia strain; P = .02, for the Shanghai strain). Receipt of a second dose of vaccine resulted in no statistically significant change in ELISPOT responses in either group, and ELISPOT values after the second dose remained lower in transplant recipients than in their siblings (P = .01, for the New Caledonia strain; P = .03, for the Shanghai strain).
The secondary objective of this study was to correlate immune response with clinical parameters, including age, CD4+ T cell count, total T cell count, time from transplantation, and degree of immunosuppression. Evaluation by Spearman rank correlation revealed no statistically significant correlation between HI antibody titer or ELISPOT response and age, CD4+ T cell count, or total T cell count. Importantly, antibody responses were not predictive of ELISPOT results. Time from transplantation, mean tacrolimus level, and mean daily prednisone dose also did not correlate with immune response. However, many subjects had undergone transplantation at least 1 year prior to the study and received relatively low dose immunosuppressive regimens.
Eight (57%) of the 14 transplant recipients who had nonprotective antibody titers to either strain at the time of enrollment did not experience seroconversion to the strain to which they were susceptible; 5 did not experience seroconversion to either strain. In addition, 5 other transplant recipients who entered the study with seroprotective titers did not experience seroconversion. This is in contrast to the majority of healthy siblings who experienced seroconversion, even in the context of seroprotective titers at the time of enrollment. Among the 13 participants who did not experience seroconversion, 11 experienced complications during the study period, including Epstein-Barr virus disease or viremia (n = 7), CMV disease or viremia (n = 3), and acute allograft rejection (n = 4) (table 2). Six subjects received non–tacrolimus-based immunosuppressive therapy. In contrast, no episodes of acute allograft rejection were observed in the 17 transplant recipients who experienced seroconversion (P = .03), and only 2 of these subjects received non–tacrolimus-based immunosuppressive therapy (P = .05). There were 9 cases of Epstein-Barr virus and/or CMV complications among the participants who experienced seroconversion, which was not statistically different from the rate among those subjects who did not experience seroconversion (P = 1).
None of the transplant recipients or their siblings experienced documented influenza virus infection during the course of the study, and there were no serious adverse events attributed to vaccination. None of the subjects disen-rolled from the study because of vaccine-related adverse events. Eleven of the transplant recipients were hospitalized during the 6-month follow-up period; 4 were hospitalized because of episodes of mild or moderate biopsy-proven acute allograft rejection, 1 was hospitalized because of exacerbation of autoimmune hepatitis, and 6 were hospitalized because of acute, febrile illnesses, including gastroenteritis and noninfluenza pneumonia. The cases of allograft rejection and autoimmune hepatitis occurred at least 6 weeks after vaccination. Results of both direct fluorescent antibody testing and viral culture for influenza were negative for all children evaluated for respiratory illnesses.
Our study represents one of the first evaluations of the cell-mediated immune response to inactivated influenza vaccine in pediatric solid organ transplant recipients and suggests a diminished T cell response in a cohort of older, minimally immunosuppressed subjects. An evaluation of the T cell response to influenza vaccination (by ELISPOT assay) that involved 65 adult renal transplant recipients also reported diminished IFN-γ secretion in response to influenza-specific antigen stimulation in the recipients, compared with healthy control subjects . Similar to our study, in that study , HI antibody titers were not predictive of IFN-γ ELISPOT results.
Although a recent trial involving healthy influenza vaccine-naive children argued the necessity of 2 doses of vaccine to achieve seroprotective antibody titers , our results suggest that a second dose of vaccine has little effect on either humoral or cell-mediated immunity. Our results may be limited by a small sample size, but they are consistent with results of a previous study of the cell-mediated response to inactivated influenza vaccination in healthy infants and children, in whom a second dose of vaccine had little effect on T cell IFN-γ secretion . IFN-γ secretion following inactivated influenza vaccination in that trial was age dependent and was most prominent in children aged 6 months to 4 years. The relationship between age and immune response to inactivated influenza vaccine is underscored by recent studies suggesting that CD80 and CD86 expression on activated monocytes, which are increased in magnitude in younger persons, are predictive of vaccine immunogenicity [31, 32]. Additional studies are necessary to clarify the association between age and vaccine immunogenicity, particularly in immunocompromised populations, and to determine whether a second booster dose is beneficial in transplant recipients.
Thirteen transplant recipients in our study did not experience seroconversion to either viral strain, including 8 subjects who demonstrated nonprotective titers at baseline. The statistically significantly increased rate of acute allograft rejection during the study period and the trend towards increased use of non–tacrolimus-based immunosuppressive therapy in this group suggest that the inability to achieve seroconversion may be a result of receipt of more-potent immunosuppressive treatments. The episodes of acute allograft rejection were treated with methylprednisolone (20 mg/kg/day for 1–3 days) and with an increase in the dosage of maintenance immunosuppressive therapy. The increased rates of Epstein-Barr virus and CMV viremia among subjects who did not experience seroconversion did not reach statistical significance, but the trend observed could be related to both virus- and drug-induced immunosuppression. Prospectively identifying cohorts of children who are unlikely to respond to influenza vaccine could also identify the children who may require antiviral prophylaxis in the context of outbreaks of influenza. This vulnerable cohort could also benefit from novel vaccination strategies [33–35].
Four transplant recipients developed mild or moderate biopsy-proven acute allograft rejection during the course of the study. The rate of allograft rejection in this cohort is consistent with national rates of acute allograft rejection during the first 2 years after transplantation . A small study involving 28 adult heart transplant recipients (14 vaccinated patients and 14 control subjects) suggested an increased risk of reversible, low-level histological allograft rejection following inactivated influenza vaccination , but a larger study involving 51 adult liver transplant recipients identified no association between inactivated influenza vaccination and acute allograft rejection or asymptomatic elevated transaminase levels . No pediatric studies have thoroughly evaluated the relationship between vaccination and allograft rejection. This is a critical area of investigation, because some transplantation centers delay influenza vaccination following transplantation to avoid vaccine-related graft dysfunction. Evidence suggests that the risk of allograft rejection following acute influenza infection outweighs the theoretical risk of allograft rejection in response to vaccination. The rate of acute allograft rejection approached 62% in a review involving 30 adult solid organ transplant recipients with influenza virus infection , and influenza-mediated allograft rejection was a well-documented cause of mortality in pediatric series . Influenza virus infection may induce allograft rejection by up-regulating inflammatory cytokines and chemokines, including IL-1, TNF, IL-6, and IL-8, leading to activation of immune mechanisms that result in cellular rejection .
The small number of subjects in our pilot study, the exclusion of infants, and the lack of age-matched control subjects may have limited the results of our trial. Most of the subjects enrolled in our study had undergone transplantation at least 6 months prior to the study and, thus, were beyond the period of highest risk for allograft rejection. Furthermore, the clinical significance of a diminished ELISPOT response is unclear, because few data exist that document the ELISPOT response to influenza vaccine in a healthy population. In contrast to HI assays in which seroprotective responses are well defined, no established ELISPOT value or fold change from baseline has been demonstrated as predictive of cell-mediated immune response to influenza.
The consistently lower cell-mediated immune responses demonstrated in transplant recipients (compared with their healthy siblings) in this study argue the need for larger and more thorough investigations of the immunogenicity of inactivated influenza vaccine in this population. It is crucial that future studies include the most vulnerable children, including infants and transplant recipients within the first year after transplantation. Ideally, future trials would use more sophisticated immunological studies to elucidate the phenotype of influenzaspecific T cells following vaccination. Larger and more-detailed studies could provide the data necessary to define a truly protective cell-mediated immune response to influenza vaccine in immunocompromised populations and would provide the framework for future studies of live attenuated vaccines. Determining the parameters that predict an effective and safe response to influenza vaccine could lead to more-effective strategies for protecting immunocompromised hosts against influenza pandemics and emerging pathogens.
We thank Dr. Adolfo Garcia-Sastre, for his generous gift of vaccine-specific influenza virus strains, and Dr. James Godbold, for his assistance in statistical analysis.
Financial support. Northeast Biodefense Center (U54 AI057158-Lipkin) and National Institutes of Health (U19 AI62623).
Potential conflicts of interest. All authors: no conflicts.