|Home | About | Journals | Submit | Contact Us | Français|
Heart failure (HF) patients (pts) are at risk for influenza despite widespread vaccination. Both humoral (antibody) and cytotoxic T-lymphocyte (CTL) responses are important for protection. We explored antibody and CTL mediated responses to the influenza vaccine in HF pts compared to healthy controls.
We studied 29 HF pts (9 ischemic, 20 non-ischemic) stable on HF therapies and 17 healthy controls. Participants had phlebotomy before and after influenza vaccination. Antibody production was measured in serum by hemagglutination inhibition assay (HIA), and CTL responses (via IFNγ and IL-10 production) were measured in isolated peripheral blood mononuclear cells (PBMCs) with ELISA.
CTL responses demonstrated increased IL-10 production in HF pts after vaccination (p=0.002), but similar IFNγ responses to healthy controls. All participants demonstrated antibody seroprotection; groups had similar rates of seroconversion (p=NS). Antibody-mediated response to the newest vaccine antigen, H3N2, was reduced in HF (p=0.009).
Patients with HF had higher vaccine induced IL-10 concentrations, suggesting a different CTL phenotype for vaccine responses. HF patients did not mount as vigorous of an antibody immune response to the newest vaccine viral strain compared to healthy individuals. These data suggest that immunologic memory may be important for vaccine protection in HF pts.
Chronic heart failure (HF) predisposes to influenza infection and its complications. Excess mortality observed during winter months in individuals with HF may be attributed to influenza. Vaccination against influenza decreases cardiac related hospital admissions, acute HF exacerbations, and all cause mortality. Despite widespread influenza vaccination programs, overall influenza-related hospitalization and death rates are rising, particularly in patients with cardiac disease. In addition to increased hospital admissions, influenza also results in longer lengths of stays and increased mortality in patients with HF compared to younger, healthy individuals. Older adults and persons with cardiac disease or other co-morbidities and treatments that render them immune-compromised are at greater risk for influenza infection despite vaccination due to reduced antibody and cell mediated responses to vaccines.[4, 5] Due to significant morbidity and health care costs, the need to improve the efficacy of influenza vaccine in patients with HF is urgent.
HF results in an upregulated sympathetic nervous system. Growing evidence shows that the sympathetic nervous system activation decreases immune response via activation and modulation of beta2-adrenergic receptors (β2-AR). Human T and B lymphocytes express β2-AR. The β2 adrenergic signaling cascade activates cAMP dependent elements on the DNA, which modulate cytokine gene transcription.[8, 9] A direct catecholamine effect through β2-AR on cytokine gene regulation decreases responses to vaccines. In vitro models show that increased β2-AR density suppressed IFNγ synthesis. Therefore, it is logical that patients with HF demonstrate reduced vaccine responses as compared to healthy, age matched controls, potentially due to up-regulated adrenergic pathways.
An inactivated trivalent influenza vaccine is recommended for those at high risk for influenza morbidity and mortality. The most widely accepted definitions of antibody response are seroconversion and seroprotection, reflecting antibody titer changes to just one of the three vaccine viral strains. Most adults develop both humoral antibody and cytotoxic T-lymphocyte (CTL) immune responses to vaccination, indicating that both T-helper type 1 (Th1) and T-helper type 2 (Th2) responses occur following influenza immunization.[11–13] Antibody titers as an indicator of vaccine efficacy and protection against influenza illness in older adults are insensitive to impaired cell-mediated immunity with disease and increasing age. One study demonstrated that antibody titers did not distinguish between HF participants who developed influenza illness and those who did not.
The CTL and humoral (antibody) responses to all three vaccine viral strains have not been examined in heart failure patients compared with controls. We hypothesized that patients with HF will mount less pronounced CTL and antibody-mediated immune responses to influenza vaccination compared with healthy individuals.
We studied patients with HF and healthy controls. Eligible HF participants had systolic or diastolic dysfunction documented by echocardiogram in previous 6 months, with American College of Cardiology(ACC)/American Heart Association(AHA) Stage C, New York Heart Association (NYHA) Functional Class I, II or III HF. All patients with HF were on stable medical therapy for at least 30 days, including target or maximally tolerated doses of angiotensin converting enzyme (ACE) inhibitors or angiotensin receptor blockers (ARBs) and beta adrenergic blockers (carvedilol 25mg twice daily or metoprolol succinate 200mg once daily), when appropriate. Exclusion criteria for patients with HF were contra-indications to ACE inhibitors or ARBs and β-blockers. Additionally, healthy control and HF patients with a history of allergic reaction to influenza vaccine, allergy to egg products, or moderate to severe acute febrile illness were excluded. The protocol was approved by the University of Wisconsin institutional review board. All participants provided written informed consent in accordance with established guidelines for the protection of human subjects.
This was a prospective, open label study in 19 healthy individuals and 32 individuals with established HF followed at the University of Wisconsin Hospital Advanced Heart Disease Program. The primary outcome variable was the difference in CTL influenza responses (via IFNγ and IL-10 production) to influenza vaccine between patients with HF and healthy controls. Participants received the 2006–07 inactivated influenza vaccine intramuscularly during October or November of 2006. Participants underwent phlebotomy prior to and two to four weeks following vaccination. The timing of these samples ensured adequate antibody and CTL responses to the vaccine after immunization.[16–18]
The influenza vaccine viral strain content changes annually to contain vaccine viruses anticipated to be the three most commonly circulating strains in a given year. The three types of virus strains included are H3N2, H1N1, and B type, classified based on viral surface antigens. For the 2006/2007 season, the vaccine contained A/New Caledonia/20/99 (H1N1)-like virus; A/Wisconsin/67/2005 (H3N2)-like virus, and B/Malaysia/2506/2004-like virus.
CTL responses were assessed with cell culture and ELISA which measured IFNγ and IL-10 concentrations. Peripheral blood mononuclear cells (PBMCs) were isolated and washed. A 1mL suspension containing one million PBMC was aliquoted into a 48-well plate. Cells were grown in culture (cRPMI-10) with no antigen, in the presence of influenza antigens (4 hemagglutination inhibition units from each viral strain), and with phytohemagglutinin (PHA) 2.5 μg/ml for 96 hours at 37°C with 5% CO2. Supernatant fluids were harvested and frozen until performance of cytokine assays.[1, 3, 7] Cytokine concentrations (IFNγ and IL-10) in the supernatant were measured using ELISA (OptEIA Sets, Pharmingen, San Diego, CA) (CV<10%) according to the manufacturer’s instructions. Samples were run in duplicate and read at 450nm using SOFTmax software for MAXline microplate readers (version 2.02). A standard curve was generated by plotting the log of the concentration of the standards against the measured absorbance. If any absorbance reading was outside the absorbance of the highest standard concentration, the supernatant was diluted and re-assayed. Samples below the level of detection were assigned a concentration of 2 pg/mL.
Antibody responses were assessed by hemagglutination inhibition assay (HIA) which measured influenza antibody concentrations following immunization. HIA was performed in duplicate using standard microtiter techniques. Briefly, influenza virus-induced agglutination of guinea pig red blood cells was inhibited by antibodies present in the human serum. Serial dilutions of the human sera were made. Titrated influenza antigen was incubated with the serum dilutions for 30 minutes. Guinea pig red blood cells (50 μl of 0.5% in phosphate buffered saline) were added and incubated for 45 minutes. The dilution of serum that no longer inhibits hemagglutination was the influenza antibody titer.[17–19] A protective antibody response to influenza vaccine was defined as an antibody titer equal to or greater than 1:40. Seroconversion following influenza immunization was defined as a fourfold increase in antibody concentrations.[20, 21]
Age, sex, and race distributions of healthy controls and HF participants were compared using t-tests or Wilcoxon rank sum tests for continuous variables and Chi-Square tests for categorical variables. The primary study outcome variable for statistical analysis was the pre- to post-vaccine difference in CTL responses (via IFNγ and IL-10 cytokine productions) between participants with HF and healthy controls. Assuming a 25% difference between groups in IFNγ and IL-10 production, 16 participants were needed in each group to achieve 80% power at the 0.05 significance level. Log-transformation was performed for both cytokine and antibody concentrations if data had non-normal distribution, and following back transformation, were reported in their original units. We compared the absolute differences of log transformed values of cytokine production elicited by influenza antigens (all three strains combined) between healthy controls and HF groups with a Wilcoxon rank sum test. Influenza antibody concentrations were also compared using absolute differences of log-transformed values (post immunization to baseline) between the healthy control group and HF group using a Wilcoxon rank sum test. Finally, we compared the seroprotection (HIA >40) and seroconversion (>4 fold increase) rates in each group using chi square tests. Data were analyzed on the intent-to-treat principle.
Thirty two HF patients (8 females, 24 males) and 19 healthy controls (8 females, 11 males) were enrolled, with a mean age of 58 ± 13 years (median 57 years, range 26 – 81 years) for HF participants, and 47 ± 10 for healthy controls (median 51 years, range 29 – 65). HF participants were older than healthy control participants (p=0.004 for comparison between groups). Among the HF group, mean ejection fraction was 39% ± 11% (median 35%, range 15%–55%), and patients were NYHA functional classes I (N=16), II (N=13), and III (N=3). Other baseline characteristics for the HF group are shown in Table 1. Three participants were lost to follow up in the HF group. In the healthy control group, one participant did not receive the influenza vaccination so only baseline data were available. No adverse reactions were reported following influenza vaccination.
T-cell data are summarized in table 2. There were no differences in absolute changes of the log-transformed IFNγ concentrations pre and post-vaccination between the HF and healthy control groups (median changes 79.8 and 89.1 pg/mL, p=0.133, figure 1). HF participants demonstrated higher IL-10 production in response to influenza vaccination compared to healthy controls (median changes 20.4 and 7.1 pg/mL, p=0.0002, figure 1).
The median ratios of post immunization IFNγ to IL10 concentrations (log IFNγ :IL10), were 1.47 for HF participants and 1.66 for healthy controls (p=0.036 by Wilcoxon rank sum test). The INFg:IL10 ratio is a measure of T-helper type 1 (INFγ) versus T-helper type 2 (IL-10) mediated T-cell responses. The significant differences between HF and healthy control groups may indicate that the HF participants demonstrated a Th2 dominant phenotype.
All participants in both groups demonstrated seroprotection to the influenza vaccine. Eighteen of 28 HF participants (64%) exhibited seroconversion compared with 14 out of 17 healthy controls (82%) (OR = 2.59 [95% CI 0.6,11.2], p=0.31). Antibody data are summarized in table 2.
Participants in the HF group demonstrated a less vigorous antibody-mediated immune response to A/H3N2, as shown in figure 2 (HF: median titers pre-vaccination 80 HAU, post-vaccination 160 HAU; healthy controls: pre-vaccination 80 HAU, post-vaccination 320 HAU, p=0.009 for comparison of response vigor between HF and healthy controls). Responses to A/H1N1 were not significantly different between the groups. Finally, responses to B/Malaysia were numerically but not statistically lower in participants with HF compared to healthy controls (HF: median titers pre-vaccination 20 HAU, post-vaccination 80 HAU; healthy controls: pre-vaccination 10 HAU, post-vaccination 80 HAU, p=0.08).
As an exploratory analysis, we examined T-cell and antibody-mediated responses within the HF group, stratified by beta adrenergic blocker (metoprolol succinate [n=10], or carvedilol [n=22]), to investigate whether differential pharmacologic effects of these two agents on the beta adrenergic receptor would influence immune responses. We did not detect an association between beta blocker and CTL responses. However, we observed poorer antibody-mediated responses to the A/H3N2 strain for participants taking metoprolol, versus those taking carvedilol, although the difference was short of statistical significance (figure 3, p=0.06).
In this study, we compared humoral and CTL immune responses to influenza vaccination in patients with heart failure compared to healthy individuals. Our results were threefold.
First, participants with HF exhibited higher IL-10 production in response to influenza vaccination compared with healthy controls. Additionally, HF participants had a lower IFNγ:IL:10 ratio compared with healthy controls, indicating a Th2 dominant phenotype. There were no differences in IFNγ production between groups. CTL responses are critical for recovery from influenza infection; particularly Th1 responses. CTL responses to influenza vaccine depend on IFNγ which in turn stimulate influenza-specific memory T cells. A shift from Th1 cytokines (involving IFNγ) to Th2 cytokines (involving IL-10) with aging has been associated with reduced CTL activity and diminished protection against influenza virus challenge.[22, 23] Although influenza-specific Th2 cytokines such as IL-10 do not promote recovery from influenza infection, these cytokines continue to be expressed in high levels at the site of influenza infection, indicating that a balance between Th1 and Th2 cytokines may be important for viral clearance. Our study suggests that Th1 responses are preserved in patients with heart failure, and that overall CTL responses are shifted toward involvement of Th2 cytokines. Our cohort was significantly younger than those included in the abovementioned CTL response study, suggesting that perhaps certain chronic conditions may influence response phenotype in addition to age.
Second, participants with HF mounted a less vigorous antibody response to the vaccine viral strain A/H3N2 compared with healthy individuals. The composition of the trivalent influenza vaccine changes with each influenza season. The A/H3N2 strain was the newest strain to be incorporated into the vaccine for the 2006/2007 season. Evaluation of influenza vaccine relies heavily on serologic responses. The most widely accepted measures that indicate appropriate immune responses are seroprotection (antibody titer ≥40) and seroconversion (4 fold increase in antibody titer). However, these require demonstration of an antibody mediated response to just one of the three vaccine viral strains. It may be advantageous to measure responses to each individual vaccine viral strain in order to gain a better understanding of high rates of influenza infection in patients with heart failure despite growing and widespread immunization rates. Although rates of seroprotection and seroconversion were similar between groups, HF patients were not able to mount as vigorous of a humoral response to the newest vaccine viral strain compared with healthy individuals. As such, measurement of seroconversion and seroprotection rates alone may not capture individuals who remain at high risk for infection due to varying responses to specific viral strains. Our results suggest that patients with HF may rely on immunologic memory to mount appropriate antibody-mediated immune responses to an antigen, whereas healthy individuals are able to mount a vigorous immune response upon first exposure to a new antigen. It is also of note that as a group, participants with HF were only mildly symptomatic, had a mean ejection fraction near 40%, and were maximized on goal doses of both ACE inhibitors or ARBs and beta blockers. Our study suggests that even patients with mild heart failure may not be adequately protected against influenza infection despite annual vaccination. Future studies should explore alternate influenza vaccination strategies for patients with heart failure.
Lastly, when stratifying the HF group by beta blocker and examining immune responses, we found a non-significant trend toward better antibody responses to A/H3N2 with participants taking carvedilol compared to those taking metoprolol. Integral to the treatment of an upregulated adrenergic system in heart failure is the use of beta adrenergic blockers, proven to significantly decrease all cause mortality. As mentioned previously, the sympathetic nervous system influences immune response via activation β2-AR. In particular, human T and B lymphocytes express β2-AR. Stimulation of the β2-AR leads to decreased immune responses, therefore high concentrations of epinephrine and norepinephrine in heart failure may inhibit production of cytokines necessary for appropriate vaccine response. Logically, it may be reasonable to predict that blockade of β2-AR affects immune response to influenza vaccine, and may therefore preserve an appropriate vaccine-induced protection to influenza infection. More complete β2-AR blockade with carvedilol may account for the antibody-mediated findings. We did not observe differences in CTL responses within patients with heart failure based on beta blocker stratification, although sample sizes were small. Although we consider this trend intriguing, we consider these data hypothesis generating, requiring confirmation in future larger studies designed to examine differences in immune responses by beta blocker. To our knowledge, this study is the first to examine an association of beta blocker on immune responses in patients with HF.
Several limitations to our study need to be considered. First, immune responses to influenza vaccination are known to decrease with age, and our healthy control group was significantly younger than our HF group. Declining antibody responses to influenza vaccine have been observed in the elderly,[26, 27] however both groups in our study had mean and median ages well below 65 years. Additionally, study participants had mildly symptomatic, stable heart failure. The generalizability of our results to a wider group of patients, especially those with more advanced HF, is unknown. We cannot rule out that our findings among heart failure patients taking carvedilol are due to chance based on our small sample size and limited power to examine an interaction of beta-blocker effects on immune function. We did not specifically measure adrenergic activity as it relates to immune function in our participants. As such, we cannot rule out other potential contributors to differences in immune responses between the two groups, such as differences in nitric oxide availability. Lastly, we were unable to correlate differences in immune responses between the study groups to subsequent rates of influenza infection within that season as we were not powered to perform these analyses. Future work should address the issue of individual strain vaccine-induced antibody responses and clinical correlations with influenza infection rates more specifically.
In conclusion, patients with heart failure demonstrated Th2-weighted cell mediated responses to influenza vaccination. Moreover, HF patients did not mount as vigorous of an antibody immune response to the newest vaccine viral strain compared to healthy individuals, despite similar rates of seroprotection and seroconversion. Immunologic memory may be important for vaccine protection among those with HF. Lastly, future explorations are warranted involving therapeutic agents for heart failure with actions at the β2-AR and potential benefits in preserving antibody-mediated immune responses in patients with heart failure.
We gratefully acknowledge Drs. Walter Kao, David Murray, Peter Rahko, and Elaine Winkel for allowing their patients to participate in this study.
Dr. Vardeny was supported by NIH (NCRR) 8K12RRO23268 and the American College of Clinical Pharmacy. Dr. Sweitzer was supported by NIH K23AG01022.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.