Factors Associated With Maintenance of Antibody Responses to Influenza Vaccine in Older, Community-Dwelling Adults

In the United States, yearly influenza vaccination begins in August or September. However, influenza season can extend into April of the following year. The duration of protection from annual vaccination in older adults is unknown, and little is known about factors associated with the maintenance of response throughout the influenza season. This study evaluated factors associated with the immune response of older adults to trivalent influenza vaccine and the maintenance of antibody responses for the duration of the influenza season.

Subjects

Subjects were enrolled at two sites, Vanderbilt University Medical Center (Nashville, TN) and Marshfield Clinic Research Foundation (Marshfield, WI), during September and October 2009. Subjects were eligible for recruitment if they were ≥50 years of age and had no contraindication to influenza vaccination. Subject recruitment included advertisements at Vanderbilt University Medical Center and letters of invitation to older adults who had received an influenza vaccine in the year prior at Marshfield Clinic.1 All subjects were vaccinated either by their usual caregiver or by the study staff. Strain components for the 2009-2010 Northern Hemisphere vaccine included A/Brisbane/59/2007-like (H1N1), A/Brisbane/10/2007-like (H3N2), and B/Brisbane/60/2008-like. Participants were given the trivalent seasonal vaccine because the 2009 H1N1 pandemic vaccine was not available.

Data Collection

All subjects donated serum pre-vaccination (S1) during September through October 2009, 21 - 28 days post-vaccination (S2) and post-influenza season, May through July, 2010, approximately 250 days (8 months) post-vaccination (S3). Study procedures, informed consent documents and data collection forms were reviewed and approved by Institutional Review Boards at each of the study sites.

Age, co-morbid conditions, sex, and race were ascertained from participant interview. Recent chemotherapy, radiation therapy, or use of immunomodulating medications were ascertained by self-report or chart review. CDC-defined high risk medical conditions were identified by self-report of organ transplantation, cancer, diabetes mellitus, splenectomy (functional or anatomic), cardiovascular disease, renal disease, sickle cell disease, chronic pulmonary disease, seizure disorder, immune deficiency, or dementia.2 Self-reported stress was determined by asking participants to respond yes or no to the question "Have you suffered psychological stress or acute disease in the past 3 months?" included as part of the Mini Nutritional Assessment Questionnaire.3 Height and weight were measured by research study staff and body mass index (BMI) was calculated as weight (kg) / height (m)2.

Laboratory Methods

Blood samples were processed, stored, and shipped by each institution's local Sample Processing Core to Battelle (Columbus, OH). Hemagglutinin inhibition (HAI) testing was performed in duplicate against the influenza vaccine strains in the 2009-2010 Northern Hemisphere influenza vaccine. Although there is debate about the best correlate of protection for influenza,4 seroprotection was defined as an HAI titer of ≥40 since it is the correlate recognized by the United States Food and Drug Administration.5 Seroconversion was measured at S2 and defined as a four-fold rise in HAI post-influenza vaccination compared to pre-vaccination or ≥40 if S1 was <10. Maintenance of antibody response was measured in participants with an S2 ≥40, and defined as an S3≥40. If duplicate HAI results were discrepant by more than two fold, a third test was performed and the minimum result was recorded.

Statistical Analysis

Multivariable logistic regression models were run for the binary outcomes seroprotection at S2 and S3 and seroconversion at S2. Age, BMI, stress, high risk medical conditions (yes/no), female gender, study site and transformed S1 titer were included in all models. Time in days from S1 to S3 was included in models with S3 related outcomes as a continuous variable. Restricted cubic splines were applied to age and BMI. All raw HAI titers were log transformed using the method outlined by Beyer,6 changing dilution titers to integers with HAI <10 coded as 0, 10 as 1, 20 as 2, 40 as 3 and so on. Interaction terms between age and high risk status, and stress or age and sex were tested and were not included based on non-significant p values of overall interaction terms. Figures were generated by predicting the probability of having a HAI titer ≥40 by logistic regression. All analyses were done using R version 2.12.2.