Factors Associated with Recurrent Influenza Infection, 2004-2015

Emma Seagle
Emory University

Emma E. Seagle1, Jennifer P. King1, Burney A. Kieke1, Jennifer K. Meece2, Edward A. Belongia1, Huong Q. McLean1
1Center for Clinical Epidemiology & Population Health, 2Integrated Research & Development Laboratory

Research area:  Epidemiology

Background: Previous influenza infection likely confers some protection against antigenically similar viruses, but the frequency and characteristics of repeated infections are not well understood. We described characteristics of persons with repeat medically attended influenza and identified factors associated with same subtype repeat infections.

Methods: We conducted a case-control analysis using data from annual influenza vaccine effectiveness studies at Marshfield Clinic from 2004 through 2015. These studies enrolled individuals with medically attended acute respiratory illness ≤7 days duration for influenza testing. For this analysis, a case was defined as an enrollee with ≥2 same subtype/lineage infections. Cases were matched to enrollees who had infection with the same subtype/lineage in the same season as the case’s first infection and had ≥1 medically attended negative influenza visit(s) in the same season as the case’s second infection. Conditional logistic regression was used to identify factors associated with repeat A/H3N2.

Results: Of the 12,950 patients enrolled, 324 (2.5%) had ≥2 medically attended influenza infections. Of those, 134 (41.4%) were repeat influenza A and 14 (4.3%) repeat influenza B infections; 15 had repeat infection with the same subtype (14 A/H3N2, 1 A/H1N1pdm09). No repeat B infections were same lineage reinfections. The odds of a repeat infection with A/H3N2 were significantly higher for those unvaccinated during the season of their first infection (odds ratio [OR]: 10.2, 1.2-1.1-90.6, p-value: 0.02). Age and degree of healthcare services utilization were not associated with A/H3N2 reinfection in this exploratory analysis. 

Conclusions: Most medically attended repeat infections were of different type/subtype/lineage, with same subtype recurrent infections uncommon. A/H3N2 comprised the majority of same subtype repeat infections, potentially explained by the severity of A/H3N2 (increasing likelihood to seek medical attention) or high antigenic drift. However, confounding may exist given A/H3N2 dominance in recent seasons.