Genotype prevalence and risk factors for severe clinical adenovirus infection, United States 2004-2006

Abstract

Background. Recently, epidemiological and clinical data have revealed important changes with regard to clinical adenovirus infection, including alterations in antigenic presentation, geographical distribution, and virulence of the virus. Methods. In an effort to better understand the epidemiology of clinical adenovirus infection in the United States, we adopted a new molecular adenovirus typing technique to study clinical adenovirus isolates collected from 22 medical facilities over a 25-month period during 2004-2006. A hexon gene sequence typing method was used to characterize 2237 clinical adenovirus-positive specimens, comparing their sequences with those of the 51 currently recognized prototype human adenovirus strains. In a blinded comparison, this method performed well and was much faster than the classic serologic typing method. Results. Among civilians, the most prevalent adenovirus types were types 3 (prevalence, 34.6%), 2 (24.3%), 1 (17.7%), and 5 (5.3%). Among military trainees, the most prevalent types were types 4 (prevalence, 92.8%), 3 (2.6%), and 21 (2.4%). Conclusions. For both populations, we observed a statistically significant increasing trend of adenovirus type 21 detection over time. Among adenovirus isolates recovered from specimens from civilians, 50% were associated with hospitalization, 19.6% with a chronic disease condition, 11% with a bone marrow or solid organ transplantation, 7.4% with intensive care unit stay, and 4.2% with a cancer diagnosis. Multivariable risk factor modeling for adenovirus disease severity found that age <7 years (odds ratio [OR], 3.2; 95% confidence interval [CI], 1.4-7.4), chronic disease (OR, 3.6; 95% CI, 2.6-5.1), recent transplantation (OR, 2.7; 95% CI, 1.3-5.2), and adenovirus type 5 (OR, 2.7; 95% CI, 1.5-4.7) or type 21 infection (OR, 7.6; 95% CI, 2.6-22.3) increased the risk of severe disease. © 2007 by the Infectious Diseases Society of America. All rights reserved.