Objectives: To determine the prevalence of eosinophilia among antiretroviral therapy (ART)-naïve patients infected with human immunodeficiency virus (HIV) and to identify variables associated with eosinophilia.
Methods: We included all ART-naïve HIV-infected patients entering into care at the Thomas Street Health Center (Houston, Texas) between February 2007 and January 2009. Eosinophilia was defined as absolute eosinophil count ≥ 400 cells per cubic millimeter. Patients with eosinophilia (cases) at baseline were matched to patients without baseline eosinophilia (controls). Clinical and laboratory data were collected for cases and controls. Variables associated with eosinophilia were evaluated by univariate and multivariate analyses.
Results: Sixty-five (9.7%) of 671 ART-naïve patients had eosinophilia. There was no difference in age, sex, race, or baseline CD4 count between patients with and without eosinophilia; however, patients with eosinophilia were more likely to have higher HIV RNA viral loads (5.05 vs 4.82 log10 copies per milliliter; P = 0.019). A total of 52 (80%) of 65 patients with eosinophilia (cases) had at least two follow-up clinic visits. They were matched to 104 controls. Skin rash was the only variable associated with eosinophilia (odds ratio 2.16, 95% confidence interval 1.04-4.47) in our multivariate analysis. Of eight cases tested, only one, from Central America, had a parasitic infection (hookworm). Thirty-eight (73.1%) patients experienced resolution of their eosinophilia by the end of the study (mean follow-up 1019 days). Resolution of eosinophilia did not differ between patients with and without HIV viral suppression.
Conclusions: Eosinophilia is not an infrequent occurrence among ART-naïve HIV-infected patients. Patients with eosinophilia are more likely than patients without eosinophilia to present with a skin rash. HIV RNA viral suppression did not necessarily result in the resolution of eosinophilia. Extensive workup for eosinophilia may not be necessary in most cases.