To determine the incidence of infection in human immunodeficiency virus (HIV)-infected patients during periods of neutropenia and non-neutropenia. To compare the infection rates in patients with HIV disease to those in a group hospitalized with neutropenia and hematologic malignancy.

Prospective observational study conducted between December 1985 and December 1987 at a university teaching hospital. Thirty patients with documented acquired immunodeficiency syndrome (AIDS) and absolute T-helper cells <200 mm/mm3. All patients had a period of non-neutropenia following a neutropenic period (neutrophils <1000 cells/mm3).

The rate of first infection during neutropenic and non-neutropenic periods for opportunistic infection and nonopportunistic infections were compared. There were no differences between infection rates for the two time periods for both types of infections. A subgroup of patient care days in which non-neutropenic days followed neutropenic days also was studied to eliminate selection bias. In this group, a comparison of infection rates also revealed no difference between neutropenic and non-neutropenic periods. An alternate analysis of the time until first infection during periods of neutropenia or non- neutropenia was done using the Kaplan-Meier product limit method. There was a longer infection-free period for the neutropenic group for opportunistic infections, but it was not statistically significant (p<.1). In addition, we compared HIV-infected patients with a group of 37 patients with neutropenia from hematologic malignancy. There was a significantly higher rate of all infections, particularly bacteremias (p<.001), in the group of patients with hematologic malignancies when compared with all subsets of patients with HIV disease.

We conclude that patientswith HIV disease and modest neutropenia do not have an increased risk of bacterial infection. The incidence of all infections is significantly greater in patients with neutropenia secondary to hematologic malignancy, (Infect Control Hosp Epidemiol. 1991;12:429-434.)