Serum immunoreactive erythropoietin levels and associated factors amongst HIV-infected children

Objectives: To determine the spectrum of serum immunoreactive erythropoietin (SIE) levels amongst HIV-infected children aged < 13 years in relation to the levels among healthy children as well as those with real failure; to examine the relationship between clinical and laboratory parameters and SIE levels. Design: A cross-sectional study with a descriptive non-interventional format. HIV-infected Canadian subjects were recruited through four tertiary Canadian and one Bahamian centre. Children with renal failure and healthy children were recruited from one of the Canadian centres. Methods: Study subjects had clinical and laboratory profiles determined at baseline and at each of five follow-up periods over 1 year. SIE levels were measured by radioimmunoassay with a normal range of 12-28 IU/l. Data handling and statistical functions were performed by the Canadian HIV Trials Network. Results: The study enrolled 133 HIV-infected subjects and 38 controls. Of these, 117 HIV-infected subjects, 24 healthy controls, and 11 controls with renal failure were eligible for analysis. The median age of infected subjects was 44 months, whereas that of healthy controls was 56 months, and 95 months for controls with renal failure. The median SIE levels were 14 and 11 IU/l for subjects with renal failure and healthy subjects, respectively. The median SIE level was 61 IU/l among zidovudine (ZDV)-treated subjects and 22 IU/l among ZDV-naive HIV-infected subjects. HIV-infected children almost invariably had SIE levels < 200 IU/I. The median SIE levels amongst HIV-infected subjects whose hemoglobin levels were < 100 g/l were 98 and 31 IU/l for ZDV-treated and ZDV-naive subjects, respectively (P = 0.002). This difference in median SIE levels between ZDV-treated subjects and ZDV-naive subjects was also observed among subjects whose hemoglobin levels were > 100 g/l (median, 58 and 15 IU/l, respectively; P < 0.001). Hemoglobin level was the most important predictor of log(10) SIE (P < 0.01 for ZDV-treated and ZDV-naive subjects). Conclusions: SIE levels amongst HIV-infected children were affected by HIV infection, use of ZDV, and presence or absence of anemia. SIE levels amongst HIV-infected children were generally lower than 200 IU/l. This characterization of SIE levels will facilitate clinical trials of exogenous recombinant human erythropoietin in HIV-infected children with anemia. (C) 1998 Lippincott Williams & Wilkins.