The present study compared two different systems of classification of patients with
common variable immunodeficiency (CVID); one based on in vitro
immunoglobulin biosynthesis; and another on CD4-naïve cell counts. Peripheral blood mononuclear cells (PBMCs) were isolated from 35 patients with CVID and 20 healthy controls. They were stimulated for the secretion of
IgM and
IgG after stimulation with Staphylococcus aureus Cowan I (SAC) upon supplementation of
interleukin-2 (IL-2) or with
pokeweed mitogen. T cell subsets were estimated by flow cytometry. By the first system, patients were classified into group A (n=18) with secretion of neither
IgG nor
IgM; into group B (n=12) with detectable
IgM but no
IgG secretion; and into group C (n=5) with
IgM and
IgG secretion similar to controls. By the second system, patients were classified into group I (n=12) with less than 109 CD4-naïve cells/mul; into group II (n=12) with CD4-naïve cells within 109-225microl(-1); and into group III (n=11) with more than 225 CD4-naïve cells/mul. All groups I-III were defective for in vitro release of
IgG and
IgM. The likelihood ratio for
splenomegaly in patients with <225 CD4-naïve cells/mul was 5.08 (p: 0.024). CD4-naïve cell counts of patients were positively correlated to serum levels of
IgG and
IgA of patients. The presented results revealed that the former system described adequately the function of B cells and the latter the clinical status of the patient. Our proposal is that both should be used for the characterization of patients with CVID.