We have previously shown that patients with SLE have significantly lower percentages and absolute numbers of NK(CD3-/CD16+56) cells in their peripheral blood compared with normals. Patients with active disease had very low levels of NK cells and the reduction was also associated with patients who had renal involvement. We have now performed a serial study immunophenotyping 11 patients with SLE and renal involvement using dual colour immunofluorescence and flow cytometry. Patients were tested every three months on an average of three occasions. As a control, nine SLE patients without renal involvement were immunophenotyped for similar intervals; 11 normal controls were also tested. Major lymphocyte subsets (T, B and NK) remained very stable during serial bleeds. However, the NK cell populations were decreased significantly in patients with renal involvement both as percentages (5 +/- 6 vs 9 +/- 5, P < 0.0001) and absolute counts (75 +/- 108 vs 109 +/- 52, P < 0.001) in comparison to non-renal patients. Analysis of disease activity using BILAG score showed an inverse correlation between renal system activity and percentage and absolute number of NK cells (P < 0.002 and 0.01, respectively). In this study we have also analysed a CD8 T cell subset which we have not studied before. We have found a significantly increased percentage of CD38+CD8+ T cells(activated CD8 subset) in patients with SLE in comparison to normal controls. We did not find any association with the CD38+CD8+ T cells and disease activity as measured by BILAG or renal involvement. NK cells are important factors in immunity against virus infections and tumour cells. CD38+CD8+T cells are increased in viral infections. We speculate that the lack of NK cells in SLE patients might have an association with increased CD38 expression.