The peripheral blood CD4 count is a useful marker of immunological status in HIV infection. However, it makes up only 1% of total body CD4 cells [Parslow T. Lymphocytes and lymphoid tissue. In: Stites D, Terr A, Parslow D, editors. Basic and clinical immunology. Appleton and Lange; 1994. p. 22-40.] has a wide intra- and inter-individual variability [Turner BJ, Hecht FM, Ismail RB. CD4+ T-lymphocyte measures in the treatment of individuals infected with human immunodeficiency virus type 1. A review for clinical practitioners. Arch Intern Med 1994;154:1561-73.] and CD4 cell pathology is just one of the immune defects caused by HIV [Fauci S. Multifactorial nature of human immunodeficiency virus disease. Science 1993;262:1011-8.]. Thus a given value should always be interpreted within a specific clinical context. We describe an individual with Pneumocystis pneumonia (PCP) who was reviewed by HIV medical services several times before the correct diagnosis was made. On each occasion the possibility of PCP was discounted as he had a well-preserved blood CD4 count. Subsequent examination of his pulmonary T-lymphocyte subsets revealed marked CD4 lymphopenia.