To determine whether respiratory virus infections (URI) are associated with exacerbation of nephrotic syndrome (NS) in childhood, a prospective two-winter study of 32 children with NS was done. We obtained pre- and post-season viral serologic studies, biweekly nose and throat viral cultures, daily urinalysis, biweekly telephone follow-up for URI and renal complaints, and clinical assessments as indicated. When a URI occurred, viral cultures were done weekly if the child was at home and twice weekly if hospitalized. Sixty-one URIs occurred; the agent was identified in 33 (51.6%) (respiratory syncytial virus 14, influenza virus five, parainfluenza virus five, varicella zoster virus four, adenovirus three, Mycoplasma pneumoniae one, and Chlamydia trachomatis one). Forty-one exacerbations occurred, 71% with URI; 29% had no URI during the preceding 10 days (P less than 0.01). Total relapse occurred in 29 of 41 exacerbations, 69% with URI and 31% without URI (P less than 0.01). Patients with unstable NS had more exacerbations than those with stable NS (15 of 19 (79%) vs four of 13 (31%), P less than 0.001) and more URI (2.32 vs 1.46 per child, P less than 0.05). Exacerbations in patients with minimal change, mesangioproliferative, and focal glomerulosclerosis occurred in 40%, 60%, and 64%, respectively. We conclude that exacerbations and relapses of childhood NS are temporally related to URI. Inasmuch as multiple viral agents were associated with exacerbations, nonspecific host response to infection, not viral antigen or antibody response, may be the link to NS.