To determine whether rimantadine can protect family members from acquiring influenza A viral illness and to assess the possible selection of drug-resistant strains of virus, we conducted a randomized, double-blind, placebo-controlled study in three communities during two influenza seasons. When influenza A occurred in a family, the members (including the index patient) were given either rimantadine (adult oral dose, 200 mg per day) or placebo for 10 days. The presence of illness was monitored by daily recording of symptoms and temperature measurements; infection was determined by isolation of the virus and by serologic studies. Among households with documented influenza A infections, symptomatic illness occurred in one or more contacts in 10 of 28 families treated with rimantadine and in 10 of 209 families treated with placebo. Asymptomatic secondary influenza A infections were found in five families assigned to receive rimantadine and in four families assigned to receive placebo. Rimantadine-resistant strains of influenza A virus (H3N2 subtype) with mutations consisting of single amino acid changes in the M2 protein (residue 27, 30, or 31) were recovered from eight index patients and five contacts treated with rimantadine. There was apparent transmission of drug-resistant strains of virus in six contacts with secondary illnesses in five families. We conclude that when index patients are treated concurrently, rimantadine is ineffective in protecting household members from influenza A infection. If rimantadine is used for both treatment and postexposure prophylaxis in families, rapid selection and apparent transmission of drug-resistant influenza A viruses can occur.