Variation in attack rates of paralytic disease by region during the 1988-1989 epidemic of type 1 poliomyelitis in Oman provided the stimulus to test the hypothesis that these observations were due to regional differences in the response of infants to trivalent oral poliovirus vaccine (OPV). Seroprevalence studies of 394 children born during the outbreak were conducted in six different regions of Oman and in two socioeconomic status (SES) groups in the capital city of Muscat; a seroconversion study was also carried out in 105 infants born after the outbreak. Seroprevalence rates by region after receipt of at least three doses of OPV ranged from 90% to 100% (median 94%) to poliovirus type 1, and from 86% to 100% (median 97%) to type 2, and from 47% to 79% (median 72%) to type 3, with the lowest rates observed in regions with the highest incidence of type 1 paralytic disease. In Muscat, seroprevalence rates were also significantly lower in low versus high SES groups (type 1: 84% versus 98%, respectively [P = 0.006]; type 3: 59% versus 86%, respectively [P = 0.001]). In the seroconversion study conducted after the outbreak, 89%, 100% and 50% of infants had detectable antibodies to types 1, 2, and 3, respectively, after four doses of OPV. Low responses to type 3 were also associated with the occurrence of sporadic cases of type 3 poliomyelitis in 1991, in spite of high rates of coverage with at least four doses of OPV (> 96%) throughout the country.(ABSTRACT TRUNCATED AT 250 WORDS)