Penicillin resistance of Pneumococci is a problem in several European countries. Therefore, we examined 510 invasive pneumococcal isolates, collected between 1992 and 1999 via a national network of diagnostic laboratories covering the entire population of Scotland, for penicillin susceptibility, in order to determine the prevalence, site of infection and serogroup/type distribution of penicillin-resistant Pneumococci in the elderly (> or = 65 y). Of the 510 isolates, 91.6% (n = 467) were from blood, 4.7% (n = 24) from other sterile sites and 3.7% (n = 19) from cerebrospinal fluid. The prevalence of penicillin non-susceptible isolates during the study period was 9%. An increase in the proportion of Pneumococci non-susceptible to penicillin was detected from 1996 onwards, from 10.8% in 1996 to 14.3% in 1999. There were 2 isolates with high-level penicillin resistance, both of which were of serotype 14, accounting for 4.3% (2/46) of all non-susceptible isolates. Penicillin non-susceptible isolates belonged to the following serogroups: 14 (32.6%); 9 (30.4%); 6 (19.6%); 23 (10.9%); and 19 (6.5%). The leading non-susceptible serotype/group varied according to the specimen type: serotype 14 for blood and serogroup 9 for all other sterile sites. Current polysaccharide and new 7-, 9- and 11-valent conjugate vaccine formulations included the serogroups responsible for all the penicillin non-susceptible isolates detected. Therefore vaccination represents the most effective strategy for decreasing the burden of drug resistance. Constant surveillance of the patterns of antibiotic non-susceptible isolates, the site of infection and the serogroup/type are necessary in order to select antibiotic therapy and establish vaccination policy for the prevention of invasive pneumococcal disease.