We carried out a study for the nonculture detection of susceptibility of Neisseria meningitis to penicillin G in three laboratories of the European Monitoring Group on Meningococci (EMGM). Thirteen clinical samples (cerebrospinal fluids) and corresponding bacterial isolates from 13 cases of invasive meningococcal infection were distributed to the three laboratories. The MICs of penicillin G were determined for the isolates. Each laboratory used an "in-house" PCR-based method to determine alterations to the penA gene, which is associated with a reduced susceptibility to penicillin G. Nucleotide sequences from the 3' end of the penA gene were also determined. We observed a good correlation between genotyping of penA and the phenotypic determination (MIC) of susceptibility to penicillin G. The results obtained by the three methods for penA in the samples correlated very well with those obtained in bacterial isolates and with sequence data. The kappa coefficient that was used to estimate the level of agreement between genotypic results varied between 0.65 and 1, indicating a good agreement. This suggests that genotyping can predict susceptibility of N. meningitidis to penicillin G. These data strongly suggest that genotyping of penA should be used to determine meningococcal susceptibility to penicillin G in culture-negative cases. Although the nucleotide sequence of penA may be the gold standard in genotyping of penA, the less expensive PCR-based approach reported in this study may be quicker when a large number of isolates and clinical samples need to be tested.