Patients hospitalized with community-acquired pneumonia caused by S. pneumoniae strains with intermediate susceptibility to penicillin according to the conventional definition respond well to treatment with adequate doses of beta-lactam antibiotics. All studies currently available comparing mortality between patients with pneumonia caused by nonsusceptible and susceptible pneumococci agree that resistance of MIC 2 mg/L is not associated independently with an increased mortality. Most but not all studies could not prove an effect of microbial resistance on morbidity. There are data suggesting, however, that pneumococcal pneumonia caused by highly resistant strains (MIC > or = 4 mg/L) does affect the outcome. Pneumococcal resistance remains a matter of concern. Most reports show an increase not only of resistance rates, but also of the proportion of highly resistant strains. The selection of initial empirical antimicrobial treatment of patients with community-acquired pneumonia should be performed judiciously. Because the serum and pulmonary levels achieved with penicillin or related drugs are several times higher than the MICs of most strains, pneumonias caused by S. pneumoniae currently defined as not susceptible to penicillin should respond well to treatment with a beta-lactam antibiotic, used in optimal doses. Consequently, there is no reason fundamentally to change the current approach to initial empiric antimicrobial treatment of patients with community-acquired pneumonia. Nevertheless, increases in resistance to macrolides may prompt a limited use of these drugs in the outpatient setting. In any case, treatment failures may occur at higher levels of resistance, and a change in the definition of susceptibility categories toward higher cutoffs for S. pneumoniae seems to be reasonable.