Infectious complications are a frequent cause of morbidity and, at many centers, the major cause of death in patients with cancer. The increased risk and severity of infectious sequelae result from profound alterations in normal host defenses that occur secondary to the underlying malignancy and the treatment thereof. During the last decade, early empiric antibiotic therapy has become standard practice in the initial management of febrile granulocytopenic patients and has contributed significantly to the improved outcome among patients undergoing cancer therapy. Although early death due to unsuspected or inadequately treated bacterial infection has been largely overcome, new problems--also with life-threatening implications--have emerged. As the use of cancer chemotherapy continues to increase, new populations of patients are being placed at increased risk of infection. Defining the host and environmental factors that contribute to this risk assumes central importance for delineating those patients who require the most intense surveillance. Changing medical practices (e.g., increased use of indwelling catheters) have contributed to the emergence of new pathogens. Recent drug developments (e.g., the third-generation cephalosporins and extended-spectrum penicillins) offer new treatment options, as well as generate controversy and confusion. For example, authorities disagree on the optimal duration and modifications in treatment that are required by cancer patients who remain granulocytopenic and who thus are at continued risk of multiple infectious episodes or superinfections. A question of current interest is whether combination therapy with synergistic agents is important in light of the development of the third-generation cephalosporins and extended-spectrum penicillins. Several of these new antibiotics have an exceedingly broad spectrum of activity that includes Pseudomonas aeruginosa, as well as Enterobacteriaceae, Serratia, Citrobacter, indole-positive Proteus, and anaerobes (including Bacteroides fragilis). However, the third-generation cephalosporins are not as active against staphylococci and streptococci as are the first-generation cephalosporins, and none is effective against enterococci. Nonetheless, these agents achieve serum levels that can be 10 to 100 times greater than the minimal inhibitory and bactericidal concentrations of gram-negative bacteria, raising the possibility that these drugs might be effective as single agents. The advantages of the third-generation cephalosporins are their minimal toxicity and long serum half-lives.(ABSTRACT TRUNCATED AT 400 WORDS)