Streptococcus pneumoniae bacteremia in patients with cancer: disease characteristics and outcomes in the era of escalating drug resistance (1998-2002).

In the current era of multidrug-resistant organisms, the clinical spectrum of Streptococcus pneumoniae infection remains unclear, especially in immunosuppressed patients with cancer. We sought to define the characteristics of pneumococcal bacteremia in patients who were receiving care at a comprehensive cancer center. All consecutive episodes of S. pneumoniae bacteremia between January 1998 and December 2002 were evaluated retrospectively. One hundred thirty-five episodes of pneumococcal bacteremia occurred in 122 patients. Sixty-three (52%) of 122 patients had hematologic malignancies; the others had solid tumors. The median Acute Physiology and Chronic Health Evaluation II score was 14 +/- 5. Twenty-four episodes (18%) occurred during neutropenia (<500 cells/microL). Sixty-five patients (53%) were receiving antineoplastic therapy, and 36 (30%) were receiving systemic corticosteroids. Twelve (41%) of 29 hematopoietic stem cell transplant (HSCT) recipients had received transplantation within 12 months of the infection diagnosis; 11 patients had graft-versus-host disease (chronic in 10). In 27 episodes (22%), S. pneumoniae bacteremia was considered as a breakthrough infection. Nine (56%) of 16 hospital-acquired episodes of S. pneumoniae bloodstream infection occurred in patients with profound neutropenia, whereas 15 (13%) of 119 episodes of community-acquired infection occurred during neutropenia (p < 0.0002). In 91 episodes (67%), patients had radiographic evidence of pneumonia. Infected catheters were associated with 21 episodes (16%). Forty-eight (36%) of 135 isolates were not susceptible to penicillin (minimum inhibitory concentration [MIC] > or = 2 microg/mL); 9 (7%) showed intermediate susceptibility to ceftriaxone (MIC >0.5 and <2.0 microg/mL). Nineteen patients (16%) died within 2 weeks of diagnosis; 18 deaths were attributed to systemic pneumococcal infection. Univariate analysis showed no significant increase in the risk of short-term death in patients with infection due to penicillin non-susceptible organisms (OR [odds ratio], 1.47; 95% confidence intervals [CI], 0.53-4.05; p < 0.46), initially discordant treatment (OR, 1.0; 95% CI, 0.62-665.4; p < 0.16), presence of pneumonia (OR, 1.19; 95% CI, 0.39-3.62; p < 0.76), neutropenia (OR, 1.0; 95% CI, 0.28-4.09; p < 0.92), systemic corticosteroid use (OR, 1.96; 95% CI, 0.69-5.60; p < 0.21), or antineoplastic therapy (OR, 1.45; 95% CI, 1.52-4.05; p < 0.47). Similarly, patients with hematologic cancers compared to those with solid cancers (OR, 1.0; 95% CI, 0.49-3.70; p < 0.56) and recipients of HSCT compared to those with no history of transplantation (OR, 1.0; 95% CI 0.59-12.71; p < 0.20) did not have a less favorable outcome. In conclusion, most pneumococcal bloodstream infections were community acquired, although hospital-acquired infections were common in neutropenic patients. It is noteworthy that initially discordant therapy, penicillin non-susceptible S. pneumoniae, and other conventional predictors of unfavorable outcome were not associated with increased mortality rates in these high-risk patients with cancer.