School of Public Health

Jason Goldman

Readily-Available Antibiotic Formulations to Improve Outcomes in Cancer Patients with Severe Sepsis and Septic Shock


Cancer patients are at high risk for severe sepsis (SS) and septic shock (SSh) and delay to effective antimicrobial therapy (ABx) is strongly associated with increased mortality. Anti-pseudomonal beta-lactam intravenous monotherapy is equally effective and less toxic than combination therapy for uncomplicated neutropenic fever, but combination therapy may be superior for more severe disease.


We implemented a clinical algorithm to simplify timely and effective empiric ABx and other resuscitative care to cancer outpatients with SS/SSh prior to hospital admission. Triple therapy with meropenem, tobramycin and linezolid or alternatives such as aztreonam for penicillin-allergic patients can be co-administered and provides broad coverage for resistant organisms typically encountered in this population. A pre-printed order form triggered dispensing of kits containing ABx, fluids and dexamethasone. We performed a retrospective cohort study to assess the impact of this strategy.


From 1/1/08 through 1/31/12, 162 patients met inclusion criteria. Median age was 53 (IQR: 42 – 63) years and 65% were male. The majority of patients (87%) had hematopoietic malignancies. 77 (48%) were hematopoietic stem cell transplant recipients and 80 (49%) were neutropenic. SSh was diagnosed in 25 patients (15%), SS in 46 (28%), sepsis in 72 (44%) alternative diagnosis in 6 (4%) and infection without systemic inflammatory response syndrome in 13 (8%). Median time from clinical encounter to ABx administration was 111 (IQR: 60 – 178) minutes, 93% had blood cultures drawn prior to ABx, 46% received dexamethasone and 99% had crystalloid infusion started before hospital transfer. De-escalation on hospital day 1 occurred in 95% of persons admitted. 44% of 25 persons with SSh received vasopressors. 71 persons (44%) had bacteremia and 18% of 93 isolates were multidrug resistant. Possible nephrotoxicity occurred in 7 patients. 30 day mortality was 6/160 (4%) including 3/71 (4%) with SS/SSh. For each hour delay to administer antibiotics, there was an18% increased risk of developing SSh or death within 30 days (95% CI: 4 – 34%), p=0.01.


A program to simplify choice of aggressive empiric ABx among cancer patients presenting to an ambulatory clinic with suspected sepsis was associated with excellent survival in those with SS/SSh, without excessive adverse events or inappropriately long empiric ABx durations.