Cullen, M.H., Billingham, L.J., Gaunt, C.H., & Steven, N.M. (2007). Rational selection of patients for antibacterial prophylaxis after chemotherapy. Journal of Clinical Oncology, 25, 4821–4828.

Adult patients with cancer receiving cyclic chemotherapy for solid tumors or lymphoma who were at risk for temporary, severe neutropenia (absolute neutrophil count [ANC] < 500/mm³) were treated with oral levofloxacin 500 mg or matching placebo daily for seven days during the expected neutropenic period. Treatment began on day 5 for regimens associated with early onset of neutropenia (e.g., docetaxel), day 8 for 14-day and 21-day cycles, and day 15 for 28-day cycles.

Patients were on study for a mean of 4.4 cycles of chemotherapy, with 45% of patients completing six cycles.

784 patients were randomly assigned to the placebo arm and received 3,459 cycles of chemotherapy (mean = 4.4 cycles per patient).

Random assignment of patients in the SIGNIFICANT trial was stratified by age (younger than age 40, 40–59, and 60 years of age and older) and cancer type (breast, testicular, small cell lung, Hodgkin disease, non-Hodgkin lymphoma [NHL], and others).

1,565 adults starting chemotherapy for solid tumors or lymphomas; eligible regimens were known to be associated with a risk of neutropenia (ANC <  500/mm³), but were not routinely given with granulocyte–colony-stimulating factor (G-CSF) support. Many different types of cancer were included.

60 oncology centers in the United Kingdom.

The study was a prospective, multicenter, randomized, double-blind, placebo-controlled trial with secondary univariate and multivariate analysis.

• The primary outcome measure was the incidence of clinically-documented febrile episodes (FEs), defined as a core temperature exceeding 38°C and attributed to infection.
• The original results demonstrated a reduction in FEs in patients receiving levofloxacin. However, the use of prophylactic antibiotics may increase the rate of antibiotic resistance, so the data from a 2005 study by the authors were analyzed to determine which patients benefit most from prophylactic antibiotics. The analysis measured rates of FEs across all cycles, calculated as the proportion of randomly assigned patients experiencing the event at least once during the chemotherapy program (per patient FE rate). The analysis separates events in cycle 1 because chemotherapy dose reductions and possibly antibiotic resistance may affect later cycles. Per cycle FE rates are calculated as the proportion of observed cycles in which an event occurs.
• Rates of hospitalization for the treatment of suspected infection (HTSI) also were analyzed. HTSI occurred in a subset of patients with FEs and also in some patients not meeting the trial criteria for an FE.
   

119 of 784 (15.2%) control group participants had at least one FE during chemotherapy.

Treatment benefit of quinolone prophylaxis was present across all cycles.

As reported in a 2005 article by the authors, the per-patient FE rate was 10.8% (84 of 781) for patients receiving levofloxacin compared with 15.2% for patients receiving placebo (119 of 784), giving a statistically significant reduction in the risk of FE (odds ratio = 0.67; 95% confidence interval [0.5, 0.91]; p =0.009).

For the first cycle only, the per-patient FE rate was 3.5% in patients receiving levofloxacin compared with 7.9%  in controls (odds ratio = 0.42; 95% confidence interval [0.26, 0.66]; p =0.0001),  whereas for non–first cycles, the per-patient FE rate was 7.8% (61 of 781) and 9.8%  (77 of 784), respectively (odds ratio = 0.78; 95% confidence interval [0.55, 1.11]; p = 0.16).

Per-cycle FE rates in cycle 2 and cycles 2–6 indicate that prophylactic benefit is gained in the small number of patients who experience an FE in cycle 1, but not in the much larger group of patients who do not experience an FE in cycle 1.

The data suggest that the benefit of antibiotics is greatest in those who experience an FE in cycle 1 because they are at higher risk of developing an FE in subsequent cycles compared with patients who do not develop an FE in cycle 1.
 

• Secondary analysis
• The primary outcome is the incidence of FEs rather than documented infections, and a reduction in documented infections may be a more meaningful endpoint.
• The sample is at low risk for febrile neutropenia. An important consideration for low-risk patients with short durations of neutropenia is whether quinolone prophylaxis is of greater benefit than the option of outpatient quinolone treatment for fever and neutropenia, should it occur. Both the National Comprehensive Cancer Network and the Infectious Diseases Society of America recommend oral quinolone-based regimens as outpatient empirical therapy for neutropenic fever in adults meeting criteria for low risk for complications.
• Use of quinolone prophylaxis may preclude later use of the regimens as empirical therapy for neutropenic fever in the same patient. Guidelines on outpatient management of adults with neutropenic fever assume that quinolones were not used as prophylaxis and the use of quinolones for prophylaxis precludes their use for treatment.