P. Furno et al., Ceftriaxone versus beta-lactams with antipseudomonal activity for empirical, combined antibiotic therapy in febrile neutropenia: a meta-analysis, SUPP CARE C, 8(4), 2000, pp. 293-301
The object of this work was to compare the efficacy of antibiotic combinati
ons including ceftriaxone with that of combinations including an antipseudo
monal beta-lactam for the empirical treatment of febrile neutropenia in can
cer patients. We identified all published randomised trials comparing two a
ntibiotic combinations differing only in the beta-lactam, being ceftriaxone
in one treatment group and an antipseudomonal beta-lactam in the other. Th
e quality of individual trials was formally evaluated. A meta-analysis was
performed using the Peto-modified Mantel-Haenszel method for combining bina
ry data. Primary analysis was done, for both febrile episodes and bacteraem
ic episodes, using failure of empirical antibiotic treatment defined as mod
ification of the initial allocated regimen or death during treatment. Secon
dary analysis was done using death from any cause in the two treatment grou
ps. Data relating to 1,537 febrile neutropenic episodes recorded in eight r
andomised clinical trial were pooled s. Overall, there were 256 treatment f
ailures out of 782 febrile episodes treated with ceftriaxone-containing com
binations (32.7%), and 243 out of 755 treated with antipseudomonal beta-lac
tam regimens (32.1%). The pooled odds ratio of failure for ceftriaxone-cont
aining combinations for febrile episodes was 1.04, with the 95% confidence
interval ranging from 0.84 to 1.29, and that for bacteraemic episodes was 0
.93 (95% confidence interval 0.58-1.49). With regard to overall mortality,
there were 54 deaths among 782 febrile episodes treated with ceftriaxone-co
ntaining combinations (6.9%) and 62 deaths among 755 febrile episodes treat
ed with antipseudomonal beta-lactam-containing regimens (8.2%). The pooled
odds ratio of death for ceftriaxone regimens was 0.84 (95% confidence inter
val 0.57-1.24). Results of this meta-analysis show that in the empirical tr
eatment of febrile neutropenia, antibiotic combinations containing ceftriax
one are as effective as those in which the beta-lactam has specific activit
y against Pseudomonas aeruginosa, such as ureidopenicillin or ceftazidime.