A comparison of ceftazidime and aminoglycoside based regimens as empirical treatment in 1316 cases of suspected sepsis in the newborn

Abstract

We report a prospective, non-blind, randomised, multicentre, parallel group, multinational investigation to compare ceftazidime to aminoglycoside based regimens as empirical treatment in 1316 cases of suspected sepsis in the newborn. In each of the 15 study centres either ceftazidime alone (CAZ) or ceftazidime + ampicillin (CAZ+AMP) was compared to an aminoglyocoside/ampicillin combination (AG+AMP). In all cases treatment was based on “an intention to treat”. Bacteria considered to be pathogenic were isolated from 176/1316 (13.4%) patients. The incidence of proven infection varied from 39% in a Yugoslav centre to 6% in a British centre; a further 489/1316 (37.1%) patients fulfilled the criteria for clinically suspected sepsis. A total of 210 bacterial isolates from 197 infection sites in 176 patients were considered to be clinically significant. The cure rate for evaluable patients with proven infection who were treated with CAZ+AMP (97%, 30/31) was significantly higher than that for the corresponding patients treated with AG+AMP (66%, 26/39), (P<0.002). The difference in cure rate between CAZ monotherapy (79%, 34/43) and AG+AMP (86%, 32/37) was not significant. Treatment failed in 28/150 (18.7%) evaluable patients. There were significantly fewer failures (P<0.001) with CAZ+AMP than with AG+AMP therapy. There were 55 staphylococcal infections. Treatment was successful in 16/19 evaluable patients treated with CAZ or CAZ+AMP and in 16/29 evaluable patients treated with AG+AMP. None of the study centres encountered problems with ceftazidime resistant bacteria. The cure rate for patients with only clinical and radiological evidence of sepsis was greater than 94% in all treatment groups. Of the study population 65 (4.9%) died, 15 deaths were attributed to infection, pathogenic bacteria were only isolated from 10. The mortality rate for infected babies was 5.7% compared to 4.8% for those without confirmed infection. All the deaths associated with infection were due to Gram-positive bacteria. This study suggests that the practice of continuing antibiotic therapy once pretreatment cultures are known to be negative should be seriously reconsidered. It is concluded that CAZ+AMP is superior to either AG+AMP or ceftazidime monotherapy for the treatment of infection in the newborn. Further studies are required to confirm these observations in neonates with proven infection.