Authors (including presenting author) :
Tsang WP (1), Chan KH (1)
Affiliation :
(1) Intensive Care Unit, Tseung Kwan O Hospital, Kowloon East Cluster
Introduction :
Severe community-acquired pneumonia (CAP) causes substantial mortality in ICU. While broad-spectrum antibiotics (e.g. amoxicillin-clavulanate) are first-line empirical therapy in guidelines, extended-spectrum antibiotics (e.g. piperacillin-tazobactam, carbapenems, vancomycin) are increasingly used for drug-resistant pathogens (DRPs) (e.g. Pseudomonas aeruginosa, ESBL+ Enterobacteriaceae, MRSA). To reduce emergence of antibiotic resistance, empirical extended-spectrum antibiotics were recommended by guidelines only in the presence of risk factors for DRPs. However, they are liberally given in practice, even when risk factors are insufficiently present, for fear that DRPs are causing severe presentation and poor outcome. Local studies are lacking.
Objectives :
To compare the clinical outcome associated with broad- and extended-spectrum antibiotics, as empirical treatment in culture-positive, severe CAP in ICU
Methodology :
It was a retrospective cohort study (n=250) of culture-positive, severe CAP patients from ICUs of Tseung Kwan O Hospital and Pamela Youde Eastern Nethersode Hospital. Patients were classified into either broad- (34.8%) or extended-spectrum (68.2%) empirical antibiotic groups, with clinical outcome compared, and adjusted with age, sex, coverage of causative pathogens, organ support and APACHEII score. Subgroup analysis was performed for hospitals, patients with and without DRPs.
Result & Outcome :
DRPs did not differ in frequency significantly between the groups (broad: 43.2%; broad: 36.7%; p=0.277), though a more severe CAP (a higher APACHEII score, use of inotrope and renal replacement therapy), was found in the extended-spectrum antibiotic group. This group achieved a higher coverage of causative pathogens (broad: 51.7%; extended: 82.2%; p< 0.001). It had an almost universal (98.8%) use of anti-pseudomonal antibiotics.
On regression, extended-spectrum antibiotics were associated with decreased ventilator days in general (adjusted incidence rate ratio [aIRR] 0.845, 95% CI [0.768, 0.93], p< 0.001), but increased ventilator days in the non-DRP subgroup (aIRR 1.19, [1.044, 1.356], p=0.009). They were associated with slightly longer ICU stay in general (aIRR 1.176, [1.08, 1.282], p< 0.001), in both hospital subgroups and the non-DRP subgroup (aIRRs around 1.2, p< 0.001 to p=0.003), but no significant increase in the DRP subgroup. No significant change was found with 30- or 60-day mortality, in general or in subgroups.
Conclusion: Despite fewer ventilator days, empirical extended-spectrum antibiotics were associated with slightly longer ICU stay, with no benefit in mortality. They should be used judiciously.
