RESUMO
OBJECTIVES: Our objective was to systematically study the influence of length of hospital stay on bacterial resistance in relevant respiratory tract isolates. METHODS: Using prospective epidemiological data from the National Swiss Antibiotic Resistance Surveillance System, susceptibility testing results for respiratory isolates retrospectively retrieved from patients hospitalised between 2008 and 2014 were compiled. Generalized additive models were used to illustrate resistance rates relative to hospitalisation duration and to adjust for co-variables. RESULTS: In all, 19 622 isolates of six relevant and predominant species were included. Resistance patterns for the predominant species showed a species-specific and antibiotic-resistance-specific profile in function of hospitalisation duration. The oxacillin resistance profile in Staphylococcus aureus isolates was constantly increasing (monophasic). The pattern of resistance to cefepime in Pseudomonas aeruginosa was biphasic with a decreasing resistance rate for the first 5 days of hospitalisation and an increase for days 6-30. A different biphasic pattern occurred in Escherichia coli regarding amoxicillin-clavulanic acid resistance: odds/day increased for the first 7 days of hospitalisation and then remained stable for days 8-30. In the adjusted models epidemiological characteristics such as age, ward type, hospital type and linguistic region were identified as relevant co-variables for the resistance rates. The contribution of these confounders was specific to the individual species/antibiotic resistance models. CONCLUSIONS: Resistance rates do not follow a dichotomic pattern (early versus late nosocomial) as suggested by current hospital-acquired pneumonia treatment guidelines. Duration of hospitalisation rather appears to have a more complex and non-linear relationship with bacterial resistance in hospital-acquired pneumonia, also depending on host and environmental factors.