Guidance of empirical antimicrobial therapy by surveillance cultures in high-risk neutropenic patients: a retrospective cohort study

Study design and setting

This retrospective cohort study was conducted in two university medical centres that had different EAT regimes for febrile episodes in high-risk neutropenic patients, in Amsterdam, the Netherlands,: the VU University Medical Centre, a 730-bed tertiary care centre (referred to as meropenem-treating centre (MTC)), and the Academic Medical Centre a 1002-bed tertiary care centre (referred to as ceftazidime-treating centre (CTC)). The different antibiotic policies in the centres are described in Additional file 1: Table S1.

Patients and data collection

We screened all neutropenic patients > 18 year, who received antibiotic prophylaxis, admitted to the haematology wards during the following period: march 2016 to april 2019 in MTC and October 2015 to december 2019 in CTC. All patients who received chemotherapy for a haematological malignancy resulting in protracted neutropenia (< 500 neutrophils/μL for > 7 days), further referred to as high-risk neutropenic patients, with at least one febrile episode were included. We excluded all patient who were transferred to another hospital during treatment. We also excluded all patients that opted out of chemotherapy before neutropenia occurred and all patients in which fever did not occur. From the included patients we gathered the following variables: age, gender, underlying hematologic malignancy, type and start date of the chemotherapeutic course, neutrophil count, white blood cell count and differential, presence of peripheral of central venous catheter, total parenteral nutrition, ICU transfer and mortality.

Definitions

A febrile episode was defined as an episode during high-risk neutropenia, in which diagnostic blood cultures were drawn due to clinical signs and symptoms of infection (fever, hypotension and/or chills) and in which empirical antibiotic therapy was initiated within 48 h (Fig. 1). Fever was defined as a temperature of ≥ 38.3 °C measured once, or ≥ 38.0 °C measured multiple times during one hour.

Fig. 1figure 1

Schematic overview FN episode.

Subsequent febrile episodes could be included when the systemic antibiotics had been halted and prophylactic antibiotics were restarted for at least 24 h. In case of clinical signs and symptoms of infection (fever, hypotension and/or chills) at least two blood cultures were drawn according to local protocols in both centres. All blood cultures drawn at onset of fever were included in the analysis. Blood cultures drawn after ≥ 24 h of empirical treatment were excluded from the analysis to ensure selection of the EAT, as treatment is often optimized thereafter based on culture results or clinical course of the infection. We included BSI with coagulase-negative staphylococci (CNS) according to the CDC criteria for laboratory confirmed bloodstream infection with common commensals, which means that CNS should be identified by a culture from two or more blood specimens collected on separate occasions [11]. IEAT was defined as EAT that did not include at least one in vitro active antibiotic against the isolated microorganism in the BSI. In case of infections with vancomycin susceptible CNS and enterococci species, administration of vancomycin within 48 h after onset of fever was regarded as adequate treatment (as per protocol in both centres). The duration of neutropenia was calculated by counting days from the last day before neutrophil levels dropped < 500/μL until neutrophil recovery (≥ 500 neutrophils/μL). When patients died during neutropenia the end of the neutropenic episode was the date of death. In the absence of an absolute neutrophil count, the white blood cell and differential were used to define neutropenic period.

Microbiological procedures

Of the included patients, all surveillance and blood-isolates gathered during the study period were extracted. In high-risk neutropenic patients from both centres, pharyngeal and rectal samples were obtained weekly, beginning one week before the start of chemotherapy until neutrophil recovery (> 500 neutrophils). In pharyngeal and rectal samples growth and susceptibility of Pseudomonas aeruginosa and aerobic gram-negative bacilli were reported. In rectal swabs drawn in CTC, growth and susceptibility of gram-negative bacilli were only reported in pathogens resistant to either cotrimoxazole, ciprofloxacin or both. In MTC, growth and susceptibility of Candida spp. were also reported, whereas in the CTC growth and names of yeast and susceptibility testing was only performed when requested by the treating physician or the clinical microbiologist.

For microorganisms without EUCAST breakpoints, or isolates without MIC values, the S/I/R interpretation as reported by the laboratory was used, if available. In this study 3GC-R was defined as: P. aeruginosa and/or Enterobacterales group I non-susceptible to ceftazidime according to EUCAST breakpoints or I + R interpretation as reported [12]. Enterobacterales group II (Enterobacter cloacae complex, Klebsiella aerogenes (formerly E. aerogenes), Citrobacter freundii complex, Hafebrileia alvei, Serratia spp., Providencia spp., Morganella morganii) were considered 3GC-R independent of MIC values due to inducible chromosomal AmpC β-lactamases [13, 14]. Extended-spectrum beta-lactamase (ESBL) production was confirmed by combination disc diffusion test with both cefotaxime and ceftazidime, with and without clavulanic acid (Rosco, Taastrup, Denmark), interpreted according to the Dutch national guideline[13]. Gram negative bacteria (GNB) were categorized as ‘carbapenem-resistant’ if reported intermediately susceptible or resistant (I + R) to meropenem or imipenem, and otherwise as ´non-carbapenem-resistant´ (i.e. carbapenem-susceptible or not reported, according to local protocols).

3GC-R colonization was defined as the presence of a 3GC-R in rectal or pharyngeal swabs obtained for surveillance purposes. Febrile episodes within 3–365 days after the retrieval of a 3GC-R in surveillance cultures were regarded as febrile episodes preceded by 3GC-R colonization. The time interval of 365 days was based upon ESBL-carriage in travellers in whom most (88.7%) decolonized within one year [15]. This relationship was investigated independent of the micro-organism involved. In other words, a 3GC-R was considered predictive of any 3GC-R microorganism, i.e. E. coli could predict a 3GC-R K. pneumoniae. When multiple isolates were available, the shortest interval between 3GC-R positive surveillance cultures and 3GC-R BSI was used to calculate the median time between (the latest known) colonization and BSI.

Data-analysis and statistics

Continuous variables were presented as mean ± standard deviation if normally distributed, or as the median and interquartile range if skewed. Categorical variables were presented as proportions and/or counts. Sensitivity, specificity, positive (PPV) and negative predictive values (NPV) were calculated using standard formulas [16]. We performed three additional sensitivity analysis. In the first, we only included surveillance cultures positive for 3GC-R with co-resistance to either cotrimoxazole or ciprofloxacine, resembling the culture methods in CTC and identifying resistance against the standard prophylactic regimens. In the second, we only included ESBL-producing Enterobacterales (ESBL-E) to confirm that the relationship between colonization and BSI was independent of our definition of 3GC-R. In the third, we shortened the time interval between surveillance culture and febrile episode to one month. We included multiple dependent febrile episode per patients, therefore for the analyses of data we used Generalized Estimating Equations (GEE) with an independent correlation structure. In our GEE analyses we estimated the correlation between BSI, 3GC-R-EP BSI and IEAT and the dichotomous clinical outcome: ICU transfer/all-cause mortality within 30 days after onset of fever. We investigated confounding by age, sex and treatment centre. The potential reduction of carbapenem use was calculated as the proportion of febrile episodes not preceded by 3GC-R colonization therefore not necessitating empirical treatment with carbapenems in a surveillance culture guided EAT approach, compared to standard EAT with carbapenem. Data analysis was performed using R Statistical Software (version 4.0.3; R Foundation for Statistical Computing, Vienna, Austria).

留言 (0)

沒有登入
gif