Bloodstream infections in allogeneic haematopoietic cell recipients from the Swiss Transplant Cohort Study: trends of causative pathogens and resistance rates

TO THE EDITOR: Bloodstream infection (BSI) is the most common infectious complication after allogeneic haematopoietic cell transplantation (HCT) affecting 20 – 60% of patients in the pre- and early post-engraftment phase [1 – 3] and later in patients acute graft-versus-host disease (aGvHD) [4]. While quinolone prophylaxis during aplasia is bene ﬁ cial to some extent, breakthrough BSIs caused by multidrug-resistant (MDR) bacteria have been reported [2, 5]. In this prospective study, we assessed the incidence and risk factors for BSIs in HCT recipients included in the Swiss Transplant Cohort Study (STCS) and the European Society for Blood and Marrow Transplantation (EBMT), the temporal dynamics of causative pathogens and resistance, and risk factors for mortality. The STCS enrolls >95% of all allogeneic HCT recipients in Switzerland [6] and the EBMT enrolls all allogeneic transplant patients in Europe [7]. From a merged dataset of the two cohorts, we included allogeneic HCT recipients aged ≥ 16 years, transplanted between September 2009 and October 2018 with signed informed consent. The study was approved by the ethics committees of all three transplant centers and the STCS scienti ﬁ c board. As per local guidelines, none of the transplant centers administered quinolone prophylaxis. Primary ﬂ uconazole prophylaxis was administered in two centers and all three initiated mold-active antifungal prophylaxis in patients with aGvHD on high-dose steroids. BSIs occurring up to 14 days before the ﬁ rst transplantation were recorded. We de ﬁ ned a BSI as all positive blood cultures with the same pathogen up to 7 days after the ﬁ rst detection. For polymicrobial


TO THE EDITOR:
Bloodstream infection (BSI) is the most common infectious complication after allogeneic haematopoietic cell transplantation (HCT) affecting 20-60% of patients in the pre-and early postengraftment phase [1][2][3] and later in patients acute graft-versushost disease (aGvHD) [4]. While quinolone prophylaxis during aplasia is beneficial to some extent, breakthrough BSIs caused by multidrug-resistant (MDR) bacteria have been reported [2,5]. In this prospective study, we assessed the incidence and risk factors for BSIs in HCT recipients included in the Swiss Transplant Cohort Study (STCS) and the European Society for Blood and Marrow Transplantation (EBMT), the temporal dynamics of causative pathogens and resistance, and risk factors for mortality. The STCS enrolls >95% of all allogeneic HCT recipients in Switzerland [6] and the EBMT enrolls all allogeneic transplant patients in Europe [7]. From a merged dataset of the two cohorts, we included allogeneic HCT recipients aged ≥16 years, transplanted between September 2009 and October 2018 with signed informed consent. The study was approved by the ethics committees of all three transplant centers and the STCS scientific board. As per local guidelines, none of the transplant centers administered quinolone prophylaxis. Primary fluconazole prophylaxis was administered in two centers and all three initiated mold-active antifungal prophylaxis in patients with aGvHD on high-dose steroids. BSIs occurring up to 14 days before the first transplantation were recorded. We defined a BSI as all positive blood cultures with the same pathogen up to 7 days after the first detection. For polymicrobial BSI, all pathogens detected within 5 days from the first positive blood culture were attributed to the same BSI. Skin commensals were considered relevant if isolated from two or more blood cultures within 48 h and the presence of clinical signs and symptoms of infection. Pathogen identification and susceptibility testing were interpreted according to the European Committee on Antimicrobial Susceptibility Testing. We compared two time periods by dividing the follow-up period in half using the date of the median number of transplantations (February 4, 2015). Neutropenia was defined as a neutrophil count <0.5 × 10 9 /L. aGvHD was graded according to standard criteria. The peri-transplantation period was defined as day -14 to day +30 after HCT, the early post-transplantation period day +31 to +100, and the late post-transplantation phase >100 days. Statistical methods are described in the Supplementary materials.
Over time, BSI frequency with Enterobacterales increased from 59 (18.3%) to 136 (29.6%) and CoNS decreased from 107 (33.2%) to 103 (22.4%), during the first to the second time period, respectively (Supplementary Table S2). No significant changes in resistance in gram-positive pathogens were noted between the two time periods (Supplementary Table S2). The decrease in gram-positive BSIs over time has also been reported by the ONKOKISS surveillance with a parallel increase in gram-negatives [10], which we could confirm in Enterobacterales. No methicillin-resistant Staphylococcus aureus and only one vancomycin-resistant E. faecium isolate were isolated. These findings support avoiding vancomycin administration as initial empirical therapy for febrile neutropenia in our setting.
In uni-and multivariable analysis, factors associated with a first BSI (Supplementary Table S3) were grade III/IV aGvHD, progression/ relapse of the hematological disease, preexisting comorbidities, and male gender. Neutrophil recovery was a protective factor for the occurrence of BSI. One-year post-transplant mortality was 26.6% (95% CI 24.3-29.06); 46 (10.2%) died within 21 days from the first BSI. The median survival time was 5.85 years (95% CI 4.2-9.1) ( Supplementary  Fig. S3) with the main cause of death being progression/relapse of the hematological condition (329; 58.6%, data not shown).
The strengths of the current study are the representative and complete data, and the analysis of the BSI according to the different HCT time periods-including a long follow-up time beyond neutropenia. Limitations are the observational nature of data, lack of individualized data on previous antibiotic exposure, and colonization with MDR pathogens.
To conclude, BSI occurred in one-third of HCT recipients, most frequently in the first 30 days post-transplantation. The observed quinolone resistance of 41.5% in Enterobacterales and 22.4% among P. aeruginosa in our study, together with resistance rates of 56 and 53% reported by transplantation centers across Europe [11] and MDR breakthrough infections [5]  Streptococcus spp.

CoNS
Other GP  the benefit of quinolone prophylaxis in similar epidemiological settings.
Considering the low rate of BSI with ESBL-producing or carbapenem-resistant pathogens (3% and <1%, respectively) among our patients, a de-escalating approach, as recommended by the ECIL-4 guidelines [12], using carbapenems for all febrile episodes might not be justified in our setting. Rapid identification of pathogens and susceptibility phenotype remain key to choose an efficient regimen at a minimal cost of selection pressure.