Pre-Transplant Colonization By a Multidrug-Resistant Gram Negative Bacteria Has No Impact on Overall Survival and Mortality after Hematopoietic Stem Cell Transplantation: A Single-Center Experience in 362 Patients

Multidrug resistant Gram-negative (MDR-GN) bacteria are responsible for severe infections in immune-compromised hosts. Clinical outcome is often poor, especially in patients undergoing hematopoietic stem cell transplantation (HSCT) with mortality rates exceeding 50%. In patients with a documented pre-transplant MDR-GN bacterial infection, HSCT is usually regarded as a potentially harmful procedure, despite its curative potential. Recently, active surveillance strategies based on rectal swab evaluation have been recommended with the aim of early identification of MDR-GN carriers, although they are not routinely performed in every transplant center.

After an Institutional outbreak of K. pneumoniae carbapenemase -producing (KPC-Kp) infections in our Centre in June 2012, we decided to start an Institutional programme of active surveillance. No patient was excluded from HSCT, despite a documented colonization. In this study, we retrospectively analysed the results of the active surveillance programme and the impact of a pre-transplant colonization on HSCT outcomes.

From July 2012 to January 2016 (43 months), 362 consecutive patients (254 undergoing allogeneic HSCT (allo-HSCT) and 108 autologous HSCT (auto-HSCT) were examined. All patients had a pre-transplant rectal swab evaluable. Blood cultures were performed at each febrile episode. Neutropenic patients developing fever were treated according to internal antimicrobial guidelines, usually with multiple targeted antimicrobial therapy in case of a previously known MDR-GN bacterial colonization. Median follow-up was of 602 days.

Pre-transplant carriers were 8% and 17% in auto-HSCT and allo-HSCT recipients, respectively. Both in auto-HSCT and allo-HSCT, 30% of carriers were colonized by KPC-Kp. Among auto-HSCT, the overall survival (OS) did not significantly differ in carriers versus non-carriers (86% vs 84% at 2 years; P=0.231). Moreover, neither transplant-related mortality (TRM) nor infection-related mortality (IRM) were significantly different among the two groups, being of 0% vs 4% at 100 days (P=0.32) and of 0% vs 2% at 100 days (P= 0.099), respectively. In allo-HSCT, the OS was not significantly influenced by the carrier status nor univariate (43% vs 50% at 2 years; P=0.091) or in multivariate analysis (HR 1.41; P=0.129). Furthermore, TRM and IRM were comparable in MDR-GN bacteria carriers versus non-carriers, being 31% vs 25% at 2 years (P=0.301) and 23% vs 14% at 2 years (P=0.304). A MDR-GN bacterial colonization was not an independent risk factor for TRM and IRM, with an hazard ratio (HRs) of 1.428 (P=0.268) and of 1.69 (P=0.215). Furthermore, we did not observe any significant variation in the incidence of severe acute GvHD when comparing carriers and non-carriers (HR 1.86; P=0.087). In multivariate analysis severe acute GvHD was the only independent factor for a worse OS, TRM and IRM ( HRs of 3.177 (P< 0.001), 7,933 (P< 0.001) and 18,690 (P < 0.001), respectively.

We registered a total of 168 Gram-negative bloodstream infections (BSIs) and 41/168 (24%) were due to a MDR bacteria, of which 6/41 in auto-HSCT patients and 35/41 in allo-HSCT recipients. Extended spectrum β-lactamases (ESBL+) E. coli (39%) and KPC-Kp (24%) were most frequently isolated from blood cultures.

Thanks to the early initiation of a combined antimicrobial therapy guided by the antimicrobial susceptibly test deriving from rectal swabs in patients developing neutropenic fever, we witnessed a significant reduction in 30-day MDR-GN bacteria-related mortality (0% for auto-HSCT and 17% for allo-HSCT patients). Noticeably, in our Center 30-day KPC-Kp-related mortality was only 30%.

Our retrospective single-centre study was able to demonstrate that, both in auto-HSCT and allo-HSCT, a pre-transplant colonization status is not significantly associated with a reduced OS, nor with a reduced TRM and IRM. Moreover, the prompt initiation of an adequate empirical therapy based on information coming from rectal swab is able to reduce MDR-GN bacteria-related mortality and to improve the overall HSCT outcome.