With the continuous improvement of technology and the expansion of donor options, hematopoietic stem cell transplantation (HSCT) has become an important, even the best treatment for hematological malignancies, some solid tumors, inherited metabolic disorders and immune diseases (Khaddour et al., 2022). However, HSCT recipients are particularly vulnerable to bacterial infection due to mucosal injury, immunosuppression, prolonged hospitalization, invasive procedures and frequent exposure to antibiotics. In recent years, most regions of the world have shown a trend of epidemiological change from gram-positive bacteria to gram-negative bacteria (GNB) (Omrani and Almaghrabi, 2017). It has been reported that approximately 15% to 74% of bacteremia in HSCT recipients is caused by GNB (Mikulska et al., 2014). In addition, several studies have confirmed that GNB infection is strongly associated with poorer prognosis. For example, Poutsiaka DD et al. reported that the mortality of HSCT recipients experiencing GNB bloodstream infection was about 59% (Poutsiaka et al., 2007). Ortega M et al. and Mikulska M et al. also found that the mortality of HSCT recipients on the 7th and 30th day after GNB bacteremia was 17%-22% and 24%-31%, respectively (Ortega et al., 2005; Mikulska et al., 2012). What’s worse, in the past two decades, the resistance rate of GNB to antibiotics has been increasing, and the emergence of multi-drug resistant (MDR), extensively-drug resistant (XDR), and pandrug resistant (PDR) bacteria has also become more and more common (Sahitya et al., 2021). Carbapenems were once considered as the last line of defense against GNB, whereas the rising drug resistance rate poses a huge challenge to researchers. Carbapenem resistant gram-negative bacteria (CRGNB) is listed as a public health threat prioritized by the Centers for Disease Control and Prevention of the United States and World Health Organization (Babiker et al., 2021).

At present, CRGNB infection has become one of the most common complications and causes of death for HSCT recipients, and the economic burden and mental stress brought by it have a negative impact on patients’ families and even the whole society. Although a number of epidemiological and clinical data related to infection in this special population have been published previously, most have focused on carbapenem-resistant Enterobacteria (CRE). However, some studies demonstrated that infections caused by other CRGNBs, such as Pseudomonas aeruginosa and Acinetobacter, are growing at an extremely fast speed, even surpassing CRE as the most dominant CRGNB isolates among transplant patients in some regions (Babiker et al., 2021).

This study analyzed the composition, distribution, drug resistance and prognosis of CRGNB infection in HSCT recipients in the hematological ward of a large general hospital in Central-south China in the past seven years, and paid special attention to the risk factors related to CRGNB acquisition and death. It provides valuable information for comprehensively understanding the characteristics of post-transplant infection, adjusting anti-infection treatment strategies and improving the prognosis of patients.

Although global attention is now focused on COVID-19, carbapenem resistance remains a serious threat to international public health, and can even be considered an “underestimated silent pandemic”. Infections caused by GNB, particularly CRGNB, are becoming increasingly problematic in HSCT recipients (Satlin et al., 2014; Su et al., 2021).

In this work, we revealed the crude infection rate of CRGNB in HSCT recipients was 7.4% (87/1173). It is well-known that geographical differences in GNB resistance are striking. In some researches, including HSCT patients, the incidence of CRGNB ranged from 0% in Sweden (Blennow et al., 2014) to 4.7-5.8% in the United States (Satlin et al., 2016; See et al., 2016) and 20.9% in Italy (Trecarichi et al., 2015). In general, carbapenem resistance in Asia was more serious than that in Europe and the United States. Further analysis showed that lung was the most common site of infection, followed by bloodstream. Here, the incidence of CRGNB bacteremia was 2.90% (34/1173), lower than the 8.10% reported by Ding LM et al. (Ding et al., 2021). This may be related to patients’ refusal to repeatedly draw blood due to personal factors such as illness, financial or psychological pressure, which leads to limited blood culture times and low positive rates. Unlike the invasive operation of blood collection, the method of obtaining sputum culture samples was relatively easy to be accepted, so the detection rate of pulmonary infection in our data was higher. The mortality of CRGNB infection calculated here was 47.1%, much higher than that of non-CRGNB infection group (25.3%), and this difference was similar to previous reports (Averbuch et al., 2017; Vinker-Shuster et al., 2019).

As for antibiotic sensitivity, our data indicated that the resistance rates of CRGNB infected by HSCT recipients to commonly used quinolones, cephalosporins, sulfonamides and piperacillin-tazobactam were all over 50%, only relatively sensitive to some aminoglycosides and TGC, but the situation was not optimistic either. Studies have suggested that TGC has good in vitro activity against CRGNB strains, with a sensitivity to carbapenem resistant Klebsiella pneumoniae and Escherichia coli of 94.2% and 100%, respectively. Moreover, it is also effective for most ventilator-associated pneumonia caused by CRAB. However, a meta-analysis, which included 13 randomized clinical trials, showed that TGC was positively associated with increased mortality (Chen et al., 2020). In particular, TGC is a bacteriostatic agent that cannot directly kill bacteria, which severely limits its therapeutic reactivity in immunosuppressed hosts (Satlin et al., 2014). Fortunately, the introduction of some new antibiotics in recent years has greatly improved the therapeutic effect of CRGNB infection. For example, Ceftolozane/tazobactam (C/T) was proved to have good activity against CRPA. In addition, compared with polymyxins and aminoglycosides, it can bring higher clinical cure rate and less acute kidney injury (Pogue et al., 2020). But even so, the activity of C/T against Acinetobacter and Stenotrophomonas maltophilia is still very limited (Yahav et al., 2020). Ceftazidime avibactam (CTV) is a β-lactam/β-lactamase inhibitor combination. At present, many investigations have confirmed that CTV has excellent performance on CRGNB, especially CRKP, infected in the population after HSCT and solid organ transplantation (Chen et al., 2021; Clerici et al., 2021). However, in an in vitro experiment, researchers found that CRGNB isolated from patients’ blood and stool samples was resistant to CTV at a rate of 91% (Niyazi et al., 2022). Furthermore, CTV has not yet shown efficacy on CRPA and CRAB (Chalhoub et al., 2018; Sanz-García et al., 2018).

Meropenem-vaborbactam (MEV) is a combination of meropenem and vaborbactam, a new generation β-lactamase inhibitor. Alosaimi S et al. pointed out in a multicenter study from the real world that MEV has positive clinical results and safety in GNB infections, including CRE and CRPA (Alosaimy et al., 2021).

Imipenem/cilastatin/relebactam (Recarbrio™) is an intravenously administered combination of the carbapenem imipenem, the renal dehydropeptidase-I inhibitor cilastatin, and the novel β-lactamase inhibitor relebactam. It has been approved for the anti-infective treatment of CRE and CRPA (Heo, 2021). However, the addition of relebactam does not enhance the sensitivity of imipenem to Acinetobacter baumannii and Stenotrophomonas maltophilia (Principe et al., 2022). Cefiderocol is a novel siderophore cephalosporin with potent in vitro activity against gram-negative pathogens including MDR Enterobacteriaceae and non-fermenting organisms, such as Pseudomonas aeruginosa, Acinetobacter baumannii, Burkholderia cepacia, and Stenotrophomonas maltophilia (Abdul-Mutakabbir et al., 2020).

In addition to the above-mentioned “new antibiotics”, some relatively “ancient” preparations have also entered people’s vision again. For example, colistin, as an antibacterial agent started to be used in the 1950s, is active against a variety of GNB such as Acinetobacter baumannii, Pseudomonas aeruginosa, Escherichia coli and Klebsiella pneumoniae. Recently, because of the emergence of “superbacteria” that are resistant to other antibiotics, colistin has been revived as a salvage treatment for MDR and XDR GNB infections (El-Sayed Ahmed et al., 2020). It should be noted that on account of the narrow therapeutic window, the nephrotoxicity and neurotoxicity caused by colistin still need special attention. Interestingly, a report from Pakistan in 2021 demonstrated that even in patients with impaired renal function, colistin can effectively eliminate bacteremia, with minimal and reversible adverse reactions (Babar et al., 2021). Fosfomycin has high activity to Enterobacteriaceae, but it is ineffective to non-fermentative bacteria due to inherent resistance (Samonis et al., 2010), and its sensitivity to anaerobic bacteria is also very limited (Boyanova, 2015). Whereas, in some studies, fosfomycin has been shown to reduce the MIC of bacteria to other antibiotics, thereby restoring their sensitivity to non-fermenting bacteria (Mikhail et al., 2019). Unfortunately, owing to the actual factors such as the time when drugs are approved for marketing, drug prices, medical insurance system, and patients’ economic affordability, many antibiotics have not been widely used in our patients. At the same time, some corresponding drug sensitivity tests were not carried out, resulting in incomplete data that could not be included in subsequent statistical analysis.

Not surprisingly, the mortality of HSCT recipients in the CRGNB infected group was significantly higher than that in the matched non-CRGNB group (47.1% vs. 25.3%). Further multivariate analysis revealed that the interval from diagnosis to transplantation > 180 days, septic shock, low platelet count and high bilirubin level at the initial stage of infection were independent risk factors for death. In previous reports, we have described the adverse effects of extended transplant interval: persistent disease state, multiple chemotherapy treatments, low immunity, poor body function, delayed donor cells implantation, etc (Liu et al., 2022). This is why both domestic and international guidelines unanimously recommend early transplantation for eligible patients (Blackburn et al., 2019). However, the fact is, due to age, personal financial situation, family support, medical environment, donor factors, etc., there are still many patients cannot receive HSCT within the optimal time range, which affects the prognosis to some extent.

The conclusion that septic shock was an independent risk factor associated with mortality in CRGNB patients is similar to previous reports (Qiao et al., 2017; Aguado et al., 2018). This seems to be expected, since septic shock is one of the most critical indicators of disease severity. This finding emphasizes the importance of early identification and treatment, which may improve survival for patients with septic shock and the resulting respiratory and heart failure (Wu et al., 2021).

The primary risk of extremely low platelet count at the initial stage of infection is bleeding, especially in important organs and intracranial hemorrhage. Meanwhile, it is also a side warning of patients’ insufficient hematopoietic recovery ability. Infection, severe bleeding, and delayed hematopoietic reconstruction undoubtedly bring great obstacles to the rehabilitation of patients after transplantation.

A high level of serum total bilirubin means poor liver function, which may be caused by underlying liver-related diseases (such as viral hepatitis), or drug-induced liver damage. But either way, it will lay a “time bomb” for the future development of hepatic veno-occlusive disease (HVOD) or liver rejection (Yoon et al., 2021). HVOD is one of the most serious complications after HSCT, and the mortality of severe patients can be up to 50% (Bajwa et al., 2017). The main clinical manifestations include fluid retention, ascites, weight gain, jaundice, hepatomegaly and pain, and even multiple organ failure. Researches learned that the risk factors of HVOD involve multiple aspects, such as patient’s primary disease, age, complications, increased serum ferritin, acute renal failure, transplantation type, pretreatment regimen, and anti-GVHD drugs (Corbacioglu et al., 2019). Therefore, avoiding hepatotoxic drugs as much as possible and strengthening liver function protection to reduce the occurrence of HVOD have a positive effect on improving the survival rate of patients after transplantation. Currently, ursodeoxycholic acid (UDCA) and defibrotide (DF) have been recommended for the prevention of HVOD (Carreras, 2012; Dignan et al., 2013). DF is also approved for the treatment of HVOD by the United States Food and Drug Administration (FDA) and the European Union (Corbacioglu et al., 2019). Although some other drugs such as prostaglandin E1 (PGE1) have also been reported to have therapeutic effects on HVOD, their exact efficacy still needs to be further verified by prospective clinical studies (Dignan et al., 2013).

Admittedly, there are some limitations to our study. First, due to the retrospective nature, we cannot perform molecular analyses to identify the types of carbapenemases produced by CRGNB organisms. Second, this is a single-center cohort study, so the results may not be popularized to other regions. Third, as previously mentioned, clinical data and efficacy assessments for some novel antibiotics are not yet available. However, this exactly reflects the actual problems faced by different regions of the world in the utilization of medical resources at this stage. By reviewing a large number of literatures, we summarized the relevant information on the use of antibiotics against CRGNB, hoping to provide reference for medical staff in the future treatment options.

In short, HSCT recipients are prone to CRGNB infection, and the drug resistance rate and mortality rate are extremely high. Interval from diagnosis to transplantation >180 days, septic shock, platelet count < 20 × 10⁹/L and total bilirubin > 34.2 μmol/L at the initial stage of infection were 4 independent risk factors associated with death. Shortening the transitional waiting time of transplant population, strengthening liver function screening and protection, appropriate platelet transfusion to reduce bleeding risk, and early identification and timely correction of shock will help improve the prognosis of patients.

The original contributions presented in the study are included in the article/supplementary material. Further inquiries can be directed to the corresponding author.

YJ and YFL were involved in the study conception and design. YunL, YiL, ZY and XC participated in the data management. All authors participated in the interpretation and statistical analysis of data. YJ wrote the draft and YFL revised it. All authors contributed to the article and approved the submitted version.