Clinical Treatment Options for Carbapenem-resistant Gram-negative Infections in China: a Single Centre Real-world Experience

Carbapenem-resistant gram-negative bacteria constitute a serious threat to public health worldwide. However, as a result of the complexities of clinical therapy, antibiotic options against carbapenem-resistant pathogens have not yet been fully standardized. Here, we conducted a retrospective study in 65,000 inpatients over a 2-year period that involved a total of 86 patients from whom carbapenem-resistant gram-negative bacteria were isolated. Monotherapy using trimethoprim/ sulfamethoxazole, amikacin, meropenem, and/or doxycycline in our hospital exhibited a clinical success rate of 83.3% for carbapenem-resistant Klebsiella pneumoniae, monotherapy using moxioxacin, piperacillin/tazobactam, cefepime, and/or ceftazidime for carbapenem-resistant Pseudomonas aeruginosa exhibited a clinical success rate of 77.7%, and monotherapy using cefoperazone/sulbactam or combination therapy with tigecycline and cefoperazone/sulbactam for carbapenem-resistant Acinetobacter baumannii exhibited a clinical success rate of 62.1%. Our ndings highlight the clinical strategies used in our hospital to successfully treat carbapenem-resistant gram-negative bacterial infections.


Introduction
Carbapenem-resistant gram-negative bacteria have become increasingly prevalent in recent years, and now constitute a serious threat to public health worldwide. The antibiotic-resistant phenotype of these bacteria leads to more severe clinical outcomes and an increased economic burden because of greater medical costs, longer hospital stays, the need for intensive care units, and increased mortality. In general, carbapenem-resistant gram-negative bacteria include carbapenem-resistant Klebsiella pneumoniae (CRKP), carbapenem-resistant Pseudomonas aeruginosa (CRPA), and carbapenem-resistant Acinetobacter baumannii (CRAB). CRKP were found in 3.2-7.1% of outpatients, CRPA ranged from 6.6-9.8% of patients on regular wards, and CRAB occurred in 11.5-13.4% of patients in intensive care units [1].
In clinical practice, carbapenem has been identi ed as a better choice for the systemic treatment of serious gram-negative bacterial infections. Given the high frequency of antibiotic resistance, carbapenem resistance poses a new and di cult challenge in therapeutic decision-making [2]. To date, several retrospective studies have been reported that consider treatment options for carbapenem-resistant gramnegative bacterial infections, and some guidance has been recommended for clinical practice [1,3,4]. However, as a result of the complexities of clinical infections, antibiotic therapy needs to be individualized to control carbapenem-resistant pathogens in accordance with different regions [5][6][7][8].
In this study, we focused on our hospital's own strategy for treating bacterial infections and tried to identify optimal antibiotic therapeutic strategies to improve clinical e cacy when treating carbapenemresistant pathogens.

Study setting
The study was performed at the Songjiang Hospital a liated to Shanghai Jiao Tong University School of Medicine (Preparatory Stage), the only regional medical centre in Songjiang district of Shanghai, comprising 1035 beds for a population of 2 million people. It is a tertiary facility with up to 3.7 million outpatients and emergency patients, 65,000 inpatients, and 43,000 surgical patients from 2018 to 2019.

Patients
We conducted a retrospective cohort study during the period from January 2018 to December 2019. All patients with clinically-isolated carbapenem-resistant gram-negative bacteria in our hospital were included in the study. A total of 86 consecutive patients were assayed for carbapenem-resistant gramnegative bacteria, of which 44 were infected and 42 were colonized with this organism. Patients were excluded if they tested positive for colonization.
According to the standards of the United States Centers for Diseases Control and Prevention [9], we assayed for healthcare-associated infections and colonization. Brie y, bacterial colonization was de ned as a positive culture from a patient without any clinical symptoms of infection, while bacterial infection was characterized by a positive culture from a patient with signs of in ammation. Infectious indices generally refer to body temperature, white blood cell (WBC) count, and C-reactive protein (CRP) and procalcitonin (PCT) levels. Carbapenem-resistant gram-negative bacteria isolated from patients were resistant to both meropenem and imipenem.
The studies involving human participants were reviewed and approved by the research committee of Shanghai Songjiang Clinical Medical College of Nanjing Medical University. The patients/participants provided written informed consent to participate in this study.

Evaluation of therapeutic e ciency
Global cure was a composite endpoint, which was de ned as clinical improvement and microbiological clearance by day 7. Clinical improvement was deemed as being afebrile for ≥ 48h, having less than 12,000 cells/mm 3 or a ≥25% reduction in the WBC count, and being hemodynamically stable without the need for vasopressors. Microbiological clearance referred to eradication of the original causative organism from subsequent bacterial cultures upon 7-day therapy, whereas microbiological failure referred to persistence of the original causative organism in the subsequent bacterial cultures by day 7. In the absence of follow-up bacterial culture data, patients who showed clinical improvement were de ned as displaying microbiological clearance [10].

Bacterial isolation
Clinical specimens were collected for bacterial culture on or after day 3 of hospitalization. Antimicrobial susceptibility data were obtained before patients received antimicrobial therapy for carbapenem-resistant gram-negative bacteria.
Multiple drug resistant (MDR) refers to the absence of susceptibility to three or more antimicrobial categories. Extensively drug resistant (XDR) refers to the absence of susceptibility to all but two or fewer antimicrobial categories. Pan-drug resistant (PDR) refers to the absence of susceptibility to all antimicrobial categories.

Results
Clinical therapy and the outcome of CRKP infection Among 12 patients for whom CRKP was detected, six were de ned as colonization and six were de ned as infection. Of note, the infectious indices, which included the WBC count, CRP and PCT levels, from six patients with CRKP were signi cantly elevated (Supplementary Table 1). The in vitro antibiotic susceptibility data against CRKP infection were summarized in Table 1 and Supplementary Table 2. CRKP was resistant to meropenem and imipenem, but highly susceptible totigecycline, amikacin, and trimethoprim/sulfamethoxazole, with sensitivity rates of 100%, 83.3%, and 50.0%, respectively. Consistent with our antibiotic susceptibility pro le, antibiotic monotherapy in our hospital showed a success rate of 83.3% (5/6) for eliminating CRKP, only one case of a urinary tract infection failed treatment with moxi oxacin ( Table 2). Five patients were successfully treated with monotherapy, two (one urinary tract and one lung infection) with trimethoprim/sulfamethoxazole, one (lung infection) with doxycycline, one (urinary tract infection) with amikacin, and one (lung infection) with meropenem, which was indicative of a positive clinical outcome. Interestingly, meropenem resistance was established for CRKP in vitro, but it showed a better clinical outcome for patients. Taken together, monotherapy with trimethoprim/sulfamethoxazole, amikacin, meropenem and/or doxycycline served as the effective therapeutic strategy for CRKP in our hospital.
Clinical therapy and the outcome of CRPA infection Among 13 patients for whom CRPA was detected, four were de ned as colonization and nine were de ned as infection. As presented in Supplementary Table 3, the infectious indices for nine patients with CRPA were signi cantly elevated. To evaluate the success of antibiotic therapy, we determined the susceptibility rates of these bacteria to antibiotics in vitro (Table 3 and Supplementary Table 4). CRPA showed resistance to meropenem and imipenem, but was relatively sensitive to cefepime, piperacillin/tazobactam, and ceftazidimeat rates of 88.9%, 66.7%, and 66.7%, respectively.
The clinical outcomes exhibited a high success rate (7/9, 77.7%) for eliminating CRPA (Table 4). Of the two cases with lung infection for whom treatment failed, one patient was treated with ceftazidime monotherapy and one patient was treated with moxi oxacin and meropenem combination therapy. As a result of their transfer to another hospital, they did not receive the alternative antimicrobial therapy. Among the seven cases that were successfully treated, ve patients with a lung infection received therapy with either moxi oxacin (n = 2), piperacillin/tazobactam (n = 2), or ceftazidime (n = 1), and one patient with a urinary tract infection received cefepime monotherapy and one patient with a blood infection received moxi oxacin monotherapy. Of note, one patient experienced a signi cant improvement in respiratory symptoms, with initial failure to moxisaxin therapy, followed by ceftazidime administration. Taken together, antibiotic monotherapy with moxi oxacin, piperacillin/tazobactam, cefepime, and ceftazidime exhibited good clinical e cacy against CRPA in accordance with our antibiotic susceptibility test.

Clinical therapy and the outcome of CRAB infection
Among 61 patients for whom CRAB was detected, 32 were de ned as colonization and 29 were de ned as infection. As presented in Supplementary Table 5, the infectious indices for 29 patients with CRAB were signi cantly elevated. As presented in Table 5 and Supplementary Table 6, CRAB exhibited resistance to meropenem and imipenem, but was relatively sensitive to tigecycline (89.6%), cefoperazone/sulbactam (55.1%), and piperacillin/tazobactam (33.5%).
The clinical outcomes exhibited a relatively high success rate (18/29, 62.1%) for eliminating CRAB infection upon monotherapy or combination therapy (Table 6)

Discussion
Our aim was to identify treatment regimens for infections caused by carbapenem-resistant gram-negative bacteria in our hospital. In this study, a total of 44 patients who infected with carbapenem-resistant gramnegative were found in 65,000 inpatients over a 2-year period, indicative of a very low infectious rate.
Furthermore, monotherapy (trimethoprim/sulfamethoxazole, amikacin, meropenem, and doxycycline) was an effective antibiotic regimen for the treatment of CRKP infection; monotherapy of moxi oxacin, piperacillin/tazobactam, cefepime, and ceftazidime was the preferred antibiotic choice for CRPA infection; and cefoperazone/sulbactam monotherapy or tigecycline combination therapy could effectively treat CRAB infections.
Carbapenem is the treatment option for patients who develop serious infections with MDR, XDR, and PDR gram-negative bacilli, but unfortunately the resistance to carbapenem signi cantly compromises this treatment choice [11]. Consequently, carbapenem resistance among gram-negative bacteria is disseminating worldwide, which poses a serious threat to current medical practices. Similar to previous reports [12,13], CRAB isolates in our study were XDR strains, showing susceptibility to only one and/or two antibiotics in vitro. Unlike previous reports [12,13], CRKP and CRPA isolates in our study were MDR, indicative of a better antibiotic choice in vitro.
In fact, the antibiotic susceptibility determined by in vitro culture does not always correlate with the success of clinical therapy in vivo [11]. Exact antibiotic doses and bacterial inocula can be easily assessed on agar plates, but this may not be replicated in patients. As a result, despite many in vitro studies of bacterial infection, the ndings do not always translate into successful treatment in clinical studies. In this study, we observed that the implementation of antibiotic regimens in CRKP and CRPAinfected patients were consistent with in vitro activity. However, a minor difference was found in CRABinfected patients. Of the 29 patients with CRAB, resistance to empirical antibiotics piperacillin/tazobactam (2/2, 100%) and imipenem/cilastatin (1/1, 100%) was shown in vitro but patients effectively recovered, and in 20 cases resistance to cefoperazone/sulbactam was shown but clinical e cacy was evident for most patients. This discrepancy maybe attributed to the different environments in vivo and in vitro, the minimum inhibitory concentrations, as well as the clinical e cacy of the cefoperazone/sulbactam combination [14,15].
Current guidance for the treatment of carbapenem-resistant gram-negative bacterial infections relies on antibiotic therapy based on polymyxins (including colistin or polymyxin B), aminoglycosides, and tigecycline [16,17]. Emerging retrospective clinical studies show that antibiotic options have not yet been fully standardized, and that treatments may need to be individualized to control carbapenem-resistant pathogens depending on different regions [1]. Papst et al. showed that in Israel, monotherapy was the preferred choice for treating CRPA and CRAB infections, whereas in all other countries, combination therapy with two drugs was standard [11]. Ceftazidime/avibactam was commonly used for the treatment of CRKP-infected patients in the USA, whereas ceftolozane/tazobactam was often used in Spain, Italy, France, and the USA [11]. Chen et al. reported that both monotherapy of colistin and combination therapy of colistin and carbapenem showed high cure rates of CRAB and CRKP [18]. Our study showed that our hospital's antibiotic treatment strategies, which included monotherapy of trimethoprim/sulfamethoxazole, amikacin, meropenem, and/or doxycycline therapy for CRKP, monotherapy of moxi oxacin, piperacillin/tazobactam, cefepime, and/or ceftazidime for CRKP, and monotherapy of cefoperazone/sulbactam or combination therapy with cefoperazone/sulbactam and tigecycline for CRAB, according to standard guidance, led to a high cure rate.
In summary, our study collates the successful treatment regimens for infections caused by carbapenemresistant gram-negative bacteria in our hospital. In the future, it will be of great interest to elucidate the molecular mechanisms of carbapenem resistance. It will also be important to further identify optimal antibiotic treatment strategies through a multi-centre study involving a greater number of patients in Shanghai.