Carbapenem-resistant Acinetobacter baumannii (CRAB) is at the top of the World Health Organization’s list of drug-resistant bacteria and poses a significant threat to human health (Willyard, 2017). It can cause severe infections including pneumonia, bloodstream infection, abdominal infection, and skin and soft tissue infection, for which almost no treatment is available (Willyard, 2017). Data from the China Antimicrobial Surveillance Network (CHINET) has shown that the percentage of imipenem-resistant Acinetobacter baumannii (AB) increased from 39.0% in 2005 to 78.1% in 2018 (CHINET, 2019). Infections caused by CRAB have high mortality due to limited antimicrobial options, and CRAB bloodstream infection tends to occur in critically ill patients along with rapid disease progression and higher mortality. A previous study showed that the 30-day mortality of patients with CRAB bacteremia in the intensive care unit (ICU) was as high as 79.8% (Kim et al., 2012). The new β-lactamase inhibitor ceftazidime/avibactam presents an opportunity to improve the survival of patients with carbapenem-resistant Pseudomonas aeruginosa and Enterobacteriaceae infections, but the susceptibility of CRAB to ceftazidime/avibactam was poor (Hsueh et al., 2019). Several new antimicrobial agents, including aztreonam/avibactam, cefiderocol, and eravacycline, might be promising therapeutic options in treating CRAB infections (Leone et al., 2019). Currently, the treatment options for CRAB infections are limited, including colistin, tigecycline, sulbactam, and aminoglycosides (Karaiskos et al., 2017). Therefore, knowing which of the existing antimicrobials to select in controlling CRAB infections is a huge challenge for clinicians. Observational studies have shown no difference in statistics between monotherapy and combination therapy in mortality and clinical efficacy in patients with CRAB infections (Karaiskos et al., 2017). There are also no convincing studies that recommend the combination of carbapenems, colistin, or sulbactam (Karaiskos et al., 2017).

The resistance mechanisms of CRAB to carbapenem mainly include the production of carbapenemases (especially class D carbapenemases) in AB, excessive expression of an active efflux pump, functional loss of outer-membrane porins (Brink, 2019). Different antimicrobial agents have different antimicrobial mechanisms and targets. Thus, understanding the antibiotic resistance mechanisms exhibited by CRAB will help clinicians to choose the antimicrobial regimens appropriately. At present, some studies have focused on the resistance mechanisms of clinical strains (Jiang et al., 2019; Li S. et al., 2019), and other studies have focused on observational research of the clinical efficacy of different antimicrobial regimens in the treatment of CRAB infections (Elsayed et al., 2019; Russo et al., 2019). Few clinical studies have compared the efficacy of the existing antimicrobial regimens based on different resistance mechanisms. Therefore, it is extremely important to know how to better select from the existing antimicrobials and optimize the treatment regimen according to the antibiotic resistance mechanisms to achieve a better clinical outcome.

This study explored the following three aspects: (1) The in vitro activities of imipenem (IPM), biapenem (BIP), sulbactam (SUL), tigecycline (TGC), or polymyxin B (PB), and the combination of IPM-TGC, IPM-BIP, PB-TGC, TGC-SUL, PB-SUL against CRAB were determined. (2) It investigated the mechanisms of carbapenem resistance in clinical CRAB strains. (3) we evaluated the relationship between resistance mechanisms, antimicrobial regimens, and clinical outcomes in patients with CRAB infections.

In this study, we found that CRAB displayed high sensitivity to TGC and PB, which is similar to that reported in many previous reports (He et al., 2015; Yazdansetad et al., 2019). Despite their in vitro activity against CRAB, the results of studies on the clinical effectiveness of TGC have been mixed, and even show a higher mortality risk (McGovern et al., 2013). A low serum concentration of TGC is inadequate to eradicate CRAB, and this limits its use in CRAB bloodstream infection. The role of TGC in treating CRAB infections remains controversial. Nephrotoxicity and neurotoxicity of colistin limited its clinical use, and patients with carbapenem-resistant bacterial infections receiving colistin monotherapy had a less than optimal outcome (Zusman et al., 2017). The combination of several antimicrobial agents produced synergistic effects in this study, especially the PB-SUL combination. Although many previous studies have demonstrated the synergistic or additive effects of an antimicrobial combination (He et al., 2015; Le Minh et al., 2015), the antibacterial activity in vitro does not necessarily reflect activity in vivo. In clinical practice, the efficacy of TGC and polymyxin was uncertain even when used in combination with other antimicrobials (Zusman et al., 2017; Doi, 2019). A previous study showed that the polymyxin-vancomycin combination produced a synergistic effect in vitro (Bae et al., 2016). Vancomycin may enhance the permeability of polymyxin on the A. baumannii outer membrane, and perhaps other antimicrobial agents, such as ceftazidime, aztreonam, and meropenem, may also have this effect, which can be studied further.

A high percentage of OXA-23-positive isolates were observed in this study. Carbapenemase production was the main resistance mechanism of CRAB to carbapenem. However, different types of resistance mechanisms in CRAB did not affect the clinical outcome. A previous study found that patients infected with bla_KPC–2-producing carbapenem-resistant Enterobacteriaceae (CRE) had poor outcomes (Wang et al., 2018). Mutations in bla_KPC–3 in CRE may lead to a decrease in the antibacterial activity of ceftazidime-avibactam (Haidar et al., 2017). Early detection of the carbapenemase type might help improve the prognosis of patients with CRE infections. With respect to CRAB, studies on the relationship between the clinical outcome and specific resistance genes are limited. A previous study showed that mortality in critical patients with A. baumannii complex bacteremia was not correlated with any specific genospecies (Lee et al., 2014). Resistance mechanisms might not play a significant role in the clinical outcomes of patients with CRAB infection, but further studies are needed.

For strains with the presence of carbapenemases, more isolates exhibited synergistic activity with a combination of PB/TGC and SUL, but there was no statistically significant difference in the synergistic effect for the strains with overexpression of efflux pump genes and functional loss of porin. Leelasupasri et al. (2018) also found that colistin combined with SUL has a good synergistic and additive effect on CRAB strains, and colistin plus SUL against CRAB seemed to be an interesting option. There are limited studies on the in vitro antibacterial activity of an antimicrobial combination based on different resistance mechanisms. An open-label, randomized controlled trial found that colistin-meropenem combination was not superior to colistin monotherapy in the cumulative survival rate (Paul et al., 2018). Although many combinations, such as SUL-TGC, SUL-carbapenems, polymyxin-carbapenems, and polymyxin-TGC, were used for CRAB infection (Chinese XDR Consensus Working Group et al., 2016), the efficacy of colistin-SUL against CRAB in clinical use also has to be evaluated. The 28-day treatment failure rate in these patients was 28.05% (23/82). The patients at an advanced age, those with MODS, those with a lower level of eGFR, and those admitted to the ICU had a poor clinical outcome, which was similar to that in previous studies (Sheng et al., 2010; Niu et al., 2018; Du et al., 2019). The treatment failure rate increased with the advanced age of patients, possibly associated with a decreased immune response to pathogens as age increases (Li M. et al., 2019). MODS, a lower level of eGFR, and ICU stay often indicate that the patients are critically ill and are more likely to die.

This study also showed that the level of PCT in the treatment failure group was significantly higher than that in the treatment success group. PCT has now been considered as a marker for early diagnosis, prognosis, and antibiotic stewardship in patients with bacterial infections, especially acute respiratory infections and sepsis (Lee et al., 2020).

Some previous studies have shown that appropriate therapy could decrease the mortality in patients with CRE bloodstream infections (Gutiérrez-Gutiérrez et al., 2017; Wang et al., 2018). In this study, we found that appropriate antibiotic therapy did not improve the clinical outcome of patients with CRAB infection. Most of the patients in the treatment failure group were in a critical condition when they developed CRAB infection. MODS is associated with very high mortality (Ramírez, 2013), and their clinical outcome was still poor even if they were given aggressive antibiotic therapy. Therefore, early identification of bacterial infection, immediate administration of antibiotic therapy, and prevention of severe infection-induced MODS are very important to reduce bacterial infection-related mortality. This study also found that the group that died had higher TGC MICs. Some patients were given TGC because polymyxin was not available in China at that time. Lower MICs of TGC can more easily meet the pharmacokinetic/pharmacodynamic parameter requirements.

In our study, multivariate analysis showed that ICU stay and PCT ≥ 2 ng/ml were closely correlated with a poor outcome of patients with CRAB infection. Previous studies showed that there were many risk factors for treatment failure in CRAB infection, including ICU stay, TGC therapy, presence of septic shock, higher acute Physiology and Chronic Health Evaluation II (APACHE II) score, central venous catheterization, inappropriate empirical antimicrobial treatment, and multiple organ failure (Sheng et al., 2010; Niu et al., 2018; Du et al., 2019). High PCT values often represent severe infection, which often occurs in critically ill patients and easily leads to death. A previous systematic review found that inappropriate empirical antimicrobial treatment was a risk factor for higher mortality in patients with CRAB infections (Du et al., 2019), which was different from that in our study. Further multicenter, high-quality, prospective studies are needed.

This study has some limitations. First, since the MIC values of TGC and polymyxin of many strains were very low, only some strains were evaluated for the in vitro antibacterial activity of an antimicrobial combination. Second, only three patients were treated with polymyxin because it was not available in our hospital at that time. Therefore, clinical data on polymyxin combined with SUL was not available. Relevant studies will be carried out in the future.

In conclusion, this study demonstrated the synergistic potential of the combination of polymyxin/TGC and SUL in vitro against CRAB presenting carbapenemase-encoding genes. The clinical outcome of patients was not associated with resistance mechanisms. The patients at an advanced age, those with MODS, and those admitted to the ICU had a poor clinical outcome. ICU stay and PCT ≥ 2 ng/ml were risk factors for treatment failure in patients with CRAB infection. Appropriate antibiotic therapy did not improve the clinical outcome of critically ill patients. Infections with higher MICs of TGC were more difficult to control. It may be more important to identify the infection early, actively control its progression to severe infection, and reduce ICU admission.

All datasets generated for this study are included in the article/supplementary material.

The studies involving human participants were reviewed and approved by the Ethics Committee of West China Hospital, Sichuan University. Written informed consent for participation was not required for this study in accordance with the national legislation and the institutional requirements.

XL and JQ conceived of and designed the study. XL supervised the experiments. JQ, RY, and CF performed the experiments. QW and CF collected the data. JQ, RY, and QW analyzed or interpreted the data. JQ wrote the draft. All authors read, revised, and approved the final manuscript.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.