摘要
(See the Major Article by Strich et al on pages 611–21.) According to the Centers for Disease Control and Prevention, 2.8 million antibiotic-resistant infections occur in the United States each year, and 35 000 people die from these infections [1]. Among these antibiotic-resistant pathogens, carbapenem-resistant gram-negative bacteria, including Enterobacterales and Pseudomonas aeruginosa, are among the most critical pathogens that threaten human health [1, 2]. The primary reasons that these bacteria are considered such grave threats are their rapid emergence around the world, the high mortality rates associated with infections that they cause, and the extremely limited options for antimicrobial therapy [3–5]. Prior to 2015, options for the treatment of carbapenem-resistant gram-negative infections were primarily limited to polymyxins (colistin and polymyxin B), aminoglycosides, and tigecycline [6]. Aminoglycoside resistance among carbapenem-resistant Enterobacterales (CRE), as well as their toxicity and diminished effectiveness for bacteremia and pneumonia, limited their use for these infections [7–9]. The therapeutic role for tigecycline was also limited because of its low bloodstream concentrations, association with increased mortality compared to other agents, and lack of antipseudomonal activity [10, 11]. Polymyxins, despite being abandoned for decades because of their toxicity profiles, thus became heavily relied upon for these infections out of desperation [12, 13]. Ceftazidime-avibactam was approved by the US Food and Drug Administration (FDA) in February 2015, representing the first ß-lactam-based therapy with in vitro activity against non-metallo-ß-lactamase-producing CRE and most carbapenem-resistant Pseudomonas aeruginosa, including strains producing Klebsiella pneumoniae carbapenemase (KPC) and OXA-48 carbapenemases [7, 14]. In this issue of Clinical Infectious Diseases, Kadri and colleagues used a national administrative database to assess the use of ceftazidime-avibactam in US hospitals during the first 3 years after its approval and compared it to colistin use [15]. They found that despite the known toxicities of colistin, colistin use was much more common than ceftazidime-avibactam throughout the first 2 years after ceftazidime-avibactam approval (2015–2016). Although colistin use declined and ceftazidime-avibactam use increased throughout the study period, ceftazidime-avibactam was still used less than colistin by the end of 2017. Additionally, because the analysis did not account for polymyxin B use, the differences between overall polymyxin and ceftazidime-avibactam use were likely greater than what were reported. The investigators also assessed factors associated with ceftazidime-avibactam therapy compared to colistin therapy. They found that patients with chronic kidney diseases were more likely to receive ceftazidime-avibactam and that ceftazidime-avibactam was administered almost exclusively at teaching hospitals and with infectious diseases consultation. The continued reliance on colistin is unfortunate because recent investigations have revealed that polymyxins are not only nephrotoxic, but they are also relatively ineffective. Recent randomized trials that compared other new gram-negative agents to polymyxins for carbapenem-resistant gram-negative infections have consistently demonstrated increased mortality and decreased clinical responses with polymyxins [16–18]. Real-world observational studies echo these data, documenting worse clinical outcomes with polymyxin-based therapy for carbapenem-resistant infections, compared to treatment with ceftazidime-avibactam or ceftolozane-tazobactam [19, 20]. Recent pharmacokinetic-pharmacodynamic (PK-PD) data provide explanations for the poor efficacy of the polymyxins. Colistin is administered as a prodrug, colistimethate, that undergoes renal clearance more efficiently than conversion to the active drug, colistin [21]. Even with optimized dosing regimens, the majority of patients with normal renal function will not achieve a colistin exposure that is correlated with bacterial killing in murine neutropenic thigh models when infected with an organism that has a colistin minimum inhibitory concentration (MIC) of 2 µg/mL (the most widely applied susceptibility breakpoint) [21–23]. Furthermore, the target colistin exposure is also the threshold above which the risk of nephrotoxicity markedly increases, highlighting the extremely narrow therapeutic index of colistin [24]. Thus, colistin is not only a poor option for patients with impaired renal function because of the risk of nephrotoxicity, but it is also a poor option for patients with normal renal function because of the low probability of achieving effective drug exposures. The finding from Kadri et al that patients without kidney disease were more likely to receive targeted colistin therapy highlights that increased education is needed regarding this limitation of colistin. In contrast to colistin, polymyxin B is not administered as a prodrug and exposure is unrelated to renal function [25]. However, preliminary PK-PD investigations in K. pneumoniae infection models suggest similar concerns about the ability of polymyxin B to reliably achieve effective exposures at maximally tolerated dosages [26]. Murine models also demonstrate that both colistin and polymyxin B are unlikely to achieve even bacterial stasis for gram-negative pneumonia, even with optimized dosing [26, 27]. This is a major limitation of these compounds when administered intravenously, and thus it is an alarming finding from Kadri et al that colistin was just as likely to be administered as ceftazidime-avibactam for targeted therapy of respiratory tract infections. It is also alarming that they found that colistin use was more likely in transplant recipients, who represent some of our most immunocompromised patients. The slow uptake of an antibiotic that is less toxic and more effective than the prior standard-of-care therapy warrants consideration of factors that contributed to this situation. It should be noted that there was a shortage of ceftazidime-avibactam in late 2016 due to a manufacturing problem [28], but this does not explain the slow drug uptake during times that the drug was available. The high cost of ceftazidime-avibactam compared to colistin ($900–$1000 vs $20–30 per day) was likely a major contributor to its limited use [29]. Neither the phase 2 nor phase 3 clinical trials of ceftazidime-avibactam included patients with carbapenem-resistant gram-negative infections [30–35]. Thus, clinical data supporting the use of this agent for infections where the drug was needed were limited during the first few years after FDA approval, posing challenges in justifying the higher cost of the drug to hospitals. Even now, despite our new knowledge of the limited effectiveness of the polymyxins, clinical data supporting the use of ceftazidime-avibactam for carbapenem-resistant gram-negative infections are limited to observational studies [19, 36, 37]. Although a recent study suggests that ceftazidime-avibactam is cost-effective compared to colistin for CRE infections based on US standards, the absence of randomized comparative data limits the strengths of this conclusion [38]. Thankfully, randomized clinical trials of carbapenem-resistant pathogens were performed with subsequent new gram-negative agents, demonstrating consistently inferior results with the polymyxins [16–18]. Another critical factor that limited the use of ceftazidime-avibactam was the long delay until availability of susceptibility testing methods for this agent. Ceftazidime-avibactam disks did not become FDA cleared and available for disk diffusion testing until >1 year after drug approval and MIC-based susceptibility testing did not become cleared and available until >2 years after drug approval [39, 40]. The inability of clinical microbiology laboratories to perform ceftazidime-avibactam susceptibility testing was a major barrier to use, particularly given that metallo-ß-lactamase-producing CRE and P. aeruginosa, and even some strains of KPC-producing CRE, are resistant to ceftazidime-avibactam [41, 42]. This delay between drug approval and the availability of susceptibility testing devices unfortunately was not unique to ceftazidime-avibactam and also occurred with other recently approved agents, such as ceftolozane-tazobactam and ceftaroline [40]. Thankfully, the device regulatory landscape has responded to this problem and substantially shortened this delay for antimicrobial agents that were approved in the last 2 years [43]. This analysis by Kadri et al highlights an important problem in our current paradigm for the development of new antimicrobial agents for MDR pathogens: even when a new agent becomes available that is a better option than the current standard-of-care, forces that include cost, formulary restrictions, antimicrobial stewardship programs, lack of susceptibility testing methods and adequately powered randomized trials for target pathogens, and slowly changing clinician practices stifle the clinical use of the agent, which in turn eliminates incentives for antimicrobial development. The situation is even worse for more recently approved agents that target CRE, such as meropenem-vaborbactam and plazomicin. US sales for meropenem-vaborbactam were approximately $8 million during the year after its approval, and US sales of plazomicin were less than $1 million [44]. These revenues are not sustainable for companies that spend more than $1 billion to develop and bring a drug to market [45]. Thus, it should not be surprising that Achaogen (who developed plazomicin) and Melinta (current sponsor of meropenem-vaborbactam) both filed for bankruptcy [46, 47]. This situation also poses great danger to the viability of other newly approved gram-negative agents, such as eravacycline, imipenem-relebactam, and cefiderocol. In summary, this paper by Kadri et al highlights the challenges involved in incorporating the use of new agents that are more effective and less toxic but also more expensive than older agents. Infectious diseases practitioners have a responsibility to advocate for the use of these agents instead of polymyxins in patients infected with appropriate pathogens, although at the same time avoiding their use in patients infected with organisms that are susceptible to other first-line ß-lactams, so that their efficacy can be preserved. Finding the right balance in stewardship of these newer agents is critical in our ongoing fight against MDR gram-negative bacteria. Potential conflicts of interest. M. J. S. has received grant support from Allergan, BioFire Diagnostics, Contrafect, and Merck, and has received personal compensation for consulting from Achaogen and Shionogi. The author has submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.