European Review for Medical and Pharmacological Sciences 2021; 25: 5275-5292 Comparison of different colistin regimens for the treatment of pneumonia caused by multidrug-resistant microorganisms: a systematic review and meta-analysis H.-M. CUI, X. LIN, Y.-Y. LIU, Y.-H. SHEN Department of Hospital Infection Management, Huzhou Traditional Chinese Medicine Hospital Affiliated to Zhejiang Chinese Medical University, Zhejiang Province, P.R. China Abstract. – OBJECTIVE: Multidrug-resistant tion (1.37) in patients receiving combined intra- pneumonia is a common cause of hospital-relat- venous/inhaled colistin than in those receiving ed morbidity and mortality across the world. The intravenous colistin alone. Additional analyses high prevalence of multidrug-resistant pneumo- revealed higher rates of nephrotoxicity (1.30) nia due to resistant gram-negative pathogens has and mortality (1.44) in patients receiving intrave- led to a re-introduction of colistin. The adverse nous colistin than in those receiving combined events associated with intravenous colistin can intravenous/inhaled colistin. be alleviated by administering the drug nasal- CONCLUSIONS: We provide evidence showing ly (i.e., inhalation) or in a combination including improved clinical, morbidity, and mortality out- both inhalation and intravenous presentations of comes in patients with multidrug-resistant pneu- the drug. A review study compared the impact of monia receiving inhaled colistin or combined in- these administration methods on clinical, mor- haled/intravenous colistin than those receiving bidity, and mortality-related outcomes in patients intravenous colistin alone. These findings should with multiple-drug resistant pneumonia. Howev- help clinicians stratify the risks associated with er, the publication of newer cohort trials, warrants different colistin administration routes to manage an update of the state of the evidence. multidrug-resistant pneumonia. To compare the clinical, morbidity, and mor- tality outcomes in patients with multidrug-re- Key Words: sistant pneumonia receiving either intravenous Drug resistance, Pneumonia, Colistin, Intravenous, colistin or combined drug presentations (ie, in- Aerosol, Mortality. haled and intravenous). MATERIALS AND METHODS: A systematic search of the academic literature was performed according to the PRISMA guidelines across five databases (Web of Science, EMBASE, CEN- Introduction TRAL, Scopus, and MEDLINE). We conducted a random-effect meta-analysis to compare out- Acute respiratory drug-resistant pneumococ- comes such as rate of clinical cure, microbio- cal infections are considered a leading cause of logical eradication, nephrotoxicity, and over- 1-3 all mortality in patients with multidrug-resistant mortality in hospital settings worldwide . Accor- pneumonia receiving either intravenous colistin, ding to the World Health Organization, the onset inhaled colistin, or a combination of those ad- of pneumonia in a medical facility is facilitated ministration routes. by the presence of multidrug-resistant pathogens RESULTS: From 963 studies, we found 16 eli- including Acinetobacter baumannii, Klebsiel- gible studies with 1651 patients (61.6 ± 7.7 years) la pneumoniae, and Pseudomonas aeruginosa4. with multidrug-resistant pneumonia who had re- 5,6 ceived either intravenous, inhaled colistin or a Epidemiological studies have reported a high combined inhaled/intravenous administration. incidence of multidrug-resistant pneumonia in Our meta-analysis revealed higher rates of clin- hospital settings, 15% to 24%; and, a recent Glo- ical cure (OR, 1.61) and microbiological eradica- bal Burden of Disease study found that almost 2.3 Corresponding Author: Yuehua Shen, B.Sc.; e-mail: [email protected] 5275 H.-M. Cui, X. Lin, Y.-Y. Liu, Y.-H. Shen million patients worldwide perish annually due to this systematic review and meta-analysis to syn- multidrug-resistant pneumonia7,8. thesize the evidence on the available literature. The prevalence of multidrug-resistant pneu- We compared the outcomes (clinical cure rate, mococcal infections has been increasing9 due to microbiological eradication rate, nephrotoxicity, and the ineffectiveness of conventional antibiotics overall mortality) of these colistin regimens in pa- against drug-resistant pathogens. Therefore, older tients with multidrug-resistant pneumonia. Our fin- polymyxin antibiotics like colistin have garnered dings should provide clinicians with a clearer view attention for the management of multidrug-resi- of the morbidity and mortality-related risks associa- stant pneumonia10-12. Studies13,14 have suggested ted with different routes of colistin administration in that the cationic colistin molecule acts by bin- patients with multidrug-resistant pneumonia. ding electrostatically to the negatively charged lipid-A of the gram-negative bacteria, thereby damaging the cell membrane’s structure of the Materials and Methods bacteria by displacing its divalent calcium and magnesium cations. This ionic change destabili- We adhered to the PRISMA (Preferred Repor- zes the phospholipid bilayer of the gram-negative ting Items for Systematic Reviews and Meta-A- bacteria ultimately resulting in a leakage of its nalyses) guidelines36 while conducting this me- cellular contents and its destruction15-17. Doshi et ta-analysis. al. (2013)18suggested that the efficacy of colistin is dependent upon its administration route. The- Data Search Strategy se authors suggested that while conventional in- We performed the literature search in five travenous colistin may not permeate to the lung scientific databases (Web of Science, MEDLINE, parenchyma, the aerosol route could allow larger CENTRAL, EMBASE, and Scopus) from incep- and quicker deposition of the drug at the infection tion till March 2021. We used a combination of site. Many animal studies have confirmed the be- MeSH keywords including “Colistin”, “inhala- neficial impact of aerosol/combination (inhalation tion”, “intravenous”, “IV colistin”, “pneumonia”, + intravenous) routes of colistin administration as “drug resistance”, “morbidity”, and “mortality”. compared to intravenous administration alone19-21; Additionally, we manually searched the biblio- however, a consensus of the evidence in humans graphy section of the included studies to identi- is lacking. fy further relevant studies. The inclusion criteria Many cohort studies22-24 have compared clini- were the following: cal, morbidity, and mortality-related outcomes in - Studies comparing clinical cure, microbiolo- patients with multidrug-resistant pneumonia recei- gical eradication, and nephrotoxicity rates in ving colistin via different routes. However, a con- patients with multidrug-resistant pneumonia sensus on the overall mortality according to these receiving either intravenous colistin or com- administration routes has not been reached. We bination regimens with inhaled colistin. found studies18,22,24-26 reporting higher mortalities in - Studies comparing the overall mortality patients receiving intravenous colistin than in those outcomes in patients with multidrug-resistant receiving other regimens and we found other studies pneumonia receiving either intravenous coli- routes27-30 reporting higher mortalities for patients stin or combination regimens with inhaled receiving colistin via inhalation/combination routes. colistin. Similarly, the clinical cure rates of the different re- - Studies with human participants. gimens also remain unclear; while some studies re- - Case-control studies, prospective cohort ported higher clinical cure rates in patients receiving trials, or retrospective cohort trials. colistin via combined/inhaled route25-27,31-33, others - Studies published in peer-reviewed scientific reported the opposite effect28,30,34. journals. To the best of our knowledge, one review stu- - Studies published in English. dy35 has compared clinical, morbidity, and mor- Two reviewers independently screened the stu- tality-related outcomes of the different colistin dies. Cases of disagreements were resolved by di- administration routes in patients with multi- scussion with a third independent reviewer. drug-resistant pneumonia. However, since the publication of that review, several high-quality Quality Assessment cohort studies22-24, 27-29, have been published and We conducted the risk of bias assessment of warrant an updated analysis. Thus, we designed the included studies using the Newcastle Ottawa 5276 Colistin regimens for pneumonia scale37. This tool evaluates the outcomes for se- imputation of studies from either side of a plotted lective reporting, confounding bias, measurement graph to identify unbiased effects. The significance of outcomes, and incomplete data availability as level for this study was determined at 5%. threats that can compromise the validity of the analysis results. Two reviewers independently were in charge of the methodological quality as- Results sessment; and again, disagreements were solved by arbitration with a third reviewer. Our search across the five academic databases provided 950 studies. We identified an additional Data Analysis 13 during the screening of the reference sections We conducted a within-group meta-analysis of the included studies. After application of our using the Comprehensive Meta-analysis version inclusion criteria, we were left with 16 studies (Fi- 2.0 software38 based on the random-effects model39. gure 1). From all the included studies 13 were re- We calculated the odds ratio to evaluate the odds of trospective cohort studies18,22-24,26,28-30,32,33,42,43,