The first identification of a colistin-resistant Escherichia coli strain, from the pig sample in China in 2015 revealed that widespread colistin use in livestock farming probably exerts selective pressure on gut bacteria, promoting the emergence and rapid spread of the plasmid-mediated mcr-1 gene associated with colistin resistance. After the first discovery of the mcr-1 gene in China in 2015, extensive studies on mcr genes from 2016 to 2020 have detected additional mcr gene variants, worldwide. From other mcr gene variants, mcr-3, -7, -8 and -10 were first detected in China, while mcr- 2, -4, -5 and -6 were first reported in European regions and the mcr-9 gene was first discovered in the US.¹⁴ All these gene variants were detected in different microorganisms isolated from different hosts, as depicted in Table 1. Among diverse bacterial species from humans, animals, and the environment, 10 mcr gene variants (mcr-1 to mcr-10) have been identified, each conferring distinct colistin resistance mechanisms among members of the Enterobacteriaceae family. Although the catalytic activity and substrate specificity differ among phosphoethanolamine transferase encoded by various mcr variants, all share a conserved function in Lipid A modification.^15,16

The phosphoethanolamine transferase (PET) enzyme modifies the Lipid A component of LPS by adding cationic groups, thereby reducing colistin binding and conferring resistance (Figure 1). Among the various mcr variants, mcr-1 is the one most extensively studied and is located on multiple conjugative and non-conjugative plasmids, including IncX4, IncH1, IncI2, IncHI1, and IncHI2, as well as on hybrid plasmids such as IncX3–X4 and IncI2–IncFIB.¹¹ Plasmid analyses revealed ISApl1, an IS30 transposase, flanking the mcr-1, as part of the composite transposon Tn6330 (ISApl1–mcr-1–ISApl1), which facilitates mcr-1 mobilisation onto IncHI2 and IncI2 plasmids.²⁷ This insertion sequence aids mcr-1 gene mobilisation and rapid dissemination of colistin resistance among different Enterobacterales species,²⁷ as illustrated in Figure 1. In the IncX4 plasmid of K. pneumoniae, the mcr-1–pap2 region forms a hairpin structure flanked by two ISApl1 elements, serving potentially as an intermediate that facilitates mcr-1 dissemination. A subsequent loss of ISApl1 in IncX4 plasmid may enhance mcr-1 gene stability and provide adaptive advantages to host bacteria under selective antibiotic pressure.²⁸ In contrast, mcr-9 and mcr-10 genes identified in 59 Enterobacter isolates were carried on a variety of Inc plasmid types, including single replicons such as IncFIB(pECLA) and IncFIB(K), and multiple replicons such as IncFIA(HI1)–IncFIB(K) and IncFIB(pECLA)–IncFII(pECLA).²⁹

Owing to the presence of co-resistant genes on plasmids and chromosomes, bacteria have developed resistance to carbapenems, colistin, and β-lactam antibiotics, leading to extensively drug-resistant (XDR), multidrug-resistant (MDR), and pandrug-resistant (PDR) phenotypes in clinical, community and environmental settings.³⁰

Genomic sequencing studies reveal that plasmids carrying the mcr-1 gene often co-harbour other resistance determinants, including carbapenemase and extended-spectrum β-lactamase (ESBL) genes. IncHI2 plasmids harbouring mcr-1 often carry additional resistance genes against multiple antibiotic classes, including β-lactams (blaTEM−1), aminoglycosides (aadA1, aadA2, aph(6)-Id), tetracyclines (tetA, tetR), macrolides (mef(B)), and sulphonamides (sul1, sul2, sul3). Other variants, such as mcr-4 and mcr-9, have been associated with co-resistance to carbapenemase genes KPC-2 and OXA-48.³¹ Coexistence of the mcr-1 gene with other antimicrobial resistant genes (ARGs) such as floR, tetA, fosA, aac-3-IV, aac(6_)-lb, aadA1, ^blaTEM has been documented in many colistin-resistant Enterobacteriaceae like Salmonella spp., E. coli, and K. pneumoniae. Such coexistence heightens the risk of the emergence of pan-drug resistance and potentially threatens the use of active therapeutics for such infections.³² Co-occurrence of distinct mcr genes, mcr-1, mcr-3 and mcr-7, in addition to different ARGs against various antibiotics comprising β-lactams (bla_CTX-M-Gp1-9, bla_VEB, bla_CMY bla_PER), tetracyclines (tetA, tetB, tetC), aminoglycosides (aac(6′)-Ib, aadA, aph(3′)-Ia, ant(2′)-Ia), quinolones (oqxA, oqxB, qnrA, qnrB, qnrS), sulphonamides (sul1, sul2, sul3), chloramphenicol (cmlA), and macrolides (ermB, mefAE) were observed in zoo animals. Co-presence of many ARGs points out that a zoo acts as a potential reservoir contributing multidrug resistance and the possibility of its dissemination in other settings.³³

The convergence of colistin and carbapenem resistance in a single isolate is of particular clinical concern. For hospitals in African settings, where newer β-lactam combinations and carbapenemase inhibitors are often unavailable or unaffordable, pan-drug-resistant infections can leave clinicians with no viable treatment option.³⁴ This scenario, already documented in intensive care unit (ICU) outbreaks across several countries, underscores why containing the spread of co-resistant strains requires urgent and coordinated action.³⁵

A systematic review conducted between 2014 and 2021 on the global prevalence of plasmid-mediated mcr genes in E. coli across 54 countries and five continents reported an overall prevalence of 6.51%. The Asian region exhibited the greatest diversity of mcr variants (mcr-1 to mcr-9, excluding mcr-2), followed by Europe, which reported five variants (mcr-1 to mcr-5). The use of colistin as a growth promoter in pig farming may contribute to the diverse distribution of mcr variants observed in these regions.³⁶ Oceania documented only the mcr-1 gene variant, while the American region reported two variants: mcr-1 and mcr-3. Africa recorded three mcr gene variants: mcr-1, mcr-5 and mcr-8. In Asia, Thailand demonstrated the highest diversity with six variants (mcr-1, mcr-3, and mcr-6 to mcr-9), whereas in Europe, Spain exhibited four variants (mcr-1 and mcr-3 to mcr-5).^36,37,38,39 Among livestock, mcr prevalence was highest in pigs (14.9%) and chickens (15.8%). Conversely, isolates from healthy humans exhibited lower mcr prevalence rates of 7.4% and clinical cases of 4.2%.³⁶ In the African region, mcr-1 to mcr-3, mcr-4, mcr-5, mcr-8 and mcr-9 variants were reported in different pathogens including E. coli, Alcaligenes faecalis Klebsiella, Citrobacter, Salmonella, Enterobacter, Pseudomonas, and Acinetobacter species, emphasising the need for strategies to constrain the spread of colistin resistance in Africa.⁴⁰ In Latin America, mcr-1 prevalence was reported in bacterial isolates from animals (8.7%), food (5.4%), and humans (2.0%).⁴¹ Environmental studies have detected colistin-resistant bacteria in migratory bird faecal samples, suggesting that birds act as vectors in the environmental circulation of mcr-1 through contact with contaminated habitats.⁴² Anthropogenic activities such as agricultural runoff and marine pollution may serve further as reservoirs and dissemination pathways for colistin resistance in aquatic ecosystems.⁴³ These findings highlight the complex interplay between human, animal, and environmental factors driving the occurrence, contamination, and global dissemination of plasmid-mediated mcr genes, raising serious public health concerns.

Data on mcr gene distribution in Africa are growing but remain fragmentary relative to other regions. In North Africa, mcr-1 has been identified in clinical E. coli and K. pneumoniae isolates from Egypt, Algeria, and Tunisia.^44,45

In sub-Saharan Africa, studies from Nigeria and South Africa have detected mcr-1 in human clinical samples and livestock. Additional mcr variants including mcr-5, mcr-8, mcr-9, and mcr-10 have been identified in Nigeria, while mcr-8.1 has been reported from Kenya, frequently alongside extended-spectrum β lactamases (ESBL)-producing strains.^40,46

The widespread informal use of colistin in African veterinary settings, combined with limited regulatory enforcement, creates conditions for ongoing resistance selection and spread.^47,48 The absence of Africa from multicentre clinical outcome studies represents a critical knowledge gap that warrants urgent attention from researchers and policymakers on the continent.⁴⁹

In Enterobacteriaceae, chromosomally mediated reduction in colistin (polymyxin) susceptibility is linked primarily to chemical modifications of the Lipid A moiety of lipopolysaccharide in the bacterial outer membrane. The addition of amino-4-deoxy-L-arabinose (L-Ara4N), phosphoethanolamine (PEtN), or both to Lipid A decreases its affinity for colistin. Environmental stimuli activate signal transduction pathways that regulate these chromosomal adaptations. These adaptive responses induce L-Ara4N and/or PEtN modification of Lipid A through overexpression of the pmrHFIJKLM (also known as arnBCADTEF or pbgPE), pmrCAB, and pmrE operons. The expression of these operons is regulated by two-component systems (TCSs), namely PmrA–PmrB and PhoP–PhoQ. In these systems, the pmrCAB operon encodes the sensor kinase PmrB and its cognate response regulator PmrA, which can be activated by the PhoP–PhoQ system.⁵⁰ In the presence of polymyxins, low Mg²⁺ concentrations, or acidic conditions, the PhoP–PhoQ TCS activates Lipid A-modifying genes, thereby reducing colistin binding. Upon activation, the sensor kinase PhoQ undergoes autophosphorylation and subsequently phosphorylates its response regulator PhoP. The phosphorylated PhoP activates PmrD, in turn stimulating the PmrA–PmrB system, facilitating Lipid A modification. Under conditions of elevated Fe³⁺ or Al³⁺ concentrations and low pH, autophosphorylated PmrB transfers the phosphate group to PmrA, initiating transcription of the arnBCADTEF (pbgPE), pmrCAB, and pmrE operons. The pmrE gene product catalyses the synthesis and transfer of L-Ara4N to Lipid A. Meanwhile, the pmrC gene within the pmrCAB operon encodes PET, which facilitates the incorporation of positively charged PEtN into Lipid A.^50,51 Together, the PhoP–PhoQ and PmrA–PmrB systems modify the Lipid A moiety with L-Ara4N and PEtN, thereby reducing its overall negative charge. This reduction in negative charge weakens the electrostatic interactions between colistin and Lipid A, providing protection under antibiotic selection pressure.⁵² In another TCS, CrrA–CrrB, a missense mutation can induce autophosphorylation of the sensor kinase CrrB, which then phosphorylates its cognate regulator CrrA, leading to synthesis of the CrrC protein. The CrrC protein, acting via the PmrA–PmrB TCS, promotes the production of L-Ara4N and PEtN.⁵³ Such TCS-regulated chromosomal mutations confer colistin resistance in various bacteria, including Enterobacter spp., K. pneumoniae, Salmonella spp., E. coli, P. aeruginosa, and A. baumannii.⁵⁴

The small transmembrane protein MgrB serves as a negative regulator of the PhoP–PhoQ TCS. Upon environmental stimulation, PhoQ phosphorylates PhoP, which activates the transcription of multiple genes, including mgrB. MgrB inhibits PhoQ activity, suppressing PhoP phosphorylation and downregulating the PhoP–PhoQ signalling cascade.⁵⁵ In K. pneumoniae, deleterious mutations in the mgrB gene disrupt its inhibitory role on the PhoP–PhoQ TCS, thereby promoting colistin resistance. Nonsense mutations, deletions, insertion sequence (IS) transpositions, and amino acid substitutions can lead to a dysfunctional mgrB, resulting in upregulation of PhoP-regulated genes that enhance Lipid A modification and confer colistin resistance.^56,57

Alternatively, deletions or insertions in Lipid A biosynthesis genes (lpxA, lpxC, and lpxD) disrupt Lipid A and, consequently, LPS production, promoting colistin heteroresistance in A. baumannii.⁵⁸ However, LPS loss reduces bacterial survival, virulence, and transmission capacity, imposing a higher fitness cost compared with PmrA–PmrB-mediated resistance.⁵⁹ In general, chromosomal mutation-mediated colistin resistance is more stable but carries a lower horizontal transmission risk than plasmid-mediated resistance.⁶⁰ Nevertheless, studies have reported vertical transmission of co-integrated mcr genes from plasmids to chromosomes through ISApl1 transposons, integrative and conjugative elements (ICEs), or insertion sequences. Following integration, these elements can be maintained stably in the chromosomes of daughter cells, as observed in resistant strains of E. coli and Ralstonia pickettii. In addition, inactivation of the chromosomal mgrB gene via ISKpn72 mobile element insertion generates more stable resistant genotypes in K. pneumoniae. Such genetic events pose significant challenges to controlling the dissemination of colistin-resistant infections.^61,62,63,64 Clonal outbreaks of chromosomally mediated colistin- and carbapenem-resistant infections in ICUs often lead to treatment failure and prolonged hospitalisation, underscoring the urgent need for effective containment strategies.^65,66,67

Both plasmid-mediated and chromosomal mutation-based mechanisms contribute to colistin heteroresistance in Gram-negative bacteria (Figure 2), collectively posing a significant healthcare burden.

Colistin is commonly used to treat carbapenem-resistant infections, either as monotherapy or in combination with other antibiotics; however, current clinical evidence remains inconclusive regarding which approach is more effective.^68,69 The emergence of colistin resistance, often alongside resistance to other antibiotic classes, poses a serious threat owing to the growing prevalence of pandrug-resistant (PDR) bacterial strains.^70,71 To address this therapeutic challenge, several strategies are being explored, including the development of novel drugs, combination therapies involving colistin and other agents, drug repurposing, photodynamic therapy, colistin-sparing regimens, nanotechnology-based delivery systems, phage therapy, CRISPR (Clustered Regularly Interspaced Short Palindromic Repeats) interference (CRISPRi), and quorum-quenching (quorum-sensing inhibition) approaches. Most of these strategies remain in the in vitro or preclinical phase, with limited supporting evidence from human clinical trials. Although these innovations show promise, further validation through well-designed clinical studies is essential before they can be implemented as alternative therapies against colistin resistance.^72,73 A colistin-sparing approach can be adopted to treat carbapenem-resistant Enterobacteriaceae (CRE) infections using newer antimicrobial agents such as cefiderocol and sulbactam/durlobactam (novel β-lactam antibiotics), plazomicin (a next-generation aminoglycoside), eravacycline (a fluorocycline antibiotic), and omadacycline (an aminomethylcycline antibiotic). Fosfomycin has also demonstrated potent antimicrobial activity against MDR Enterobacteriaceae-related infections.^74,75 These alternative antibiotics can be administered either as monotherapy or in combination with other agents to reduce dependence on colistin for treating CRE infections and to limit the emergence of further colistin resistance.⁷⁶

Combination therapy involving colistin with carbapenems or β-lactamase inhibitors has demonstrated synergistic effects against MDR and XDR Acinetobacter baumannii infections, as well as significant in vitro antimicrobial activity.^76,77,78,79 Colistin combined with antimicrobial peptides or Ethylenediaminetetra acetic acid (EDTA) has also shown effectiveness against colistin-resistant and MDR bacterial infections.^80,81 Novel compounds currently under investigation include Fluopsin C, a secondary metabolite derived from Streptomyces and Pseudomonas species, which has shown broad-spectrum activity against both Gram-positive and Gram-negative drug-resistant bacteria.^82,83 Similarly, Terrain, a metabolite extracted from Aspergillus terreus, has demonstrated potent antibacterial activity.⁸⁴ In drug-repurposing approaches, niclosamide combined with colistin has exhibited synergistic antimicrobial effects against colistin-resistant A. baumannii and K. pneumoniae.⁸⁵ Emerging genetic and molecular strategies, including CRISPR/Cas9-based methods, phage genome engineering,⁸⁶ and quorum-quenching techniques,⁷³ hold promise for overcoming colistin resistance and multidrug-resistant infections.

Colistin resistance poses a major global health challenge, contributing to treatment failures, delays in effective therapy, prolonged hospital stays, and elevated mortality rates. A. baumannii, K. pneumoniae, and E. coli are the primary contributors to MDR infections associated with colistin resistance. Several studies have reported that pathogens demonstrated either colistin sensitivity, intermediate susceptibility or resistance against colistin along with carbapenem resistance, contributing multidrug resistance, or extensive drug resistance. These infections result in diverse clinical outcomes, including recovery, mortality, ICU admission, and extended hospitalisation, as summarised in Table 2.^{87,88,89,90,91,92,93,94,95,96}

Several phenotypic approaches are available for antimicrobial susceptibility testing, broadly categorised as dilution-based, diffusion-based, gradient-based, and automated methods.⁹⁷ For colistin specifically, CLSI (M100, 2025)⁹⁸ and EUCAST (2024)⁹⁹ endorse only the broth microdilution (BMD) method as the reference standard for susceptibility testing in Enterobacterales and other Gram-negative bacilli. Diffusion-based approaches, including disk diffusion and gradient strip tests (E-tests), are not recommended for colistin because of poor agar diffusion of the large polymyxin molecule and consequent poor reproducibility. Laboratories should therefore rely on BMD for routine colistin testing and confirm borderline results with complementary methods to ensure accuracy in clinical reporting.¹⁰⁰

Molecular and genomic techniques enable direct detection of genes responsible for colistin resistance. Commonly used methods include polymerase chain reaction (PCR), whole-genome sequencing (WGS), and DNA microarray analysis.¹⁰⁸

Recent advances have introduced emerging tools such as CRISPR-based assays, nanopore sequencing, and rapid phenotypic antibiotic susceptibility testing (AST) methods.

Despite advances in diagnostic technologies, global surveillance of colistin resistance remains inconsistent. In low- and middle-income countries, inadequate laboratory infrastructure, limited resources, insufficient funding, and a lack of governance and informatics-based systems contribute to under-detection and inconsistent reporting of resistance data. Harmonising data collection, improving comparability, and supporting timely public health interventions require increased participation of international health organisations, such as the WHO Global Antimicrobial Resistance and Use Surveillance System (GLASS), and the establishment of regional genomic monitoring networks. Effective control of AMR is not possible without addressing knowledge gaps related to both treatment practices and surveillance systems. Limitations in monitoring colistin resistance, both locally and globally, arise from multiple factors. These factors include gaps in treatment knowledge, inadequate transmission surveillance, and insufficient preventive strategies.¹²⁴

To preserve the efficacy of colistin against infections caused by carbapenem- and β-lactam-resistant bacteria, its use as a prophylactic agent and growth promoter in livestock has been restricted or banned in many countries.¹²⁷ In addition to the rapid spread of plasmid-mediated colistin resistance, chromosomal mechanisms also contribute to the stable transmission of resistance-conferring mutations.⁸ The presence of colistin resistance in agricultural, veterinary, and environmental sources presents significant challenges to both clinical and non-clinical settings. Therefore, a One Health approach should be implemented at local, national, and global levels to address this issue. Given the link between antibiotic use and the emergence of colistin resistance, it is essential to regulate its use in both clinical and non-clinical settings.¹²⁸ Controlling its spread through food, animals, humans, and the environment requires robust surveillance and effective preventive strategies.

In the African context, practical implementation of One Health surveillance could draw on structures that already exist. The Africa CDC AMR Surveillance Network provides a framework for coordinating human health data across the continent, while the animal health programmes and national veterinary services of AU-IBAR offer entry points for agricultural and livestock monitoring.^129,130,131

Regional economic communities including ECOWAS (Economic Community of West African States), EAC (East African Community), SADC (Southern African Development Community), and IGAD (Intergovernmental Authority on Development) could play a role in harmonising policies across national borders. Livestock, and particularly small ruminants and poultry, are central to the livelihoods of large populations across Africa, and their role as reservoirs of colistin-resistant organisms owing to informal antibiotic use means these animals cannot be excluded from any meaningful surveillance framework.^132,133

Rapid detection and the implementation of effective treatment regimens, guided by clinical assessment of colistin resistance, are essential for diagnostic and therapeutic strategies. Effective surveillance must integrate clinical, veterinary, and environmental data within a One Health framework to monitor the global emergence of resistance. In addition, global efforts must include social awareness initiatives and sustainable financial support to implement this approach effectively.^8,127

Implementing the recommended BMD method for colistin susceptibility testing presents real practical difficulties for many African laboratories. The method requires cation-adjusted Mueller–Hinton broth, sterile 96-well microtitre plates, and accurately calibrated bacterial inocula, all of which may be costly, intermittently available, or simply absent in public-sector facilities. In this context, many clinical laboratories continue to rely on disk diffusion or E-test methods, largely because these methods are cheaper and more accessible, despite the fact that neither is recommended for colistin by CLSI or EUCAST.^99,100

Beyond the gold standard method, access to molecular platforms for mcr gene detection is limited across much of the continent. Trained microbiologists and laboratory scientists are in short supply, external quality assurance programmes for colistin testing are not widely available, and supply chains for reagents are frequently disrupted. The lateral flow mcr-1 immunoassay offers a lower resource option for detecting the most prevalent variant, although it does not detect other mcr genes.¹⁰⁷ Strengthening African laboratory capacity will require targeted investment in training, affordable equipment, and integration into regional reference networks, as well as support for participation in WHO GLASS (Global Antimicrobial Resistance and Use Surveillance System) reporting.