Growing levels of resistance to available antimicrobial medicines are causing tens of thousands of deaths each year across the world. By 2050, the overall costs associated with antimicrobial resistance (AMR) could reduce global gross domestic product (GDP) by 2 to 3.5%. One concern is the development of resistance to the carbapenem antibiotics among Gram-negative bacteria, in particular, the carbapenemase-producing Enterobacteriaceae (CPE). Enterobacteriaceae are the source of community- and hospital-acquired infections and commonly cause opportunistic infections, including pneumonia, and sometimes death. CPE are resistant to nearly all available antibiotics, with the exception of colistin. Emerging resistance to colistin therefore has troubling implications for patient care.

Carbapenemase enzymes confer resistance to most β-lactam antibiotics, including carbapenem, making them particularly dangerous because they can inactivate a wide range of different antibiotics. Accurate detection of carbapenemase producers is therefore essential for infection control and management of antibiotic therapy. Carbapenemases are divided into three main classes, which differ in their mechanism of action. Class A includes the most common carbapenemase, Klebsiella pneumoniae carbapenemase. Class B carbapenemases are metallo-lactamases. An important member of this class is New Delhi metallo-betalactamase (NDM), which has spread widely since it was first described in 2009. Class D includes carbapenem-hydrolyzing oxacillinase-48 (OXA-48), which is being reported with increasing frequency in outbreaks and case reports across the world. OXA-48 is particularly difficult to detect.

Over the past 2 years, Enterobacteriaceae that produce OXA-48 and NDM have rapidly spread through Europe. In 2015, 13 of 38 countries in Europe reported endemicity or interregional spread of CPE, compared with 6 countries in 2013. Concerns have also been raised over extensively resistant Pseudomonas aeruginosa in Belarus, Kazakhstan, and Russia. Almost all of these isolates were resistant to all antibiotics except colistin; a few were resistant even to colistin.

The resistance to colistin is particularly troubling. This polymyxin antibiotic was first used more than 50 years ago to treat Gram-negative infections, but its use declined on account of its severe effects on kidney function. Broad-spectrum beta-lactams (such as third-generation cephalosporins and carbapenems) were less toxic and easier to administer. However, as the incidence of multidrug-resistant infections has increased, so has—out of necessity—the use of colistin.

Colistin resistance arises through at least two known mechanisms, as illustrated by a recent report by Giani et al. on a large hospital outbreak of CPE Klebsiella pneumoniae infections in Italy. Resistance to colistin resulted in 93 bloodstream infections that were very difficult to treat. The outbreak was triggered by increased colistin consumption, underscoring the role of selective pressure generated by antibiotic use. However, the further development and spread of colistin-resistant mutants were mainly attributable to chromosomal mutations. This mechanism of colistin resistance facilitated the spread of resistant organisms within and between hospitals, but did not lead to the transfer of resistance between strains. Reinforcement of infection prevention and control measures was sufficient to limit the spread of these infections.

However, Liu et al. have recently documented a new mechanism of resistance to polymyxin antibiotics (colistin and polymyxin B). They have discovered a plasma-mediated polymyxin resistance gene, mcr-1, in Escherichia coli isolated from animals housed in an intensive pig-farming operation in Shanghai. Further testing showed a high prevalence of the gene in E. coli isolates from raw pig and poultry meat at slaughterhouses and supermarkets in four provinces in China. The gene was also found in samples collected from hospital patients infected with E. coli and K. pneumoniae in two hospitals in two provinces. Researchers in Denmark and France have confirmed the presence of the mcr-1 gene in stored samples of bacteria obtained from patients and poultry meat. These observations show that mcr-1 is not confined to China.

These findings are deeply worrisome. Resistance to polymyxin antibiotics resulting from chromosomal mutations generally cannot spread to other bacteria. However, as Liu et al.‘s study confirms, colistin resistance is transmissible between different bacteria because it involves a different mechanism, plasmid-mediated transfer. E. coli and K. pneumoniae are already increasingly resistant to all other available antibiotics; the transfer of colistin resistance to such bacteria can thus result in untreatable infections.

The study by Liu et al. leaves many questions unanswered, including the exact relationship between colistin use and the resistant bacteria found in animals, meat, and humans. This gap highlights the need for much better monitoring and reporting on the nature and extent of antibiotic use in animals and for further investigations. Nevertheless, there is a compelling public health need to take action now to slow the development of AMR.

In May of 2015, the World Health Assembly adopted a Global Action Plan against antimicrobial resistance. The plan reflects a global consensus on approaches to the problem but also a strong recognition that efforts to reduce AMR must extend beyond health. AMR cannot be addressed without the involvement of other sectors such as agriculture, development, economics, foreign affairs, and industry. The engagement of a broad range of sectors will require much greater awareness of the dangers posed by AMR, as well as political support.

Stewardship of existing antimicrobials within health and agriculture will be crucial for preserving the effectiveness of current antibiotics. At the same time, the ongoing evolution of new patterns of resistance will inevitably require the development of, and access to, new antimicrobial agents. Addressing AMR globally and across multiple sectors is a daunting challenge, but the consequences of not doing so will be far worse.

This article is reprinted from Science magazine, DOI: 10.1126/science.aad9450.