What Is Antimicrobial Resistance?
Antimicrobial resistance (AMR) is a growing issue across the globe. It occasionally makes the news, particularly when new "superbugs" appear. But you may be wondering what antimicrobial resistance actually is.
The term "antimicrobial" covers antibiotics, antifungals, antivirals, and antiparasitics; the drugs used to treat bacterial, fungal, viral, and parasitic infections respectively. Resistance is the capability of a microorganism to survive exposure to drugs that would normally kill them or stop their growth. Importantly, antimicrobial resistance is a property of a microorganism –– not of humans. Humans do not become resistant, certain microorganisms do.
Resistance is not an all-or-nothing property, but rather a spectrum. A microbe may be fully susceptible, partially susceptible, or fully resistant to a given drug, and that status is drug-specific. The same microbe can be susceptible to one drug and resistant to another, which is why treatment depends on knowing both the microbe and its resistance pattern.
The Resistance Terminology Ladder: MDR, XDR, and PDR
Resistance is described in tiers based on the number of drugs an organism is resistant to both within and across different drug classes. Clinicians and public health systems use a graded vocabulary because the degree of resistance changes the clinical stakes.
Single-class resistance is the term used to describe a microbe that is resistant to a single drug in a single antimicrobial drug class. However, it is not uncommon that when a microbe evolves to be resistant to a certain drug that other drugs within the same class are also affected. This is called cross-resistance, in which all drugs in a specific class become ineffective because they work through the same or similar method to kill the microorganism.
A microorganism is considered multidrug resistant (MDR) when the microorganism is resistant to at least one drug across three different drug classes. Extensive drug resistance (XDR) means that susceptibility remains in only one or two drug classes. Lastly, the most serious form of antimicrobial resistance is pan-drug resistance (PDR), which is when a microorganism is resistant to all known antimicrobial drugs. The term "superbug" has been applied in the media to different microbes that may be MDR, XDR, or PDR. The microorganisms most associated with these tiers are often summarized by the ESKAPE grouping.
- Enterococcus faecium
- Staphylococcus aureus
- Klebsiella pneumoniae
- Acinetobacter baumannii
- Pseudomonas aeruginosa
- Enterobacter species
These specific bacteria are classified as such because they are particularly difficult to treat. Depending on the specific strain, these bacteria may be classified as MDR or XDR, with a few specific strains exhibiting PDR. These bacteria are common in hospitals and can be difficult to eliminate from medical equipment and other surfaces due to their ability to create biofilms.
How Does Antimicrobial Resistance Develop?
Resistance develops when microbes acquire the ability to survive a drug, which can occur through two different processes: spontaneous genetic mutations or horizontal gene transfer. Drug exposure then selects for the microbes carrying the resistance genes, which multiply and spread.
There are two distinct genetic origins, and the difference matters for detection later. Mutational resistance arises when a random change occurs in the microorganism's genetic code. As microorganisms reproduce, they make new copies of their DNA. This process isn't error free. While many mutations are neutral or sometimes harmful to the cell, other times they can be beneficial. The mutations that make a microorganism resistant to a drug can do so by changing the drug's target so that the drug cannot appropriately bind, altering the microorganism's cell wall so that the drug cannot easily enter the cell, create pumps that remove the drug from the cell, or creates a new enzyme that can make the drug harmless.
The other way resistance arises is through acquired resistance. Acquired resistance occurs when a microbe gains ready-made resistance genes from other microbes through a process called horizontal gene transfer, sometimes referred to as lateral gene transfer. This is one of the main drivers of the spread of antimicrobial resistance, so it is important to understand.
Rather than being located in a microbe's core genes, resistance genes frequently travel on what are called mobile genetic elements. These elements include plasmids (small loops of DNA that move between bacteria), transposons ("jumping genes"), and integrons (which capture and stack multiple resistance genes). A single plasmid can carry resistance to several drug classes at once and hop between species, which is how multidrug resistance can spread across a hospital or a microbiome quickly. Importantly, bacteria do not need to be of the same species to share mobile genetic elements with one another. This is why it is important to understand resistance in the broader context of the microbiome.
The Molecular Mechanisms of Resistance
Bacteria resist drugs through four main strategies: destroying or modifying the drug, altering the drug's target, keeping the drug out (or pumping it back out), and bypassing the process the drug attacks.
These four mechanism classes explain most clinically important resistance, and each maps to specific, named genes and enzymes that laboratories look for.
| Mechanism | How it works | Examples | Drugs affected |
|---|---|---|---|
| Enzymatic inactivation | The microbe produces enzymes that chemically disable the drug | Beta-lactamases, including extended-spectrum beta-lactamases (ESBLs) and carbapenemases (e.g. KPC, NDM, OXA-48) | Penicillins, cephalosporins, carbapenems |
| Target modification | The drug's binding site is altered so the drug no longer attaches | mecA → altered penicillin-binding protein in MRSA; vanA/vanB → altered target in vancomycin-resistant enterococci | Beta-lactams; vancomycin |
| Reduced accumulation | The microbe pumps the drug out (efflux pumps) or blocks it from entering (porin loss) | Efflux systems; outer-membrane porin changes | Multiple classes |
| Target bypass / protection | The microbe uses an alternative pathway or shields the target | Alternative metabolic enzymes; target-protection proteins | Sulfonamides, others |
The reason this table matters: a resistance mechanism is only detectable by a genetic test if the test knows to look for the responsible gene or mutation. Enzymatic mechanisms driven by an acquired gene (like a carbapenemase) are relatively straightforward to detect by their gene. Mechanisms driven by a subtle mutation, by loss of a structure (porin loss), or by regulation of expression can be harder to read from DNA alone, though there has been much progress in this area through the application of machine learning, a type of artificial intelligence that helps with prediction.
The Drivers of Resistance
Biology sets the stage; human behavior sets the pace. Known accelerators of antimicrobial resistance include over-prescription of antibiotics, incomplete or unnecessary courses, heavy antimicrobial use in agriculture and livestock, gaps in infection prevention, and — most relevant to diagnostics — empiric prescribing without knowing the microorganism or its resistance profile. Each unnecessary exposure applies selection pressure that favors resistant survivors. That final driver is the hinge between the resistance problem and the diagnostic solutions discussed next.
How Is Antimicrobial Resistance Detected?
Resistance is detected by identifying the microorganism causing an infection and determining which drugs it responds to. Two paradigms exist: phenotypic testing, which measures how the microbe behaves against drugs, and genotypic testing, which reads the microbe's genes for resistance markers.
The two paradigms answer the clinical question differently. Phenotypic testing asks, does this drug actually stop this organism? Genotypic testing asks, does this organism carry the genetic machinery for resistance? Those are related but not identical questions — a distinction that becomes the crux of interpretation. Culture and susceptibility testing remain the phenotypic reference standard, while genotypic methods such as PCR and next-generation sequencing are increasingly available.
Phenotypic Testing: Culture and Antibiotic Susceptibility Testing (AST)
Phenotypic testing grows the organism and exposes it to drugs to see which inhibit it. The key measurement is the minimum inhibitory concentration (MIC) — the lowest drug concentration that stops visible growth — which is then classified as susceptible, intermediate, or resistant.
After an organism is isolated from a sample (blood, urine, tissue), it is exposed to a range of drug concentrations. Going a layer deeper into how the result is produced: the minimum inhibitory concentration (MIC) is read off then compared against a breakpoint — a threshold set by standards bodies (CLSI in the US, EUCAST in Europe). This is then used to categorize the microbe as susceptible, intermediate, or resistant. Common methods of antibiotic susceptibility testing include broth microdilution (the reference method for MIC), disk diffusion, and gradient strips.
The strength of phenotypic AST is that it measures the behavior of the microbe against the drug. It captures resistance regardless of which underlying mechanism caused it, including mechanisms nobody has characterized yet. However, it also has limitations. It requires the microbe to grow, so microbes that grow slowly or not at all in culture may take longer or cannot be tested at all. It also occurs in a laboratory setting, which may not accurately reflect how the microbe actually behaves in the human body, or in vivo. While often useful, it can take several days to return results. During this time a patient may be taking a drug, prescribed empirically, that does not work against the microbe causing their symptoms and actively contributes to the spread of antimicrobial resistance.
Genotypic Testing: Reading Resistance from DNA
Genotypic testing detects mutations or mobile genetic elements from a sample using molecular methods, such as targeted PCR or next-generation sequencing, without needing to grow the microbe first.
Instead of watching growth in the presence of a drug, genotypic methods look for the genetic determinants of resistance. They differ mainly in breadth:
- PCR amplifies and detects specific, known resistance genes or mutations. It's fast and sensitive but only finds what it's designed to look for. Moreover, it may detect resistance genes present in microbes that are not actually causing the patient's symptoms, unnecessarily limiting treatment options.
- Clinical metagenomic sequencing, a type of next-generation sequencing, reads all the DNA in a sample directly and compares the genetic information found to a database. This allows for broader coverage of the microbes that may be present as well as more detailed characterization of their resistance profile. See our clinical metagenomics overview for a deeper dive.
The complete set of resistance genes in a sample is termed the "resistome" — the total genetic content with the potential to confer resistance. The appeal of culture-independent genotypic testing is speed and reach: it can flag resistance markers without waiting for growth and can detect microbes that are difficult or impossible to culture. Similar to phenotypic testing, interpretation can be difficult as the presence of resistance markers does not necessarily mean the microbe is resistant.
| Method | What it measures | Needs culture? | Speed | Breadth | Key limitation |
|---|---|---|---|---|---|
| Phenotypic AST | Actual growth in the presence of a drug | Yes | 2–5 days | Varies | Misses non-culturable or slow growing microbes, may not reflect what occurs in the body |
| PCR | Microbes + specific known genes/mutations | No | 1–2 days | Narrow | Only finds what it targets, interpretation and gene–organism linkage are hard |
| Clinical metagenomic sequencing | Microbes + resistance genes from sample | No | 2–3 days | Broadest | Interpretation and gene–organism linkage are hard |
Why Phenotypic and Genotypic Resistance Testing Do Not Always Predict Resistance
Detecting a resistance gene shows the potential for resistance, not a guarantee of it. A gene may be present but not active, resistance may come from a mutation the test doesn't cover, or a brand-new mechanism may be invisible to the database. This is why genotypic and phenotypic testing may be used together by clinicians to more confidently determine which drug may work against a specific microbe.
Here are some of the reasons phenotypic testing, and to a greater extent genotypic testing, may not always predict resistance:
- Presence ≠ expression. A gene can be present in the DNA but silent or poorly expressed, so the microbe may still be susceptible. Genotype describes capability; phenotype describes behavior. Moreover, the microbe may behave one way in culture and another way in the body. This can happen because of differences in nutrient availability, pressure from other microorganisms, or other environmental factors that alter a microbe's ability to express resistance.
- Mutational resistance is easy to miss. Resistance caused by single nucleotide changes in a microbe's DNA, called a point mutation, is only detectable if the test specifically covers that mutation. PCR panels and databases used in sequencing are incomplete for the full universe of resistance-conferring mutations.
- Novel or uncharacterized mechanisms are invisible. A genotypic test can only report what its database knows. A new mechanism, or a variant not yet curated, can produce a falsely reassuring "no resistance genes found."
- Multifactorial resistance resists prediction. When resistance results from several mechanisms combined (an efflux pump plus reduced permeability plus a weak enzyme, say), no single gene predicts the phenotype cleanly.
- Degree matters, and genes don't quantify it. Phenotypic AST yields an MIC — how resistant a pathogen may be. For PCR, gene presence is usually binary and doesn't capture the magnitude that can determine whether a higher dose is still viable. For clinical metagenomic sequencing, more comprehensive analysis can be performed but only if the database supports it.
Concordance between genotype and phenotype is genuinely high for certain well-characterized microbe–drug–gene combinations and much lower for others, which is why genotypic methods, though increasingly available, still lack fully standardized quality and interpretation frameworks. It's also why, in the sequencing context specifically, reliably detecting AMR markers from sequencing data has historically been more difficult than identifying the microbe itself. The practical upshot for both patients and providers: molecular resistance detection is a powerful, fast-moving complement to phenotypic testing — especially valuable for hard-to-culture microbes and rapid triage — but more often than not it does not replace antimicrobial susceptibility testing across the board. However, Biotia's tools are addressing the challenges in interpretation associated with clinical metagenomic-based detection of antimicrobial resistance.
AMR Detection through Biotia's Clinical Metagenomic Sequencing and Bioinformatics Platforms
Biotia is a leader in the emerging field of clinical metagenomics, a type of next-generation sequencing that reads all microbial DNA in a biological sample and compares it to a curated reference database containing the genetic information of thousands of microorganisms.
Biotia's diagnostic platform, BIOTIA-ID, refers to how samples are tested through clinical metagenomic sequencing in the laboratory. Biotia's bioinformatics platform is called BIOTIA-DX, and refers to how the sequencing data generated from BIOTIA-ID is compared to our reference database. Through this end-to-end process, Biotia is able to both identify pathogens and characterize their resistance profile.
Because our approach is genotypic, one might think our diagnostic and bioinformatic solutions are subject to the same limitations outlined above: presence ≠ expression, mutational resistance is challenging to detect, novel or uncharacterized mechanisms may be missed, multifactorial resistance can be challenging to predict, and the degree of resistance is not readily quantifiable based on gene detection. However, our highly-skilled researchers in both clinical microbiology and bioinformatics addressed these challenges when building out these tools.
- Presence ≠ expression. For our clinical diagnostic product, the BIOTIA-ID Urine Test, rather than including all potential resistance genes or point mutations, the clinical report has only included resistance markers that are highly correlated with phenotypic expression. This correlation was established through the use of publicly available AMR databases in addition to proprietary data generated in our laboratory. This strongly increases the actionability of the antimicrobial resistance report when a pathogen is detected.
- Mutational resistance is easy to miss. A major strength of using clinical metagenomic sequencing is the control it provides over the amount of data generated from a sample. This allows us to achieve a higher resolution of the microbe and its resistance profile to more accurately identify specific strains as well as instances of mutational resistance.
- Novel or uncharacterized mechanisms are invisible. This is an inherent limitation to genotypic testing that cannot be overcome via sequencing or bioinformatic optimizations. As our understanding of antimicrobial resistance grows, our reference database is updated to include new mechanisms identified for different microbes.
- Multifactorial resistance resists prediction. Multifactorial resistance has been overcome due to the advantages offered by clinical metagenomic sequencing. Our bioinformatics pipeline includes a machine learning step that has been trained on public and proprietary AMR data, including the specific species, their resistance markers, as well as their antibiotic susceptibility testing results. The result is that when multiple resistance markers are present, we can more accurately predict whether the combination will be phenotypically resistant.
- Degree matters, and genes don't quantify it. Similar to the above, through our reference database and the use of machine learning, we are able to understand in greater depth the likelihood of phenotypic resistance based on the presence of other genes that may regulate the expression of resistance genes.
While still not perfect, through a combination of sequencing and bioinformatic optimizations, we have reduced the impact of the limitations posed by using clinical metagenomic sequencing to profile antimicrobial resistance. In fact, our efforts have been recognized by some of the leading AMR researchers internationally. For two consecutive years, our software BIOTIA-DX has been awarded "Best Prediction Accuracy" in the Critical Assessment of Massive Data Analytics (CAMDA) AMR Challenge. We continue to conduct research on how to best leverage genomic information to predict antimicrobial resistance, as well as best practices for integrating clinical metagenomic sequencing results into routine healthcare practice.
Why Detecting Resistance Matters: Treatment and Stewardship
Detecting resistance lets clinicians match treatment to the specific microbe and its resistance profile instead of guessing, improving the odds of effective first-line therapy and supporting antimicrobial stewardship.
Undetected resistance means a patient may spend days on a drug the organism shrugs off, prolonging illness and, in serious infections, worsening outcomes. Accurate detection enables targeted therapy and reduces reliance on broad-spectrum agents that themselves drive further resistance. This is what antimicrobial stewardship is all about: better diagnostic information leads to better prescribing practices, which slows resistance emergence across the population.
