How to Read and Interpret an Antibiogram

An antibiogram is a table that shows, for each bacterial species commonly cultured at a hospital, the percentage of isolates that tested susceptible to each antibiotic over a defined period, usually a year. Reading one is straightforward once you know the layout: bacteria line the rows, antibiotics fill the columns, and the cell where they intersect tells you the likelihood that a given bug will respond to a given drug. The real skill is knowing what those percentages do and do not tell you, and how local context changes their meaning.

What the Grid Actually Shows

A standard antibiogram is built from cumulative susceptibility testing performed in a hospital’s microbiology laboratory. Each bacterial isolate recovered from a patient sample is tested against a panel of antibiotics. The lab determines whether the organism is susceptible, intermediate, or resistant to each drug, typically by measuring the minimum inhibitory concentration (MIC), which is the lowest concentration of an antibiotic that prevents visible bacterial growth in a controlled setting.¹ Those individual results are then pooled across all patients over the reporting period to produce a single summary number for each bug-drug pair.

The number in each cell is a susceptibility rate expressed as a percentage. If the cell for Escherichia coli and ciprofloxacin reads “72%,” that means 72 percent of E. coli isolates tested at that facility during the reporting year were susceptible to ciprofloxacin. Higher is better. A clinician treating a suspected E. coli infection who wants good odds of choosing an effective drug before culture results come back will scan across that row and look for the highest numbers.

Most hospitals update the antibiogram annually. The Clinical and Laboratory Standards Institute (CLSI) provides consensus guidelines on how labs should collect, store, analyze, and present these data, including sample templates for the final report.² In practice, many clinicians encounter the antibiogram as a laminated card, a wall poster, an intranet page, or a tab inside their electronic health record.

Using the Numbers for Empirical Therapy

The main reason antibiograms exist is to guide empirical antibiotic selection, the choice you make before you know exactly what organism is causing the infection or which drugs it responds to.³ When a patient shows up with a urinary tract infection, pneumonia, or bloodstream infection, clinicians need to start treatment right away. Culture results can take two to three days. The antibiogram gives a probabilistic snapshot of what drugs are likely to work at that particular institution.

A common rule of thumb in infectious disease practice is that empirical therapy should aim for a susceptibility rate of at least 80 percent for mild infections and at least 90 percent for serious ones like bloodstream infections or hospital-acquired pneumonia. If your hospital’s antibiogram shows that only 65 percent of Pseudomonas aeruginosa isolates are susceptible to a fluoroquinolone, that drug is a poor empirical bet for suspected Pseudomonas infections. You would look for an alternative with higher coverage.

That said, the antibiogram is one input, not the whole picture. Previous culture results from the same patient, recent antibiotic use, and local resistance patterns all factor in.⁴ A patient recently treated with a cephalosporin is more likely to harbor a resistant organism, so even a high susceptibility rate on the antibiogram might not apply to that individual.

The First-Isolate Rule

One of the least intuitive but most important rules behind antibiogram construction is the “first isolate” rule. CLSI guidelines recommend that only the first isolate of a given species per patient per analysis period be included, regardless of what body site it came from or what its resistance profile looked like.⁵ If a single patient with a chronic wound has Pseudomonas cultured from the same site six times in a year, only the first culture counts.

The reason is statistical. Chronically infected patients tend to harbor resistant organisms, and their repeated cultures can dramatically skew the data. In one study comparing first-isolate methodology against total-isolate methodology, the total number of isolates was roughly 1.7 times higher than the first-isolate count, and species like Acinetobacter baumannii showed ratios as high as 3.4 to 1.⁵ Including all those duplicate isolates would artificially deflate susceptibility percentages because resistant bugs get cultured more often than susceptible ones.

When you look at an antibiogram, it is worth knowing whether the lab followed this rule. Most accredited hospital labs in the United States do, but not all institutions worldwide follow the same methodology. If the antibiogram does not specify, ask the microbiology lab.

Why Hospital-Wide Numbers Can Mislead

A standard hospital-wide antibiogram pools data from every unit: emergency department, medical floors, surgical wards, and intensive care units (ICUs). The problem is that resistance patterns vary widely across those settings. ICU patients tend to have longer hospital stays, more prior antibiotic exposure, and more device-related infections. Their organisms are often more resistant than those found on general medical floors.

Research consistently shows relevant differences in susceptibility rates across different hospital departments and specimen types.⁶ A study of surgical and trauma ICU patients found that for key respiratory pathogens like Pseudomonas, Acinetobacter, and Staphylococcus, their patients showed different susceptibility profiles compared to the hospital-wide respiratory antibiogram.⁷ In practice, the hospital-wide number might tell you that 82 percent of Pseudomonas is susceptible to piperacillin-tazobactam, but in the ICU that rate could be considerably lower.

Using a unit-specific and culture-type-specific antibiogram for empirical therapy in the ICU would improve the likelihood that patients receive adequate initial treatment.⁸ Some hospitals now produce separate antibiograms for the ICU, for bloodstream isolates, or for urine isolates. If your facility offers these, they will almost always be more useful than the hospital-wide version for guiding a specific clinical decision.

Combination Antibiograms

A traditional antibiogram lists susceptibility to individual drugs. But for organisms like Pseudomonas aeruginosa, where single-agent resistance rates can be high, clinicians often use two-drug combinations. A combination antibiogram calculates the probability that at least one of two drugs will be active against a given isolate, even if the isolate is resistant to the other. This is possible because resistance to different drug classes is often independent, meaning the same bug that resists a cephalosporin might still respond to an aminoglycoside.

In a large analysis of over 11,000 non-duplicate P. aeruginosa isolates from 304 U.S. hospitals, single-agent susceptibility rates ranged from about 73 percent for fluoroquinolones to 85 percent for piperacillin-tazobactam. Combination regimens improved coverage, with piperacillin-tazobactam plus an aminoglycoside reaching the highest rate at about 93 percent. Adding an aminoglycoside consistently outperformed adding a fluoroquinolone.⁹ Even so, no combination in that dataset consistently achieved the 95 percent coverage threshold many institutions aim for, underscoring how challenging empirical Pseudomonas therapy can be.

At individual hospitals the story can look different. One university teaching hospital found that combination regimens containing amikacin provided coverage rates above 98 percent for P. aeruginosa, while those containing levofloxacin stayed in the low 90s.¹⁰ This is exactly why local data matters. National averages cannot substitute for your own institution’s numbers.

If your antibiogram only shows single-drug susceptibilities, you can estimate combination coverage yourself. The simplest method: if drug A has 80 percent susceptibility and drug B has 70 percent, and resistance to the two drugs is independent, then the chance that an isolate is resistant to both is 0.20 × 0.30 = 0.06, meaning combined coverage is roughly 94 percent. In reality, resistance patterns are not perfectly independent, so this calculation is an approximation. Formal combination antibiograms built from isolate-level data are more accurate.

Selective and Cascade Reporting

You might notice that certain antibiotics are missing from your antibiogram or that a culture report seems incomplete. This is often intentional. Many hospitals use a strategy called selective or cascade reporting as part of their antimicrobial stewardship programs. The idea is to suppress results for broader-spectrum drugs when a narrower-spectrum agent is effective, steering clinicians toward the narrowest appropriate therapy.¹¹

A typical cascade works like this: if an Enterobacterales isolate is susceptible to ceftriaxone (a third-generation cephalosporin), the lab does not release the susceptibility results for cefepime (a fourth-generation cephalosporin) or meropenem (a carbapenem). Only when ceftriaxone tests resistant does the lab reveal the cefepime result, and only when cefepime also tests resistant does meropenem appear.¹² The goal is to keep the broadest-spectrum drugs out of sight and out of mind unless they are genuinely needed.

For the antibiogram reader, cascade reporting means the denominator for a broad-spectrum drug may be different from what you expect. Meropenem susceptibility might look lower on the antibiogram than it truly is, because the only isolates tested against it (or at least reported) were the ones already resistant to narrower drugs. If you see an unexpectedly low susceptibility rate for a carbapenem or a broad-spectrum cephalosporin, cascade reporting may be the explanation. Check with your lab or pharmacy to understand which suppression rules are in place.

Tracking Resistance Over Time

Beyond guiding day-to-day prescribing, antibiograms serve as a surveillance tool for monitoring resistance trends. Comparing this year’s antibiogram with the previous year’s can reveal emerging problems: maybe E. coli susceptibility to trimethoprim-sulfamethoxazole dropped five percentage points, or Klebsiella susceptibility to carbapenems started to slip. Some institutions produce multi-year trend analyses specifically for this purpose.¹³

When reviewing trends, keep a few things in mind. Small year-to-year fluctuations in susceptibility rates, say two or three percentage points, can be statistical noise rather than a real shift, especially for organisms with low isolate counts. CLSI guidelines recommend that a susceptibility rate only be reported when based on at least 30 isolates of a given species.² If the antibiogram shows a number for an organism with fewer than 30 isolates, treat it cautiously. Some hospitals footnote these numbers or omit them entirely.

Larger shifts, on the order of 10 percent or more over a couple of years, deserve attention and possibly investigation. They can reflect changes in the patient population, new admissions from facilities with high resistance, outbreaks of a resistant clone, or changes in prescribing habits. Sometimes they reflect a change in methodology rather than biology, which brings us to a common source of confusion.

CLSI Versus EUCAST Breakpoints

The susceptibility percentages on an antibiogram depend entirely on where the line is drawn between “susceptible” and “resistant.” Two major organizations set these lines: CLSI (used mainly in the Americas and parts of Asia) and EUCAST (the European Committee on Antimicrobial Susceptibility Testing, used in most of Europe and increasingly adopted elsewhere). Their breakpoints are not always the same.

When a hospital switches from one standard to the other, or when breakpoints are revised within the same standard, susceptibility rates on the antibiogram can change overnight without any actual change in the bugs. A study examining the impact of switching from CLSI to EUCAST breakpoints found that for certain drug-organism pairs, the differences were substantial. For example, Pseudomonas aeruginosa susceptibility to meropenem was reported as 88.1 percent under CLSI 2009 guidelines but only 78.3 percent under EUCAST 1.3 guidelines, a nearly 10-point drop driven purely by where the cutoff was placed.¹⁴

A separate analysis at a hospital in Laos found that adopting EUCAST breakpoints would have significantly altered susceptibility reporting for first-line agents against the most frequently isolated gram-negative pathogens, predominantly reclassifying isolates from susceptible to intermediate or resistant.¹⁵ The practical lesson: if your antibiogram looks dramatically different from last year’s, check whether the lab changed breakpoint standards before concluding that resistance has shifted.

Inpatient Versus Outpatient Antibiograms

Most hospital antibiograms reflect inpatient cultures, but infections treated in outpatient clinics and emergency departments draw from a different pool of organisms. Community-acquired urinary tract infections, for instance, tend to involve less resistant bacteria than hospital-acquired ones. A hospital’s inpatient antibiogram might show E. coli susceptibility to ciprofloxacin at 70 percent, while the same organism in primary care outpatient cultures is susceptible 85 percent of the time.

Some healthcare systems have started producing outpatient-specific antibiograms to address this gap. A comparison of inpatient and outpatient antibiograms from two different health systems found no consistent pattern in susceptibility differences across settings, meaning you cannot simply assume that outpatient rates are uniformly higher. The researchers concluded that outpatient-specific antibiograms would more accurately inform empirical prescribing in the primary care setting.¹⁶

Regional variation adds another layer. A multicentric study of community-acquired urinary E. coli across different regions in India found that susceptibility patterns varied enough between regions that a single national antibiogram would have been misleading for local prescribing decisions.¹⁷ If you are prescribing in a community setting, try to find an outpatient or community antibiogram rather than relying on your nearest hospital’s inpatient data.

Common Gaps in Clinician Understanding

Despite being a fundamental stewardship tool, antibiograms are not as widely understood as you might expect. In a study of 30 physicians at a teaching hospital in Sri Lanka, only 12 reported any familiarity with the concept, and just one demonstrated comprehensive understanding of how to apply one clinically. Most had never seen an antibiogram in practice. However, once the format was explained, the majority found it clear, and almost all expressed strong agreement that antibiogram training would improve their patient care.¹⁸

This finding likely resonates beyond Sri Lanka. Many medical training programs spend relatively little time on antibiogram interpretation, and the practical nuances described in this article, first-isolate methodology, cascade reporting, unit-specific stratification, and breakpoint standards, are rarely covered in formal coursework. If you are a clinician or pharmacy professional encountering an antibiogram for the first time, or revisiting one after years of ignoring it, you are far from alone.

Digital Tools and Personalized Antibiograms

The traditional antibiogram is a population-level tool. It tells you what drugs work against the average isolate at your institution but says nothing about the specific patient in front of you. A growing area of development is the personalized antibiogram, which uses machine learning to incorporate patient-specific factors like prior cultures, recent antibiotic exposure, comorbidities, and hospital unit to generate individualized susceptibility predictions. Researchers have proposed embedding these prediction models directly in electronic health records, where they could compute and display a susceptibility probability score for each antibiotic in real time.¹⁹

Short of full machine learning integration, many hospitals have already built digital versions of their antibiograms into clinical decision support tools. One pediatric hospital developed an online platform called “Antibiogram+” that combined institutional susceptibility data with empiric antibiotic recommendations, dosing guidance, and treatment duration suggestions for common conditions, all accessible through the electronic health record and hospital intranet.²⁰ These tools represent the direction the field is heading: from a static annual table to a dynamic, context-aware resource that meets clinicians inside their workflow.

For now, though, the printed or PDF antibiogram remains the workhorse at most facilities. Learning to read it well, understanding what the percentages mean, recognizing when a hospital-wide number does not apply to your patient, and knowing the methodological quirks that can distort the data, is one of the simplest high-yield skills in antibiotic prescribing.