Microbiology in Practice: An FAQ on Diagnostic Principles and Standards

What sample should be submitted for culture and why?

A proper sample is critical to obtain clinically useful results. Apart from collection of feces and urine, there are three main sampling methods: tissue biopsy, fine needle aspirate and swab.

• Tissue is preferred, whenever possible. When tissue is available, the area should be cleaned and debrided so that affected but viable tissue is sampled. Histology should be performed concurrently. 
• Fine needle aspirates may be useful with deeper infections, where surgical intervention is not being performed. Ideally, multiple aspirates should be collected. Cytology should be performed in parallel with culture.
• Swabs can be useful but are more prone to user error and contamination. When swabs are used for sample collection, the site should be cleaned and debrided, so that a clean bed of infected but viable tissue is accessible. The swab should be pressed onto the surface, with pressure helping drive bacteria-laden fluid up from the infected area.
• Pus should not be sampled, apart from discrete, closed abscesses. Even then, it is preferable to culture the abscess wall after incision, drainage and lavage. Pus is a by-product of the infection, not the infection site.3 Pus can capture local contaminants, complicating interpretation of results. 

Is indicating “urine” adequate for the laboratory to know whether to use urine or serum breakpoints? 

Urine breakpoints should only be used for lower urinary tract disease. If pyelonephritis is present or urine is being cultured in a potentially septic patient, antimicrobial activity in tissue is critical. Treating pyelonephritis based on urine breakpoints could lead to use of an ineffective drug. Even with lower urinary tract disease, consideration of the specific location of infection is needed. If there is concern about the bladder wall (e.g., polypoid cystitis, recurrent cystitis with a markedly thickened bladder wall), then it is possible that antimicrobial activity in tissue is more important than in urine. It is important to identify the best way to provide the information required for the laboratory to apply correct breakpoints.  Using the correct test code and providing specimen details to indicate kidney or bladder wall may be required if obtaining tissue or serum levels on a urine sample are important for a patient.

What is an MIC?

The minimum inhibitory concentration (MIC) is an in vitro measure of the susceptibility of a bacterium to an antimicrobial. It is the lowest concentration of an antimicrobial that inhibits visible growth of a bacterium in vitro.

There are various methods to determine MIC. Diagnostic laboratories most commonly use broth microdilution testing. That involves adding a standard amount of bacteria to small wells that contain culture broth with a specific concentration of antimicrobial. Usually, a series of 2-fold dilutions is used (e.g. 0.5, 1, 2 and 4 ug/ml). The lowest concentration of the drug that prevents visible growth in the culture well is the MIC. 

What does an MIC tell us?

The MIC tells us the antimicrobial concentration that inhibits the bacterium in vitro. This is a very useful clinic guide but depends on the bacterium/antimicrobial interaction to be similar in vivo and the drug concentration that is expected at the site of infection must be taken into consideration. This is discussed below under Breakpoints.

Does the lowest MIC mean the best antimicrobial? 

No. The MIC is an indication of how much drug is required to inhibit the bacterium. Other factors influence how effective and appropriate that drug would be. Of particular importance is the drug level that is expected to be present at the site of infection. A bacterium could have a low MIC to Drug A and a high MIC to Drug B, but if Drug A only achieves low levels at the site of infection, based on dosing or drug penetration, it may not be an effective option. Antimicrobial selection should not be based on picking the drug with the lowest MIC on the list of results.

I often see < or > with MIC results. Why is that?

A limited range of antimicrobial dilutions are tested. Since a culture plate can only accommodate a set number of wells, there is a need to balance practicality and depth of testing. Therefore, diagnostic laboratories tend to test a narrow range of concentrations that correspond to the most important concentrations for deciding whether a bacterium is susceptible or resistant. If a bacterium is quite susceptible (it is inhibited by all concentrations that are tested) or quite resistant (it grows in all well), then < or > are used to indicate that the reported MIC is not the absolute number.

For example, if a drug is tested at 0.5, 1, 2 and 4 ug/ml, and there is no growth, the MIC is 0.5 ug/ml or lower.  We don’t know if 0.25 ug/ml or lower would have also inhibited growth. This would be reported as < 0.5 ug/ml.  Conversely, if the bacterium grew in all wells, the MIC must be greater than 4 ug/ml. We don’t know if it is truly 8 or 8000 ug/ml, so this is reported as > 4 ug/ml.

What else is done for susceptibility testing besides MICs?

Disk diffusion is another common method. This involves placing an antimicrobial-impregnated disk on a culture plate that has been seeded with the bacterium. Higher concentrations of the drug are present close to the disk, as the antibiotic diffuses a short distance. The zone of inhibition, the area where there is no visible bacterial growth is measured and compared to a breakpoint, which is discussed more below. 

How does the laboratory determine if a bacterium is resistant or susceptible to an antimicrobial? 

MIC values or disk diffusion zones of inhibition indicate how much antimicrobial is required to inhibit bacterial growth. Treatment success also depends on the antimicrobial concentration that is expected at the site of infection. If antimicrobial concentration is high, a bacterium with a relatively high MIC might be inhibited, whereas it wouldn’t be if the drug concentration was lower. For interpretation, the MICs or disk diffusion zones of inhibition are compared to a breakpoint. 

For example, the breakpoint for cefpodoxime and E. coli is < 2 ug/ml for susceptible, 4 ug/ml for intermediate and >8 ug/ml for resistant.  Therefore, if the MIC of 2 ug/ml or lower would be reported as susceptible. A bacterium of 8 ug/ml or higher would be reported as resistant. Similarly, for disk diffusion, the breakpoint for E. coli and cefpodoxime is > 21 mm for susceptible, 18-20 mm for intermediate and <17 mm for resistant. A bacterium with a zone of inhibition of 17 mm would be reported as resistant. One with a zone of 21 mm or greater would be susceptible.  

Are there standard approaches to breakpoints?

Laboratories should use validated breakpoints from reputable organizations. For veterinary medicine, the main source of breakpoints and other testing guidance is from the Clinical and Laboratory Standards Institute (CLSI).¹ 

A major change in CLSI breakpoints was made recently. Why does that happen? 

Changes in breakpoints reflect advancement of knowledge, such as better understanding of the pharmacokinetics of an antimicrobial or the MIC distribution of a bacterium. It also may involve creating breakpoints for bacterium/drug/species combinations that were not previously covered. It can cause short term confusion and challenges for diagnostic laboratories but reflects advancement in testing and patient care. 

Labs vary in how quickly they make changes when these standards are updated, and the best commercial labs will find it important to make these changes. Veterinarians should ask their labs if or when they are going to adopt the changes.

Are all breakpoints the same?

No. Breakpoints are specific to each bacterium/drug/animal species combination. 

What is “I”? Can I use that drug in a patient.

“I” stands for intermediate susceptibility, meaning the bacterium’s MIC is slightly above the threshold for being considered susceptible. While it’s not low enough to have confidence in the drug under normal circumstances, it’s close enough that the drug may still work in certain clinical situations. This requires getting more drug to the site of infection than was anticipated based on the breakpoint. This can potentially be done through increasing the dose (at least doubling the dose).  Increasing the dose is feasible for drugs that have wide margins of safety so we can safely use a higher extra-label dose, or where the dose used to set the CLSI breakpoint is at the low end of a licensed range. 

For example, an E. coli isolate form a dog with an MIC for amoxicillin of 0.5 ug/ml would be reported as intermediate for infections involving tissue (not urine). The breakpoint is based on a dose of 11 mg/kg twice a day. Since we have a wide margin of safety for most beta-lactams, we can increase the dose. If the dose is doubled to 22 mg/kg q12 hours, the bacterium would be assumed to be susceptible, since the breakpoint for susceptible was one 2-fold dilution lower than the intermediate breakpoint. For a drug like amoxicillin, using q8h (or even q6h) dosing can also be considered. 

Additionally, there are some drugs that will achieve higher concentrations in certain tissues than they do in serum, such as macrolides (e.g. azithromycin) in lungs. If it is known that higher (at least double) drug levels are achieved in a certain tissue, an intermediate drug would likely be a viable treatment option.

What is SDD?

SDD stands for “susceptible, dose dependent”. This is a new category that is now part of CLSI guidelines.  SDD is used for a subset of bacterium/drug combinations where there is an ability to safely use higher doses to enable treatment of bacteria with specific MICs. To some degree, it is an evolution of the “I” result, where specific guidance is provided to indicate that the bacterium is considered susceptible to a specific, higher dose. For example, for enrofloxacin and E. coli from dogs, the following breakpoints are used:

• < 0.06 ug/ml: susceptible at 5 mg/kg q24 hours
• 0.12 ug/ml: SDD, susceptible at 10 mg/kg q24h
• 0.25 ug/ml: SDD, susceptible at 20 mg/kg q24h
• 0.5 ug/ml: resistant

SDD dosing is currently only available for a small number of bacterium/drug combinations. On lab reports, it would typically be indicated by SDD (instead of S, I or R), with a notation indicating the drug dose for that MIC. 

Are there different breakpoints for different types of infection or infection sites?

As discussed above, some drugs will achieve higher (or lower) concentrations at different body sites. For example, beta-lactam antimicrobials (e.g. amoxicillin) are excreted at high levels in urine. Therefore, a bacterium that might not be inhibited in tissue might be effectively treated in urine because of the higher drug level. For that reason, there are some breakpoints specifically for bacteria from urine. For example, an MIC of < 0.25 ug/ml is required to call E. coli from tissue/serum susceptible amoxicillin. However, the MIC can be higher, up to 8 ug/ml, and still be considered susceptible in lower urinary tract disease, as efficacy is based on urine drug levels. 

This highlights the importance of understanding where the disease is. For example, you might have E. coli from urine with an MIC of 4 ug/ml. We have to know where the suspected disease is. If it is cystitis, it would be reported as susceptible (MIC is less than the < 8 ug/ml breakpoint) because of the high drug levels in urine. However, if the animal had pyelonephritis, tissue drug levels are the key. That means the breakpoint would be <0.25 ug/ml and the bacterium would be considered resistant. This is why simply reporting “urine” as the specimen is not ideal.  Working with your lab to determine how best to indicate which breakpoints are required for the sample is important in these situations.

Are there CLSI breakpoints for all bacterium/drug combinations?

No. CLSI only has breakpoints for select bacterium/drug combinations for specific animal species. These tend to cover the most common and important pathogens and drugs. However, breakpoints are lacking for various bacterium/drug combinations in different species. Among companion animals, there are more CLSI breakpoints for dogs than cats, and no breakpoints for other small animals. This is often because of a lack of data about drug pharmacokinetics and bacterial MIC distributions, something that hampers assessment of optimal approaches.  

What do laboratories do it there are no CLSI breakpoints for a specific bacterium/drug/animal species?

Human breakpoints are sometimes used,  and while they can be useful differences in pharmacokinetics and bacteria mean that less confidence can be had in this approach. Laboratories often have internal breakpoints which are based on evaluation of their own available data (e.g. MIC distribution in bacteria from their laboratory), adapted from other laboratories, or obtained from research papers that suggested a breakpoint or extrapolated from related bacteria.  These breakpoints can be useful but are associated with more uncertainty.

Why might there be situations where bacteria are seen or expected (i.e., an abscess) but there is no growth on culture? 

“No growth” results may occur in situations where it is quite convincing that a bacterial infection is present. Some bacteria do not grow well or at all under normal laboratory conditions. Some bacteria might die during transportation, particularly anaerobes. “No growth” results must be evaluated considering other supporting evidence for a bacterial infection (e.g. cytology), specimen quality, specimen handling, time from collection to submission and likely pathogens (e.g. are fastidious organisms possible causes). No growth simply reflects the culture yield from the specimen as it was received at the laboratory. It does not definitively indicate the status of the patient.

Alternatively, “no growth” may truly reflect the status of the patent, such as with sterile abscesses or where infections were falsely suspected based on spurious results of other testing (e.g. strain debris misinterpreted as bacteria on cytology).

The laboratory has indicated “normal flora” or “mixed flora”.  Shouldn’t the lab report all bacteria identified and test them further? 

Many body sites have an abundant and diverse microbial population, including large numbers of non- or minimally pathogenic species. Standard laboratory practices are for laboratories not to report expected commensal bacteria or bacteria that are likely to be contaminants.^2,3 Therefore, culture results that suggest that a bacterium was present but was not reported further indicate a laboratory following standard practices. It is important to choose a laboratory that uses these practices.

If the clinician has questions about why a bacterium wasn’t reported the diagnostic laboratory should provide clarification. In some situations, additional testing may be indicated, especially when more clinical or specimen information is available. However, this should be reserved for situations where it is deemed that the bacterium may be clinically relevant. Otherwise, clinicians should not press diagnostic laboratories to report information that is considered clinically irrelevant. This could lead to unnecessary treatment, inappropriate treatment or delay further investigation into the actual cause of disease. 

The laboratory indicated a bacterium is expected to be susceptible to certain drugs, but susceptibility testing was not performed. Why weren’t the drugs tested, and should I ask for testing? 

Some bacteria have very predictable susceptibility to antimicrobials, with little or no concern about acquired resistance.  It is an acceptable (and expected) practice for laboratories to not perform susceptibility testing on these organisms.^2,3 This practice is only done on a defined set of bacteria/drug combinations where there is strong evidence supporting predictable susceptibility. 

There may also be situations where susceptibility testing is not performed because of methodological challenges. Some bacteria grow poorly under normal laboratory conditions and this hampers susceptibility testing. In these situations, laboratories will often report what the bacterium is expected to be susceptible to, based on available research.  While this is a reasonable assumption and can help guide clinical decisions, it should not be taken as a guarantee of susceptibility. 

Why are so few drugs reported on some cultures?

Some bacteria are intrinsically resistant to a range of antimicrobials. When a bacterium is known to be resistant, testing is not recommended. Rather, laboratories are instructed to either not report the drug or to indicate that the bacterium should be considered resistant to the drug. Common examples are Enterococcus spp and Pseudomonas spp. A limited range of drugs is reported because other commonly tested drugs are not applicable for these bacteria. Not reporting the results in these situations is consistent with CLSI guidance.