Microbiology: Gram stain, culture and sensitivities

A microbiology result is not a single answer; it is a conversation between the laboratory and the bedside, written in three layers. The first layer is the direct examination, usually a Gram stain, which is available within an hour of the sample arriving. The second layer is the culture, which names the organism and how much of it is present. The third layer is antimicrobial susceptibility testing, which states, for each drug, how little or how much of it is needed to stop the organism growing. A reader who learns to read all three layers, in that order, can usually decide what to treat, what to ignore, and what to check again before the sample was ever sent.

This page is written for a first-time reader. It covers how to collect the specimen so that the report is worth acting on, how to read a Gram stain and a culture report, how to separate a contaminant from a true pathogen, how minimum inhibitory concentrations and breakpoints turn into the letters S, I and R, and how to read a report that names resistant organisms such as extended-spectrum beta-lactamase producers, methicillin-resistant Staphylococcus aureus and carbapenemase producers. It closes with a short discussion of antibiotic stewardship. Throughout, the units printed on your own laboratory's report, and the clinical picture, take priority over anything written here. National algorithms for collection, identification and reporting, such as the UK Standards for Microbiology Investigations published by the UK Health Security Agency, are adapted locally, so local laboratory practice and the named guideline body for your setting always take priority. ^[10]

By the end you should be able to:

1. Collect blood, sputum, wound and urine samples so that the culture has the information it needs. ^[6][7][8][1]
2. Read the Gram stain and predict which organisms the culture is likely to grow. ^[11][9]
3. Separate a contaminant from a true pathogen using site, organism, burden and clinical context. ^[6]
4. Explain how a minimum inhibitory concentration becomes an S, I or R report, using EUCAST and CLSI definitions. ^[3][4][5]
5. Recognise an ESBL, MRSA and carbapenem-resistant organism report, and choose the next step of therapy. ^[2]
6. Apply the result to a treatment decision without overtreating a colonised or contaminated sample. ^[2][9]

A microbiology test measures whether living organisms are present in a specimen, and, if so, which ones and how susceptible they are to drugs.

Specimen collection. The specimen is the whole test. Send blood in the correct volume and into the correct bottle, and the yield rises in a stepwise fashion; send too little and a real bacteraemia is missed. In published work, detecting organisms rose from 65 to 75.7% sensitivity with 20 mL of blood to 80.4 to 89.2% with 40 mL and 95.7 to 97.7% with 60 mL, and only 19.7% of routine bottles were filled with the recommended 8 to 10 mL. ^[6] Collect two separate blood culture sets from different sites if the patient is stable enough, so that a true pathogen grows in both while a skin contaminant grows in only one. ^[6] Always take cultures before antibiotics if the clinical situation allows, because the detection rate fell from 9.0% of bottles in patients not yet on treatment to 2.8% in those already treated. ^[6]

Sputum is not whatever is coughed up. A purulent portion is Gram stained, and high-quality samples are those with fewer than 10 squamous epithelial cells per field, or a ratio of leukocytes to squamous cells of 10 or more together with more than 5 organisms per field at 100 times magnification. ^[7] A sample that fails those criteria is probably saliva and grows oral flora; the laboratory should ask for another. ^[7]

Wound and soft tissue samples are better as tissue than as a surface swab. Surface cultures are useful for identifying organisms on the wound surface, but a deep infection is usually sampled by tissue or pus from the leading edge, because a swab misses what lies below and is less sensitive for true infection. ^[8]

Urine must be a midstream clean-catch sample taken before antibiotics in pregnancy and in men with urinary tract infection, and sent for culture and susceptibility in those groups. ^[1] A sample taken after the first antibiotic dose, or from a bag in a baby, tells you about the antibiotic rather than the organism. ^[1]

Direct examination. The Gram stain is the fastest microbiology result. It separates bacteria by their cell wall: Gram-positive bacteria stain purple and Gram-negative bacteria stain red after counterstain, and some organisms are Gram-variable. ^[11] Morphology is then described, such as cocci in clusters, cocci in chains, Gram-negative rods, yeasts or mixed flora. ^[9] Those shapes imply different empirical treatment questions: Gram-positive cocci in clusters, Gram-negative rods and yeasts point toward very different empirical regimes and infection control actions. ^[9]

Culture. The culture report says whether growth occurred, how quickly, how pure it was, and which species was identified. Modern laboratories use automated incubators and matrix-assisted laser desorption ionisation, which shorten the time to identification, but they do not remove the need to think about whether the organism found is the cause of the illness. ^[9]

Antimicrobial susceptibility. The laboratory tests how the organism grows at different concentrations of each drug. The result is expressed as a minimum inhibitory concentration, the lowest concentration of the drug that prevents visible growth in vitro, usually in milligrams per litre, and as an interpretation of that number. ^[4][5]

Rapid diagnostics. Rapid tests add pieces at different points: Gram stain is first; direct mass spectrometry and multiplex molecular panels identify the organism and selected resistance genes earlier; rapid phenotypic susceptibility testing and antigen or lateral flow tests add more information later. ^[9] They are adjuncts that change a decision only when the result reaches a clinician who can act on it. ^[9]

Microbiology does not have a reference interval in the usual sense, because a healthy person normally has no organisms in sterile sites such as blood, cerebrospinal fluid or deep tissue, and non-sterile sites such as urine or sputum always carry flora. The numbers below are the interpretive thresholds that the named sources quote, and they should be read against the printed ranges and clinical context on your own report. ^[1][7][6]

Ranges, thresholds, methods and reporting conventions differ by laboratory, age, sex and method, so the laboratory's own range always wins. ^[6][1][7]

The "values" in a microbiology report are qualitative in most places, so this table lists what a positive or negative culture, or an unusual result, most often means, and what to consider next. ^[1][6]

Named combinations of findings change the interpretation more than any single line.

• Culture negative in blood but the patient is clearly septic: previous or ongoing antibiotics, fastidious organism, wrong number of sets, or inadequate volume. ^[6] Repeat with full volume and before further doses if possible. ^[6]
• Gram-positive cocci in clusters on a Gram stain from a blood culture: suggests Staphylococcus aureus, with important management and infection-control implications, although only culture gives the species. ^[9]
• Coagulase-negative staphylococcus in one of two blood culture sets with no intravascular catheter and no clinical evidence of infection: most likely a contaminant. Treat only when both sets are positive, the same species is grown, or there is a prosthetic valve or device. ^[6]
• Gram-negative rods in urine with symptoms of cystitis: an enteric organism such as Escherichia coli is the likely cause, and the culture should guide whether the empirical drug chosen is still active. ^[1]
• ESBL-producing Escherichia coli or Klebsiella in a urinary isolate: carbapenems are the most reliable treatment for invasive infection; uncomplicated cystitis may be managed with nitrofurantoin, trimethoprim-sulfamethoxazole, single-dose aminoglycosides, pivmecillinam or sulopenem depending on susceptibility, availability and renal function. ^[2]
• Carbapenem-resistant Enterobacterales on a report: both porin-mediated and carbapenemase-mediated mechanisms exist, and the distinction matters for infection control and for therapy. ^[2]
• MRSA on any report: infection-control precautions are needed and the patient should be treated with an anti-MRSA agent according to the site of infection and local guidance. ^[2][5][9]
• Antibiotic before the culture: a major fall in yield and a risk of a false-negative report. ^[6]
• Asymptomatic bacteriuria over 100,000 CFU/mL in a pregnant woman: treat it, because it is a risk factor for pyelonephritis and premature delivery. ^[1] The same count without symptoms in a non-pregnant adult usually does not need treatment. ^[1]
• Mixed heavy growth from a superficial wound swab: usually surface colonisation; a deep tissue or biopsy sample is needed to diagnose true infection. ^[8]
• S, I and R together on one report: they are separate categories. Do not add S and I together as non-susceptible; I means susceptible with increased exposure. ^[3]

The values below are invented for teaching and are labelled as illustrative. They are not real patient data, and a real report replaces every number and category with the issuing laboratory's own working. Every rule, threshold and mechanism used in the interpretations comes from the cited sources. ^{[1][2][6][3][8]}

1. A colony count of less than 1,000 CFU/mL is below the greater-than-100,000 CFU/mL level that NICE quotes as significant for asymptomatic bacteriuria. ^[1]
2. The Gram stain of uncentrifuged urine shows no organisms, which supports the culture rather than contradicting it. ^[1][11]
3. With no symptoms reported, this is not asymptomatic bacteriuria either, because the count is below the threshold. ^[1]
4. No antimicrobial treatment is indicated by this report alone.

Conclusion: this is a negative urine culture with no significant bacteriuria, so the next step is clinical reassessment and no antibiotic, repeating the sample if urinary symptoms develop or persist.

1. Growth in two of two independently collected blood culture sets, within a short time of each other, makes a skin contaminant unlikely and supports true bacteraemia. ^[6]
2. Gram-negative rods on the Gram stain are the first organism-level result, and they predict an enteric pathogen before the species is confirmed. ^[9][11]
3. Ceftriaxone, cefepime and aztreonam are all resistant, which is the expected pattern for an ESBL-producing organism. ^[2]
4. An ESBL is an enzyme that inactivates most penicillins, cephalosporins and aztreonam, while ESBL-producing organisms generally remain susceptible to carbapenems. ^[2]
5. Meropenem and amikacin test susceptible, so a carbapenem is the most reliable treatment for invasive infection, with the aminoglycoside as a possible adjunct according to renal function and local policy. ^[2]
6. Source control and infection precautions are needed alongside antibiotics. ^[9]

Conclusion: this is an ESBL-producing Escherichia coli bacteraemia with a carbapenem and amikacin susceptible, so the next step is to switch to a carbapenem, ensure source control, and stop the empirical cephalosporin.

1. Surface swabs are useful for identifying organisms on the wound surface, but a deep infection is best diagnosed from tissue or pus from the leading edge. ^[8]
2. Mixed growth of more than 3 species with heavy total counts is the typical pattern of surface colonisation rather than a single infecting organism. ^[8]
3. A colonised wound can still grow a genuine pathogen, so the result is not meaningless, but it cannot prove that any one of these organisms is causing deep infection. ^[8]
4. Anaerobic culture was not requested, yet pressure ulcers and deep soft tissue infections often need it. ^[9]
5. The report alone should not be used to prescribe antibiotics; a tissue biopsy or needle-aspirated pus would be more informative. ^[8]

Conclusion: this swab shows surface colonisation rather than definitive wound infection, so the next step is to obtain a deep tissue or pus sample before choosing antibiotics, and to treat according to the clinical picture in the meantime.

• Acting on a contaminant as if it were infection. A single set growing coagulase-negative staphylococci, Cutibacterium acnes, Corynebacterium, Bacillus (not anthracis) or Micrococcus in a well patient usually represents skin flora introduced at collection. ^[6] Repeat before treating. ^[6]
• Sending an inadequate blood volume. Most routine bottles contain less than the recommended 8 to 10 mL, and yield rises steeply with volume. ^[6]
• Starting antibiotics first. The detection rate falls several fold once antibiotics are on board. ^[6] Take cultures before the first dose when it is clinically safe. ^[6]
• Treating saliva as sputum. A sample with more than 10 squamous epithelial cells per field and no leukocyte predominance represents mouth flora, and its culture is not worth acting on. ^[7]
• Reading a swab as a diagnosis. A superficial swab grows what is on the surface; a deep infection needs tissue. ^[8]
• Treating a threshold as a magic number. Asymptomatic bacteriuria is defined by more than 100,000 CFU/mL, but significance in a symptomatic patient, in pregnancy, in catheterised patients and in men is judged clinically and by method. ^[1]
• Misreading I as R. Susceptible increased exposure is not resistant; reporting S plus I as non-susceptible loses clinically useful information. ^[3]
• Quoting a breakpoint from the wrong table. Breakpoints are organism, drug, dosing schedule and committee specific; EUCAST and CLSI differ for some combinations and the current editions must be used. ^[4][5]
• Treating ESBL by the name rather than the report. An ESBL-producing organism is usually carbapenem-susceptible, but the exact drug, the site of infection, pharmacokinetics and local policy decide the choice, not the label alone. ^[2]
• Ignoring infection control. ESBL, carbapenem-resistant and MRSA reports carry a contact-precaution implication, not only a drug choice. ^[2][5][9]
• Forgetting that rapid tests are adjuncts. A rapid result that does not reach a clinician in time to act is an expensive delay. ^[9]
• Comparing methods or laboratories. Serial monitoring of any susceptibility trend needs the same laboratory and method, because committees and editions change. ^[4][5]

• A microbiology report has three layers - direct examination, culture, and susceptibility - and they are read in that order. ^[6][9]
• The specimen is the test: the volume, site, timing relative to antibiotics and collection method decide whether the report is trustworthy. ^[6][7][8][1]
• A Gram stain is the fastest result and sorts organisms by cell-wall class and morphology. ^[11][9]
• Contaminants and pathogens are separated by the organism, the site, the burden, and more than one sample agreeing. ^[6]
• The minimum inhibitory concentration is a laboratory measurement; the S, I and R letter is a clinical interpretation of it against a breakpoint. ^[3][4][5]
• Under EUCAST, S means susceptible at standard dosing, I means susceptible with increased exposure, and R means resistant even with increased exposure. ^{[3] [4]}
• Breakpoints are committee specific, organism specific, version specific and dose specific, and EUCAST and CLSI are not identical for every drug-bug pair. ^[4][5]
• ESBL-producing organisms inactivate most penicillins, cephalosporins and aztreonam but usually remain carbapenem susceptible. ^[2]
• Carbapenem-resistant and ESBL reports need both targeted therapy and infection control. ^[2] MRSA follows the same principle with anti-staphylococcal therapy. ^[5][9]
• Antibiotic choice follows the report, the site of infection, pharmacokinetics and local guidance rather than the organism name alone. ^[2]
• Stewardship is part of the result: do not treat colonisation, do not treat asymptomatic bacteriuria in a non-pregnant adult, and review with the culture. ^[1][2]
• Rapid tests are adjuncts that save time only when a clinician can act on them. ^[9]