This site is intended for health professionals only

At the heart of general practice since 1960

Read the latest issue online

CAMHS won't see you now

How to make best use of microbiology

Consultant medical microbiologist

Dr Marina Morgan on which tests to do and how to interpret them

n average medical microbiology laboratory processes around 100,000 GP and hospital inpatient specimens a year. Urine testing accounts for the bulk of the workload, while antenatal and immune status serology for hepatitis B makes up much of the rest. Latterly, MRSA screens have contributed massively to workload lately, but it is likely pragmatism will prevail ­ management curbing the incessant rise.

The almost mystical world of microbiology often leaves clinicians baffled, and 'reading by rote'. At the risk of sounding patronising, the following is an attempt to express some simple principles that make interpretation of results easy.

Following bacterial infection, the neutrophil polymorph response is manifested by an increase in peripheral blood neutrophil count, or by the formation of pus. Pus consists of polymorphs that have gobbled up the bacteria and then died of indigestion.

Beware the neutropenic patient, who despite active infection will produce no pyuria (dipsticks for leucocyte esterase negative) sputum or pus in wounds. Most laboratories issue a specimen collection manual to guide clinicians which specimens to send and how to take them: the Exeter version can be found at

Urine culture and

sensitivity (C&S)

Some GPs culture all patients' urines, whereas others culture only when empirical trimethoprim therapy fails. Urine from all children, pregnant women and immunosuppressed patients should be cultured, irrespective of dipstick results.

For the vast majority of patients with symptoms of uncomplicated UTI and positive dipsticks, trimethoprim empirically is entirely reasonable.

Dipstick testing is a useful diagnostic aid, providing tests include leukocyte esterase as well as nitrites. Nitrite testing relies on a good protein diet, and organisms such as candida and Enterococcus faecalis cannot convert nitrate to nitrite.

Some antibiotics produce false nitrite and haemoglobin tests. Leucocytes in urine may indicate vaginal discharge, or presence of a catheter. Sterile pyuria and haematuria need investigating for tuberculosis, stones or tumour as clinically appropriate.

Interpretation of

an MSSU result

True mid stream specimens of urine are hard for females to produce, most are actually 'clean catch' specimens. Any clean and dry receptacle can be used to collect the urine before transfer into the narrow-mouthed specimen container. Direct microscopy gives a white blood cell count, which if >100/mm3 is termed 'significant' pyuria, indicative of infection. Presence of red cells, except for menstruation, is always abnormal. A pure culture of >105 organisms indicates an infection. Difficult interpretations fall between these (see box 'Urine results for interpretation').

Skin cells contaminating urine on collection account for 'heavy mixed growths', and advice to repeat the test.

Bacteruria and pyuria are inevitable with catheterised patients. Catheter specimens of urine (CSUs) should be cultured if there are clinical signs of infection attributable to UTI. Nursing home-generated CSUs account for a massive overuse of antibiotics, predisposing to C. difficile and MRSA colonisation. There is no such thing as a 'routine' CSU. Systemically unwell patients should receive a stat dose of gentamicin (80-120mg) pending overnight culture results, but if oral antibiotics are deemed sufficient, a quinolone is essential to cover pseudomonas.

Never treat on the basis of a CSU result more than two days old, since organisms change in variety and sensitivity roughly every three days.

Sputum specimens

Most sputum specimens are actually mucky saliva, or produced by patients on antibiotics. A gram stain is useful here, since pus cells with intracellular organisms reflect the bacteria reaching the depths of the airways before being consumed by polymorphs, and hence likely pathogens. A gram film with 'mixed organisms' or 'epithelial cells +' reflects oral contamination. Especially common are MRSA and candida, transferred from the throat to the sputum as it rolls by up into the mouth.

Whether these warrant treatment depends on clinical signs of infection, but their presence often heralds later invasive disease. Sensitivities should be noted to aid empirical prescribing in the event of deterioration, pending later cultures.

Oral thrush is a clinical diagnosis. A culture is only necessary for sensitivity testing in treatment failures.

Stool specimens

A bugbear of most labs is specimens in unsuitable containers, and the worst examples always occur with stools. Receptacles such as margarine pots inevitably leak. Specimens labelled 'stool' that are obviously urine, and minuscule amounts of dried stool mummified in layers of toilet paper, should be discarded.

A positive culture is legally unnecessary to notify food poisoning; the history is sufficient.

Follow up 'clearance' cultures after initial isolation of salmonella or campylobacter are technically unnecessary, since legally even food handlers can return to work once diarrhoea has clinically settled. Exceptions to this are S. typhi (typhoid) and E. coli 015, where a series of negative stools are mandatory.

Ideally, all gastroenteritis in children should be investigated microbiologically in case of E. coli 0157, but of course this is not practical. It is essential to culture stools from all patients, particularly children, with bloody diarrhoea.

If 'tropical travel' is omitted from the clinical details, cholera will be missed, since a special selective medium is necessary for culture. 'Ova, cysts and parasites' are a particularly time-consuming examination, but rewarding to treat. A diarrhoeal specimen from a tropical traveller that yields no pathogens is worth repeating in case of intermittent excretion.

Although threadworms may be detected microscopically in stool, a moistened perianal swab in a specimen pot containing some fluid for centrifugation is more aesthetic and hygienic than sellotape slide.

Genital specimens

Swabs, particularly high vaginal swabs, should never be 'dry' but always sent in a protective moist and usually charcoal-containing medium.

'Candida spp' really means a 'non-Candida albicans', implying fluconazole may not be the most appropriate antifungal. Almost all C. albicans are sensitive to fluconazole but other species are often not. If clinically relevant, it would be worth identifying further with biochemical and sensitivity tests. Recalcitrant symptomatic candidiasis warrants culturing with a request for full identification and sensitivity testing. The swab can be used for candida of any species.

Vulvo-vaginitis in pre-pubertal girls is usually due to group A ?-haemolytic streptococci (up to 55 per cent in Exeter) and very rarely due to candida. A third have a history of threadworms. Preputial or pericatheter discharge swabs from catheterised patients are rarely useful, except to discover MRSA colonisation that is probably better unsuspected.

It is vital to use correct swabs and transport media for chlamydial ELISA testing, and it is good practice to confirm all positives with a more sensitive second test (immunofluorescence or IF) on the same specimen. Swabbing a ruptured vesicle with a dry swab, then breaking it off into viral transport medium (commercially available) for culture, enables differentiation of the herpes group causing infection. Virus transport media contains antibiotics so is unsuitable for bacterial culture.

Alternatively, confirmation of the mere presence of herpes virus particles is easily achieved by pressing a glass slide to the vesicle, air drying the fluid, then sending for electron microscopy (EM). Only culture differentiates the two types of herpes simplex, and EM will not discriminate between these and varicella.

Wound swabs

'Screening swabs' for staphylococcal or streptococcal colonisation are best done using a swab pre-moistened with sterile water, to maximise the pick-up of organisms from dry skin/wounds.

Specimens requesting immediate MRSA screen are processed with only MRSA in mind. Infected wounds that may have MRSA are best sent requesting culture and sensitivity, since MRSA will automatically be detected, and other pathogens found too. Pus is always preferable to a swab, yields more anaerobes on culture, and should be sent in a pot, never in a syringe.


The aphorism 'you only find what you look for' is particularly true in serology. Clinical details dictate the serological tests performed. The lab protocol usually includes a batch of tests appropriate for certain symptoms. A 'respiratory screen' would include chlamydia, mycoplasma, Q fever and, in season, influenza and respiratory syncitial virus. A 'lymphadenopathy screen' would include CMV, Epstein Barr virus, toxoplasma and often cat scratch disease serology.

Exposure of a non-immune pregnant woman to varicella can result in fetal varicella syndrome. In the absence of a history of chickenpox and if there is any doubt about immunity, stored antenatal blood can be retrieved for testing for varicella antibodies. .

HIV, hepatitis B and C testing can be done on one sample of blood. Although routine, antenatal tests are genuinely high-risk specimens ­ despite the principle of universal precautions ­ and should be labelled as such. Most laboratories require a specific form to be completed indicating the patient has given informed consent for an HIVtest.

C-reactive protein (CRP)

Usually performed by the biochemistry department, this test is far more useful in assessing infected patients than peripheral blood white cell count. It is much underused by GPs, and overused inappropriately by hospital doctors.

The CRP is just one of the 'acute phase proteins' that contributes to the erythrocyte sedimentation rate. With a half-life of 24-36 hours, a single high level can be helpful, suggesting a bacterial infection. A CRP series, initially two-three times per week, is used to monitor the effect of antibiotic treatment of chronic bacterial infections such as osteomyelitis, or endocarditis. Viral infections have no effect on the CRP.

In diagnostic conundrums, a CRP may be diagnostic. An old woman with a CRP of 80-100, an ESR of 80-100 and haematuria is highly likely to have a chronic bacterial infection such as tuberculosis or sub-acute bacterial endocarditis.

In the absence of any antibiotic treatment, a normal CRP, haematuria and a very high ESR would point to autoimmune or malignant pathology (such as hypernephroma) rather than bacterial infection.

Most seriological testing is not worth doing until some time into the infection, often 10 days, and early samples are stored as 'baselines'. Your laboratory may offer IgM testing for some infections so it is worth consulting its user manual.

How to make best use of


Urine results for interpretation


Finding  · 10-100 polymorphs/mm3 l<10 polymorphs/mm3=""  · 100="">

 · Red cells absent  · Red cells ++  · RBCs + squames absent

 · Squames + (or 'epithelial  · No squames  · Culture >105 organisms

cells present') of Pseudomonas and Candida

 · Heavy mixed growth of  · Culture-pure growth of

>105 organisms/mm3 coliform; 105/mm3

Interpretation  · Contaminated, repeat Infected, possible renal infection, Colonised catheter, no

 · Probably a female urine, low white count probably treatment unless patient unwell,

likely vaginal discharge due to drinking lots of water clinically attributable to UTI

contaminating, accounting

for polymorphs

Micro hints for interpretation and treatment

 · Treat patients not results.

 · Just because a lab gives sensitivities, it does not mean the microbiologist believes the organisms warrant antibiotic treatment.

 · Any Staphylococcal isolate reported as flucloxacillin sensitive will be sensitive to co-amoxiclav and cephradine. All MRSA will be resistant to these.

 · Otitis externa, or 'mucky ears' growing pseudomonas, enterococcus, candida and aspergillus, need aural toilet not topical antibiotics or antifungals.

 · Problematic bacterial vaginosis (Gardnerella) in pregnancy can be treated reasonably successfully with amoxycillin

(up to 80 per cent).

Inappropriate specimens

 · Inadequate labelling or mismatched details (legally these specimens should be discarded).

 · Inadequate clinical details ('tropical travel' is insufficient: country and dates are more helpful since differing incubation periods help us decide what to test for).

 · Nappies/toilet paper with faecal smears.

 · Rectal swabs for gastroenteritis.

l'Routine' leg ulcer swabs.

l'Routine' catheter urine specimens.

l'Routine' stool cultures for salmonella, campylobacter clearance.

 · Urinary catheter tips.

 · Swabs for viral culture not in virus transport media.

 · Female urine for chlamydial ELISA (too many contaminants).

 · Swabs from dermatophyte infections for fungal culture.

Investigations of fever following tropical travel

 · Never forget malaria: protean manifestations of infection.

 · Three EDTA samples to haematology for thick and thin blood films remain the minimum initial investigation of choice.

 · Malaria serology ­ unhelpful as a diagnostic aid.

 · Stools for parasites and culture, mention countries visited on request form.

 · CRP, normal levels (<10g )="" suggest="" viral="" infection,="" 30-50g/l="" occasionally="" with="">

l'Tropical' serology ­ diagnostic yield for schistosomiasis, amoebiasis and so on depends on exposure to risk, and time since exposed.

 · Up to 1 per cent of acute illness and PUO following tropical travel is due to HIV seroconversion: suspect if rash and atypical lymphocytes on film.

Inappropriate requests

 · Malaria serology: very rarely indicated.

 · Brucella serology: no new cases acquired in the UK in years.

 · Chlamydia serology for diagnosis of current infection (should be reserved for investigation of infertility).

 · HIV serology of donor following inoculation injury on call

(post-exposure prophylaxis should be started on basis of risk assessment, never wait for results).

Rate this article 

Click to rate

  • 1 star out of 5
  • 2 stars out of 5
  • 3 stars out of 5
  • 4 stars out of 5
  • 5 stars out of 5

0 out of 5 stars

Have your say