J Clin Pathol 2001;54:911–919 911

Best Practice No 167 J Clin Pathol: first published as on 1 December 2001. Downloaded from

The laboratory diagnosis of urinary tract infection

J C Graham, A Galloway

Abstract women develop a UTI during their lifetime; the Urinary tract infection is common, and it incidence increases at puberty and remains is not surprising that urine specimens high throughout adult life, only after the age of make up a large proportion of those sam- 50 years is a similar incidence seen in males. ples submitted to the routine diagnostic UTI accounts for approximately 23% of all laboratory. Many of these specimens will hospital acquired infections.2 Although the show no evidence of infection and several incidence of infection is high, most specimens methods can be used to screen out received will show no evidence of infection and negative samples. Those that grow bacte- several methods have been developed to screen ria need to be carefully assessed to out negative samples to minimise expense and quantify the degree of and improve turnaround times. These will also be hence clinical relevance. To influence reviewed. treatment, a final report should be pro- Most infections at all ages are the result of duced within 24 hours of specimen re- enteric bacteria, especially , ceipt, with turnaround times continuously which colonise the perineum and then ascend monitored. Much work needs to be done to the urethra to multiply and infect the bladder, http://jcp.bmj.com/ determine the cost eVectiveness involved kidney, and adjacent structures.3 The most in processing urine specimens and the common site of infection is the bladder. evidence base for the final report pro- Haematogenous infection of the urinary tract vided. occurs most notably with tuber- (J Clin Pathol 2001;54:911–919) culosis and Salmonella spp, and direct introduc- Keywords: laboratory diagnosis; urinary tract infection tion of organisms during instrumentation of the urinary tract is also well recognised. Struc- tural and functional abnormalities result in a on September 27, 2021 by guest. Protected copyright. The aim of the microbiology laboratory in the wider range of possible infecting organisms. management of urinary tract infection (UTI) is Urinary tract infection may occur with or to reduce morbidity and mortality through without symptoms; the latter is known as accurate and timely diagnosis with appropriate antimicrobial sensitivity testing. Although opti- covert or asymptomatic bacteriuria. Because mal specimen collection, processing, and urine must pass through the distal urethra and interpretation should provide the clinician with in women over the perineum, it may become a precise answer, no single evaluation method contaminated by the normal flora of these is foolproof and applicable to all patient regions. of more than one bacterial groups. In practice, laboratories will not be able strain suggests such contamination, but even to approach each specimen individually, and when a single strain is isolated, quantitative standard operating procedures are generated to culture is required to determine whether it cover the processing of most samples, with the indicates true bacteriuria. Kass, in his original aim of detecting the abnormal presence of bac- studies validating the midstream urine speci- teria and fungi within the urinary tract.1 The men (MSU), showed that 95% of hospitalised Clinical Microbiology interpretation of results requires an under- patients with acute pyelonephritis had more 5 Laboratory, Royal standing of the limitations of local laboratory than 10 colony forming units (cfu)/ml of Victoria Infirmary, protocols and of the clinical context in which urine, whereas only 6% of asymptomatic Queen Victoria Road, patients had this degree of bacteriuria.45 Newcastle upon Tyne the specimen was taken. NE1 4LP,UK Urine specimens make up a large proportion Subsequent studies have shown that lower bac- J C Graham of the samples submitted to a routine diagnos- terial counts can be important; this applies A Galloway tic laboratory. A large laboratory may examine both to men and symptomatic women in whom 200–300 urine samples each day. This heavy 30–50% have fewer than 105 organisms/ml.67 Correspondence to: Dr Galloway workload reflects the frequency of UTI both in These cut oV values can be applied to all angela.galloway@ general practice and in hospital settings. In rapidly growing bacteria but not fungi or nuth.northy.nhs.uk children, infection is more common in young fastidious organisms. Patients with frequency Accepted for publication girls, except in the neonatal age group, where dysuria syndrome in whom urine cultures show 31 May 2001 boys predominate. It is estimated that 20% of no appreciable growth should be investigated

www.jclinpath.com 912 Graham, Galloway

for other agents that cause non-specific urethri- used by patients with neurogenic bladder and tis, such as Chlamydia trachomatis.8 It should be these specimens should be treated as CSUs. noted that above the distal urethra the urinary Increasingly, samples are being received tract is normally sterile, and any bacteria from patients with urological problems, includ- isolated from urine samples taken directly from ing patients with ileal conduits and those who the bladder, ureter, or kidney must be viewed have undergone bladder reconstruction, and J Clin Pathol: first published as on 1 December 2001. Downloaded from as clinically relevant. The detection of poly- these can present diYculties with interpret- morphonuclear cells (pyuria) and red blood ation. Specimens should only be taken if there cells (haematuria) in urine is useful for the are clinical signs of infection (for example, diagnosis of infection or other renal tract malaise, pyrexia, or vomiting) and they should pathologies. be obtained via careful catheterisation of the Each laboratory should aim to have available stoma or reconstructed bladder using an asep- a final report with microscopy, culture, and tic technique. The interpretation of the cul- sensitivity on a substantial number of speci- tures should then be as for any catheter speci- mens (for example, > 90%) within 24 hours men. after the receipt of the specimen, and laborato- Suprapubic aspirates (SPAs) were often ries should monitor their turnaround times as obtained from babies and young children and part of quality assurance. The ability to screen are still considered the “gold standard” and are out negative specimens as quickly as possible used in diYcult cases. Any isolate should be should also be considered; ideally this is done considered clinically relevant. Obtaining an at source. The physician can make a clinical SPA involves an invasive procedure; however a judgement as to whether a negative screen (for sterile adhesive bag or pad (that is, a sanitary example, by reagent strip testing or bedside towel with the appropriate absorption charac- microscopy) is suYciently accurate to rule out teristics, lining a nappy) is much simpler but infection in that individual patient, given the can still achieve a definitive answer in 50–75% limitations of the method used. of cases.9 After cleansing the perineum, the This broadsheet deals specifically with the baby is maintained in an upright position until diagnosis of bacterial UTI causing cystitis or urine is passed into the bag, pad, or alterna- pyelonephritis. It does not include the labora- tively via a clean catch into a container. tory investigation of prostatitis, infection with Urine samples collected from the ureter (at mycobacteria, or the examination of urine for cystoscopy) or from the kidney (via a nephros- parasites. tomy) should be treated in the laboratory as a fluid from a sterile site and all bacterial and Collection and transport of specimens fungal growth viewed as clinically relevant. For Rigorous care during the collection of urine is the diagnosis of prostatitis, urine specimens http://jcp.bmj.com/ vital to prevent contamination by commensal may be collected after massage of the prostate flora, especially in female patients and children. via the rectum because this is said to release Most samples are MSUs, and patients should any sequestered bacteria or inflammatory cells be given clear instructions on discarding the from this site. first part of the stream before collection in an As can be seen from the above, not all speci- appropriate sterile container. Female patients men types are the same and correct interpret-

should be instructed to part the labia while ation of urine cultures requires accurate data on September 27, 2021 by guest. Protected copyright. passing urine to avoid contamination. The ini- being clearly present on the request form. tial few millilitres of urine wash away distal Urine will permit growth of bacteria and if urethral organisms and hence the MSU is rep- there is to be a delay in transport (> 2 hours) resentative of bladder urine. It requires good or in setting up cultures in the laboratory then control of micturition and an adequate volume it should be stored refrigerated at 4°C; this will of urine in the bladder. It may prove diYcult to also preserve the white cell count.10 11 Boric get such a sample in the elderly or those with acid is often used to retain the bacterial count, hip joint problems. Early morning samples may but its antibacterial activity can reduce the harbour greater bacterial counts but are less number of organisms present, especially if an amenable to outpatient clinical practice. How- inadequate volume of urine is dispensed into ever, they are recommended for the diagnosis the container.12 Hence, rapid transportation/ of renal . processing with refrigeration (if necessary) is Catheter specimens of urine (CSUs) are the preferred method. A or dip inocu- often obtained from patients with long term lum is useful in overcoming delays in culturing indwelling catheters. Bacteria are frequently urine specimens, but because no microscopy recovered but only a few are important and can be performed on the specimen, it is most samples should only be taken when signs and useful for the follow up of patients. It is more symptoms such as fever, loin pain, or suprapu- expensive than routine methods. bic pain suggest infection. Urine should be aspirated directly from the catheter using a Initial processing of specimens sterile needle and syringe and then placed in a A clear specimen of urine is unlikely to grow sterile container. Bacteria multiply in catheter bacteria in great numbers but a cloudy bags so specimens from this site are unsuitable. specimen can result from bacteria, crystals, or Temporary catheterisation was often used to leucocytes. Because gross visual examination diagnose urinary infection in women but cannot always be relied upon, several rapid because of the risk of bacteria being introduced screening methods have been developed; these into the bladder this method is no longer are best performed at the bedside and accu- acceptable. Intermittent catheterisation may be rately detect pyuria or the presence of bacteria.

www.jclinpath.com The laboratory diagnosis of urinary tract infection 913

Recurrent infection as saprophyticus and enterococci. Atypical presentation Post-treatment The test is best performed on early morning Diabetes mellitus General practice Pregnancy Pre-operative screening urine and, although specificity is high, sensitiv- Symptomatic Immunocompromised Asymptomatic children ity ranges from 35% to 85%. Combined with children Renal tract abnormalities Elderly care the leucocyte esterase test, the sensitivity rises to 70–100% with only a small decrease in spe- J Clin Pathol: first published as on 1 December 2001. Downloaded from cificity.14

Bed side Dip stick –ve microscopy screening Discard OTHER METHODS USED Glucose oxidase +ve +ve Glucose is usually present in urine at low con- Isolated cystitis centrations, and bacteria will utilise this energy source; hence, a positive test is indicated by the Microscopy and culture Empirical treatment absence of glucose. The test is quite sensitive Figure 1 Guidelines on the use of dipsticks in clinical but false positives can occur along with false practice. negatives, in patients with diabetes mellitus, or However, culture remains the preferred in those patients infected with bacteria that do method of detecting and quantifying bacterial not metabolise glucose. growth. activity REAGENT STRIP TESTING Catalase is found in most bacteria that cause Several tests are available, those most com- UTI (but not streptococci) and in associated monly used to assist in the diagnosis of urinary somatic cells (leucocytes and erythrocytes). It tract infection are protein, blood, leucocyte is measured with a tube based assay (API Uris- esterase, and nitrate reductase.13 Their princi- creen; BioMérieux, Lyon, France) and has pal role is in excluding infection; combining the been promoted for screening asymptomatic last two tests has a negative predictive value of populations. Field trials have given variable 98% but much lower positive predictive value results.17 (38%). Strips may be read visually or by machine. A suggested protocol for their use is Bioluminescence giveninfig1. UTIscreen (Coral Biomedical, San Diego, California, USA) is a semiautomated system Protein that measures bacterial ATP. A releasing agent A positive urine test for protein is a poor indi- is added along with the firefly enzyme luciferin- cator of infection on its own, with a high rate of luciferase and the sample is incubated at room http://jcp.bmj.com/ false positives and negatives; however, it may temperature for 15 minutes, during which time indicate several other renal pathologies, includ- the integrated light output is measured. The ing glomerulonephritis and pre-eclampsia. specificity of this method at detecting > 105 cfu/ml is 70%, sensitivity 96%, negative Haemaglobin predictive value 98%, and positive predictive Haematuria may also be detected in UTI but value 55%. Hence, its principal role is in can also result from a variety of conditions,

rapidly screening out negative specimens. False on September 27, 2021 by guest. Protected copyright. including calculi and neoplasia. The strips negatives may be seen with candida and detect peroxidase activity of haemoglobin and enterococcal infection and false positives with myoglobin, but because ascorbic acid can gross haematuria.18–20 inhibit peroxidase reactions, false negatives may ensue. Haemolysis may result in negative Malthus microscopy. The Malthus system (Malthus Ltd, Stoke on Leucocyte esterase Trent, UK) measures conductance between A chloroacetate stain reacts with the enzyme two electrodes immersed in media. Growth is leucocyte esterase found in primary neutrophil indicated by a change in electrical impedance granules. The detection of pyuria by this test is and has been used for the detection of bacter- reasonably sensitive (72–97%) and may be aemia. Its use in detecting bacteriuria is limited more accurate than microscopy because en- and early evaluations showed that no single zyme activity is still retained when white cells medium allows the detection of all bacteria. have disintegrated. False negatives or reduced The system is not suitable if a preservative such reactions occur in the presence of ascorbic as boric acid is used because this alters acid, boric acid, doxycycline, cefalexin, gen- conductance. If the cut oV time is set at 2.5 tamicin, nitrofurantoin, glycosuria, urobilino- hours (somewhat longer than most rapid gen, or high concentrations of protein; false screening methods), 80% of true positives are positives occur with clavulanic acid, imipenem, detected. However, there is a high false positive 21 or contaminated specimens.14 15 More recently, rate and it has not gained wide acceptance. strip testing of urinary lactoferrin, a protein found in neutrophil nuclei and granules, has Electrochemical detection method been suggested as an alternative to leucocyte This technique, which determines molecular esterase testing.16 hydrogen production using a platinum elec- trode, was first applied to the detection of col- Nitrate reductase (Greiss test) iforms in water. In one study, 94% of positive This enzyme reduces nitrate to nitrite and is urines (> 105 cfu/ml) were detected at four present in coliforms but not other bacteria such hours, but it should be noted that certain

www.jclinpath.com 914 Graham, Galloway

Table 1 Common causes of “sterile” pyuria MICROSCOPY Detection of pyuria Female genital tract infection/non-specific urethritis in male A urinary excretion rate of more than 4 × 105 patients Prostatitis leucocytes/hour is found in 96% of patients Neoplasia of the renal tract with symptomatic bacteriuria, but in only 10% Renal calculi J Clin Pathol: first published as on 1 December 2001. Downloaded from Catheterisation of patients with covert bacteriuria. This is more Renal tuberculosis readily determined by finding > 10 Fever in children, independent of cause leucocytes/ml in uncentrifuged urine; in symp- Prior antimicrobial chemotherapy tomatic infection counts are often much higher than this. However, it should be noted that microorganisms such as staphylococci, strepto- pyuria only indicates inflammation and does cocci, Acinetobacter spp, and Candida albicans 22 not always mean infection. Table 1 lists the require longer detection periods. causes of “sterile” pyuria. Pyuria and/or bacte- ria on microscopy are highly suggestive of UTI Enzyme immunoassay and are useful criteria to select specimens for An indirect immunoassay has been developed direct sensitivity testing.27 However, the ab- to detect specific to antigens on sence of pyuria does not exclude infection bacteria most commonly associated with UTI. because patients with neutropenia may have an Preliminary evaluation found the test to be inadequate white cell response to infection. labour intensive, costly, and to have relatively Alkaline urine, such as that encountered with low sensitivity and specificity (62% and 65%, Proteus spp infection, results in white cells dis- respectively).23 integrating before microscopy being per- formed. Pyuria is considered by some to be a Microcalorimetry 28 Minor changes in temperature as a result of poor predictor of infection. bacterial metabolism can be used to indicate high numbers of bacteria in a urine specimen. Detection of haematuria No large scale studies of this technique have Haematuria is commonly seen in acute cystitis been published.24 but is not diagnostic of that condition. It is rarely seen in other dysuria syndromes but is Photometry often seen in non-infective renal disease. Table Detects bacterial growth based on changes in 2 indicates the importance of other elements light transmission and is more rapid than con- seen on urine microscopy.29 ventional culture techniques. However, if the interval for detection is prolonged to detect certain pathogens such as Pseudomonas aerugi- Detection of bacteriuria http://jcp.bmj.com/ nosa, then there is an increase in false positive Microscopy of uncentrifuged, unstained urine results. Several automated systems are avail- will detect more than 104 bacteria/ml of urine. able.25 Sensitivity increases if the urine is centrifuged and/or Gram stained; this also permits prelimi- Colorimetric filtration nary identification of the pathogenic bacteria. Urine is passed through a filter that traps cells, These methods are labour intensive and 18 Safronin O dye stains these cells which can impractical for routine specimens. on September 27, 2021 by guest. Protected copyright. then be detected visually or using a semiauto- mated colorimetric system. False positive results caused by cellular elements often occur and sensitivity rates fall dramatically when try- Methods ing to detect < 105 cfu/ml. MICROSCOPY For routine use, a semiquantitative method Turbidimetric screening using a flat bottomed microtitre tray and an Measuring turbidity using a double beam inverted microscope is recommended. Pipette turbidometer can screen out negative samples. a fixed volume (80 µl, now the recommended At a 94% sensitivity, 55% of samples are screen volume) from a well mixed sample of urine and negative.26 dispense into a microtitre tray.30 After leaving to settle for 10 minutes, examine each Table 2 Importance of elements seen on direct microscopy of urine well—for example, with a ×32 objective and Elements Importance ×10 eyepiece (FOV number 18)—count the number of white and/or red blood cells, and Cells White blood cells Urinary tract infection or inflammation multiply this by a conversion factor to deter- (see causes of sterile pyuria (table 1)) mine the number of cells/ml or cells/litre. This Red blood cells Urinary tract infection or inflammation method is useful for detecting casts and can be Calculi or neoplasia Eosinophils Acute interstitial nephritis integrated into a multipoint inoculator system. Epithelial cells Contamination of specimen Cell numbers may be more accurately deter- mined by using a counting chamber, but this is Casts Hyaline Normal finding in concentrated urine too labourious for routine diagnostic work. Granular Renal parenchymal disease (non-specific) Reporting accurate numbers of white cells is Red blood cell Glomerulonephritis, vasculitis probably not necessary and bands of reporting, White blood cell Interstitial nephritis, pyelonephritis Epithelial cell Acute tubular necrosis, interstitial nephritis, glomerulonephritis such as 0, < 10, >10, < 50, > 50, < 200 and Crystals Several diVerent types of crystals may develop in the urine including uric > 200, may be more practical. The reporting of acid, calcium phosphate, calcium oxalate, cystine and sulphur. Their red cells as present or absent is usually evaluation is outside the remit of the laboratory suYcient for most clinical practice.

www.jclinpath.com The laboratory diagnosis of urinary tract infection 915

AUTOMATED MICROSCOPY patients. An appropriate selective medium, The DiaSys R/S 2000 is an automated system such as mannitol salt agar with oxacillin (2 mg/ for the microscopic examination of urine. It litre), should be used. simply draws a sample from a specimen Media and culture conditions to aid the iso- container to an optical slide assembly on the lation of more fastidious organisms such as microscope stage. In comparative studies it has lactobacilli, corynebacteria, Gardnerella vagina- J Clin Pathol: first published as on 1 December 2001. Downloaded from a high degree of correlation with standard lis, Mycoplasma spp, and Haemophilus spp may microscopic examination. Laboratory costs can be worth pursuing for symptomatic patients be reduced by adopting this method.31 with pyuria and negative routine culture. The An image processing computer (Yellow use of multipoint inoculation to perform IRIS; International Remote Imaging Systems, routine anaerobic culture has been suggested Chatsworth, California, USA) is able to recog- by some authors as being clinically valuable.35 nise diVerent particle sizes and can be used to A microtitre method may be used in the identi- analyse stop motion pictures from a video fication of enterobacteriaceae, and this tech- camera on a flow microscope. Several types of nique may also be applied to sensitivity testing cellular elements present in uncentrifuged using an automated reader.36 urine are recognised; however, the application of this technology is limited by cost.32 Standard loop method A sterile calibrated loop (1, 2, or 10 µl) is BACTERIAL CULTURE dipped into the urine. The is then The most common cause of urinary tract inoculated and spread. Up to four samples may infection is E coli. In hospital practice, other be inoculated on to a 9 cm plate using one of bacterial species commonly seen include en- the smaller loops. The bacterial count is calcu- terobacter, klebsiella, proteus, pseudomonas, lated from the number of cfu on the plate after enterococci, and staphylococci.33 Staphylococ- overnight incubation and the quantity of urine cus saprophyticus is a common cause of infection originally inoculated. If a 1 µl loop has been in young sexually active women. The labora- used one colony on the plate represents 103 tory must also quantify culture results to deter- organisms/ml in the original specimen. If the mine the clinical relevance of an isolate. Table detection of lower counts is to be achieved a 3 shows a comparison of the organisms isolated larger volume of urine must be used. in the community with those found in hospital. Filter paper method CULTURE METHODS A strip of commercially prepared sterile filter Before culturing the urine should be mixed by

paper is dipped to the prescribed mark and http://jcp.bmj.com/ inverting the container. then, after the removal of excess fluid, touched against an agar plate. Growth is recorded after Choice of media overnight incubation: 25 colonies of bacilli and The media chosen must be able to support the 30 of cocci represent 105 cfu/ml.37 growth of urinary pathogens and possible con- taminants, inhibit Proteus spp from swarming, Multipoint inoculation and distinguish lactose and non-lactose fer- Several specimens are analysed in parallel by menters. Cysteine lactose electrolyte deficient this method, which takes a fixed volume of on September 27, 2021 by guest. Protected copyright. medium (CLED) fulfils these criteria. Culture urine that is then inoculated on to several plates should be incubated overnight at 35– plates. Twenty specimens can be inoculated on 37°C in air. More recently, a new chromogenic to each 9 cm plate. Usually this method will agar has been described for the detection of only detect > 104 organisms/ml. Each plate urinary tract pathogens that may provide better may contain a variety of antibiotics or bio- diVerentiation of bacteria than conventional chemical reagents so that direct sensitivity pro- 34 media. Anaerobes rarely cause UTI but files can be determined and the provisional culture should be considered in a selected identification of most isolates is possible.38 The group of patients, such as those with persistent method may be automated using a Mastascan 35 pyuria or foul urine and symptoms of UTI. In (Mast Group Ltd, Bootle, UK). immunosuppressed patients (including those on intensive care or neonatal units), culture for Other methods Candida spp should be performed because the These include pour plates (serial dilutions of urine may be positive before, or may indicate, urine incorporated into 50°C agar) and roll the development of fungaemia. Screening for tubes (urine poured into an agar coated plastic methicillin resistant Staphylococcus aureus tube); neither of these techniques is suitable for should include urine samples from catheterised routine laboratory use. Table 3 Organisms causing urinary tract infections in the 33 community and in hospital IDENTIFICATION OF BACTERIA Clear protocols for the identification of bacte- Microorganism Community (%) Hospital (%) ria and fungi should be in place. In general, it is Escherichia coli 69.4 50.8 adequate to report coliforms as such without Proteus mirabilis 4.3 5.1 full identification. Proteus spp are Klebsiella/Enteobacter spp 4.7 7.3 spp 5.5 11.9 positive and resistant to nitrofurantoin. Pseu- Staphylococcus spp 4.0 8.4 domonas aeruginosa is an oxidase positive 0 11.1 lactose non-fermenter, resistant to most first Others 12.1 5.4 line antibiotics. If clinical details suggest that a

www.jclinpath.com 916 Graham, Galloway

non-lactose fermenting coliform may be a Sal- evidence of infection (pyuria and/or bacteriu- monella spp and the urease and oxidase tests are ria) so as to permit rapid reporting. This negative then slide agglutinations with “O” and method may be more representative than pick- “H” antisera should be performed from ing individual colonies for subculture, particu- cultures on blood (or other non-selective) agar larly given the heterogenous nature of urinary plates. Any positive results should be followed tract infection.39 The degree of pyuria that trig- J Clin Pathol: first published as on 1 December 2001. Downloaded from up with biochemical confirmation. If typhoid gers the performance of direct sensitivity fever is suspected, 5–10 ml of uncentrifuged testing should be decided locally depending on urine should be inoculated into double the patient group examined. It is suggested that strength selenite, incubated overnight at 37°C all urines that show bacteria on microscopy in air, then subcultured on to deoxycholate cit- and those with pyuria > 100 white cells/mm3 rate agar, which in turn is incubated overnight should be tested. Recent recommendations for at 37°C in air. Any suspect isolate should be disc content and zone size interpretation have dealt with in containment level 3 accommoda- been published by the British Society for Anti- tion. microbial Chemotherapy.40 One advantage of using blood agar alongside Each sensitivity report is tailored to guide CLED is that Gram positive bacteria are more clinicians to the most appropriate agents and it easily characterised. If uncertainty exists, a is often necessary to suppress antibiotics if the catalase test will distinguish streptococci (nega- isolate is not deemed to be clinically relevant. tive) from staphylococci (positive). Staphyloco- Suppressing antibiotic sensitivities on the ccus aureus is DNase, slide, and tube results of positive specimens may be a particu- positive. Staphylococcus saprophyticus can be larly useful way of educating users that identified by its resistance to and treatment of a positive catheter urine is not this makes a useful distinction from other normally warranted. In addition, the presence coagulase negative staphylococci, which are or absence of pyuria may be used to decide usually only important in specific situations which sensitivities are reported. However such as instrumented or catheterised patients. because the definition of UTI is based on bac- If the appearance of the colony is typical of terial counts and not on the presence or Enterococcus faecalis report the organism as absence of pyuria, performing sensitivities such; if uncertain, perform a bile aesculin test. should be related to the number of bacteria â-Haemolytic streptococci can be readily iden- present and the relevant clinical situation. Spe- tified by Lancefield group testing. cific agents may be unsuitable in particular Other isolates are identified using standard situations—for example, the reporting of intra- laboratory techniques, all multiply antibiotic venous antibiotics to general practitioners— resistant organisms need to be fully identified. and certain antibiotics are relatively contra- http://jcp.bmj.com/ Fungi need only be identified if there is evi- indicated in pregnancy. If sensitivity testing is dence to suggest that the isolate is clinically not performed (for example, on mixed cul- relevant. Because cut oV values vary from tures) then culture plates should be kept for author to author, we recommend that repeat five days so that further testing may be sampling is performed to determine that there performed if necessary. is persistent funguria (catheters should be

changed). Candida albicans is germ tube on September 27, 2021 by guest. Protected copyright. positive and usually sensitive to fluconazole, Interpretation and reporting of culture itraconazole, and amphotericin. Sensitivity results testing and the identification of other candida The interpretation of culture results can be and fungal species are only necessary in considered as more of an art than a science. A selected patients, such as those who are urine culture result depends on so many severely immunocompromised. variables, such as appropriate collection, trans- The detection of antimicrobial substances is port, and the limits of the methods of not routinely recommended but should be detection. The reliability of single positive incorporated in the multipoint set to exclude urine culture in diagnosing UTI is only 80%, false negative culture results. The detection of rising to 90% if a repeat culture shows identical coated bacteria in urine is not recom- results.5 Traditionally, > 105 bacteria/ml of mended in the routine diagnostic laboratory urine showing a single isolate is taken to but may be useful to distinguish between upper indicate bacteriuria and distinguishes infection and lower urinary tract infection in selected from contamination in asymptomatic patients. patients. This degree of bacteriuria is usually used in surveillance and epidemiological studies to SENSITIVITY TESTING allow standardisation of data. Mixed culture The choice of agents to test will depend upon with a predominant organism should also be local antibiotic policies and resistance patterns. considered as clinically relevant, although the In general, the primary agents tested target possibility of contamination exists. Counts as coliforms and enterococci, and second line lowas102/ml in symptomatic women are sensitivities need only to be performed if less relevant when enterobacteriaceae are grown, common bacteria or resistant isolates are but this is not necessarily the case with other encountered. The suggested first line agents microorganisms.6 A count of 103/ml is viewed include amoxicillin, trimethoprim, cefalexin as the lower limit of clinical relevance in symp- (or other oral cephalosporins), nitrofurantoin, tomatic men.7 Therefore, pure culture of even a co-amoxiclav, and ciprofloxacin. Urine is used low count of bacteria should always be consid- as the primary inoculum when there is ered as potentially important and sensitivity

www.jclinpath.com The laboratory diagnosis of urinary tract infection 917

Table 4 Common causes of falsely low bacterial counts in However, for most CSUs sensitivity testing on urine specimens several isolates is unnecessary provided culture plates are available for five days in case Colonisation without infection Dilutional through excessive rehydration symptomatic infection develops. Exposure to Acidification or alkalinisation of urine antibiotics in hospital favours alteration of Urinary frequency J Clin Pathol: first published as on 1 December 2001. Downloaded from Concurrent use of a systemic antimicrobial agent and/or other regional flora and the acquisition of resistant growth inhibitor strains which, as a consequence of cross infec- Use of a topical cleansing agent with antimicrobial activity tion, may result in outbreaks. To avoid unnec- during specimen collection Haematogenous infection of the urinary tract essary treatment with antibiotics, suppressing Obstruction of the renal tract distal to the site of infection results of antibiotic sensitivity is recommended Infection with a fastidious or slow growing organism with an appropriate comment such as “catheter associated bacteriuria does not require treat- ment unless there are clinical signs of infec- testing performed if there are appropriate clini- tion” and if appropriate add “sensitivities avail- cal details. Table 4 lists the causes of low bacte- able on request”. rial counts in urine specimens. These recom- mended cut o values are taken from carefully V SPECIMENS FROM CHILDREN conducted studies and in routine diagnostic These may be from a pad specimen or a urine work specimen collection, storage, and trans- bag and therefore contamination can readily portation may be suboptimal. Each laboratory occur. Negative cultures or a slight growth should define which groups of patients or may be diagnostically useful. However, posi- wards or departments warrant additional work tive cultures should be confirmed by a repeat or consideration of low bacterial counts so that specimen, more than 103 cfu/ml is taken as appropriate sensitivity testing and reporting relevant, especially if there is a single isolate.28 can occur. In general, repeat culture should be Further investigation using a DMSA requested before treatment is started in any (99Tc-dimercaptosuccinic acid) scan or ultra- patient in whom the diagnosis of UTI is sound is warranted in children under 5 years to doubtful, or if contamination is suspected. exclude urinary tract abnormalities that might When doubt exists about a culture result, the be amenable to surgery.43 Routine screening of comment “Please repeat if clinically indicated” urine in infants or children is not recom- may be added to the standard comment. Table mended, although if renal abnormalities are 5 indicates a possible schedule for reporting suspected, including those infants who have specimens. Bacterial counts or the presence or been exposed to cocaine in utero, then screen- absence of pyuria cannot be used to localise the ing should be performed.44

level or severity of infection; however, other http://jcp.bmj.com/ laboratory based tests may be used—for exam- URINARY TRACT INFECTION IN PREGNANCY ple, quantification of antibody coated bacteria Asymptomatic bacteriuria occurs in 2–10% of 41 42 in urine or serum C reactive protein values. pregnant women. In the absence of specific treatment it tends to persist and in one third of CATHETER SPECIMENS those aVected progresses to cause acute Asymptomatic bacteriuria commonly develops pyelonephritis.45 Infection may be complicated and approximately 95% of patients will have by low birth weight and prematurity, pre- on September 27, 2021 by guest. Protected copyright. notable bacteriuria (> 105 cfu/ml) after cath- eclampsia, maternal anaemia, amnionitis, and eterisation for two weeks or more. Overt infec- intrauterine death. The treatment of asympto- tions with fever, acute pyelonephritis, and matic bacteriuria is thought to reduce these bacteraemia occur in both short and long term risks.45 The optimal method for screening is catheterised patients. Covert infection occurs urine culture, which should be repeated to in the long term catheterised patient with exclude contamination. Reagent strip testing symptoms such as catheter obstruction, calculi only is less eVective in identifying those formation, periurinary infection, and chronic patients who will develop pyelonephritis, but it interstitial nephritis or pyelonephritis. Speci- does oVer considerable cost savings. In the mens should only be sent when these condi- future, other rapid methods described above tions are suspected clinically. These patients may be useful screening tests to complement may have polymicrobial bacteriuria and should urine culture. not be dismissed, especially if repeat cultures Symptomatic urinary tract infection during from correctly taken specimens are positive. pregnancy usually presents in the second Table 5 Guide to reporting positive urine cultures

Count No. of isolates Specimen type Sensitivity required Comments

>105 cfu/ml 1 or 2 MSU Yes Report with sensitivities CSU Yes Report organism, suppress sensitivities unless appropriate clinical details, CSU comment >2 MSU and CSU No Report as contaminated specimen—repeat if clinically indicated 104–105 cfu/ml 1–2 (not skin flora) MSU and CSU No Report as “no significant growth”, unless the patient is immunocompromised, on relevant antibiotics, or has raised WCC (>50) 1–2 (skin flora) No No significant growth >2 No No significant growth <104 cfu/ml Any MSU and CSU No If > 100 WCC and no antibiotics stated or detected consider culture for fastidious organisms, etc 1–2 Yes Immunocompromised or MSU and clear evidence of symptoms (lower limits of 102 cfu/ml for women with the urethral syndrome and 103 cfu/ml for men) cfu, colony forming units; CSU, catheter specimens of urine; MSU, midstream specimens of urine; WCC, white blood cell count.

www.jclinpath.com 918 Graham, Galloway

trimester. Unlike asymptomatic bacteriuria it catheter should be changed whenever possible. has a low relapse rate after treatment and pro- Some authors recommend alkalinisation of the gression to pyelonephritis is uncommon. Be- urine but consideration should also be given to cause pregnancy itself may result in frequency systemic fluconazole or amphotericin bladder and nocturia it may be diYcult to distinguish washouts or irrigation. Sometimes, cystoscopy symptomatic from asymptomatic bacteriuria. will provide visual proof or histological evi- J Clin Pathol: first published as on 1 December 2001. Downloaded from Several antibiotics used to treat urinary tract dence of candida infection, thereby avoiding infection should be avoided during pregnancy. the need for repeat sampling. Urethral candi- These include aminoglycosides, quinolones, diasis occurs as an extension of candida vagin- tetracyclines, and trimethoprim (first trimes- itis in women and in men as a result of sexual ter). Laboratory reporting of sensitivity data contact.46 should reflect this. Quality issues OTHER PATIENT GROUPS Laboratories should seek Clinical Pathology Asymptomatic bacteriuria in the elderly is not Accreditation (CPA (UK) Ltd) or other appro- associated with an increased mortality rate or priate accreditation for all clinical microbiology with morbidity, such as hypertension or renal work. Inherent in this is the performance of dysfunction. UTI may present atypically—for internal audit, internal quality assessment, par- example, as falls, immobilisation, or ticipation in the National External Quality confusion—and although urine culture should Assurance Scheme (NEQAS), and involve- be considered as part of any geriatric assess- ment in internal quality assessment of near ment, routine screening is not recommended. patient testing facilities. Involvement in the Incidental detection of bacteriuria in men war- “Q” probes and Q Track system of the College rants further investigation, in particular for of American Pathologists allows comparative prostate or bladder outflow problems. Infection audit with other laboratories. Q probes specific is more common in diabetic women, but not in to the diagnosis of UTI include studies on diabetic men or school age diabetics, and rou- transport, handling of urine samples, and urine tine screening in the absence of symptoms is culture contamination.10 47 In addition, The not recommended. Urinary tract infection is Clinical Benchmarking Company (UK) has seen more frequently after renal transplanta- looked at specific UTI projects with a view to tion: it occurs in up to 79% of patients and is improving cost eVectiveness.48 Evidence based often asymptomatic. It can result in graft practice for the reporting of urine culture dysfunction, so routine screening, as part of a results is still unfortunately lacking. regular review, is recommended. All isolates need to be carefully evaluated and repeat Conclusion http://jcp.bmj.com/ cultures requested as appropriate. Screening Because the sensitivities of uropathogens are for UTI is also recommended in patients who relatively predictable, uncomplicated cystitis is require urological procedures, including extra- often treated with empirical short course anti- corporeal shock wave lithotripsy. Many of these biotics without urine culture. If, however, anti- patients may have uteretic stents in place that microbial prescribing is to focus on the also predispose to infection. treatment only of true bacterial infections, as

Human immunodeficiency virus infection, has been recommended in the recent National on September 27, 2021 by guest. Protected copyright. particularly when CD4 counts are below 200 Health Service Executive guidance,49 then cells/mm3, appears to be associated with an urine culture should be performed. The first increased risk of bacteriuria. Other immuno- step in controlling the workload of the suppressed patients (for example, transplant diagnostic laboratory is to have clear and patients, patients on chemotherapy, or high appropriate guidelines for sending urine cul- dose corticosteroids) also have an increased tures from symptomatic and asymptomatic risk of UTI. patients, both in general practice and in hospi- tals, and to be certain that bedside screening CANDIDURIA tests are applied as appropriate. Specimens The kidney is involved in 90% of patients with should be collected and transported in a disseminated candida infection and candiduria correct manner, so that contamination and is an early indicator of systemic candidosis. bacterial overgrowth are minimised. The use of However, Candida spp may colonise the automation will allow large numbers of speci- perineum and urethral meatus resulting in mens to be processed with reduced technical contamination of urine during collection. input, but results are not as well standardised Treatment is started on the basis of other risk as they are for quantitative culture, which per- factors for disseminated infection. Clinically mits the detection of low numbers of bacteria, relevant candiduria in other situations is more mixed samples, and sensitivity results. Thus, diYcult to define and, in the absence of a rec- urine culture will remain the “gold standard” ognised clear cut oV in colony count that until these issues can be dealt with. distinguishes between contamination and in- fection, repeated isolation of candida is a useful 1 Maskell R. Laboratory diagnosis. In: Urinary tract infection in clinical and laboratory practice. London: Edward Arnold, guide in deciding on further evaluation. Risk 1988:26–42. factors for funguria include urinary tract 2 Emmerson AM, Enstone JE, GriYnM,et al. The second national prevalence survey of infection in hospitals— abnormalities, diabetes mellitus, antibiotic overview of the results. J Hosp Infect 1996;32:175–90. treatment, and immunosupression. Chroni- 3 Sussman M. Urinary tract infections. In: Collier T, Balows A, Sussman M, eds. Topley and Wilson’s microbiology and cally catheterised patients may develop asymp- microbial infections, Vol 3, 9th ed. London: Arnold, tomatic candiduria and, in this situation, the 1998:601–21.

www.jclinpath.com The laboratory diagnosis of urinary tract infection 919

4 Kass EH. Asymptomatic infections of the urinary tract. 30 Shepherd M. A revision of the microtitre tray method of Trans Assoc Am Physicians 1956;69:56–63. urine microscopy. PHLS Microbiology Digest 1997;14:231– 5 Kass EH. Bacteriuria and the diagnosis of infections of the 7. urinary tract. Arch Intern Med 1957;100:709–14. 31 Stevens M, Livsey SA, Williams D, et al. Evaluation of the 6 Kunin CM, Van Arsdale White L, et al. A reassessment of R/S 2000 automated microscopy system. Medical Devices the importance of “low-count” bacteriuria in young women Directorate Evaluation Report. London: Her Majesty’s with acute urinary symptoms. Ann Intern Med 1993;119: Stationery OYce, 1993.

454–60. 32 Roe CE, Carlson DA, Daigneault RW, et al. Evaluation of J Clin Pathol: first published as on 1 December 2001. Downloaded from 7 Lipsky BA, Ireton RC, Fihn SD, et al. Diagnosis of bacteriu- Yellow IRIS. An automated method for urinalysis. Am J ria in men—specimen collection and culture interpret- Clin Pathol 1986;5:661–5. ation. J Infect Dis 1987;155:847–54. 33 Grunberg RN. Changes in urinary pathogens and their anti- 8 Stamm WE, Wagner KF, Ansel R, et al. Causes of the acute biotic sensitivities, 1971–1992. J Antimicrob Chemother urethral syndrome in women. N Engl J Med 1980;303:409– 1994;33(suppl A):1–8. 15. 34 Samra Z, Heifetz M, Talmor J, et al. Evaluation of urine of 9 Vernon S, Redfern A, Pedler SJ, et al. Urine collection on new chromogenic agar in detection of urinary tract patho- sanitary towels. Lancet 1994;344:612. gens. J Clin Microbiol 1998;36:990–4. 10 Howanitz PJ, Saladino AJ, Dale JC. Timeliness of urinalysis. 35 Bannon J, Halem MH, Noone M. Anaerobic infection of the A college of American Pathologists Q-probe study of 346 urinary tract: are they being missed? J Clin Pathol 1998;51: small hospitals. Arch Pathol Lab Med 1997;121:667–72 709–10. 11 Hindman R, Tronic B, Bartlett R. EVect of delay on culture 36 Courcol RJ, Deleersnyder H, Roussel-Delvallez, et al. Auto- of urine. J Clin Microbiol 1976;4:102–3. mated reading of a microtitre plate: preliminary evaluation 12 Gillespie T, Fewster J, Masterton RG. The eVect of in antimicrobial susceptibility tests and Enterobacteriaceae specimen processing delay on borate urine preservation. J identification. J Clin Pathol 1983;36:341–4. Clin Pathol 1998;52:95–8. 37 Leigh DA, Williams JD. A method for the detection of 13 Jones C, Bennitt W, Halloran SP. Evaluation of urine reagent significant bacteriuria in a large group of patients. J Clin strips. Medical Devices Directorate Evaluation Report. Pathol 1964;17:498–503. London: Her Majesty’s Stationery OYce, 1998. 38 Faires M, George R, Jolly J, et al. Multipoint techniques for 14 Woolhandler S, Pels RJ, Bor DH, et al. Dipstick urinalysis urine specimens. In: Multipoint methods in the clinical labora- screening of asymptomatic adults for urinary tract tory. PHLS British Society for Microbial Technology, disorders. JAMA 1989;262:1215–24. London: PHLS, 1991:75–9. 15 Beer JH, Vogt A, Neftel K, et al. False positive results for 39 Freeman R, Sisson PR, Burdess B. Heterogeneity within leucocytes in urine dipstick test with common antibiotics. apparently pure cultures of Escherichia coli freshly isolated BMJ 1996;313:25. from significant bacteriuria. J Med Microbiol 1996;45:349– 16 Arao S, Matsuura S, Nonomura M, et al. Measurement of 52. urinary lactoferrin as a marker of urinary tract infection. J 40 Andrews JM, for the BSAC working party on susceptibility Clin Microbiol 1999;37:553–7. testing. BSAC standardised disc susceptibility testing 17 Stevens M, Mitchell CJ, Livsey SA, et al. Uriscreen test for method. J Antimicrob Chemother 2001;48(suppl 1):43–57. urinary tract infection. Medical Devices Directorate Evalua- 41 Galloway A, Green HT, Menon K, et al. Antibody coated tion Report. London: Her Majesty’s Stationery OYce, bacteria in urine of patients with recent spinal cord injury. 1993. J Clin Pathol 1990;43:953–6. 18 Pezzlo MT. Detection of urinary tract infections by rapid 42 Galloway A, Green HT, Windsor JJ, et al. Serial concentra- methods. Clin Microbiol Rev 1988;1:268–80. tion of C-reactive protein as an indicator or urinary tract 19 Pezzlo MT, Ige V, Woolard AP, et al. Rapid bioluminescence infection in patients with spinal injury. J Clin Pathol 1986; method for bacteriuria screening. J Clin Microbiol 1989;27: 39:851–5. 716–20. 43 Friedman AL. Urinary tract infection. Curr Opin Pediatr 20 Males BM, Bartholomew WR, Amsterdam D. Leukocyte 1998;10:197–200. esterase-nitrate and bioluminescence assays as urine 44 Gothbrath-Flaherty EK, Agarwal R, Thakar R, et al. screens. J Clin Microbiol 1985;22:531–4. Urinary tract infections in cocaine-exposed infants. J Peri- 21 Smith TK, Eggington R, Pease AA, et al. Evaluation of the natol 1995;15:203–7. Malthus 128H microbiological growth analyser for detect- 45 Berrington AW, Bint AJ. Diagnosis and management of uri- ing significant bacteriuria. J Clin Pathol 1985;38:926–8. nary tract infection in pregnancy. Reviews in Medical Micro- 22 Lamb VA, Dalton HP, Wilkins JR. Electrochemical method biology 1999;10:27–36. http://jcp.bmj.com/ for the early detection of urinary-tract infections. Am J Clin 46 Munoz P, Gijon P, Alcala L, et al. How to manage Pathol 1976;66:91–5. candiduria. Rev Med Microbiol 1999;9:225–31. 23 Stevens M, Mitchell CJ. Evaluation of the URISTAT test for 47 Valenstein P, Meier F. Urine culture contamination. Arch the diagnosis of urinary tract infection. Medical Devices Pathol Lab Med 1998;122:123–9. Directorate Evaluation Report. London: Her Majesty’s 48 Lauder L. Quality assurance. In: Galloway M, ed. Clinical Stationery OYce, 1993. director of pathology. Tackling the role. Oxford: RadcliVe 24 Beezer AE, Bettelheim KA, Newell RD, et al. The diagnosis Medical Press, 1999:81–93. of bacteriuria by flow microcalorimetry, a preliminary 49 NHS Executive. Resistance to antibiotics and other anti- investigation. Science Tools 1974;21:13–15. microbial agents. HSC 1999/049, March 1999. London: 25 Pezzlo MT, Tan GL, Peterson EM, et al. Screening of urine NHS publications. cultures by three automated systems. J Clin Microbiol 1982; on September 27, 2021 by guest. Protected copyright. 15:468–74. 26 Livsey SA. Turbimetric urine screening. Br J Biomed Sci Further reading 1995;52:71–3. 27 Vickers D, Ahmad T, Coulthard MG. Diagnosis of urinary Eisenstadt J, Washington JA. Diagnostic microbiology for tract infection: fresh urine microscopy or culture. Lancet bacteria and yeasts causing urinary tract infections. In: Mobley 1991;338:767–70. HLT, Warren JW, eds. Urinary tract infections. Molecular 28 Robbins DG, White RHR, Rogers KB et al.Urine pathogenesis and clinical management. Washington: ASM Press, microscopy as an aid to the detection of bacteriuria. Lancet 1975;i:476–8. 1996:29–66. 29 Barri YMH, Shah SV. Approach to the patient with renal Lewis DA. Bacteriology of urine. In: Hawkey PM, Lewis DA, disease. In: Andreoli TE, ed. Cecil essentials of medicine,5th eds. Medical bacteriology: a practical approach. Oxford: Oxford ed. Philadelphia: WB Saunders, 2001:232–7. University Press, 1989:1–19.

www.jclinpath.com