Scand J Clin Lab Invest 2000; 60: 1 ± 96

European Urinalysis Guidelines SUMMARY

These European Urinalysis Guidelines are given Classi®cation of examiniations under the auspices of the European Confedera- Examinations have been re-classi®ed into four tion of Laboratory Medicine (ECLM). hierarchical levels based on accuracy of mea- surements (chemistry on Page 12, particle Medical needs for urinalysis analysis on Page 23, microbiology on Page Urinalysis should always be performed on the 31). In addition, the previous literature on the basis of medical need, and appropriate exam- visual appearance and odour of urine is inations for various clinical populations and presented (Page 13). presentations should be determined by cost/ bene®t analysis. On this basis, selection indica- Chemical methods of examination tions for urinalysis are suggested for detecting Principles and performance criteria of multiple diseases of the kidneys or urinary tract (Page 6). test strips are reviewed (Page 13). A nitrite test The referral for detailed urine examinations should not be used alone in detecting urinary should include an adequate description of the tract infections because of its low sensitivity. specimen type, and should inform the labora- Quali®ed procedures for measurement are tory of the clinical need in order to facilitate recommended (Page 56). Pregnancy examina- correct selection of examination procedures and tions are reviewed brie¯y (Page 18). interpretation of results (Page 6). This medical Quantitative chemical measurements are dis- information is mandatory when the choice is cussed in detail, mainly with respect to protein being made between minimum and optimum measurements (Page 18). Measurements asses- procedures for different specimens. The mini- sing volume rate (diuresis) are also summarized mum information needed is outlined (Page 48). (Page 19). Reference intervals, existing reference Requirements for computerized information materials (calibrators) and measurement proce- systems are discussed (Page 49). dures are quoted (Page 57). Automation can be applied in centralized Patient preparation laboratories after appropriate evaluation of analytical equipment (and pre-analytical proce- Preparation for a high-quality urine specimen dures). Local diagnostic requirements guide the should start on the previous evening and should manner of implementation of point-of-care be as uniform as possible to allow standard methods, as well as manual or automated interpretation of results (Page 7). Details of procedures in different laboratories. patient preparation before specimen collection is given, ending with quali®ed specimens when- Particle analysis ever possible (Page 8). First and second morning urine specimens are de®ned (Page 7). Clinically signi®cant particles in urine are reviewed and classi®cation is divided into basic and advanced levels (Page 20), one of Collection of specimens, preservation and which should be selected by each laboratory (or transport by its subunit, such as emergency services versus Recommended methods of urine collection are regular-hour working personnel). listed (Page 9), including detailed illustrations For routine particle identi®cation, a standar- for the collection of mid-stream specimens dized procedure with phase-contrast microscopy (Page 50, 91 ± 96). Speci®cations for collection or supravitally stained urine sediment is recom- and transport containers are given (Page 10), mended (Pages 23, 62). Morphological criteria of and preservation procedures for chemical mea- particles are given as investigated under a cover- surements, particle analysis and microbiological slip with a known volume of urine (Page 64). examinations are listed (Page 51). Urine cytology for investigation of cancer cells is 1 2 ECLM ± European Urinalysis Group considered to require special expertise. For Automation of bacterial cultures may be microbiology, different procedures apply because considered in large microbiological laboratories bacterial infection usually needs to be detected in based on evaluation of instrument performance ordinal scale. That is why the slide-centrifugation against an appropriate comparison method. technique and Gram staining are recommended as the comparison method (Page 23). Stepwise strategies in urinalysis Instruments supplement clinical particle ana- Cost-bene®t analyses should guide screening lysis by reducing manual work (Page 25). policies. Multiple examinations recommended Evaluations against an appropriate comparison for sieving of general patient populations method are recommended. should include measurements of leukocytes, Microbiology methods bacteria, erythrocytes, albumin (protein), and of quantity for volume rate (~ diuresis; such as It is not necessary to request microbiological creatinine or other) (Page 34). In acute cases, examination of urine in all clinical situations. measurements of glucose, ketone bodies and pH Low-risk symptomatic patients consist of adult can be of additional value. A ¯ow-chart is given females with recurrent dysuria/urgency symp- for additional investigations in the case of toms without fever, and without known diseases positive ®ndings (Page 36). It includes questions predisposing to urinary tract infection (Page of clinical need. 26). Urine cultures from these cystitis patients Use of a ¯ow-chart is suggested to reduce the are discouraged for routine purposes ± they are number of less important urine cultures (Page necessary, however, for epidemiological pur- 38). After clinical evaluation, a highly speci®c poses. Rapid methods for detecting bacteriuria laboratory examination should be undertaken can be used in acute treatment decisions (Page as a ®rst step. Remaining negative specimens 26) after appreciating the possibility of false- shall then be examined with a sensitive pro- negative ®ndings. All other symptomatic cedure to exclude true negative cases. The patients belong in the high-risk group, whose development of new examinations with high specimens should be sent for bacterial culture, sensitivity (and speci®city) for rapid bacterial including identi®cation of species and anti- detection is encouraged to reduce the number of microbial susceptibility testing. routine cultures (Page 34). Asymptomatic We provide a new classi®cation of bacteria individuals should not be screened for bacter- based on the uropathogenicity and frequency of iuria, with the exception of pregnant women bacteria in the aetiology of urinary tract and some other speci®ed patient groups (Page infection (Page 26, Table IX). Limits of 38). Increased local co-operation with clinical signi®cant bacteriuria are de®ned more precisely units and detailed information from individual than previously depending on the type of specimens is needed to ®rst de®ne and then specimen, clinical presentation and organism select the minimum operating procedures or the (Page 30, Table XIII). For selected patient enhanced procedures for each specimen in groups, large 10 mL or 100 mL inocula are routine work-¯ow. needed to reach the required sensitivity. A new Increased detection sensitivity for early renal standardized unit for reporting quantitative damage is recommended because of the prog- cultures is recommended based on the develop- nostic signi®cance in diabetes mellitus, probably ing particle identi®cation techniques, adapting also in hypertension, and in monitoring patients international standardization of nomenclature receiving nephrotoxic drugs (Page 39). New and volume: colony forming bacteria/litre patient groups should be incorporated into (CFB/L) (Page 28). strategies using these sensitive measurements Details of different culture procedures are when prevention of long-term deterioration of given (Page 68). Procedures for species identi- renal function is anticipated. ®cation are quoted for routine microbiology laboratories (Page 72). Antimicrobial suscepti- Quality assurance, including analytical quality bility testing is reviewed, with acceptance of speci®cations direct susceptibility testing for rapid test positive cases (clear positives) in experienced Procedures for quality assurance and new hands (Page 32). analytical quality speci®cations are suggested European Urinalysis Guidelines 3 for rapid (ordinal scale) examinations (Page 42), document is limited to the most commonly particle analysis (Page 44) and microbiological requested examinations. Some analytes are examinations (Page 45). For ordinal-scale discussed in terms of specimen collection only, examinations they are based on detection and since the precise medical needs and actual con®rmation limits, as compared with a quan- analytical methods concerned would have titative reference procedure (Page 53), or made this document too large for everyday calculation of agreement based on k statistics use. The present recommendations have made (Page 55). Analytical quality speci®cations for use of some of the existing national guidelines protein and other quantitative chemical mea- [2 ± 7] as appropriate. Certain recent analytical surements are summarized based on earlier techniques, such as ¯ow cytometry and nucleic literature (Page 44). Implementation of elements acid ampli®cation, have widened the perspec- of a structured quality system (good laboratory tives of information obtainable from urine. The practice (GLP) with standard operating proce- latest technology is referred to but not described dures (SOP) as documented in a quality in detail. manual) is encouraged even in small labora- Target audiences: These guidelines are aimed tories, based on national co-operation (Page mainly at laboratory professionals in small and 40). general laboratories and at points-of-care. Spe- These guidelines should impact on the cial features from both clinical chemistry and validation of local working practices. In addi- microbiology are included in the appropriate tion, the principles of quality embodied in these sections. The text is divided into a general guidelines may assist independent authorities section and an appendix containing detailed charged with accrediting laboratories. methodological descriptions, to make it acces- sible to clinicians and others not directly concerned with such detail. Wherever the 1. INTRODUCTION location, the performance of examinations must be determined by clinical requirements 1.1. Foreword and this means close co-operation between the traditional laboratory disciplines and clinicians. An effective diagnostic strategy from urine The large number of specimens and the need for should be based on standard procedures emergency reports increase the pressure for for collection, transport and analysis. These rapid and reliable measurements. At the same standardized procedures are required to pro- duce consistent reference intervals and decision time, opportunities for cost-containment must limits for the harmonized interpretation of be identi®ed. These guidelines will help establish results. Europe has no consensus standard quali®ed procedures for laboratory examina- procedures. Standardization is essential not tions in these environments. only for interpretation of results in individual Process: This document is based on the work patients, but also for epidemiological studies, of the European Urinalysis Group (EUG), for determining which populations should be established under the auspices of the European screened for urinary abnormalities, and for the Confederation of Laboratory Medicine procedure to be followed when an abnormal (ECLM) in 1997 [8, 9]. The EUG has result is found [1]. In addition, laboratories collaborated with the Working Party on want to accredit their urine diagnostics by Urinalysis of the European Society for Clinical comparing their methods with acceptable refer- Microbiology and Infectious Diseases ences. These guidelines therefore intend to ®ll (ESCMID) in the preparation of the guidelines. this gap by summarizing available knowledge The ``European Urinalysis Guidelines'' are into one consensus practice for urinalysis (or based on both medical needs and available urine analysis) in Europe. methods and technologies. In collaboration Selected list of analytes: The terms ``urin- with many experts from most European coun- alysis'' and ``urine analysis'' are synonymous. tries (see the list below) the present paper The combination of analytical procedures used attempts to create consensus practice for in practice is changing and varies in different urinalysis. The EUG wishes to express its clinical situations. To form a basis, this gratitude to the various sponsors listed in 4 ECLM ± European Urinalysis Group

Section 1.2, without whose help this document Prof. Dr. Walter G. Guder could not have been produced. Institute for Clinical Chemistry StaÈdtisches Krankenhaus Bogenhausen Englschalkinger Strasse 77 D-81925 Munich 1.2. Members of the ECLM-European Germany Urinalysis Group and ESCMID Working Phone: z49 89 9270 2280 Party, Other Contributors and Sponsors Fax: z49 89 9270 2113 Members of the ECLM ± European Urinalysis E-mail: [email protected] Group Dr. Timo Kouri (Chairman) ESCMID Working Party on Urinalysis Centre for Laboratory Medicine Tampere University Hospital Vanya Gant, London, United Kingdom P.O. Box 2000, FIN-33521 Tampere (Moderator) Finland Hans Hallander, Stockholm, Sweden Phone: z358 3 247 5484, Fax: z358 3 247 5554 Francis O'Grady, Nottingham, United E-mail: Timo.Kouri@tays.® Kingdom Marc Struelens, Brussels, Belgium Dr. Giovanni Fogazzi Divisione di Nefrologia e Dialisi Other contributors Ospedale Maggiore, IRCCS Amores C, Jaen, Spain Via Commenda 15, I-20122 Milano Anyszek Tomasz, Krakow, Poland Italy Arnadottir M, Reykjavik, Iceland Phone: z39 02 5503 4557 Aspevall O, Huddinge, Sweden Fax: z39 02 5503 4550 Baerheim A, Bergen, Norway / Gent, Belgium E-mail: [email protected] BeÂne M-C, Nancy, France Blaseio U, Hamburg, Germany Dr. Vanya Gant Blondiau R, Erembodegem, Belgium Department of Clinical Microbiology Bondarenko I, St Petersburg, Russia University College Hospital Colombo JP, Bern, Switzerland Grafton Way, London WC1E 6DB Delanghe J, Gent, Belgium United Kingdom Dybkaer R, Copenhagen, Denmark Phone: z44 171 380 9913 Edel H, Munich, Germany Fax: z44 171 388 8514 Emanuel V, St Petersburg, Russia E-mail: [email protected] Fang-Kirchner S, Vienna, Austria Fenili D, Romano di Lombardia, Italy Dr. Hans Hallander Fernandes B, Toronto, Canada Swedish Institute for Infectious Disease Control Fuellemann R, Mannheim, Germany Unit of Quality Assurance Garcia JL, Sevilla, Spain SE-17182 Solna Gatermann S, Bochum, Germany Sweden Grubb A, Lund, Sweden Phone: z46 8 457 2490 Fax: z46 8 3025 66 GyoÈry AZ, Sydney, Australia E-mail: [email protected] Hesse A, Bonn, Germany Hoeiby N, Copenhagen, Denmark Dr. Walter Hofmann HofgaÈrtner W, Oxford, United Kingdom Institute for Clinical Chemistry Hyltoft-Petersen Per, Odense, Denmark StaÈdtisches Krankenhaus Bogenhausen Ito K, Yokohama, Japan Englschalkinger Strasse 77 Itoh Y, Osaka, Japan D-81925 Munich Ivandic M, Munich, Germany Germany Kermes K, Tartu, Estonia Phone: z49 89 9270 2282 Khorovskaya L, St Petersburg, Russia Fax: z49 89 9270 2113 Kimling H, Mannheim, Germany E-mail: [email protected] Kokot F, Katowice, Poland European Urinalysis Guidelines 5

Kontiainen S, Helsinki, Finland BIPM International Bureau of Weights and Kratochvila J, Nymburg, Czech Republic Measures Kutter D, Junglinster, Luxembourg BSAC British Society for Antimicrobials and Lapin A, Vienna, Austria Chemotherapy Lipsic T, Bratislava, Slovak Republic CAMP Christie-Atkins-Munch-Petersen factor Malakhov V, Moscow, Russia CAP College of American Pathologists Naber KG, Straubing, Germany CDC Centers for Disease Control (U.S.A.) Naegele U, Mannheim, Germany CEN European Committee for Standardization Newall R, Newbury, United Kingdom CFB Colony-forming bacteria (instead of the Newman DJ, London, United Kingdom traditional colony-forming unit, CFU) Parker D, Elkhart, IN, U.S.A. CLED Cystine-lactose electrolyte de®cient (agar) Peheim E, Bern, Switzerland CNS Coagulase-negative staphylococci Penev MN, So®a, Bulgaria CRM Certi®ed Reference Material Piemont Y, Strasbourg, France CV Coef®cient of variation (relative standard Ponticelli C, Milano, Italy deviation) Price C, London, United Kingdom EC Enzyme Commission (number of classi®ca- Pugia M, Elkhart, IN, U.S.A. tion) Rada F, Tirana, Albania ECCLS European Committee for Clinical Recio F, Sevilla, Spain Laboratory Standards Rowan RM, Glasgow, United Kingdom ECLM European Confederation of Laboratory Rowlands T, Meylan, France Medicine Salum T, Tartu, Estonia EN European Standard Sandberg S, Bergen, Norway EQA External Quality Assessment Selgren S, Elkhart, IN, U.S.A. EQAS External quality assessment scheme (or Siitonen A, Helsinki, Finland survey) Sikirica M, Zagreb, Croatia ESCMID European Society of Clinical Micro- Steinhausen R, Mannheim, Germany biology and Infectious Diseases Tetz VV, St Petersburg, Russia EUG European Urinalysis Group Tzontcheva A, So®a, Bulgaria FDA Food and Drug Administration (U.S.A.) Verhaegen J, Leuven, Belgium FESCC Federation of European Societies of Zaman Z, Leuven, Belgium Clinical Chemistry FN False negative Sponsors: FP False positive A. Menarini Diagnostics GBS Group B streptococci Bayer Diagnostics Inc. GFR Glomerular ®ltration rate BD Microbiology and Vacutainer Systems Inc. GLP Good Laboratory Practice Bio-Rad Laboratories Inc. hCG Human chorionic gonadotropin Labquality Ltd HPF High-power ®eld (in a microscope) ROCHE Diagnostics GmbH IEC International Electrotechnical Commission Sarstedt AG & Co IFCC International Federation of Clinical SYSMEX EUROPE GMBH Chemistry IQC Internal quality control The text does not necessarily re¯ect the detailed ISO International Organization for Standardi- opinions of any of the contributors or sponsors zation since it is the product of a consensus process. IUPAC International Union of Pure and Applied Chemistry IUPAP International Union of Pure and 1.3. Abbreviations Applied Physics IVD In Vitro Device AACC American Association for Clinical JCCLS Japanese Committee for Clinical Chemistry Laboratory Standards AMA Antimicrobial activity kDa Kilodaltons ASM American Society for Microbiology KIA Kligler Iron Agar 6 ECLM ± European Urinalysis Group

LC Con®rmation limit epidemiology of target diseases should be LD Detection limit considered: for example, screening for incipient NCCLS National Committee for Clinical nephropathy by detecting albuminuria in Laboratory Standards patients with diabetes mellitus is recommended NIST National Institute of Standards and world-wide [10, 11]; on the other hand, that for Technology schistosomiasis in children should be restricted NRSCL National Reference System for Clinical to those countries where this parasite is endemic Laboratories [12]. General indications for urinalysis are given OIF oil immersion ®eld in Table I. OIML International Organization of Legal Clinical presentations in these individuals can Metrology vary widely from asymptomatic ambulatory RBC red blood cells patients to high-risk immunosuppressed indivi- RBP retinol-binding protein duals with life-threatening complications. No RCF relative centrifugal force age range is exempt. Clinical need may dictate ROC receiver operating characteristic (curve) an urgent examination with a turnaround time RPM rotations per minute of less than 0.5 ± 2 h, rather than a con®rmatory SOP Standard Operating Procedure examination the result of which comes too late SRM Standard Reference Material for decision-making. The repertoire of local SRGA-M Swedish Society of Medicine's Refer- laboratory or point-of-care environments will ence Group for Antibiotics also in¯uence the selection of laboratory TQM Total quality management examinations that can be performed. Some of TSIA Triple Sugar Iron Agar these considerations are discussed in Section 8 URL Upper reference limit (of a speci®ed (Stepwise strategies in urinalysis). reference interval) VIM International vocabulary of metrological Request and report information. The format of terms the request for and the report of urinalysis is VP Voges-Proskauer test in¯uenced by the site of examination: at WBC white blood cells point-of-care, specimen collection and analysis results can be documented directly into the patient record, whereas a remote laboratory 2. MEDICAL NEEDS FOR always needs a paper or computerized request. URINALYSIS The request reaching the laboratory may initi- ate a stepwise procedure agreed locally for a After a long history of clinical urinalysis there is particular patient group. Such pre-determined a need to update the medical relevance of strategies maximize diagnostic yield while different investigations of urine. Cost/bene®t maintaining cost-ef®ciency. analyses should guide the implementation of The importance of adequate clinical and examinations for various populations. The specimen-related information for correct selec-

a TABLE I. Medical indications for urinalysis.

(1) Suspicion or follow-up of symptoms or situations suggesting the possibility of urinary tract infection (2) Suspicion or follow-up of non-infectious renal disease, either primary or secondary to systemic diseases, such as rheumatic diseases, hypertension, toxaemia of pregnancy, or to the adverse effects of drugs (3) Suspicion or follow-up of non-infectious post-renal disease (4) Detection of glycosuria from speci®ed patient groups, e.g., individuals admitted to hospital for various medical emergencies, or from pregnant women (5) Follow-up of only selected diabetes mellitus patients, e.g., children at home, to detect morning glycosuria and ketonuria in addition to blood glucose measurements (6) Detection or follow-up of selected metabolic states, e.g., vomiting and diarrhoea, acidosis/alkalosis, ketosis, or recurrent urinary stone formation

a If understood widely, urine quantities are measured in diagnostics of several endocrine, metabolic and inherited diseases, pregnancy, drugs of abuse, etc., most of which were not discussed in these guidelines which focus mainly on diseases of kidneys and urinary tract. European Urinalysis Guidelines 7 tion of examination procedures and interpreta- compliance and rapid transportation to the tion is generally underestimated. Suf®cient laboratory can be organized. detail is seldom documented for urinalysis Second morning urine is a single specimen specimens. The minimum information is pro- voided 2 ± 4 h after the ®rst morning urine. In posed in the Appendix (Annex 10.1). contrast to the ®rst morning urine, its composi- The analysis report may be a structured note tion may be affected by prior ingestion of food entered directly into the patient's record when and ¯uids and by movement. However, it may examination is performed at point-of-care. The be more practical for ambulatory patients. To remote laboratories must use a standardized increase the sensitivities of bacterial culture and report format with de®ned units, reference particle counting, the quality of the second intervals and report style (Appendix, Annex morning urine should be improved by allowing 10.1.3). The purpose of a report is to guide the ingestion of only one glass of water (200 mL) clinician towards rational and evidence-based after 22:00 on the previous evening and diagnostics and therapeutic management. It extending this abstinence up to the time of must be technically correct and must provide specimen collection. A bladder incubation time unambiguous information. exceeding 4 h is possible with this ¯uid restric- tion. Postural proteinuria cannot, however, be prevented and should be further investigated by examining a ®rst morning urine sample if 3. PATIENT PREPARATION diagnostic problems occur. If these standardized collection instructions have not been followed, 3.1. De®nitions of urine specimens based on the second morning urine is classed as a timing ``random'' specimen. Timed collection of urine is collected at a The following timed types of urine specimens speci®ed time in relation to another activity, are described by modifying the de®nitions e.g., therapy, meals, daytime or bed-rest. A 24-h quoted in textbooks [13 ± 15] and national urine collection contains all portions voided over guidelines [2 ± 7]. The time of specimen collec- 24 h. A timed 24-h collection can be started at tion must be recorded both on the examination any time of the day by emptying the bladder request and on the subsequent report to aid in and noting the time. All urine during the next the correct interpretation of ®ndings. 24 h is then collected and preserved as appro- Random urine is a portion of single voided priate for each analyte (see Appendix, Annex urine without de®ning the volume, time of day 10.3.4). or detail of patient preparation. This is usually Timed overnight urine is collected by emptying the unavoidable case in acute situations. the bladder just before going to bed, noting the Random urine specimens are associated with exact time, and then collecting all urine portions many false-negative and some false-positive during the bed-rest period. At the end of the results. period, the last portion is collected, the exact First morning urine is the specimen voided time recorded and the total volume of overnight immediately after an overnight bed-rest before urine noted. The specimen or a representative breakfast and other activities. This is also called aliquot is then sent to the laboratory. early morning urine. It is recommended that the early morning urine be voided after an 8-h period of recumbency, and after not less than 3.2. Patient preparation before specimen 4 h storage time in the urinary bladder (even if collection the bladder was emptied earlier during the The patient should be told why a urine night). This has been traditionally recom- specimen requires to be examined and given mended as the standard specimen for urinalysis, instructions on how it should be collected (see because it is more concentrated than day urine Section 4.1). Ideally, the instructions should be and allows time for possible bacterial growth in given both orally and in written form accom- the urinary bladder. This specimen is most panied by illustrations where possible to ensure easily collected from hospitalized patients, but uniformity of the collection procedure (see may be collected even at the patient's home if Appendix, Annex 10.3.1). Because the same 8 ECLM ± European Urinalysis Group specimen is often used for both microbiological provided by diet (e.g., salt and phosphate), and chemical measurements, the instructions but increases the excretion of metabolites should combine both requirements. associated with catabolism, e.g., ketone bodies and ammonia (16). In detecting glycosuria, a Transmission of pre-analytical information. carbohydrate meal before the specimen collec- Adequacy of patient preparation and type and tion improves the clinical sensitivity (postpran- success of specimen collection can be coded dial urine). In general, fasting before urinalysis on the label on the sample tube. When avail- is designed to reduce diuresis. The preparation able, this information should ultimately be of patients for standardized fasting blood transmitted to the patient records together specimens can be combined with that for with the results of examinations, e.g., ``quali- urinalysis. ®ed''/``standard'', or ``random''/``non-stand- ard'' specimen, to increase the reliability of 3.3.2. Exercise and body posture. Wide biologi- interpretation. Thus, signi®cant deviation from cal variation in urine composition is related to ``standard'' should be recorded on the request physical activity and body posture. Exami- form or label added to the container as nation of the morning urine and avoidance of ``non-standard''. strenuous physical exercise minimizes these in¯uences. Urinary calcium excretion increases 3.3. In¯uence and interference factors more than twofold on immobilization of a patient by bed-rest (18). Exercise may increase Biological (in vivo) factors, changing the true the amount of body constituents excreted by concentration of a measured component, cause increasing glomerular ®ltration as a result of problems in the interpretation of laboratory increased blood pressure (e.g., appearance or results, although the measurement process itself increase in albuminuria or haematuria). is correct. These are called in¯uence factors (discussed in detail below). In addition, there are other factors that 3.3.3. Incubation time in the bladder. To technically interfere with the analytical demonstrate reliable bacterial growth, classical method applied (called interference factors, microbiological advice has been to allow bac- 16). The latter are particularly important with teria a log phase of growth by incubating the non-speci®c analytical methods, e.g., those used urine in the bladder for 4 ± 8 h [19]. Urine is a in traditional test strip ®elds (see Annex 11.1), good culture medium for many bacteria. but also in other chemical measurements from Using the classical Griess's examination urine (17). As a classical example from micro- (applied in the nitrite examination pad on test biology, the specimen for bacterial culture strips), it has been shown that the ®rst morn- should be obtained before antimicrobial treat- ing urine specimen is more sensitive in detect- ment is initiated to allow bacterial growth. ing asymptomatic bacteriuria among pregnant women than later specimens [20]. In most 3.3.1. Volume rate (diuresis) and fasting. recent guidelines for the evaluation of new Many urine constituents change in concentra- anti-infective drugs, no incubation times are tion when the urine volume rate (excretion mentioned [21]. The urgency of micturition rate of water~diuresis) alters due to variation associated with acute infection will often not in ¯uid intake, reduction of renal concentrat- permit suf®cient bladder incubation time ing ability or ingestion of diuretic substances. before voiding, resulting in false-negative cul- If sensitive screening is needed, a low volume tures. For chemical analyses, no particular rate (20 ± 50 mL/h) is desirable to produce incubation time is necessary. In studying cell highly concentrated specimens. This is best and cast morphology, the best results are achieved in morning urine. Documentation of obtained with a short incubation time of 1 ± the urine concentration is recommended, e.g., 2 h, provided that a high volume rate does density (speci®c gravity or weight), osmolality, not lead to false-negative results (rare particles or a related analyte, which allow calculations are seen more often in concentrated urine spe- of analyte/reference ratios (see Section 5.4.2). cimens). For patients without urgency or Starvation decreases urinary constituents acute symptoms, advice concerning bladder European Urinalysis Guidelines 9 incubation time, and a record of that time, is tegration of diagnostically valuable formed still recommended to reach the highest sensi- elements. tivities [22]. Mid-stream urine (or clean-catch urine) characterizes the middle portion of a voided 3.3.4. Contamination. Urine specimens should specimen. The ®rst portion of urine is not be collected free of internal and external con- collected, because it is always contaminated by taminating ¯uids. Sexual intercourse should be commensal urethral ¯ora in both sexes (if not avoided for 1 day before specimen collection purposely collected, see below). Minimizing because of the resulting increased amounts of contamination of specimens by commensal proteins and cells. bacteria is especially important, and requires Urine from males is usually contaminated detailed patient advice. It is important to with small amounts of secretory products from wash the introitus around the urethra in the prostate. Excretion of N-acetyl-b,D-glucos- females, and the glans penis in males with aminidase, a marker of renal tubular function, water only, before micturition. This reduces is increased for 3 days after ejaculation [23]. false-positive urine cultures by 20% or more Seminal ¯uid may contaminate urine after [24]. The use of antiseptics ± and soaps (with normal ejaculation, but also in diseases asso- variable additives) ± is not recommended as ciated with retrograde ejaculation to urinary this may affect bacterial viability [25]. The last bladder. portion is also left over after collecting Vaginal secretions or menstrual blood may 50 ± 100 mL of urine in the container. Detailed contaminate urine from females. This may be instructions for mid-stream urine collection are misleading and can be minimized by tamponing included in the Appendix (Annex 10.3.1). the vagina when acute symptoms necessitate Technically satisfactory collection of urine examination of urine. Pregnancy is associated from children and from women especially in with physiological pyuria. the late stages of pregnancy can be dif®cult. The use of different sterile devices may help women to urinate more easily into a collection con- tainer. 4. COLLECTION OF SPECIMENS, First-void urine characterizes the ®rst portion PRESERVATION AND TRANSPORT of urine voided at the beginning of micturition. This is the optimal sample for the detection of Urinalysis may be requested on specimens Chlamydia trachomatis by nucleic acid ampli®- obtained by voiding (micturition), by catheter- cation. It is NOT suitable for routine micro- ization, needle puncture, through a postopera- biological culture as urethral organisms usually tive urostomy, or by using different collection contaminate it. vessels, such as bags or special receptacles for Single catheter urine is collected after insert- bed-bound patients. The most commonly ing a sterile catheter into the bladder through obtained specimen is the mid-stream urine. the urethra (straight or ``in-and-out'' catheter- Requesting clinical staff must document the ization). For children without urinary control, method of collection to allow correct inter- this is one of the methods of con®rming or pretation of results. excluding the presence of urinary tract infection [26]. Indwelling catheter urine is collected at 4.1. Procedures for urine collection, specimen replacement or by sterile puncture of an types indwelling catheter. Urine analysis specimens The microbiological requirements usually must not be taken from the collection bag of a determine the details of any collection permanent indwelling catheter. method, because urine intended for microbio- Suprapubic aspiration urine is usually col- logical examination must avoid or minimize: (1) lected by sterile aspiration of urine through the contamination of specimens by commensal abdominal wall from a distended bladder. The bacteria, (2) growth of bacteria following bene®t of this technique is that it allows a clear- specimen collection, (3) damage or death of cut decision on the presence or absence of diagnostically relevant bacteria, and (4) disin- urinary tract infection [27]. Detailed instruc- 10 ECLM ± European Urinalysis Group tions are provided in the Appendix (Annex ± identi®cation of any bacteria to low levels 10.3.3). The risk of bladder colonization by (w104 CFB/L) suprapubic aspiration is lower than that by ± large inocula of urine to ensure accurate in-and-out catheterization. Miniaturization of plate counts (10 ± 100 mL) measurement techniques by laboratories may ± report simultaneously relative concentrations nowadays allow several different examinations at each anatomical site from the 5 mL of urine obtained typically by ± antibiotic sensitivities of any bacteria grown the suprapubic aspiration technique. Bag urine is widely collected from small The lower urinary tract may be the subject of infants, but carries a high probability of investigation, typically the prostate. For diag- contamination with skin organisms. The entire nosis of chronic bacterial prostatitis, the Meares genital region should be washed carefully with and Stamey collection procedure is still recom- water. A sterile collection bag is applied and the mended: ®rst and mid-stream portions of urine should be collected, then drops expressed with urine ¯ow checked frequently. The collection bag should be in situ for a maximum of 1 h, prostate massage, and the ®nal specimen voided after which the probability of contamination after prostatic massage [29, 30]; (see Appendix, increases greatly. Negative culture results reli- Annex 10.3.2). ably exclude urinary tract infection [26]. Border- line results need to be re-investigated from a 4.2. Containers suprapubic aspiration or single catheterized urine specimen. Sterile sampling of urine is important for Diapers or nappies are sometimes suggested urine microbiology, but may in¯uence some as collecting tools for clinical specimens from chemical measurements. Sterility of collection babies [28]. Despite occasional promising vessels means that the interior of the unopened results, especially for chemical constituents and unused container is free from interfering measured by qualitative strip examinations, it microbial contaminants. The European Phar- is hard to recommend this method due to high macopoeia de®nes sterility as reduction of variability in the ®bre construction of different bacterial growth to a probability of one brands and inability to measure the urine survivor in one million, e.g. after irradiation volumes voided. Morphological and in parti- [31]. Container manufacturers must document cular microbiological investigations are impos- their product's compliance with the intended sible with this procedure. clinical use (see Section 4.2.1), rather than Urostomy specimens (urines from ileal con- following strictly the Pharmacopoeia de®nition duit) are frequently obtained after bladder only. It is to be remembered that a particle surgery. Paediatric and adult patients with analyser or a nuclear ampli®cation method will dilated ureters may be given bilateral ureter- detect even dead bacteria. Furthermore, since ostomies. Chronic infection and bleeding at the waste is an increasingly important problem site of the stoma are common. Cleaning the globally, the development of new, environmen- stoma and discarding the ®rst portion of urine tally safe materials is encouraged for all obtained through a sterile catheter of suitable disposable containers. Finally, containers size ensures specimen quality. should comply with the European directive for in vitro medical devices [32].

Specimens to localize the site of urinary tract 4.2.1. Collection containers. infection. The collection of speci®c segments Single-voided specimens: The design of col- of urine ¯ow may help in de®ning abnormal lection containers should enable detection of areas of the urinary tract that may need uro- uropathogenic bacteria even in special situa- logical attention. Then, a dialogue between tions, i.e. at as low as 104 CFB/L level the requesting clinician and the laboratory is (equivalent to 101 CFU/mL) [5, 33]. The pri- essential before the procedure in order to mary collection container should be clean and guarantee the following: have a capacity of at least 50 ± 100 mL, with ± origin of the specimen (left/right ureter, an opening of at least 5-cm diameter to allow bladder, etc.) easy collection of urine by both men and European Urinalysis Guidelines 11 women. The container should have a wide The examination tubes used for samples for base to avoid accidental spillage and should quantitative chemical analysis should keep the be capped so that it can be transported and specimen intact and the cap should remain stored without leakage. The container and its closed upon freezing and during centrifugation cap should be free from interfering substances up to 30006g (relative centrifugal force, RCF). and should not absorb or change the urine The size, structure and length of the secondary constituents to be measured. Those parts of container vary depending on the needs of the the container and its cap which come into diagnostic procedures. contact with the urine should not contribute to microbial contamination after specimen col- 4.2.3. Labelling. All specimen containers must lection. be labelled with a tag that remains adherent Timed collections: For many chemical con- during refrigeration. The label should include stituents, quantitative excretion rates are im- a bar code (if used), a code of the exami- portant. A container designed for a 24-h nation requested, patient identi®cation and or overnight urine collection should have a requesting unit, as well as details of collection capacity of 2 ± 3 L. The container should be time, method of collection, and any additional constructed from materials that prevent (a) pre-analytical information in coded form. adherence of urine constituents, (b) exposure of Details of any preservative should be shown urine to direct light, which might alter clinically on a separate label and include any appropri- signi®cant metabolites, (c) contamination from ate hazard symbol. Labelling should not pre- the exterior when closed. Stabilizers usually vent a clear view of the specimen. The label prevent metabolic and other changes of urine must be placed on the container, not on the constituents. The container should allow for use cap. of recommended preservatives (Appendix, When body ¯uids are mailed to a distant Annex 10.3.4). laboratory, additional biohazard labels should Secondary containers (for basic urinalysis, be added and the packages must comply with usually examination tubes) should be easily the European standard EN829 (34). ®lled from the primary container without risk of spillage. The tube should be translucent to 4.3. Preservation and transport allow a clear view of the sample. The time elapsing between voiding and 4.2.2. Transport, storage and analytical con- examination of urine is a major obstacle to tainers. Specimens can be transported in their diagnostic accuracy in most laboratories. Inves- primary containers or divided into aliquots tigations performed at point-of-care are not that may range from 1 to 100 mL for chemi- subject to this delay but may suffer from cal and morphological investigations. For analytical problems. Precise collection times microbiological analysis, 1 ± 3 mL of urine in must be documented and delays exceeding the a clean container is suf®cient. For large speci®ed limits should be stated on reports. laboratories, a standardized vessel with a For many chemical constituents examined with volume of 3 ± 12 mL is essential for automated test strips no preservatives are needed, provided analytical systems. the analysis is performed within 24 h and the Examination tubes for test strip measure- tube has been refrigerated. If the specimen ment, particle counting or urine culture should contains bacteria and has not been refrigerated, keep the specimen suitable for analysis at false-positive nitrite or protein results may be z20³Coratz4³C as speci®ed. obtained using multiple test strips. In practice, Traditionally, urinalysis tubes have been strip examination should be performed on-site conical to allow decanting of supernatant when rapid or refrigerated transportation is after centrifugation. A more accurate sediment not possible. Preservation may be critical, volume is obtained by suction of the super- since some preservatives interfere with enzy- natant. Therefore as an alternative to the matic measurements (see Appendix, Annex conical shape, a round-bottom tube may be 11.1). considered for easy resuspension of the sedi- For quantitative chemical measurements it is ment when using standardized equipment. known that several speci®c proteins are unstable 12 ECLM ± European Urinalysis Group in urine but preservatives can inhibit their for successful preservation without bacterial degradation [35 ± 37]. The effect of storage inhibition. It is suggested that containers may be method-dependent, since an earlier containing boric acid should be ®lled to the radioimmunoassay for albumin did not ®nd indicated line to achieve a correct borate any effect on storing specimens at ± 20³C for up concentration. The specimen should be exam- to 6 months [38]. In the present guidelines, ined within 24 ± 48 h of production. It should be previous data are quoted [3, 39] as modi®ed noted that borate may inhibit growth of where necessary by technological development Pseudomonas spp. (Appendix, Annex 10.3.4). For particle examination the specimen should be refrigerated if not examined within 1 h, 5. CHEMISTRY EXAMINATIONS although precipitation of urates and phosphates will occur in some specimens. If precipitation 5.1. Hierarchy of methods (measurement disturbs interpretation, a new specimen should procedures) be examined atz20³C to avoid their artefactual generation. The longer the delay, the more In these guidelines, laboratory methods are likely are elements to lyse, especially when the classi®ed in four levels of performance based on urinary pH is alkaline and the relative density is accuracy of measurement. This hierarchy was low (especially true with children, often produ- adapted from earlier de®nitions of method cing a large diuresis). The WBC counts may be performance in clinical chemistry [47] to help questionable after 2 ± 4 h, even with refrigera- discussions aimed at better accuracy in tion [40]. Traditionally, ethanol (50% volume urinalysis. fraction) is used to preserve the cells but this Level 1: Rapid methods that ideally provide only partially prevents lysis of red and white the user with a fast, reliable response to an blood cells. To avoid shrinkage, polyethylene individual patient and easy handling of equip- glycol (2% mass fraction, low molecular mass ment, usually in primary care laboratories and such as Carbowax1) may be included in the at points-of-care. These typically include, for ®xative (this is called Saccomanno's ®xative example, urine test strip measurement, simple [41]). On mixing equal parts of sample and urine microscopy or dip-slide culture. Results ®xative, the particles should then be stable for 2 may be expressed on an ordinal scale. weeks. Alternative ®xatives also exist [42]. Level 2: Routine or ®eld methods used in Commercial preservatives, such as formalde- clinical laboratories because of medical need for hyde-based solutions, buffered boric acid and quantitative or specialized methods, which then formate-based solutions, and mercuric chloride- require experienced personnel and often a based tablets are also available. They have centralized site. These are often mechanized or gained renewed interest following the develop- automated methods. Examples include for ment of automated systems. Fixatives may example routine identi®cation of several uro- be adapted after evaluation with the new pathogenic bacterial species, immunochemical technology. quantitation of numerous urinary proteins, or Specimens requiring microbiological investiga- advanced urinary particle identi®cation. tion must be collected in a clean container and Level 3: An intermediate category of ``qua- examined in the laboratory within 2 h [5, 6]. li®ed comparison methods'' is proposed. The They should be refrigerated at 4³C without techniques are analytically more accurate and preservative if delay w2 h is expected. Then precise than Level 2 methods, but are not they should be examined within 24 h. If delay is applicable for routine use because they may be unavoidable and refrigeration not possible, too time-consuming or expensive. Level 3 needs containers pre-®lled with boric acid preservative to be considered and created when no Level 4 alone [43] or in combination with formate or methods exist and methods at Levels 1 or 2 other stabilizing media [44 ± 46], ideally in a need to be evaluated. Sometimes the Level 3 lyophilized form, may be used. Boric acid will methods may be designed ad hoc for evalu- stabilize white cell number and bacterial ation purposes only. Often, Level 1 may be concentration in urine held at z20³C for evaluated with Level 2 methods when all the 24 h. Boric acid concentration may be critical pitfalls of the methods of Level 2 are under- European Urinalysis Guidelines 13 stood. Level 3 methods may be considered as an The odour of normal urine is aromatic of intermediate stage in reference method devel- undetermined source. Infected urine may be opment. ammoniacal or fetid. Some metabolic diseases Level 4: The most analytically accurate have characteristic urine odours (Table III; methods are called reference and de®nitive from reference 14). methods (systematically called primary refer- ence measurement procedures). A de®nitive 5.3. Rapid chemical examinations (Level 1) method is designed to give the true value of the measured component [48, 49]. It is ``an Because many urinary tract diseases present analytical method that has been subjected to acutely, there is a need for rapid diagnostics thorough investigation and evaluation for frequently at points-of-care. Then, the ®rst sources of inaccuracy, including non-speci®city. urinalysis is often an ordinal scale (``semi- The magnitude of the imprecision and bias quantitative'') measurement. Depending on the (uncertainty) of the de®nitive method is local environment, these may be performed at compatible with its stated end purpose. The points-of-care or in laboratories in primary to true value is obtained from a mean value of tertiary care hospitals as part of a more detailed the measurements'' [50]. Very few de®nitive examination of urine. methods are available for clinical laboratories [51]. 5.3.1. Multiple test strips. Multiple test strips A reference method is de®ned as a ``thor- (dipsticks; of®cially called ``multiproperty'' oughly investigated measurement procedure, strips) have been designed to detect several of clearly and exactly describing the necessary the following components: leukocytes, bacteria conditions and procedures, for the measurement (nitrite), erythrocytes, protein (albumin), of one or more property values that has been glucose, ketone bodies, pH, relative density, shown to have trueness of measurement and bilirubin, urobilinogen and ascorbic acid. precision of measurements commensurate with A minimum combination depends on the its intended use and that can therefore be used intended use; general approaches are suggested to assess the accuracy of other measurement in Section 8 (Stepwise strategies in urinalysis). procedures for the same property (-ies), parti- The technology and principles employed in cularly in permitting the characterization of a test strips have been widely studied. Recently, reference material'' [48, 52, 53]. The European an interesting British evaluation of multiple Committee for Standardization (CEN) has also strips from seven companies was published, described guidelines for the characterization of assessing suitability for point-of-care diagnos- these procedures [54]. Reference methods tics by visual reading [57]. should be used in the evaluation of trueness The limitations of strip technology are well of routine methods. For urinalysis and bacterial known [15], but are compiled here in Appendix, culture, these are usually lacking or are in the Annex 11.1.1. However, any new drug should process of development. This is why a new be considered as a source of interference. The lower level (~Level 3) in the hierarchy was need for quality of test-strip measurements is re- introduced. Level 3 methods may be developed enforced in these guidelines by suggesting new to Level 4 after adequate descriptions of analytical quality goals (Section 9.2) and performance. evaluation tools (Annex 11.1).

5.3.1.1. Different properties of the strips (see 5.2. Visual inspection (colour and turbidity) also Appendix, Annex 11.1 for methodological and odour of urine detail). The most traditional urinalysis was based on Leukocytes (Esterase): The presence of leu- human senses. Visual inspection is often under- kocytes in urine is associated with both urinary taken and reported by the patient but should tract infection and non-infectious renal diseases. not be omitted by the laboratory. A summary Leukocytes are detected on the basis of indoxyl of the most common causes for abnormal esterase activity (deriving from neutrophil appearance of urine is given in Table II granulocytes and macrophages) released from (compiled from references 14, 55 and 56). lysed cells on the test pad. The analytical 14 ECLM ± European Urinalysis Group

TABLE II. Characteristic appearances of urine.

Appearance Cause Remarks

Colourless Dilute urine Polyuria, non-fasting specimen Cloudy, turbid Phosphates, bicarbonates, urates Leukocytes, RBC, bacteria, yeasts, spermatozoa, mucin, crystals, pus, tissue, faecal contamination, radiographic dye Rectovesical ®stula possible Milky Pyuria Infection Chyluria Lymphatic obstruction Paraf®n Vaginal cream Blue-green Biliverdin Pseudomonas infection Small intestine infections Drugs: arbutin, chlorophyll, creosote, Mouth deodorants indicans, guaiacol, ¯avins, methylene blue, triamterene, enteral nutrition (if blue dye added) Yellow Flavines (acri¯avine, ribo¯avine) Vitamin B ingestion Yellow-orange Concentrated urine Yellow foam Urobilin, bilirubin Rhubarb, senna Drugs: Salazosulfapyridine, Phenacetin, Alkaline pH pyridine derivatives, rifampicin Yellow-green Bilirubin-biliverdin Yellow foam Ribo¯avin Thymol Yellow-brown Bilirubin-biliverdin Beer brown Drugs: Nitrofurantoin Red or brown Haemoglobin, RBC Positive strip result, menstruation Myoglobin Positive strip also; muscle injury Methaemoglobin Acid pH Bilifuscin Result of unstable haemoglobin Urobilin Porphyrin May be colourless Beets, rhubarb, carotene Alkaline pH Fuchsin, aniline derivatives Foods, candy Drugs: aminophenazone, aminopyrine, antipyrine, bromsulphthalein, cascara, chinine, chloroquine, chrysarubin, hydroquinone, L-Dopa, naphthole, phenytoin, metronidazole, nitrite, nitrofurantoin, phenacetin, phenolphthalein, phenothiazine, salazosulfapyridine, senna, thymol Red-pink Urate May be associated with (massive) crystalluria Red-orange Drug: Rifampicin Red-purple Porphyrins May be colourless Brown See above Brown-black Methaemoglobin Blood, acid pH Homogentisic acid Alkaptonuria (alkaline pH) Melanin/melanogen Rare Darkening upon standing Porphyrin, homogentisic acid, melanogen, serotonin Drugs: cascara, chlorpromazine, methyldopa, metronidazole, phenacetin, imipenem sensitivity of the strip is about 80 ± 90% at the should be reached. Speci®city at a detection detection limit of 206106 WBC/L against limit of 206106 WBC/L is about 80 ± 90%. chamber counting of fresh uncentrifuged speci- Specimens with lysed cells, which are micro- mens. At 1006106 WBC/L, a sensitivity of 95% scopically classi®ed as negative, are partially European Urinalysis Guidelines 15

TABLE III. Characteristic odours of metabolic dis- the culture method, depending on the patient eases. population and the cut-off limit for positive 8 Odour Disease culture (usually chosen to 10 CFB/L or 105 CFU/mL) [58, 59, 60]. The speci®city of Sweaty feet Isovaleric acidemia and glutaric this ®eld for bacteria is high (w90%). acidemia Performance of test strips in detecting Maple syrup Maple syrup urine disease bacterial urinary tract infection must be inter- Cabbage, hops Methionine malabsorption Mousy Phenylketonuria preted carefully, since the selected cut-off for Rotting ®sh Trimethylaminuria signi®cant counts in culture (decreased to Rancid Tyrosinemia 106 CFB/L when patients with acute symptoms are evaluated), type of specimens and the studied patient population affect the results responsible for reduced speci®city. The esterase obtained [61]. The combined positivity ``either ®eld does not detect lymphocytes. Subtilisin of nitrite or leukocyte result positive'' is generally known activity may be used as a quality control useful [Table IV]. Speci®city of the combination solution. is reduced compared to the nitrite examination Bacteria (Nitrite): Nitrite examination is alone, because not all patients with leukocyturia based on activity of the nitrate reductase that have bacteriuria. The clinical sensitivity of the is present in most Gram-negative uropathogenic test-strip measurement improves if only patients rods, such as E. coli (Griess's examination). with pyuria are considered to have infection Nitrate reductase is, however, lacking from [62 ± 64]. This can be accepted in patients with a some common uropathogens, e.g. Enterococcus low risk for urinary tract infection only, since spp. and Staphylococcus spp., and will therefore high-risk patients may have signi®cant bacter- not be detected whatever their urinary concen- iuria without leukocytes in the urine (see tration. The positive detection of bacteria Section 8 for strategic approaches). requires, in addition, nitrate in the patient's Erythrocytes (Pseudoperoxidase): The pres- diet (vegetables), its excretion into urine and ence of red blood cells (RBCs), haemoglobin or suf®cient bladder incubation time. The analy- myoglobin in urine is seen either in dotted tical sensitivity of the method is variably (cells) or homogeneous appearance of colour on reported as being between 20% and 80% against the reagent pad. The examination relies on

TABLE IV. Performance of multiple test strips in combined detection of bacteriuria (either leukocytes or nitrite positive).

Patient population Specimen Cut-off limit CFB/L Sensitivity Speci®city References (prevalence of bacteriuria) (CFU/mL) (%) (%) and notes

Meta-analysis Not speci®ed 108 (105) 80 ± 90 80 ± 60 66, 67; numerous (not given) publications exist Mixed hospital patients Morning urine 108 (105)858068 (26%) Mixed hospital patients Not speci®ed Overall w105 (w102); 84 83 63 (20%) only range 105±6 (102±3) 25 not given Mixed hospital patients Morning urine 106 (103) 85 50 69; using (24%) re¯ectometry Ambulatory adult Untimed Overall w105 (w102); 65 98 70 patients, symptomatic only range 105±6 (102±3) 40 90 (38%) Symptomatic children Untimed 56107 (56104) 88 75 71, fairly high (10%) cut-off limit Symptomatic children, Untimed, 5 6 107 (56104)799872 age 1 ± 24 months catheter (5%) specimen Pregnant women, Not speci®ed 108 (105) 40 ± 50 95 73 74 asymptomatic (2 ± 5%) 16 ECLM ± European Urinalysis Group pseudoperoxidase activity shown by the haem (mostly albumin) at a concentration of about moiety of haemoglobin and of myoglobin. 0.2 g/L [75]. It is less sensitive to mucoproteins Unfortunately, this activity degrades rapidly and low molecular weight protein, and insensi- (in a few days) even when the specimen is tive to Bence-Jones proteins. The quantitative refrigerated, and is sensitive to various comparison method (Level 2 or Level 3) used to preservatives. RBCs re¯ect pre-renal, renal or evaluate the performance of the test strip (Level post-renal disease, but also occur in certain 1) has a clear in¯uence on the analytical physiological conditions, such as menstruation sensitivity and speci®city obtained. or strenuous exercise. Myoglobin can be For early detection of glomerular damage, demonstrated in patients with muscle necrosis sensitive immunochemical [76, 77] and other (crush syndrome), rhabdomyolysis (alcohol, [78, 79] rapid albumin methods have now been neuroleptic malignant syndrome or cocaine introduced and may gradually be incorporated abuse) or polymyositis. Haemoglobin without into multiple strips. They should allow detec- red cells may be detected in haemolytic states tion of albumin at 20 mg/L concentration (so- and in patients with pre-renal, renal and post- called microalbuminuria) at points-of-care as an renal haematuria if the cells have been alternative to quantitative laboratory measure- destroyed (either in vivo or in vitro) by delay ments if a rapid report is needed. in investigation. The analytical sensitivity of the Relative volumic mass (Relative density; test strip is 70 ± 80% at 106106 RBC/L against speci®c gravity): Positive or negative results chamber counting on fresh uncentrifuged speci- from all urinalysis examinations should be mens [65]. Speci®city of RBC detection is related to the state of water excretion (volume reduced when compared with chamber counts rate). The density obtained with the chemical because RBCs lyse easily in urine. test-strip ®eld is an arbitrary result allowing a Protein (Albumin): Total urine protein is a rough estimate of urine concentration only [81]. mixture of high molecular weight proteins (e.g. Osmolality should be measured from urines of albumin, transferrin, intact immunoglobulins, patients in intensive care or under parenteral a2-macroglobulin) and low molecular weight nutrition. proteins (e.g. a1-microglobulin, retinol-binding Creatinine: Creatinine measurement has been protein, immunoglobulin light chains) sieved traditionally used to estimate excretion rates by from plasma, proteins from the kidney (Tamm- relating urine concentrations of proteins [82], Horsfall protein) and those from the urinary hormones [83, 84] or other analytes to that of tract. The different causes of proteinuria are water in single-voided specimens. Its limitations listed in Table V. The traditional test-strip ®eld are discussed in Section 5.4.2. A new measure- is 90 ± 95% sensitive to clinical proteinuria ment based on a copper complex has been

TABLE V. Causes of proteinuria (modi®ed from 80).

Main groups Classi®cation Examples

Intermittent Functional Fever proteinuria Exercise proteinuria Congestive heart failure Epileptic seizures Orthostatic Occurs in upright position only Persistent Pre-renal Immunoglobulin heavy and light chain excretion (~Bence-Jones proteinuria) Myoglobinuria (in rhabdomyolysis) Haemoglobinuria (acute haemolysis) Renal Glomerular Albuminuria in IgA nephropathy Tubular Low-molecular-mass proteinuria caused by nephrotoxic drugs Mixed (glomerular Excretion of plasma proteins in advanced renal disease and tubular) Post-renal Urinary tract infection; prostatic or bladder disease; vaginal discharge European Urinalysis Guidelines 17 introduced in test strips to measure albumin/ [89] is particularly reduced in dilute and alkaline creatinine ratios from patient urines [79]. urines, typically in children with urinary tract Knowledge on clinically important interferences infection [40]. Casts are also lost in alkaline has not yet accumulated. urine [90]. Measurements of urine pH are Glucose: Examinations of urine glucose con- needed for the diagnosis of acid-base distur- centrations have largely been replaced by bances (such as hypokalaemic alkalosis with measurements of blood glucose concentration paradoxical aciduria), or when acidi®cation or [85, 86]. Urine glucose measurements may be alkalinization of urine is associated with speci®c advocated to check for inappropriate use of diseases (such as renal tubular acidosis, renal blood examinations, or for those patients stone disease), or during the elimination of unwilling to use blood sampling in addition to speci®c drugs (e.g. cytotoxic drugs). laboratory monitoring of glycohaemoglobin Bile pigments: Measurements of urinary

HbA1c concentrations [87]. In less favourable urobilinogen and bilirubin concentrations have economical situations, urine glucose monitoring lost their clinical signi®cance in the detection of is better than no monitoring. Since the ®nding liver disease following the introduction of liver of glycosuria may reveal patients with juvenile enzyme examinations from blood. They may be or maturity-onset diabetes mellitus, a screening useful, however, in differentiating icteric policy using test strips continues to be impor- patients or in detecting alcoholic liver disease tant, especially in acutely ill patients (if blood in non-laboratory environments. glucose is not used). Routine urine glucose Ascorbic acid: Ascorbic acid (vitamin C) screening of pregnant women is useful if interferes with the measurement of several combined with measurements of body mass test-strip analytes (see Appendix, Annex index (BMI) [88]. The analytical performance of 11.1.1). Because many patients ingest vitamin glucose screening by test strip is highly satis- C in large quantities (w1 g/day), measurement factory, partially because in the important acute of ascorbic acid concentration in urine helps in cases heavy glycosuria exists. identifying those patients with false-negative Ketone bodies: Ketone bodies (acetoacetate, test-strip results. The other, more direct, b-hydroxy butyrate and acetone) are excreted approach should be to try to develop test into urine in diabetic acidosis, during strenuous strips insensitive to interference by ascorbate. exercise, fasting, during enteric in¯ammations or periods of vomiting. The chemical reaction 5.3.1.2. Instruments used for multiple strip used is sensitive to acetoacetate and acetone, examinations. The manufacturer's instructions but not b-hydroxybutyrate. Ketone bodies need must be strictly followed for optimal results not be measured as part of general urinalysis, when reading test strips (Appendix, Annex but serve to classify or treat speci®ed patient 11.1.6). Instruments (rather than the naked populations, such as patients admitted as eye) are recommended for reading a multiple emergencies (especially paediatric patients), property strip, whether in the laboratory or at juvenile-onset diabetics or patients with toxae- point-of-care [91], because observer-related mia of pregnancy. After institution of insulin major errors occur frequently in practice and and ¯uid therapy in diabetic hyperglycaemia are not traceable afterwards. It is also clear and ketosis, tissue b-hydroxybuturate converts that all laboratory examinations, including back to acetoacetate, leading to transient those performed at point-of-care sites, should increase in urine acetoacetate excretion despite meet the required quality [92]. an improved clinical situation. Slight ketosis is The selection of different instruments is detected even after overnight fasting, indicating determined by the local diagnostic require- an acceptable clinical sensitivity. Unspeci®c ments. Centralized laboratories providing a reactions occur with the traditional examination 24-h service tend to automate the analysis of principle (see Appendix, Annex 11.1.1). large numbers of specimens, while point-of-care pH: Urinary pH varies between 5 and 9. sites show an increasing interest in improving Concentrated morning urine is usually acidic. the quality of single rapid patient investigations. Urines from children are often alkaline. Bac- Fully automated test-strip readers are fre- teria metabolizing urea to ammonia may also quently considered in large laboratories. Auto- increase the pH of urine. Survival of leukocytes mated urinalysis aims to improve the precision 18 ECLM ± European Urinalysis Group and accuracy of results at higher level than that 5.4. Quantitative chemical measurements achieved by visual or semi-automated methods. For many urine components, a quantitative Turnaround time, cost containment and safety result helps in the diagnosis and follow-up of of the working environment are important patients. This used to involve timed 24-h or issues in routine work-¯ow as well. A combina- other collections of urine with calculated tion of results from rapid chemical measure- excretion rates of the analytes. As a practical ments and automated particle analysis is under alternative, a reference measurement to adjust development [93]. for water excretion (osmolality, creatinine or equivalent) and calculation of analyte/creatinine 5.3.2. Pregnancy examinations. Detection of or analyte/osmolality ratios has already been pregnancy is achieved by speci®c measure- devised for measurements of protein excretion. ments of human chorionic gonadotropin This quantitative approach is typically per- (hCG) secreted by the developing placenta formed in centralized laboratories (Level 2 [94]. The appearance and rapid increase in the methods), but may be used in primary care concentration of hCG up to 10 000 IU/L in laboratories or even point-of-care sites with serum and urine during early pregnancy make methods suitable for Level 1 (e.g., albumin/ it an excellent diagnostic marker. The exami- creatinine ratio) when the cost/bene®t ratio is nations are performed from either blood or favourable. Several analytes measured quantita- urine specimens. The sensitivity to detect tively in speci®c clinical situations are listed in extra-uterine pregnancies is recommended the Appendix (Annex 10.3.4) in relation to using blood specimens, while urine is used for specimen collection. Detailed discussion was rapid examinations to detect mainly normal considered to extend these guidelines too much. pregnancies. Different de®ned sensitivities should be considered depending on the intended application, since the sensitivity at 5.4.1. Proteins. Measurement of total protein 25 IU/L is suf®cient to detect pregnancy excretion has been traditionally recommended already after 3 ± 4 days of implantation [95], for detecting renal disease [97]. In most cases i.e., before the missed menstrual period. A the urine contains albumin that may be meas- reduced sensitivity limit of 200 ± 500 IU/L may ured quantitatively or detected by test strip. be of value if only detection of normal preg- Glomerular nephropathies are characterized nancies is needed. This will delay the detection by increased excretion of albumin, transferrin of pregnancy but increase the speci®city of and in the advanced stage by high molecular obtained results. Detail of measurements is mass proteins such as IgG [98]. It is already given in the Section 11.3. widely accepted that an early glomerular dis- ease, such as incipient nephropathy in diabetes 5.3.3. Other rapid examinations. Test strips mellitus, can only be detected with sensitive with isolated ®elds to show the presence of measurements of (micro)albuminuria (analyti- glucose and/or ketone bodies, albumin or cal sensitivity down to 5 ± 10 mg/L [10, 99]). albumin/creatinine ratio only are indicated in Albumin excretion rate is elevated years screening or follow-up of some de®ned patient before reduction in glomerular ®ltration rate groups, such as certain patients with diabetes is detected by an increase in serum creatinine mellitus, pregnant women or renal patients. concentrations [100]. Increased urinary albu- Pre-eclamptic toxaemia patients should be min is now considered predictive of mortality screened by blood pressure measurements and morbidity in both non-insulin-dependent rather than by conventional urinary protein and insulin-dependent diabetes mellitus [101]. measurements [96]. However, proteinuria Albuminuria is the strongest known predictor assessment is important in treatment decisions of cardiovascular disease in non-insulin-depen- for pre-eclamptic patients. Rapid detection of dent diabetes mellitus [102]. The detection of bacteriuria is discussed in detail in Section albuminuria is a risk factor in non-diabetic 7.4. The visual reading of one or two ®elds hypertension [103 ± 105], vascular disease [106] on a strip is easier than of multiple test strips. and a predictor of heart failure [107]. Albumi- Rapid examinations for drugs of abuse are nuria also predicts mortality in elderly age in not included in these guidelines. general [108]. European Urinalysis Guidelines 19

For detection of tubular disease, a sensitive required for osmolality measurements, very strategy has also been proposed using measure- few laboratories have applied analyte/osmolal- ments primarily of a1-microglobulin [109], also ity ratios in their clinical practice [125]. Urine called protein HC [110, 111]. Tubular dysfunc- osmolality is also related to diet and ingestion tion can also be detected with other excreted of salts. markers, e.g., retinol-binding protein (RBP) [112 ± 115] and enzymes such as N-acetyl-b,D- Relative volumic mass (old terms: relative den- glucosaminidase (EC 3.2.1.30) originating from sity; speci®c gravity). Relative volumic mass kidneys [116 ± 118]. The earlier b2-microglobulin (traditional name: relative density [126]) is the is too labile in acid urine. Enzyme measure- preferred term over speci®c gravity, since the ments have gained only limited clinical value rede®nition of litre in 1964 to equal 1 cubic because the interpretation of variably elevated decimetre of water. Since the reference density excretion rates has been dif®cult among large (reference volumic mass) is the maximum den- patient populations. For example, in the sity of water at z4³C(r0~0.999972 kg/L), common paediatric urinary tract infections, there is no practical difference between volu- elevated excretions of N-acetyl-b,D-glucosami- mic mass and relative volumic mass of urine nidase, b2-microglobulin or a1-microglobulin within clinical medicine. A more signi®cant have no additional diagnostic information uncertainty of clinical results is derived from after measurements of body temperature [119, inaccurate measurement procedures in clinical 120]. This may re¯ect the dynamic metabolic laboratories. Urine relative volumic mass is states of tubular cells, high regenerative capa- closely related to osmolality [127]. The corre- city of kidneys as well as their ability to adapt lation between relative volumic mass and despite the destruction of some nephron seg- osmolality decreases in disease [128] because ments [121]. Renal tubular enzymes may, relative volumic mass depends on the con- however, be useful in assessing individuals centration of electrolytes, glucose, phosphate, exposed to nephrotoxic drugs or heavy metals carbonate and occasionally excreted iodine- [122]. containing radiocontrast media (after radio- Clinical speci®city and prognostic signi®cance logical investigations), while osmolality is in follow-up must generally be assessed to allow dependent on urea, ammonia and electrolytes. targeted application. A sensitive measurement is justi®ed with speci®c suspicion of renal disease, Creatinine. Correction of diuresis using urinary i.e., for patients with chronic diseases associated creatinine concentration to calculate analyte/ with renal complications, such as diabetes or creatinine ratios has gained general acceptance hypertension, and similar states. Since tubulo- despite certain theoretical problems [10, 11]. interstitial damage may cause end-stage renal Creatinine measurement is easily performed diseases that have currently remained unnoticed and only minimally affected by protein- by using mostly screening with albumin strips, a containing diet. Creatinine excretion suffers revised screening strategy should be considered from inaccuracies related to body weight, age, for other patient populations as well [123]. gender [129, 130] and renal function, i.e. tubu- lar secretion in uraemia [131]. Chronic diseases 5.4.2. Quantities of volume rate (diuresis). such as hypo- and hyperthyreoidism may also Osmolality. Assessment of renal concentrating affect it. Analyte/creatinine ratios should capacity is classically evaluated in kidney dis- always be measured as part of quantitative eases. The preferred quantity related to measurements if timed collections overnight or volume rate (diuresis) is urine osmolality for 24 h are to be avoided. because it is a solute property [124]. Osmolal- ity should also be measured when other com- Conductivity. Conductivity is a new parameter ponents in urine need to be related to variable that is brought to clinical laboratories and excretion of water in order to estimate excre- intensive care units with novel instruments. It tion rates. This is essentially true for both is related to osmolality, since both are depen- renal and lower urinary tract diseases, and dent on concentration of salts in urine. Con- both formed elements and dissolved chemical ductivity seems to correlate to osmolality constituents. Because a separate instrument is fairly well, even better than creatinine [69, 20 ECLM ± European Urinalysis Group

128, 132]. However, the exact clinical role of ®nding with unknown clonality. Test strip is this measurement awaits further clinical usually negative for the light chains. experience. Size fractionation of urinary proteins on sodium dodecyl sulphate-acrylamide gel electro- phoresis [141 ± 143] has been used to classify 5.4.3. Diagnosis and secondary prevention of pre-renal, renal and post-renal proteinuria renal stone formers. Some quantitative mea- before quantitative measurements of speci®c surements from urine help in diagnosing and proteins became widely available. Small labora- classifying patients with a recurring tendency tories may use this examination if a quantitative to renal stone formation [133]. After analysis result is not needed. of the stone by X-ray or the infrared method, quantitative measurements of urine volume, relative density, pH, calcium, urate, citrate, oxalate and creatinine are recommended as 6. PARTICLE ANALYSIS minimum measurements from 24-h collection [134 ± 136]. Optionally, measurements of urin- 6.1. Clinically signi®cant particles in urine ary magnesium, inorganic phosphate, ammo- Leukocytes: Granulocytes are the most fre- nia and cystine are recommended in addition. quent leukocytes detected in the urine of New technology to detect lithogenicity of patients with urinary tract infection due to urine has also been proposed based on X-ray common organisms, and may also be seen in microscopy [137, 138]. Serum (or plasma) con- other conditions such as glomerulonephritis, centrations of intact parathormone (primary interstititial nephritis and aseptic cystitis. The hyperparathyreoidism), calcium, urate, inor- appearance of lymphocytes in urine is asso- ganic phosphate and creatinine are of value, ciated with chronic in¯ammatory conditions, as well as examination of acid-base homeosta- viral diseases and renal transplant rejection. sis in blood. Dietary background must be Macrophages (mononuclear phagocytes, histio- known and understood when interpreting cytes) appear fairly often in urine of patients these results. Microscopic analysis of urinary with urinary tract infection. They are also crystals is discussed under particle analysis suggested to re¯ect, e.g., in¯ammatory activity (Section 6.1). The German experience suggests of renal disease [144]. Eosinophil granulocytes the importance of investigations of urolithiasis may occur in several disease states; they are no patients not responding to non-speci®c treat- longer seen solely as markers of acute inter- ment, since new stone attacks could be pre- stitial nephritis caused by drugs such as beta- vented by 46% in the follow-up [139]. lactamic antibiotics [145]. Erythrocytes: Haematuria remains a major sign of urinary tract and renal diseases. It may 5.5. Electrophoretic examinations also re¯ect a general bleeding tendency. Hae- Electrophoretic techniques are essential in maturia for physiological reasons (strenuous detecting the clonality of serum paraproteins exercise) and vaginal contamination (menstrua- (M components) and monoclonal immunoglo- tion) should be avoided ± if possible ± during bulin light chains in urine (Bence Jones careful patient preparation. The appearance of proteins). Immuno®xation of the urine or RBC in urine re¯ects the origin of bleeding: serum helps in de®ning the isotype of the dysmorphic erythrocytes (red cells with abnor- myeloma clone. Some authors recommend mal size or shape) suggest renal disease, whereas immuno®xation electrophoresis of all specimens RBC with normal morphology usually originate because of the sensitivity of the method [140]. from the lower urinary tract [146 ± 148]. The The speci®city must then also be understood morphology of RBC in urine is valuable in because of the existence of monoclonal gam- evaluation of patients with isolated haematuria, mopathies of undetermined signi®cance. As an because it can determine if the subsequent alternative to screening and in monitoring, diagnostic work-up is urological or nephrologi- turbidimetric quantitation of urinary k or l cal [149]. The examination technique needs chain excretion is also suggested [111], com- special training and is best performed with bined with immuno®xation in case of a positive phase contrast microscopy [147]. A subgroup of European Urinalysis Guidelines 21

TABLE VI. Levels of microscopy differentiation in clinical urinalysis.

Basic level Advanced level in addition

Red blood cells (RBC) Detailed subclasses erythrocytes: dysmorphic erythrocytes White blood cells (WBC)/Granulocytes Differentiation of leukocytes: Granulocytes, lymphocytes, macrophages (monocytes and eosinophils) Epithelial cells: From non-squamous epithelial cells: Squamous epithelial cells Squamous epithelial cells Non-squamous~small epithelial cells Renal tubular epithelial cells Transitional epithelial cells (super®cial and deep) Intestinal epithelial cells (occurring after bladder surgery) Atypical cells (experienced cytopathologist) Casts: From non-hyaline casts: Hyaline casts Erythrocyte, granulocyte casts Non-hyaline casts Renal tubular cell casts Hyaline, granular, waxy, fatty casts Bacteria and yeast-containing casts Haemoglobin and myoglobin casts Bilirubin casts Bacteria Gram-staining characteristics of bacteria (microbiology laboratories) Yeasts, Trichomonas Schistosoma haematobium (in appropriate geographical locations) Spermatozoa Spermatozoa Artefacts (hair, paper and textile ®bres, Artefacts and mucus as on the left starch, glass) and mucus Lipids: Lipids, in addition to droplets: Droplets (isolated and aggregated) Oval fat bodies (lipid-laden tubular cells), cholesterol crystals Crystals: Additional rare crystals: Urate, oxalate (mono- and dihydrated), Drugs, cystine, leucine, tyrosine, 2,8-dihydroxyadenine, xanthine phosphate and cystine abnormally shaped RBC (acanthocytes or G1 considered and reported by a general laboratory cells~urine red cells with blebs) has also been examining urine cells at an advanced level. A described recently [150 ± 152]. Because of their repeat urine specimen freshly collected and ®xed characteristic shape, identi®cation of acantho- according to a standardized procedure is needed cytes is helpful in de®ning a glomerular in cytopathological investigation. Rapid exam- haematuria. inations for bladder tumour cell antigens may Epithelial cells: Detached epithelial cells in the enter clinical practice when problems with their urine may help localize urinary tract disease. speci®cities are solved. The appearance of squamous epithelial cells Renal tubular epithelial cells are mainly found from the outer genitalia or distal urethra serves, in the urine of patients with acute tubular however, as a marker of a poor collection necrosis, acute interstitial nephritis for any technique, except during pregnancy when cause and acute cellular rejection of renal allo- epithelial cell exfoliation is increased. graft [154, 155]. Tubular cells are also found in Transitional epithelial (urothelial) cells derive patients with active proliferative glomerulone- from the multilayered epithelium lining the phritides [156], glomerular diseases associated urinary tract from the calyces of the renal with nephrotic syndrome and some metabolic pelvis to the bladder in the female and to the storage diseases, such as Fabry's disease [157]. proximal urethra in the male. These cells are Since renal tubular cells are dif®cult to frequently found in urinary tract infection and distinguish from transitional cells by unstained in non-infectious urological disorders. bright-®eld microscopy, many laboratories have The detection and de®nitive diagnosis of used the term ``small round epithelial cells'' to atypical cells is dif®cult [153] and should be describe both. This guideline recommends the undertaken by experienced cytopathologists. A use of correct cytological terminology for cell suspicion of atypical or malignant cells may be types if possible after appropriate training and 22 ECLM ± European Urinalysis Group experience. If a laboratory decides to remain at urate or oxalate, or by in vitro changes in the the basic level of differentiation for practical temperature (refrigeration) or pH of urine. reasons, the term small epithelial cell is accepted Detailed investigation for crystals in all (Table VI). specimens is unwarranted. However, detection Casts: Casts are formed in the distal tubules of crystals has clinical value in subpopulations and collecting ducts [158] from aggregation and of recurrent renal stone formers [163] (see also gel-transformation of the ®brils of Tamm- Section 5.4.3). They may also be signi®cant for Horsfall glycoprotein (uromucoid). This mater- some patients with acute renal failure [163]. In ial is produced by the cells of the ascending such cases, crystalluria is the marker of a major limb of Henle's loop and forms the hyaline disorder and is diagnostically important. The matrix of most casts [159]. A precipitate, i.e. a most typical examples are acute uric acid cast, is formed when the concentration of nephropathy and ethylene glycol poisoning, dissolved organic and inorganic material which is associated with calcium oxalate exceeds the saturation point of the normally monohydrate crystalluria. Therapeutic drugs colloidal solution. Within casts, plasma pro- possibly crystallizing in urine include sulpha- teins, lipids, different types of cells, micro- diazine, triamterene, acyclovir, indinavir and organisms (bacteria or yeasts), pigments vitamin C (see Annex 12.1.3 for details). All the (haemoglobin, myoglobin, bilirubin) and crys- above circumstances are suggested by the tals may be found. Some hyaline casts may be ®nding of either massive or atypical crystalluria, present in the concentrated morning urine of including crystalline casts. When there is a high healthy individuals. Casts, however, usually clinical suspicion of their relevance, a speci®c re¯ect the presence of renal disease. They are request should be made to the laboratory for speci®c but not sensitive markers of renal investigation of urine crystals. disease. Cystinuria can be detected by ®nding hex- Lipids (fat): Lipids are found in urine when agonal plain crystals in urine (prevalence from plasma lipoproteins leak through the damaged 1:2000 in England to 1:100 000 in Sweden [133, basement membranes of glomeruli. As lipopro- 164]). A rapid procedure (cyanide-nitroprusside tein particles are larger than protein molecules, examination) is also available [165]. To see lipiduria is typical of patients with heavy cystine crystals by microscopy, the specimen proteinuria. Lipids are usually detected as should be acidi®ed with acetic acid (at pHv6) droplets, isolated or in clumps, lipid-laden and kept refrigerated overnight to allow crystal- tubular epithelial cells (``oval fat bodies''), lization to occur. Rare 2,8-dihydroxyadenine cholesterol crystals or lipid-containing casts. crystals occur in a genetic de®ciency of adenine Microbes: Different microscopy methods vary phosphoribosyltransferase enzyme [166]. in sensitivity and their ability to detect and Another very rare xanthine crystalluria occurs permit the identi®cation of bacteria. All methods in de®ciency of xanthine oxidase; the crystals are relatively insensitive for the detection of are easily confused with urate [135]. Tyrosine bacteria at concentrations below 108/L (105/mL), and leucine crystals are associated with a severe although the presence of bacteria below this liver disease. Measurements of urinary (and concentration is important in selected groups of plasma) concentrations of amino acids are patients, whether symptomatic or not [160, 161]. recommended for con®rmation of inborn A separate Gram stain may be performed in errors of amino acid metabolism. microbiological laboratories capable of inter- preting it. Other organisms, such as yeasts and Levels of differentiation Differentiation of the Trichomonas vaginalis are common in the urine above-mentioned particles by microscopy can of women with vaginitis. Schistosoma may also be divided into basic and advanced levels be found on urine microscopy [162]. Urine (Table VI). The basic level for routine urine microscopy is not sensitive enough to exclude microscopy is a positive, speci®c identi®cation urinary tract infections. of the usual formed elements (left column). Crystals: In most instances, the ®nding of Some laboratories or clinical units report only crystals is not clinically relevant, since they may leukocytes, erythrocytes and bacteria. This occur as a consequence of transient super- should be said clearly in the speci®cations saturation caused, for instance, by food rich in given by the laboratory. The advanced level European Urinalysis Guidelines 23 of urine microscopy usually provides evidence should be related to the original volume of of renal damage (right column, usually in che- urine (in litres, L). Counting of native urine mical laboratories or nephrological units), or avoids the error created by centrifugation. A speci®es microbes in more detail, e.g. based suf®cient volume is needed to detect rare on Gram staining characteristics (microbiolo- particles related to renal damage (see Section gical laboratories). The suggested order of 9.4). Particle concentrations at the low positive process is that laboratories and nephrological range or upper reference limit in health can vary units ®rst decide between these two levels and up to 100% owing to the effect of pre-analytical then de®ne the local detailed procedures to be variables. In regard to this ®gure, a reliable followed depending on the needs for urine comparison level of enumeration of particles microscopy, be that in a clinical nephrology has a negligible bias and imprecision to allow unit or in a general or specialized laboratory. correct classi®cation of patients' results (see Annex, Section 12.1.1 for further detail). Automated instruments have improved pre- 6.2. Different techniques of visual microscopy cision by counting many more cells than visual methods. They may be used for enumerating 6.2.1. Hierarchy of microscopy methods. Refer- elements for which their method has been ence method for urine microscopy (Level 4) validated. should provide both correct identi®cation of different particles and their accurate quantities when measuring urine. Currently, no such 6.2.2.2. Bacteria counting (Slide centrifugation method exists. Standardization of the method z Gram staining procedure). The slide centri- used is essential for improving accuracy and fuge method with Gram staining offers the limit of detection [2, 167, 168], independently highest analytical sensitivity and reproducibil- of the intended level of ®nal performance. ity in the rapid detection of bacteria in urine, Special attention should be paid to different although it is not a quantitative method. It sources of error [169, 170]. The centrifugation can therefore be used in Level 3 comparisons. 8 step with removal of supernatant is a major When a cut-off value of 10 colony-forming tool for concentration of the urine specimens, bacteria/litre (CFB/L) in culture is taken, this but also a major source of errors. method has a sensitivity of 98% and a speci®- city of 90% against the culture [173]. These performance ®gures were con®rmed with spe- 6.2.2. Principles of comparison methods (Level 3). ci®c patient populations in later studies, show- 6.2.2.1. Particle counting. Identi®cation: Parti- ing an overall sensitivity of 63% and cle identi®cation needs an optical method to speci®city of 91% at 106 CFB/L [174]. The discern formed elements from their ¯uid back- main disadvantage to this robust and consis- ground and a differentiation method to allocate tent method is that it is very labour-intensive. these elements into correct classes (Table VI). Details of the method are described further Bright-®eld microscopy of unstained prepara- under microbiological methods (Appendix, tions is inadequate for detection of bacteria, Annex 13.5.1). RBC and hyaline casts, and therefore for advanced differentiation. For this reason, either supravital staining [171, 172] or phase-contrast 6.2.3. Routine identi®cation methods for urine microscopy [56], or both, is recommended. particles (Level 2). Detection of microbes by Gram staining and Standardized urine sediment under a cover- other special procedures, such as nucleic acid slip. Standardized centrifuged urine sediment ampli®cation methods, are needed in the micro- should be investigated under a de®ned size biology laboratory. Identi®cation of elements coverslip with a known volume of specimen in such as eosinophils, monocytes or all macro- the view ®eld (Appendix, Annex 12.1.2). This phages may require speci®c methods as well as is recommended as a routine visual procedure experienced evaluators. The latter may, how- in examination for kidney-related urine parti- ever, be beyond the general need of rapid differ- cles for the following reasons: entiation of urinary particles. (1) The sediment method detects particles Counting: Concentrations of urinary particles indicating renal damage. With uncentrifuged 24 ECLM ± European Urinalysis Group specimens, the investigator misses rare elements fuged urine for diagnosis of urinary tract (most importantly casts and renal tubular infection in young women has also been epithelial cells) that are clinically signi®cant developed [177]. Uncentrifuged urine may be but can be concentrated at low-speed centrifu- accepted because bacteria are not concentrated gation. To allow screening of a large volume, very much at low-speed centrifugation. i.e. up to 10 ± 20 mL under a coverslip, a low- Reporting at ordinal scale is usually suf®cient. power (e.g. 6100) magni®cation should always Details of Gram staining are described further be used in parallel to high-power (usually under microbiology methods (Appendix, 6400) optics. It is understood that centrifu- Annex 13.5.2). gation methods are never quantitative in counting RBC and WBC that are variably Chamber counting of uncentrifuged specimens. lost during centrifugation. The procedure is In some microbiology laboratories, simple then prone to high uncertainty of certain microscopy is used to improve the detection particle counts despite standardization. of urinary tract infection only [178, 179]. (2) Differentiation of important formed ele- Chambers were advocated originally to pro- ments (~ detection of renal damage) is easier vide more precise leukocyte counts than those through a thin ¯uid layer under a coverslip obtained from sediment preparations, a ®nd- (40 ± 50 mm) than through the thicker ¯uid layer ing of 10 or more leukocytes 6106/L repre- of a chamber (usually §100 mm high). A senting signi®cant and abnormal levels of thicker ¯uid layer may require more focusing leukocyturia [180 ± 182]. Because centrifugation levels to be examined. results in a variable 20 ± 80% loss of RBC and WBC [170], no sediment method can be con- Urine sediment counted in a chamber. Counting sidered as reference for quantitative urinary concentrated, centrifuged sediments in a cham- particle counting. Analytical sensitivity for ber has been advocated by some investigators bacteria will be poor at lower concentrations [168, 175] because of the large counting when compared with bacterial cultures. Perfor- volume (3.2 mL), which results in a more pre- mance ®gures for bacteriuria for techniques cise count and possibly a higher sensitivity using uncentrifuged samples of urine are very compared to uncentrifuged specimens counted variable, depending on the criteria used for in a chamber. The easier performance of the positivity in culture [183]. They will depend coverslip method on centrifuged sediment also on operator experience, whether bacilli or makes this a preferred routine method over cocci are present, and on the degree of inter- chamber counting despite the lower precision. ference by debris. Sensitivity at the 107 bac- The reference intervals of chamber counts (the teria/L (104 bacteria/mL) level usually exceeds number of particles seen in the chamber) in 80%. health vary because of the different concentra- The major disadvantage of chamber counting tion factors used in different systems [167, is that it tends to be time-consuming for routine 176]. laboratory practice. Without staining or phase contrast optics, differentiation of particles is not Gram staining. Correctly performed, exami- suf®cient for renal elements. The advantage of nation of urine using a microscope will detect precise, quantitative measurement of elements is bacteria which are either nutritionally fasti- outweighed by cost as a routine procedure, but dious and will not easily grow on standard chamber counting is useful as a comparison media, or which will only grow anaerobically, method in instrument evaluations. but are nevertheless responsible for urinary tract infection in a minority of patients. Filtration of urine particles. Filtration of urine Microbiology laboratories may bene®t from particles on supportive media allows recovery their capability of interpreting Gram staining of essentially all particles existing in a de®ned when trying to reduce the total work-¯ow. volume of urine if remaining intact. This prin- This staining has the added advantage of ciple has been used to improve the sensitivity immediately ascribing a Gram status to the to detect urine casts using a 3-mm pore size, infecting organism. A speci®c model incorpor- and starting from 20 ± 100 mL of urine [184]. ating the results of Gram staining of uncentri- Another publication used ®ltering to examine European Urinalysis Guidelines 25 physiological concentrations of RBC in urine with an inverted microscope because it is as detected by scanning electron microscopy rapid, cheap and occupies very little bench [185]. There is also a commercial version of space (a basic inverted stage microscope ®tted ®ltration which uses disposable syringes and with a 96-well microtitre plate holder). A special equipment [176]. The losses of material ®xed volume of urine (usually 60 ± 80 mL) is during ®ltration depend on the pore size and placed in a well of a 96-well ¯at-bottom pressure applied on the particles. Staining of microtitre plate and allowed to settle. The the preparations is needed for proper identi®- well is then examined using an inverted micro- cation of particles, which may also cause lysis scope. The presence or absence of white cells, of some cells. bacteria and red cells is noted on an ordinal scale. The technique requires operator train- 6.2.4. Rapid microscopy methods (Level 1). ing, as differentiating (e.g.) cocci from debris Non-standardized urine sediment under a cover- may be dif®cult. Disposable plastic 96-well slip. Traditional urine sediment is still investi- plates may not be available to all laboratories; gated in many laboratories by pouring it is however perfectly possible to cold disin- supernatant after centrifugation, inserting a fect, wash and thoroughly dry plastic plates drop of sediment under a coverslip and inves- and re-use them (see detail in Appendix, tigating an unknown volume under a micro- Annex 13.5.3). scopic view®eld. Without staining and using Test strip is used in many clinical environ- bright-®eld technique some of the elements ments to detect haematuria, pyuria or bacter- are not seen; with little knowledge of the dif- iuria without microscopy [71]. ferentiation of particles others are missed in classi®cation. In repeated routine process, this procedure gives results associated with disease 6.3. Instrumental particle analysis conditions in an arbitrary ordinal scale. Because of a wide uncertainty of results and Mechanization of urine particle analysis reduced sensitivity in detecting essential parti- commenced with the introduction of automated cles, non-standardized sediment procedure is microscopes [186] and ¯ow cytometry devices not recommended. [187, 188] about 20 years later than for blood cell counting. New technology now offers the Microtitre tray method. Microbiology labora- opportunity to replace the basic level of particle tories with a high throughput of samples analysis (see Table VI). Renal damage is may consider the microtitre tray method partially detected by identifying some casts

TABLE VII. Clinical presentations for rapid bacteriuria detection.

(1) Classical frequency/dysuria syndrome in young, low-risk women if clinically needed (2) Emergency medical services, as a ®rst rapid diagnostic examination (3) Screening for selected asymptomatic individuals, e.g., women in the antenatal clinic (4) Selecting specimens for extended investigation in the laboratory (Sections 7.3.6 and 8.2)

TABLE VIII. Medical indications for urine culture.

(1) Suspicion of acute pyelonephritis or febrile urinary tract infection (2) Suspicion of hospital-acquired urinary tract infections (possibility of reduced antibiotic sensitivity) (3) Suspicion of urinary tract infection in patients with a predisposing disease, such as patients with diabetes or anomalies of the urinary tract, recurrent stone disease, or immunocompromised state (4) Patients failing ®rst line antimicrobial chemotherapy (5) Febrile patients with indwelling catheters (6) Clinical suspicion of urinary tract infection in men (symptomatic) (7) Clinical suspicion of urinary tract infection in pregnant women (symptomatic) (8) Suspicion of urinary tract infection in children and adolescents (symptomatic) 26 ECLM ± European Urinalysis Group and small epithelial cells [69] and microcytic examinations for every patient suspected of (small) erythrocytes [189]. Rational combina- having urinary tract infection. A simple division tion of automated and visual particle analysis of the patients into usual, uncomplicated cases with chemical measurement and bacteriological suspected for lower urinary tract infection and procedures is crucial in the new urinalysis work- special or problematic cases will improve the ¯ow strategy. ef®ciency of clinical laboratory practice.

7.1.1. Clinical presentations requiring rapid 7. MICROBIOLOGY examination for bacteriuria. Rapid examina- EXAMINATIONS tions are recommended in the situations listed in Table VII. In recurrent acute lower urinary 7.1. Medical indications for microbiology tract infection from low-risk women (item 1), investigation of urine no laboratory examinations are usually neces- The aims of urine bacterial culture are (a) to sary when the symptoms are clear-cut [5, 6, identify aetiological agents of urinary tract 190]. If symptoms remain unclear, rapid meth- infection, i.e. relevant pathogens but also ods for detecting bacteriuria help in the dif- mixed ¯ora as a sign of contamination, (b) to ferential diagnosis of patients with medical estimate the concentration of bacteria, (c) to emergencies. Before classifying otherwise offer susceptibility testing for antimicrobial healthy women into this group, anatomic treatment, and (d) to follow the effects of abnormalities in the urinary tract should be antimicrobial treatment during the course of considered (Table VIII). Usually asympto- urinary tract infection. In clinical practice, matic bacteriuria should not be sought nor however, it is not necessary to perform all treated to avoid enrichment of multi-resistant

TABLE IX. The pathogenicity and frequency of micro-organisms in midstream urine.

Pathogenicity in the Frequency (percent of isolates) urinary tract A. Common B. Fairly common C. Uncommon D. Rare (w10%) (1 ± 10%) (0.1 ± 1%) (v0.1%)

I. Primary pathogens E. coli S. saprophyticus E. coli CO2-dependent, Salmonella spp.a (Leptospira, mycobacteria) II. Secondary pathogens Enterobacter spp., Citrobacter spp., Corynebacterium Enterococcus spp., M. morganii, urealyticum, Klebsiella spp., P. vulgaris, Haemophilus spp.b P. mirabilis, Serratia spp., Pneumococcib P. aeruginosa S. aureus III. Doubtful pathogens GBSc, Yeast, Acinetobacter spp., A great number CNS (others)d Pseudomonas spp., of reported cases Stenotrophomonas have been published maltophilia with exceptional cases of infections caused by other species IV. Usually urethral or a streptococci, Bi®dobacterium spp., genital ¯orae Gardnerella vaginalis, ``Diphtheroid'' rods, etc. Lactobacilli, etc.

a Low concentrations are reported even if they are most likely caused by contamination during specimen collection. b Most often isolated from children. c GBS~group B streptococci (S. agalactiae). d CNS~coagulase-negative staphylococci, urease-forming isolates or isolates found in patients with indwelling catheters have increased signi®cance. e No identi®cation and susceptibility testing (only exceptionally, if especially indicated). European Urinalysis Guidelines 27 bacterial strains. However, screening of TABLE X. Examples of bacterial concentrations selected clinical populations, such as pregnant expressed as traditional colony-forming units and recommended colony-forming bacteria. women, is warranted (Section 8.2.3). Conventional unit Standardized unit 7.1.2. Indications for urine bacterial culture (CFU/mL) (CFB/L) with identi®cation of species and susceptibility 1103 testing. Urine samples should be sent to the 103 106 bacteriology laboratory for quantitative cul- 105 108 ture and susceptibility testing from the patient groups listed in Table VIII. 7.2.1. Classi®cation based on uropathogenicity. The classi®cation of pathogens causing urinary 7.1.3. Indications for urine bacterial culture tract infections is given in Table IX. Different after completed treatment. Low-risk women groups of micro-organisms have been classi- who become asymptomatic after treatment for ®ed into 16 categories based on four degrees acute cystitis do not need follow-up with a of pathogenicity (I ± IV) and frequency in clin- post-treatment urine culture. ical populations (A ± D) [7]. The cut-offs between the degree of uropathogenicity and species frequency should not be interpreted as 7.2. Microbes of the urinary tract strict and unchangeable. The table is merely The presence of organisms in urine per se is an attempt to organize many years' experience not diagnostic of infection. Species such as in clinical microbiology [5, 6, 191, 192, 193]. Escherichia coli and Staphylococcus saprophyti- Pathogenicity may be classi®ed as follows: cus seem to be more aggressive than others and may therefore be identi®ed as primary patho- gens (see below). Other factors, such as whether I. Primary pathogenic species: Species that have urine is analysed as part of a screening the ability to cause urinary tract infection in programme, bladder incubation time, coexisting individuals with normal urinary tracts. This symptoms of infection, pregnancy and age, all group consists of E. coli and S. saprophyticus. in¯uence the speci®city attached to the presence Also found in this group are certain rare of any concentration of bacteria in urine. In primary pathogens and other species which addition, the largest variable that is impossible should be reported according to national to control accurately is the technique of the regulations, e.g. Salmonella spp. ``mid-stream urine'' at the time of obtaining the II. Secondary pathogenic species: Species that urine sample, which often is not perfect. Despite seldom cause primary infection in patients with instructions, a proportion of samples will normal urinary tracts, but often occur in contain commensal contaminants which may hospital-acquired urinary tract infection. Kleb- be present in large enough numbers to make siella spp. (species), Enterobacter spp., Proteus interpretation dif®cult. Diagnostic rules there- spp., Morganella morganii, S. aureus and fore depend on whether bacterial growth is pure Enterococcus spp. are a few common examples. or mixed. This emphasizes the importance of a III. Doubtful pathogenic species: Skin ¯ora and laboratory system that insists on, and takes other species. Those sometimes colonize in account of, clinical details for individual speci- hospitalized patients and cause hospital-acquired mens. It is not possible to provide effective urinary tract infection. The culture ®ndings can patient management if the results of urinary only be considered relevant if suprapubic investigation are not interpreted in the light of aspiration has been performed. Even if coagu- clinical information. lase-negative staphylococci (CNS) occasionally Speci®c bacteria, e.g., those causing tubercu- cause urinary tract infection, ®ndings are often losis, leptospirosis, salmonellosis or sexually a sign of contamination and have a low transmitted diseases, such as N. gonorrhoeae or predictive value. C. trachomatis, and fungal infections need IV. Urethral/genital ¯ora: Species belonging to special examination methods not discussed in normal urogenital ¯ora. These often contam- detail in these guidelines. inate voided and drawn urine samples. They are 28 ECLM ± European Urinalysis Group

TABLE XI. Diagnostic performance of different cut-offs for ``signi®cant'' coliform bacteriuria in women with acute voiding dif®culties (160).

Mid-stream urine (CFB/L) Sensitivity Speci®city Predictive value

Positive Negative

105 0.95 0.85 0.88 0.94 106 0.81 0.90 0.90 0.82 108 0.51 0.99 0.98 0.65

examined for antimicrobial susceptibility only TABLE XII. Possible causes of low bacterial con- in exceptional cases (when especially indicated). centrations in mid-stream urine. This group has been introduced to give more Early stage of infection detail for clinical interpretation and to minimize High volume rate (diuresis) false positives. Urgency symptoms (short bladder incubation time) Presence of antibiotics in urine 7.2.2. Background for limits of clinically signi®- Low pH in urine Slow bacterial growth cant bacteriuria. Contaminated specimen 7.2.2.1. Unit recommended for expressing bac- terial concentrations. New technologies for direct particle counting and development in v107 CFB/L indicated contamination during standardization of nomenclature ask for a sample collection, whereas bacterial concentra- rede®nition of the unit used to report bacter- tion in the interval of 107 ± v108 CFB/L was ial concentrations. The classical ``colony-form- dif®cult to interpret. Despite the fact that the ing unit/millilitre'' (CFU/mL) has the pitfalls criteria were developed for acute pyeloneph- that the word ``unit'' should be used as an ritis and asymptomatic bacteriuria in women, abstraction rather than count of visible things they began to be used generally, even for such as bacteria, and millilitre is not the stan- symptomatic lower urinary tract infection. dardized volume to express concentrations. Only the discovery that S. saprophyticus was a Therefore, bacterial particle concentrations are common cause of seasonal, aggressive cystitis recommended to be expressed as bacteria/litre in younger women resulted in the cut-off of (L) in a way similar to cells in body ¯uids. signi®cant growth being lowered to When grown on agar, the living bacteria §107 CFB/L in these cases. appear as colonies, corresponding to a stan- dardized unit colony-forming bacteria/litre 7.2.2.3. Signi®cance of low bacterial concentra- (CFB/L). Equivalence is given in Table X. tion in urine. Stamm et al. examined 187 sexu- ally active young women with dysuria and 7.2.2.2. Traditional Kass's criteria. Evaluating urinary urgency [160]. Cultures of mid-stream urine culture ®ndings has long been domi- urine samples were compared to urine cultures nated by Kass's criteria for signi®cant bacter- obtained through suprapubic aspiration or iuria. Kass found that 95% of women with urethral catheterization. ``Coliform'' (the term pyelonephritis had §108 CFB/L (§105 CFU/ is not de®ned in Stamm's article and most mL) of one bacterial species in a clean-catch likely refers to Enterobacteriaceae or lactose- mid-stream urine, and that such a ®nding in positive Enterobacteriaceae) bacteria were iso- two consecutive mid-stream urine specimens in lated from bladder urine in 98 (52%) women, asymptomatic women would, with 95% prob- whereas non-coliform bacteria such as S. ability, give the same result in a third mid- saprophyticus, S. aureus and enterococci were stream urine specimen [194, 195]. Thus, in cultured in 26 (14%). The women who had order to diagnose asymptomatic bacteriuria ``coliform'' bacteria in bladder urine were with reasonable accuracy, §108 CFB/L of the further analysed regarding the number of same bacterial species was necessary in two CFB/L. If 108 CFB/L mid-stream urine was mid-stream urine samples obtained within an used as a cut-off for ``signi®cant'' bacteriuria, interval of a few days. Kass also showed that the sensitivity was 51% and the negative pre- European Urinalysis Guidelines 29 dictive value was 65% (Table XI). Speci®city gated and reported should vary according to was high. If, on the other hand, a cut-off of the local case mix and should not interfere 105 CFB/L mid-stream urine was used, the with the needs of speci®c patients. sensitivity was 95% with a negative predictive value of 94%, whereas speci®city declined 7.2.3. Laboratory-related decision limits for from 99% to 85%. Among the 48 who had ``signi®cant'' bacteriuria. The following decision v108 CFB/L mid-stream urine of ``coliform'' limits for diagnostic laboratory workup are bacteria, at least one additional bacterial spe- proposed (Table XIII) [7]. They should guide cies was isolated in 35 cases, i.e. there was the practical process. Consideration has been mixed ¯ora. given to symptoms, bacteria category, number Thus, low cut-off of ``coliform'' bacteria in of species isolated, method of specimen collec- mid-stream urine more accurately predicted tion and gender. One should note the uncer- bladder infection in symptomatic women than tainty of quantitation when using routine loop in asymptomatic. The main differences between cultures. A count of 36106 CFB/L (3 colonies Kass's and Stamm's studies were the prevalence on a plate with a 1 mL loop) is more than of infection in the patient populations, as well as 106 CFB/L, but reproducibility may be within the presence of symptoms (dysuria). Thus, the a con®dence interval of 1 ± 106106 CFB/L in chosen cut-off for ``signi®cant'' bacteriuria should routine work. Few laboratories report one be adjusted to the population under examination. colony (1 colony with a 1 mL loop ~106 CFB/ Many additional studies support the observa- L). It is obvious that the symptoms and clinical tion that low bacterial concentrations of E. coli status of patients are important when evaluat- in particular have diagnostic relevance, even in ing culture results. This implies that consider- mixed ¯ora [33, 196 ± 205). Findings of E. coli ably greater attention must be paid to the have been interpreted as the ®rst phase in information given on the request and report urethritis in an ascending infection. Causes of forms (Appendix, Annex 10.1). low bacterial concentrations in mid-stream The cut-offs for symptomatic urinary tract urine are given in Table XII. Short bladder infection caused by primary pathogens (E. coli incubation time (see also Section 3.3.3) is a less and S. saprophyticus, Table IX) are set at important factor than the lowered cut-offs for §106 CFB/L in mid-stream urine specimens. signi®cance in symptomatic individuals. On the For secondary pathogens (group II), lower cut- basis of available information, the cut-offs offs of §107 CFB/L mid-stream urine for could be set even lower than those proposed women and §106 CFB/L mid-stream urine in Table XIII, but with the risk of a lower for men are recommended. For men, the risk speci®city, leading to a decreased predictive of contamination during specimen collection is value of positive results for clinical urinary tract much smaller, increasing the speci®city of low infection. The importance of speci®city should bacterial concentrations. For doubtful patho- be stressed in diagnostic practice. genic bacteria (group III) a cut-off of §108 CFB/L in mid-stream urine is proposed 7.2.2.4. Mixed cultures. Most acute uncompli- for pure culture in symptomatic patients. cated urinary tract infections result from one If symptoms are absent, the classical bacterial species. The isolation of more than criteria apply for asymptomatic bacteriuria: one organism from a single specimen of urine §108 CFB/L of the same bacterial species (in must always be interpreted in the light of (a) two consecutive mid-stream urine specimens, whether one organism is dominant, (b) how or, alternatively, a positive speci®c rapid the sample was collected (chronic catheteriza- examination and §108 CFB/L mid-stream tion versus mid-stream specimen), (c) the pres- urine in one sample). ence of other features suggesting either true When two species are present in mid-stream infection (presence of WBC) or contamination urine, it is proposed that only these be regarded (presence of squamous epithelial cells), and in symptomatic patients and that the cut-offs (d) clinical signs, symptoms and history. True are set at §106 CFB/L for primary pathogens infection with two species (even three occa- and at §108 CFB/L for secondary pathogens. sionally) may occur. Policies concerning how When more than two species are isolated from such mixed growths are subsequently investi- mid-stream urine, only ®ndings of primary 30 ECLM ± European Urinalysis Group

TABLE XIII. Suggested limiting concentrations of bacteria colonies justifying identi®cation and susceptibility testing in the laboratory.

Symptomsa and Inoculum, Species typeb and number Signi®cant colony concentration specimens min volume CFB/L (CFU/mL)

Mid-stream urine specimen: Yesa 1 mLI 1±2c 106 (103) II 1 107 (women) (104) II 1 106 (men) (103) II 2 108 (105) III 1 108 (105) Noa I ± III 1 108 (105) Yes (Special) 10 mLd I1±3c 105 (102) Suprapubic aspiration specimen Yes or no 100 mLe I ± IV 1 ± 2 104 (101) Specimen from cystoscopy or single urethral catheterisation: Yes or no 10 mLd I ± III 1 ± 2 105 (102) Specimen from indwelling catheter: Yes 1 mL I ± III 1 ± 3f 107 (104 No 1 mL I ± III 1f 108, f (105, f)

a Yes~The patient has symptoms, No ~ No symptoms, or no information about symptoms. b Suggestive category based on growth characteristics (see Table IX). Species of normal urogenital ¯ora (IV) are examined for susceptibility only if especially indicated. c Usually, only one species is identi®ed if 2 ± 5 similar colonies grow (as locally agreed) and antimicrobial susceptibility is examined. Occasionally, two species may be identi®ed for speci®c patient populations. Three or more species are usually reported as ``mixed culture'' and considered as contaminants. Susceptibility testing of isolates from mid-stream urine specimens as well as other detailed strategic decisions need local clinical and microbiological consultation. d In routine workup, a 1-mL loop is practical. However, in speci®c patient groups, such as in patients with certain urological diseases, or in a precise evaluation of patients with simple cystitis, a result at §105 CFB/L (102 CFU/mL) and a statistically reliable culture result at 106 CFB/L (103 CFU/mL) may be clinically signi®cant. This is obtained only by using a 10-mL loop. This sensitized culture procedure should be specially requested to avoid inadvertent extra work and costs caused by routine application of a 10-mL loop for all specimens. e Suprapubic aspiration specimens should be cultured from a 100 mL inoculum to reach the highest sensitivity, since w104 CFB/L (101 CFU/mL) and a statistically reliable culture result at 105 CFB/L (102 CFU/mL) may be signi®cant. f Three species are isolated on special request only from symptomatic patients (suspicion of pyelonephritis or urosepsis). Susceptibility testing from catheter specimens is done only for E. coli and Gram-negative bacteria if they are present at concentration of 107 CFB/L (104 CFU/mL) per species, or more. For asymptomatic patients, a higher limit of signi®cant growth (108 CFB/L) is suggested for Gram-negative bacteria. pathogens are reported. When no primary The reduced cut-offs for ``signi®cant'' pathogens grow, ``mixed culture'' is noted. bacteriuria necessitate cultures performed on Lower cut-offs are proposed for samples larger volumes of urine. With inoculation of obtained via suprapubic aspiration, cystoscopy 10 mL, one colony would then correspond to and single (in-and-out) urethral catheterization. 105 CFB/L. When culturing suprapubic aspira- Samples from patients with indwelling catheters tion specimens, a 100-mL inoculum is recom- are handled depending on symptomatology: In mended to approach the sensitivity §104 CFB/ symptomatic patients all isolates are considered L. A statistically reliable positive result (using for identi®cation and susceptibility testing. In the limit of 10 colonies on a plate) corresponds asymptomatic individuals the analysis is focused to 106 CFB/L (10 mL inoculum) and 105 CFB/L on Gram-negative rods. (100 mL inoculum), respectively. Findings of bacteria belonging to the normal ¯ora of the urethra and genitalia (group IV) are 7.3. Bacterial cultures reported as specimen contamination. Further examination should be done only if especially The culture methodology is structured into indicated after individual evaluation. three levels to satisfy different clinical and European Urinalysis Guidelines 31 bacteriological needs (see Section 5.1 for general speci®city. Details of the comparison method de®nitions): Quali®ed comparison methods for bacterial culture are explained in Appendix, (Level 3; due to the lack of documented Annex 13.1. reference methods), quantitative ®eld methods (Level 2) and ordinal scale or rapid methods 7.3.2. Routine cultures on plates. Common (Level 1). Individual laboratories and their sense is obviously needed in a clinical bacter- customer clinicians must decide, on the basis iological laboratory to ensure both high clini- of local patient populations and resources, the cal sensitivity and high speci®city of routine way in which urine cultures should be organized reports. This may be in¯uenced by the micro- locally. biology traditions and costs of health care in different countries. The ideal in terms of 7.3.1. Comparison method for bacterial culture patient preparation, specimen collection and (Level 3). A quali®ed comparison method transport, and ®nally in the analytical process (Level 3) is needed to establish traceability of may not be attainable. Previously suggested local variations in technology. This method signi®cance limits and volumes of inocula can also be chosen for individual patients or (Table XIII) serve as goals for optimal prac- patient groups. The indication for more tice. Expert microbiology advice is necessary extensive culture is usually evaluation of in locations possessing only minimal results from routine methodology, or from resources. other examinations suggestive for urinary Inoculum: A1-mL disposable loop inoculum tract infection, such as those from rapid is accepted for routine practice although it is examinations. less than optimal. On special request, higher Inoculum: Since the cut-off for detection of volumes should be used as recommended. bacteriuria is reduced, a greater culture inocu- Statistical reliability starts at 107 CFB/L with lum is necessary. With a precisely pipetted a1-mL inoculum, but this is also in¯uenced by 10-mL inoculum, a culture result of 106 CFB/L the accuracy of the ¯uid volume in the loop. is equivalent to 10 colonies on the agar surface, The statistical unreliability of 2 ± 3 colonies whereas 105 CFB/L equals only 1 colony on a growing on a plate is understood and consid- plate. At 10 colonies, a minimum imprecision ered when expressing the limits of signi®cant CV~30% (SD~d10) is due to the uneven growth in Table XIII, but accepted in routine distribution of bacteria in the urine. For more practice. critical applications, such as evaluation of a Culture: Quantitative culture should be per- particle analyser, a more accurate procedure is formed at least on a relatively non-selective agar needed, e.g., by pipetting 100 mL volumes of plate, such as Cystine-Lactose Electrolyte De®- serially diluted specimens (see Annex 13.1.). cient (CLED) medium. Alternatives to CLED Culture: The quantitative culture should be agar also exist [5]. In addition, culture on blood performed on CLED agar in aerobic, on blood agar is recommended as an optimum approach agar in anaerobic, and haematin agar in CO2 ([6, 7]; see Appendix, Annex 13.2). Incubation conditions for 48 h, as opposed to the 24-h for 24 h is suf®cient though not optimal as incubation of CLED agar in the routine routine. It can be less reliable when fastidious method. Comparisons with blood agar are organisms cause infection or when mixed particularly important to describe and follow culture is obtained. Inoculation of four different the reliability of the routine culture method for specimens on a single plate (``quarter-plate accreditation of microbiological diagnostic pro- technique'') is prone to contamination and cedures. The blood agar under anaerobic reading problems. It is, however, understood conditions is recommended in parallel for the that it appears advantageous particularly in following reasons: it serves as a quantitative large laboratories with high throughput. Stra- reference, reveals false-negative results on tegies to reduce the number of non-signi®cant CLED plates, improves the isolation of bacterial cultures are highly encouraged to Gram-positive bacteria, assists with the identi- improve the quality of those cultures that are ®cation of isolated species and helps in clearly indicated. differentiating contaminants from uropatho- The uncertainties of the routine process genic species, thereby increasing diagnostic should be controlled by the comparison 32 ECLM ± European Urinalysis Group method (Section 7.3.1). The addition of blood Problems with quantitation: Inoculum on a agar incubation in a CO2 atmosphere for 1z1 dipslide is not reproducible. Any colony count days will increase both sensitivity and speci®- is therefore uncertain. At high bacterial city. Depending on the success of these adjust- concentrations, con¯uent growth necessitates a ments, the routine process may be considered as subculture for identi®cation. Mixed ¯ora may a quantitative method (Level 2) or an ordinal also be missed. scale method only (Level 1). Details of routine To reduce the risk of false interpretation, culture are included in Appendix, Annex 13.2. dipslide-culturing should be viewed as an ordinal scale method for bacteriuria, combined 7.3.3. Dipslide cultures. Dipslide cultures can with the possibility of identifying E. coli. It may be considered suitable for identifying negative be possible to satisfactorily handle up to 90% of specimens and those with signi®cant growth ambulatory urine cultures requested for sus- of E. coli at ordinal scale level (Level 1). For pected urinary tract infection with this simpler reliable results, reading by laboratory profes- rapid procedure. Details of the method are sionals, or special training to clinical person- described in the Appendix, Annex 13.3. nel is recommended. The urine dipslide system has developed from 7.3.4. Enrichment cultures. Enrichment cultures devices originally designed over 30 years ago from urine using bottled broth media are dis- [206]. It presently consists of agar media coating couraged because no quantitative results are both sides of an immersible plastic paddle that obtained. Speci®city is poor due to inevitable can be dipped in urine at point-of-care, and sent overgrowth of commensal organisms. to the laboratory in a sealable universal container. Media usually consist of a relatively 7.3.5. Identi®cation of species. A proposal for non-selective agar such as CLED on one side identifying species of urinary isolates in routine and a Gram-negative selective medium such as diagnostics is compiled in the Appendix, MacConkey on the other. There may also be a Annex 13.4. It is based on the practice where third section of agar that detects beta-glucur- cultures are performed on CLED and blood onidase activity by producing brown or black agar. The tabular data suggest at least the mini- colonies [207]. These devices provide compar- mum criteria for identi®cation of each species. able results to those obtained using standard loop inoculation methods [208], and may be 7.3.6. Antimicrobial susceptibility testing. The useful where delays between the point-of-care goal of susceptibility testing is to allow the and the laboratory are either unavoidable, clinician to choose the correct antibiotic for unpredictable, or both, and where no facilities individual urinary tract infections, and to help for refrigeration are available. In two studies, in investigating the reason for treatment fail- the selective agar has been shown to have a high ure. The antibiotic sensitivity of pathogenic sensitivity (95.5%, 92%) and speci®city (97.2%, organisms should usually be examined when 100%) for detecting E. coli in urine [207, 209]. isolated in clinically signi®cant concentrations. These studies were performed in clinical micro- To reduce the workload in laboratories, it is, biological laboratories ± and did not evaluate however, suf®cient to rely on the predicted the interpretation of dipslide cultures by antibiotic sensitivity of individual bacterial spe- primary health care personnel. cies when community-acquired lower urinary Problems with species identi®cation: Estimat- tract infection is suspected in low-risk female ing growth on CLED compared to MacConkey patients, and the information on local resis- agar in order to differentiate Gram-positive tance in strains is available (then, even bacter- from Gram-negative bacteria has far too many ial culture may be unnecessary). The antibiotic sources of error to be useful in practice. For sensitivity patterns of individual species vary example, enterococci and staphylococci can considerably according to location, patient grow weakly on MacConkey, fungi grow on case mix and background antibiotic usage. MacConkey and certain Pseudomonas species do not grow on MacConkey agar. In addition, Choice of method. Standardized disc diffusion the dipslide technique does not allow identi®ca- is the frequently used method. Non- tion of Staphylococcus saprophyticus. standardized disk diffusion, so-called ``direct European Urinalysis Guidelines 33 antimicrobial susceptibility testing'', is used in selection of antibiotics are advised to increase many laboratories to increase the speed of clinical effectiveness of antimicrobial suscep- diagnostics. Comparisons have shown a 96% tibility testing and to direct clinicians to the use agreement with the standardized practice of inexpensive and effective antibiotics least [210 ± 212]. The method can be useful pro- likely to generate resistance. vided adequate exclusion criteria are in place. Trained personnel should be responsible for the method and it should not be used on 7.3.7. Detection of antimicrobials in urine. Anti- mixed culture or on low bacterial concentra- microbial drugs can lead to failure of bacterial tions (v 108 CFB/L) because the results are growth despite the presence of large numbers not clear-cut. Thus, rapid methods can assist of bacteria in the urine. If bacteria or leuko- by allowing laboratories to incorporate direct cytes are detected by microscopy, confusion susceptibility testing into their work-¯ow only occurs if no subsequent growth is seen. We for those specimens with abnormalities likely suggest the following as the most economic to represent a true infection (Fig. 1). approach: Request forms should be designed so Several alternative microdilution methods are that the presence and type of antibiotic therapy commercially available for automated or semi- can be documented by the requesting physician automated susceptibility testing. Some of these at the time the specimen is sent (see Appendix, provide acceptable accuracy in comparison with Annex 10.1.1). This will eliminate the necessity the traditional reference methods for antimi- to perform any examinations. crobial susceptibility testing; they also provide If clinical information on ingested antibiotics results within a working day or sooner. by the patient cannot be easily transferred, a Further speci®c instructions concerning the surrogate method is to measure concentrations procedures for antimicrobial susceptibility test- of the most common anti-microbial agents in ing are beyond the scope of this document. urine. Laboratories may also choose to detect Guidelines exist that describe how such exam- antimicrobial activity (AMA) by placing a drop inations should be performed, such as the British of urine on a lawn of sensitive E. coli; a zone of document [213], the NCCLS document [214] and growth inhibition will form in a positive case on reports of the Swedish Society of Medicine's the site where the urine was placed. Another Reference Group for Antibiotics (SRGA-M) option is Bacillus subtilis (ATCC 6633) grown on [215], and German guidelines [216]. Local Mueller-Hinton agar [6]. A report stating that recommendations on the practice and minimum antimicrobial activity was detected is then given.

FIG. 1. Application of direct susceptibility testing in microbiological investigation of urine specimens. 34 ECLM ± European Urinalysis Group

7.3.8. Automation of bacterial culture. For Measurement principles. Multiple test strips large microbiology laboratories, automated including the nitrite ®eld are described in Sec- urine culture may be an option in addition to tion 5.3.1. Several new technical developments the methods used for rapid detection of have been suggested [63, 67, 70, 93, 132, 219 ± potentially positive cultures. Several auto- 222]. These include enzymatic examinations mated systems exist which satisfy the require- based on adenosine triphosphatase/luciferase, ments of Level 2 cultures. Advances in catalase or glucose oxidase, ®ltration methods, technology mean that laboratories, especially advanced microscopy of Gram stained smears large ones, will be tempted to assess analysers and automated ¯ow cytometry. Some of these that can replace expensive manual work. methods are well suited to point-of-care test- Finally, it is recommended that any new auto- ing, but sensitivity and speci®city of procedures mated technique be assessed against accepta- must be investigated and quoted in relation to ble Level 3 comparison methods. Evaluation the patient population studied. Others require should preferably be three-way, where perfor- laboratory-based experience for their perfor- mance is also compared to the currently used mance, interpretation, or both. method (usually not Level 3). Urine particle analysis is discussed in Section 6. Bacterial analysis may theoretically adopt either direct particle counting (of living and 7.4. Bacterial detection by non-culture (rapid) dead bacteria) or automated rapid culturing of methods urine in selected media. Laboratory staff must There is a need for high performance rapid remember that these analysers may detect methods for the detection of bacteria in urine. particles and not growth. In microbiology This applies for the routine laboratory handling evaluation, the arising new view is that some of large numbers of specimens, for emergency specimens may contain true, aggressive patho- diagnostics and for detection of asymptomatic gens that do not grow on conventional agar bacteriuria in selected patients groups, such as plates. There may also be the issue concerning pregnant women. Development of analytically the presence of antibiotics in the urine that sensitive and speci®c rapid procedures for inhibit growth. Finally, it should be noted that detection of bacteria is encouraged both for automated techniques assigning Gram status to adults and children [217]. organisms were very helpful since they could Sensitivity view: A high performance high- predict effective antibiotic treatment. throughput screening procedure with low false- negative rates (ideally v10%) would identify true negative specimens and allow signi®cant reduction in costly and unnecessary urine 8. STEPWISE STRATEGIES IN culture. Health care savings would then be URINALYSIS achieved even in the face of an appreciable 8.1. Examinations for general patient false-positive rate. The validation of method populations performance for detection of bacteria at con- centrations less than 108/L (v105/mL) becomes Most of the costs arising from screening very important. programmes result from the need for con®rma- Speci®city view: Acutely ill patients need an tion of positive ®ndings. Screening of com- examination with high speci®city (preferably pletely unselected individuals, i.e. at general w90%) to demonstrate the presence of bacteria epidemiological level, is discouraged except for reliably when the clinical presentation is not research purposes. Selected asymptomatic indi- obvious from symptoms and signs. In these viduals may be investigated if justi®ed by cost/ cases, a positive speci®c result from a rapid bene®t analyses, e.g., pre-school children or diagnostic procedure supports a correct im- pregnant women [88, 223]. Examinations for mediate treatment decision, while the rest of the diseases in kidneys and urinary tract can be cases await results from bacterial cultures [218]. recommended for many clinical populations A typical question is how does a method cor- (~patients attending health care services in rectly classify contaminants and debris as hospitals or at ambulatory clinics because of negatives. their symptoms or diseases), but not for all European Urinalysis Guidelines 35 patient groups. Use of urine test strips, as well despite a standardized specimen) it may be as of other laboratory measurements, should appropriate to perform additional investiga- always be associated with prognostic signi®- tions to detect or isolate organisms with special cance [224]. The outlined strategy has been growth requirements (Chlamydiae, gonococci, derived from earlier recommendations [2, 123, Mycobacterium tuberculosis and fungal infec- 225, 226]. tions) by collecting another appropriate specimen In addition to the minimum properties shown and investigating it accordingly. (Fig. 2), acute cases generally bene®t from General screening of asymptomatic indivi- measurements of urinary glucose, ketone bodies duals for bacteriuria should be avoided. and pH (urine stone formers). The strategy is Bacterial culture is initiated only if the only an average general approach and not asymptomatic bacteriuria is going to be treated sensitive to the individual needs of patients. with high probability (Section 8.2.3 for detailed These strategic proposals are designed to patient groups). improve the general ef®ciency of urinalysis. Clinical decisions should be based on direct 8.1.2. Detection of proteinuria. Transient pro- patient data; laboratory and imaging methods teinuria is a fairly common ®nding among provide ancillary information only. The recom- acutely ill patients [227] even if detected at the mendations presented describe some approaches traditional 0.2 g/L albumin level. In these based on the diagnostic performance of the cases, an occasional positive result may lead methods available. The term ``®rst visit'' refers to inappropriate investigations. Exclusion of to the initial diagnostic investigations; the transient proteinuria is accomplished by follow-up investigations should be based on repeated measurement and review of anamnes- evaluation of detailed needs. Consultation of a tic data. nephrologist, clinical specialist for infectious Quantitative speci®c measurements are gen- diseases, internist, paediatrician or another erally not part of a primary screening process specialist may be necessary to plan all medically for general patient populations; they are second- required investigations. A multiple urine exami- line investigations. Total albumin or protein nation should be performed at point-of-care or quantitation can be requested ± if not already in the laboratory as locally agreed. the ®rst-line method ± as albumin or total protein/creatinine ratio (or albumin or protein/ 8.1.1. Detection of urinary tract infection. It is osmolality ratio), or as measurement of 24-h important to clearly differentiate high-risk protein excretion rate. Orthostatic (postural) patients from low-risk (recurrent female) cysti- proteinuria can be identi®ed by separate day tis patients when applying any diagnostic stra- and overnight collections. Selected patient tegies for symptomatic patients suspected for groups do require a sensitive protein determina- urinary tract infection (see Section 7.1 for tion to detect renal damage already in the ®rst de®nitions and 8.2 for strategic proposals). line. In addition to albumin and a tubular Acute cases may also bene®t from a labora- marker protein, immunoglobulin light chain tory result within 0.5 ± 2.0 h, while a highly determinations may be important (see Section speci®c result, such as that from bacterial 8.3). culture, serves to ®nally classify cases even after 1 ± 2 days. 8.1.3. Detection of haematuria. Exclusion of Some clinical patients may have general or other coloured substances (red beets, por- vague symptoms that were not associated with phyria, drugs; Table II, page 14) should urinary tract infection initially. These cases then always be a ®rst thought. Particle analysis is appear as unexpected ®ndings. Exclusion of needed to obtain a count of RBC excretion urinary tract infection at a cut-off of 108 CFB/L from a standardized specimen. Renal elements is possible from concentrated morning speci- may be seen on urine microscopy. The RBC mens with a sensitivity of 90% by rapid methods should be assessed for isomorphism, dys- (see Sections 5.3.1 and 7.4). Urine microscopy morphism and the presence of acanthocytes if should be performed to detect possible renal no proteinuria is present. This must be per- involvement if relevant. In cases of so-called formed in a laboratory capable of distinguish- ``sterile pyuria'' (no usual uropathogens grown ing the different types of RBC morphology. 36 ECLM ± European Urinalysis Group

FIG. 2. Urinalysis examinations for general patient populations. European Urinalysis Guidelines 37

An alternative approach is differentiation of and for excluding them from further investiga- the bleeding site based on urinary IgG/albu- tions for urinary tract infection, whilst it is min and alpha-2 macroglobulin/albumin ratios understood that such sieving strategies must not as measured in a centralized laboratory [123]. be applied in high-risk cases. Urine sediment examination only occasion- ally offers additional information after a test- 8.2.1. Symptomatic low-risk patients. Low- and strip measurement in unselected emergencies high-risk patients with respect to urinary tract [71], but it helps in predicting renal damage if infection were de®ned in Section 7.1. Low-risk investigated carefully from a standardized patients with suspicion of lower urinary tract specimen [22]. When the test-strip result is infection can be examined as follows: No negative for erythrocytes, leukocytes and investigations are needed if the diagnosis is albumin/protein, sediment microscopy is not clear from the symptomatology. Empirical generally warranted [228, 229]. Automated treatment can be justi®ed with known local particle counting possesses higher precision epidemiology of community-acquired infection. than visual urine sediment examination, but it If the symptoms are not clear, a speci®c may not detect all signi®cant renal elements examination (speci®city w90 ± 95% for uro- [69]. pathogens) allows a rapid con®rmation of A clinical (urological and/or nephrological) bacteriuria and justi®es treatment. Sympto- consultation is usually needed to search for matic low-risk cases remaining negative with possible local or systemic diseases causing the speci®c rapid examination (during emer- haematuria. The presence of a haemorrhagic gency hours) should be treated or investigated diathesis should not be forgotten. Urothelial by a culture method after obtaining a standar- and renal cancers often present as unexpected dized morning specimen. haematuria. On the contrary, prostatic cancer Follow-up of low-risk patients should be rarely presents with haematuria or proteinuria. organized as follows: if no symptoms re- Detection and follow-up of prostate cancer is main after treatment, no further examination undertaken by measurements of serum PSA is needed. If symptoms persist, bacterial (prostate-speci®c antigen), currently as total culture with antibiotic susceptibility testing is and free PSA molecules. Screening for bladder warranted. tumour cell antigens in urine is under develop- Epidemiology of uropathogens: The pre- ment. Experienced cytopathologists must inter- requisite for treatment of urinary tract infection pret urothelial tumour cell cytology. without bacterial cultures is an epidemiological knowledge of uropathogens and their antimi- crobial susceptibilities within a local commu- 8.2. Detection of urinary tract infection, nity. This gives valuable information on speci®c populations relapsing infections often due to reduced The suggested sieving strategy to reduce the antibiotic susceptibility of the species. National number of bacterial cultures in patients sus- and regional efforts to create and maintain pected for urinary tract infections (UTI) is these data are encouraged. illustrated in Fig. 3. Cultures of low-risk patients are not needed [207, 219] with given 8.2.2. Symptomatic high-risk patients. Speci- precautions. mens from symptomatic high-risk patients This strategy can be dif®cult to follow exactly should always be cultured for uropathogenic because of problems with speci®city, such as bacteria both to con®rm the diagnosis and to contamination and false-positive reactions, and ensure treatment (with the help of antimicro- sensitivity, because rapid detection methods for bial susceptibility testing). Urgency of micturi- uropathogenic bacteria are not perfect to date. tion may prevent suf®cient bladder incubation Despite this, such strategies should help in time, leading frequently to false-negative reducing traditional cultures markedly, but results ± even with culture methods. Signi®- should remain sensitive to needs of problematic cant growth may be as low as 105 ±106 CFB/ or speci®c cases. For example, it may be L (corresponding to 102 ±103 CFU/mL), occa- acceptable to use a highly speci®c rapid sionally needing a culture with a 10-mL (or examination for identifying positive patients even 100-mL) inoculum to reach this sensitivity 38 ECLM ± European Urinalysis Group

FIG. 3. A sieving strategy to reduce requests for bacterial cultures. reliably. The speci®city of low-count ®ndings Exceptions to be treated may include preg- should be considered. Local microbiological nant women (a prevalence of about 5% for consultation is recommended to establish asymptomatic bacteriuria), some patients rational routine processes. undergoing urogenital surgery (if the indwelling catheter is going to be used postoperatively) or 8.2.3. Asymptomatic bacteriuria. The need for receiving organ transplant, and many immuno- investigation of asymptomatic individuals is compromised individuals. During pregnancy, questionable. Liberal treatment of patients bacteriuria is treated to prevent symptomatic with asymptomatic bacteriuria by using anti- infection and premature birth [230, 231]. A microbials results in development of multi- rapid method, reaching analytical sensitivity resistant strains that replace the less danger- w90 ± 95% at 108 CFB/L (equal to 105 CFU/ ous original species. mL) when compared to bacterial culture is European Urinalysis Guidelines 39 recommended, with con®rmatory culture of diagnosis is often based on renal imaging and positive cases. biopsy. Secondary renal diseases are more common than primary due to the high prevalence of diabetes mellitus, hypertension and systemic 8.3. Sensitive detection of renal disease ± rheumatic diseases. For speci®ed asymptomatic speci®c populations patient populations, sensitive screening pro- Table XIV lists examinations used to detect grammes should be established when improve- renal disease in patient populations with high- ment in life expectancy or cost containment is risk for renal damage. The minimum and anticipated. Currently, such a screening pro- optimum lists refer to economical possibilities gramme is established for incipient diabetic by different health care service providers. nephropathy by sensitive measurements of Primary renal disease may present with albuminuria. Renal involvement in hyper- symptoms. Haematuria and albuminuria, some- tension should be screened at least by measure- times even leukocyturia may be demonstrated ment of total protein excretion (24-h collection by sieve measurements; in addition, quantitative or total protein/creatinine ratio) if sensitive total protein excretion (measured as a 24-h albuminuria measurements are beyond existing excretion rate, or protein/creatinine ratio of resources. Renal disease secondary to multiple morning urine) is usually elevated. In tubular myeloma may be detected by ®nding immuno- disease, defect in urine concentrating capacity globulin light and occasional heavy chains in may be demonstrated (measured by relative urine in association with glomerular or tubular density or osmolality). Albuminuria is a major proteinuria. feature in most glomerular diseases. Monitoring The role of urine particle morphology is to for side effects from clearly tubulotoxic drugs, con®rm the presence of renal damage (casts, such as aminoglycoside antibiotics, many cyto- renal tubular cells, dysmorphic red cells) toxic drugs and some non-steroidal anti-in¯am- because non-renal proteinuria due to, e.g., matory drugs is possible with measurements of fever, epileptic seizure or cardiac insuf®ciency is also frequent. Occasionally, renal elements urinary a1-microglobulin, retinol-binding pro- tein (RBP) and other markers (see Section may occur in urine due to secondary damage 5.4.1). caused by abdominal or retroperitoneal (surgi- Measurements of GFR are important in renal cal or gynaecological) disease, or hypotension disease. Urine creatinine measurement is also due to cardiogenic shock. The analyses of both needed for calculation of creatinine clearance. urine proteins and formed elements complement Positive ®ndings from these investigations each other in detection and follow-up of renal should be followed by more speci®c diagnostic damage. The former are more sensitive but less procedures. In central hospitals, de®nitive speci®c, while the latter are more speci®c but less sensitive markers of renal damage.

TABLE XIV. Measurements for sensitive detection of renal disease in urine.

Minimum measurements 9. QUALITY ASSURANCE Urine albumin (20 ± 30 mg/L) and total protein (quantitative, sensitivity 50 mg/L) 9.1. General principles Urine microscopy Optimum quantitative measurements additionally Urinalysis faces many challenges when its Urine albumin (sensitivity 2 ± 10 mg/L) quality is being de®ned formally. While trace- Urine a1-microglobulin (sensitivity 2 ± 10 mg/L), ability and description of uncertainties are well retinol-binding protein, or another tubular marker established in clinical chemistry and haematol- protein Urine creatinine ogy, these concepts are still premature to the (or another reference quantity for volume rate) semi-quantitative or ordinal scale examinations Optional urinary proteins in special cases used in urinalysis. For morphological analysis, Ig, k and l light chains, a2-macroglobulin systematic peer reviews are being used in Other diagnostic examinations (based on clinical presentation) cytopathology to reach the highest possible agreement between observers [232 ± 234). In 40 ECLM ± European Urinalysis Group microbiology, error in the identity or in a count address systematically all facets of laboratory obtained from cultured colonies may be dif®cult activity. Accreditation based on such interna- to set into an acceptable statistical framework. tional standards as EN45001 and ISO/IEC Quality assurance practices must also encom- Guide 25, or the later ISO17025, should be pass point-of-care testing [235 ± 237]. Finally, considered for all facets of urinalysis as well. the increasing costs of health care tend to These guidelines encourage local establishment challenge the acceptable minimum requirements of quality systems even in small European [238]. These guidelines are sensitive to such laboratories working under the supervision of pressures and are designed to propose practical specialized laboratories. Quality systems need solutions. regional co-operation and support. Publications already exist at a general level [239 ± 241] and speci®cally for quantitative serum chemistry [242]. A Scandinavian group 9.1.2. Quality manual. General principles of has advised on assessment of analytical quality good laboratory practice apply to all clinical speci®cations from a theoretical point of view laboratories. The quality system is described [243]. In urinalysis, the healthy concentrations in the quality manual of a laboratory. The or excretion rates of many substances or following contents were suggested by a particles are traditionally reported as ``nega- Scandinavian group [248]. As an example, tive'' rather than numerically. The performance some speci®c views are reminded for micro- goals at detection limits must take into account biological investigations as performed in the large biological variations that occur at low general laboratories under the supervision and concentrations characteristic for the upper level support of a local microbiology laboratory. of the reference range in health. Adherence to the international vocabulary of 9.1.2.1. Description and identity of laboratory metrological terms (VIM) is encouraged [126, ± In most European countries, national gov- 244 ± 246], although this was not done thor- ernments have designated, by decree, certain oughly in these guidelines owing to their only microbiology laboratories to investigate infec- gradual acceptance in routine work. Substance- tions of epidemiological importance. National based SI units (e.g. mmol/L) should replace regulations may also apply to other microbio- conventional mass-based (e.g. mg/dl) units in logical examinations, such as cultures from expressing results. The volume unit litre (L) has urine. been used throughout this document, including bacterial colony counts (particle concentra- tions), as an example of this standardization. 9.1.2.2. Quality policy. (Modi®ed from the The term accuracy now encompasses both EN45001 suggestion for a policy statement) precision and trueness of measurements; and, ``The aim of the laboratory is to provide inversely, inaccurate measurements may be clinically useful information through laboratory erroneous because of systematic effects (bias) examinations of samples from patients, taking or random effects (imprecision). into account the allocated resources. The quality policy is implemented by the following 9.1.1. Quality systems. The concept and prac- means: tice of Total Quality Management (TQM) is Proper specimen collection, labelling and now commonplace in many clinical labora- transport tories. Laboratory practice can be divided into Reliable analytical work three stages (pre-analytical, analytical and Turnaround time within speci®ed limits post-analytical) all of which must be quality Results reported in a clear format and assured in order to obtain reliable chemical supplemented with relevant explanatory infor- and microbiological results for patient care mation (either together with the result or in a [247]. separate laboratory handbook) The essential components of quality systems Appropriate communication is maintained may be created nationally with reference to with clinical customers to ensure correct published national guidelines. Organization of integration of laboratory information into Good Laboratory Practice (GLP) should evaluation of patients''. European Urinalysis Guidelines 41

9.1.2.3. Staff and education. Training and cer- quality of equipment should be maintained ti®cation of competence: The necessary educa- and its performance monitored daily. In urine tional requirements for certi®cation of culture, this includes, e.g., incubators with specialist competence should be documented. their required temperatures and atmospheres, Training and certi®cation should be carried and culture plates. out in consultation with the microbiology 9.1.2.8. Safety and environment. Regular hand- laboratory. Competence should be updated ling of biologically hazardous material needs regularly (e.g., yearly) and after long absences special attention in routine procedures. The (e.g., absences for longer than 6 ± 12 months). details of safety management should be orga- 9.1.2.4. Management and quality assurance. nized and explained. The documentation and procedures describing The facility must comply with current safety the quality system in detail must only be legislation. changed with the authority of the Head of the 9.1.2.9. Research and development. The quality laboratory or persons nominated to authorize manual should state how a laboratory is such changes. These documents must form a developing its routine programmes. Possible laboratory manual (with separate appendices), participation in basic research should be relevant sections of which are issued to staff. quoted as well. When changes are made, all working copies of the modi®ed document must be updated. 9.1.2.10. Measurement procedures. Reference The old copies must be stored (one archival materials and methods produced by manufac- copy) or destroyed (at the working sites). turers must be clearly described to allow audit Continuous quality improvement is a result and traceability of measurements. of professional external and internal audits, in Detailed operating procedures for any inves- addition to development of analytical tech- tigation must be in place, as must instructions niques. Audits should be arranged on a regular for use of equipment. basis, e.g., once a year. Level of identi®cation of uropathogenic The laboratory manager must have access to species must be discussed with the local advisory documents issued by relevant bodies microbiology laboratory: a small laboratory including health notices, safety information may decide to report only negative culture bulletins, health equipment information, health results or to undertake a preliminary typing of building notes and health equipment notes. positive cultures in agreement with the support- ing microbiology laboratory. 9.1.2.5. Records, maintenance and archiving. Reliable microbiological results are often Final patient records and quality assurance dependent on detailed adherence to speci®c data should be archived according to national procedures. Procedures and interpretation or institutional regulations, typically 3 ± 15 charts should be followed strictly. The medi- years, to allow later investigation for clinical, cally responsible physician or other specialized epidemiological and, occasionally, for legal clinical microbiologist, who should be available, purposes. Laboratories should also keep must have approved these. records on their maintenance. Both commercial kits and media developed in the laboratory should be veri®ed in a traceable 9.1.2.6. Facilities. Careful planning is needed manner (standardized procedures and strains in siting a microbiology laboratory to ensure obtained from international collections). stable environments (temperature, atmosphere, It is important that the number of examina- humidity, etc.), minimum contamination risks tions is large enough to maintain the level of between specimens and safety of personnel. skill (approximate minimum is about 20 urine Facilities must be available which are cultures/week). designed to be safe for the processing and containment of speci®c pathogens. 9.1.2.11. Request protocols, sample collection and handling of laboratory samples. Instruc- 9.1.2.7. Equipment and material. In both tions for the receipt and recording of speci- manual and automated laboratory work the mens must be clear and unambiguous. 42 ECLM ± European Urinalysis Group

Detailed clinical information, including cur- should explain detailed procedures for speci- rent antimicrobial treatment as well as the real men collection and transport of microbiologi- time and anatomical localization of the cal investigations, including those for urine obtained specimen (sample), is crucial for collection. Interpretation of usual culture interpretation of bacterial cultures. In urine results, antimicrobial susceptibility testing and cultures, the collection procedure is of similar rapid microbiological methods should also be importance (see Annex 10.1 for detail). included in the handbook. Physicians with specialist competence in 9.1.2.12. Veri®cation of results. Data transfer clinical bacteriology should be available for to individual patient records may be vulnerable clinical consultation. in manual work with plates. Procedures for specimen identi®cation, veri®cation of results and reporting on paper or by electronic data 9.2. Test strips and equivalent rapid processing systems should be explained. examinations 9.2.1. Trueness of measurements. The ordinal 9.1.2.13. Analytical control. A method must be scale (semi-quantitative) measurements are set in place for monitoring results, with the usually expressed as categorized data. Perfor- ability to recognize those where failures have mance can be described as sensitivities and arisen due to laboratory or human error. speci®cities, i.e., as maximal allowable frac- External quality assessment (EQA): labora- tions of analytically false-positive (FP) or tories should participate in external quality false-negative (FN) measurements against best assessment programmes. The purpose of EQA practical comparison methods. Reference is to assess how well a laboratory ful®ls the methods, according to a strict de®nition, are analytical quality speci®cations of bacterial currently unavailable. culture (Section 9.5.2). It is recommended that evaluation data be Current results of EQA must be displayed for classi®ed within two limits obtained from the all to see if they wish and a ®le of all results for comparison method: a detection limit (LD; from the past few years must be held and available the point where the ordinal scale method starts for inspection. to give positive results) and a con®rmation limit Internal quality control (IQC) by recom- (LC; from the point where all ordinal scale mended strains should con®rm the daily results should be positive). These delineate the process. ``grey zone''. From experience with test-strip IQC of even dipslide cultures should be technology, it is recommended that the ratio organized with the microbiology laboratory. between the concentrations LC/LD~5 (see Organized submissions of sample cultures Appendix, Annexes 11.1.2 and 11.1.3 for (dipslides and plates) to the supporting micro- detail). Optimal trueness of measurements is biology laboratory for veri®cation are also very suggested to be a FP rate v10% at LD and a useful (see Section 9.5.2 and Annex 13 for FN v5% at LC (compared with the most detail). accurate, closely related method) (Tables XXIII & XXIV). In many situations, or with a less 9.1.2.14. Report. The procedures for issuing optimal comparison method, a minimum per- reports must be clear and unambiguous. formance is acceptable (see Appendix, Annex Delay in reporting results from bacterial 11.1.1 for principles of measurements). With cultures should be avoided and must satisfy many comparisons, such as in leukocyte and the clinical need for treating the patient (see erythrocyte detection, the different principles of Section 9.5.3). measurement (enzyme activity versus chamber Reasons for any delay in reporting results counting) must be understood for correct beyond the agreed limits must be identi®ed, as interpretation. The same applies to comparisons must the likely clinical impact of such delay. between bacterial culture and rapid chemical Corrective measures must then be applied. examinations, such as the nitrite or similar examinations. 9.1.2.15. Co-operation with the clinicians, ward For more than two ordinal scale categories, staff and patients. Laboratory handbooks agreement should be calculated using k statis- European Urinalysis Guidelines 43 tics, thus subtracting random agreement from span. Dilutions of control solutions (with buffer observed data [249, 250] (see Appendix, Annex or pooled human negative urine) help in 11.1.4). The simple k coef®cient is used when following performance at concentrations in there are only two or three classes. For multiple the low positive range, where few stable control (4 ± 5) classes, the agreement should be calcu- solutions are available. However, since there is lated based on weighted k coef®cients. In this a matrix problem in test-strip reading, stable case the sum of expected disagreement exceeds low positive control solutions would be better 100% because of the squared weighting factors. than daily dilution of high positive control Thus, the goal must be tighter (Table XV). It is solutions with aqueous buffers. possible to arti®cially increase the k value by subdividing the results into six or more classes, but then the goal for a minimum performance 9.3 Quantitative chemical measurements should be even higher (0.75 ± 0.8). The quality These guidelines acknowledge the efforts of speci®cations are based on common sense: any the IFCC in creating reference materials for clinical laboratory examination should classify clinical chemistry [251]. For most urinary more than half of the non-random cases constituents, genuine reference materials and correctly. Arbitrarily, this is equivalent to k reference measurement procedures are awaited. w0.6, although optimally the value should be Co-operation between manufacturers in prepar- w0.8 if achievable. ing calibrators for their quantitative tech- nologies is greatly appreciated in achieving 9.2.2. Precision. Success in reproducing ordinal comparable concentrations in clinical labora- scale measurements over time is expressed as tory practice. Recommendations exist for ana- the fraction of correct results within binomial lytical quality speci®cations for routine con®dence intervals. It is recommended that quantitative serum chemistry [242]. These specimens with an original probability of close should be implemented in both internal quality to 100% for the measured category be used, control [252] and external quality assessment since the follow-up is then most precise due [253]. The formulae suggested to describe the to a narrow con®dence interval (see Appendix, maximum allowable total analytical error for Annex 11.1.5.) As an option, despite the ®nal serum chemistry are quoted in Table XVI. For ordinal scale output of patient results, a valid reference methods serving serum chemistry, a continuous signal, e.g. a re¯ectance value maximum of half these errors is proposed [254]. obtained from the measuring device, can be For urine chemistry, these proposals should followed, allowing a normal calculation of be adjusted to re¯ect the fact that changes in performance limits. The low positive range pathological states may be in a logarithmic scale (1z) is more important than the high positive compared to the low or almost negative range (3z) in rapid examinations. concentrations seen in health. Even if chemical For good control, reproducibility (day-to-day analytes are excreted in urine in large quantities, variation) should remain within the binomial they often occur in skewed distributions. Since con®dence limits in internal quality control. the same measurement is often used for both Stable control solutions are already available monitoring and diagnostic testing, a combina- for test-strip reading. Daily implementation tion of quality criteria should be considered. should then be the rule to allow judgements on the levels of assays within a reasonable time 9.3.1. Calibration of the measurements. Cali- bration of urinary total protein is preferably TABLE XV. Analytical quality speci®cations ex- performed using the CRM 470 (albumin) pressed as Kappa coef®cients. standard (see Appendix, Annex 11.2.3). Glu- Optimum Minimum cose measurements can be calibrated with the SRM909 material obtained from the National k coef®cient w0.8 w0.6 Institute of Standards and Technology [255]. (simple) (2 ± 3 classes) There is also a reference method for urine k coef®cient w0.9 w0.7 (weighted) (4 ± 5 classes) glucose measurements approved by CDC/ FDA/AACC/NRSCL [256]. Calibration of 44 ECLM ± European Urinalysis Group

TABLE XVI. Analytical quality speci®cations for compared to the levels of clinical albuminuria quantitative routine serum chemistry. or proteinuria that are up to 100-fold (10,000%) higher in nephrotic syndrome (albumin/creati- MONITORING (precision is important): nine ratio from 1 to 100 g/mol; total protein TEMon¦Ksi, where TEMon~total error in moni- toring, and si~within-subject biological standard from 0.1 up to 10 g/day). deviation Analytical performance should allow detec- DIAGNOSTIC TESTING (trueness/total accuracy is tion of microalbuminuria cases and monitoring important): of diseases based on these ®gures. As an TEDT¦Jsc, where TEDT~total error in diagnostic approximation, the following performance spe- 2 2 testing, and sc~d(si zsg )~composite biological ci®cations are proposed for urine proteins standard deviation; sg~between-subject biological standard deviation (Table XVII). In addition to total protein, analytical quality speci®cations for maximum allowable impreci- creatinine measurement can be performed sion and deviation (bias) have been calculated using the SRM914a serum calibrator from the and reported for calcium, creatinine, P-amylase, same source [257]. phosphate, potassium, sodium, urate and urea based on 24-h collections of urine [259].

9.3.2. Analytical quality speci®cations. Analyti- cal quality speci®cations for monitoring are 9.4. Particle analysis derived from within-subject biological varia- 9.4.1 Statistical uncertainty. Counting at low tion of urine constituents. In the case of pro- particle concentrations needs special attention teinuria, the average coef®cient of variation of because of statistical uncertainty. Particle one subject (Cv ) is 40 ± 50% at the within-subject counting follows the Poisson distribution upper reference limit in health [258, 259]. For [261]: creatinine, it is about 20% [109, 260], giving an average CV for albumin/creatinine ratio of s~dm; and about 40%, and maximally 80% [258]. This CVm(%)~100%/dm~100%/dT/n within-subject variation results in an estimate ~CVT(%)6dn, and of maximally allowable imprecision ¦20 CVT~100%/dT~CVm/dn ( ± 40)%. It would allow discrimination where T~ total amount of counted cells and between twofold and threefold changes in n~number of counted unit volumes (tradi- excretion rates (analyte/creatinine ratios) in tionally squares in a chamber or high-power multiple monitoring of the same patient, by ®elds under a coverslip); m~mean count~T/ using the equation for critical difference n, s~standard deviation, CV ~coef®cient of D ~a6CV, where a~2.77 at a statistical m C variation of the mean count, CV ~coef®- probability pv0.05. T cient of variation of the total count. In discrimination between health and disease At low particle concentrations, imprecision of (diagnostic testing), the average Cv between-subject the total count becomes critical: If only 1 mLis is about 25% for total protein, 75% for albumin counted at a concentration of 36106 particles/ and about 25 ± 30% for creatinine in 24-h L, the result has a theoretical CV ~60% with output [259, 260]. When calculated as ratio to T ideal procedures. Five microlitres of the same creatinine concentration in single-voided urine, suspension gives CV ~26% and counting only about 70% for total protein/creatinine and 40% T 10 mL gives CV ~18%, permitting the calcula- for albumin/creatinine ratio have been reported T tion of an estimate for the mean value. [109]. The composite biological standard devia- tion (coef®cient of variation) is then about 50 ± 90% for proteinuria, depending on the analyte 9.4.2 Suggested analytical quality speci®cations. and method of calculating the excretion rate. Quantitative counting: Urine particle counting The estimate of maximally allowable relative may bene®t from comparison with the quality deviation (bias) would be then ¦12 ± 25%. The speci®cations described for haematological validity of these ®gures must be judged against analysers with their already proven clinical the low level of physiological proteinuria accuracy [262]. In urinary tract infections, a European Urinalysis Guidelines 45

TABLE XVII. Analytical speci®cations for trueness (expressed as relative deviation) and reproducibility (expressed as imprecision) in quantitative urine chemistry, with special reference to proteinuria.

Property Performance level

Optimum Minimum

Trueness, maximum deviation from the target value 10% 25% Reproducibility, maximum imprecision CVday-to-day 20% 50%

10-fold difference (from 10 to 1006106 WBC/ uated (EQA schemes). Each site should docu- L) exists between non-infected and infected ment its personnel training. urines (see Section 6.2.3). Variation in daily volume rate (diuresis) also affects particle con- 9.5. Microbiology examinations centrations easily by a factor of 2 (100%). 9.5.1. Good microbiological laboratory practice. For trueness, an optimum performance Good Laboratory Practice should be imple- should have a relative deviation (bias) v30%. mented in all clinical laboratories, including This is important in particular for evaluation chemistry, microbiology and general labora- purposes. A minimum performance with a tories. Currently, this is often described in a relative deviation v50 ± 100% can be accepted local quality manual. Some details of such a for most routine applications, and in particular quality manual were suggested in Section 9.1.2 6 at particle concentrations below 3610 /L urine. using an example of microbiological investiga- To achieve these ®gures, one has to count tions as performed in general laboratories several microlitres of urine from each sample. under the supervision of a local microbiology In the future, these speci®cations may be laboratory. replaced by more accurate calculations. The volumes of counting chambers or other 9.5.2. Analytical quality speci®cations for bac- similar devices should have an error v10% to terial culture. These speci®cations relate to the be negligible. The manufacturer should specify microbiological analysis of urine in a Level 2 the volume accuracy for the device. facility. The performance at Level 1 may Routine sediment examination (ordinal scale require further adjustments, e.g., at points-of- speci®cations): The generally used standardized care (see Section 7.3 for speci®cations at this sediment technique essentially provides ordinal level). The given speci®cations do not neces- scale analysis of urine particles. This can be sarily ful®l the requirements for the Level 3 controlled against counts of uncentrifuged speci®cations (designed as benchmarks for urine within a chamber or by an instrument. comparative trials and quality control). Varia- In practice, delays in investigation after tion of exact techniques and technology is micturition, losses during centrifugation, suc- permissible and indeed necessary at Level 2, tion of supernatant and inadequate review of the full coverslip area appear to contribute TABLE XVIII. Maximum allowable false-negative most to inaccuracy of results, despite a rates in urine microscopy. standardized procedure. Awareness of the Particle Particle Maximum allowable many possibilities for systematic error helps type concentration false-negative in designing the most accurate procedure for (6106/L) rates (%) routine sediment analysis. Categorized perfor- mance criteria are tentatively suggested below RBC 10 20 100 5 as an experienced opinion (Table XVIII) that WBC 20 10 can be replaced by more accurate speci®cations 200 5 in the future. Bacteria 10 20 Identi®cation of clinically signi®cant urine 100 5 Casts 10 10 particles should be internally reviewed (peer 50 5 review by staff members) and externally eval- 46 ECLM ± European Urinalysis Group because of cost and work-¯ow restrictions, the TABLE XIX. Maximum allowable false-negative rapid performance of emerging technology culture results of uropathogens. and individual patient case-mix. Colony concentration Allowable false negative rate 9.5.2.1. Acceptable false-negative rates for detection by routine culture. Maximum allow- 108/L or more 2% able false-negative rates are suggested for 107 ±108/L 5% detection and identi®cation of established pri- 104 ±107/L 10% mary, secondary and doubtful pathogens, either in pure culture or as part of a mixture, identi®cation for its own sake. It also represents as de®ned by organisms in the common, fairly that which is practically achievable, rather than a common and uncommon groups (as de®ned in ``wish list'' which would require more resources Table IX) in Table XIX. These are based on and time. Individual laboratories must decide expert opinion. whether their particular case-mix requires addi- 9.5.2.2. Acceptable routine level of identi®ca- tional investigation beyond the suggested level. tion. Laboratories working at Level 2 should As an example, isolation of yeast from urine is be able to isolate ± by culture from clinical more relevant in large centres with signi®cant samples ± urine organisms at any concentra- transplantation, intensive care and immuno- tion above those set out in Table XIII, and to suppressed patient activity; the division into identify the species to the given minimum albicans and non-albicans species relates to level (Fig. 4). predicted sensitivity to certain antifungal This is not an exhaustive list of organisms agents. Finally, there is a wide choice of agar recognized as causing urinary tract infection. media that can be used for isolating these Instead, this level of identi®cation of species is organisms. A reduction in the number of media based on a combination of the prevalence and routinely used for urines (such as CLED only) pathological signi®cance of their isolation, as well may be acceptable. Suggestions for readily as the relation between the organism group and performable examinations to achieve such iden- both antibacterial sensitivity and possible resis- ti®cation are given in Appendix, Annex 13.4. tance acquisition. The list is therefore driven by Many other methods may be equally appropriate clinical relevance rather than by microbiological and can be found in microbiology textbooks.

FIG. 4. Minimum level of bacterial identi®cation in clinical urine specimens. European Urinalysis Guidelines 47

9.5.3. Acceptable turnaround times from speci- Point-of-care units and small laboratories men receipt. The requirement for Level 2 are encouraged to consult with colleagues in relates to the importance of speed in terms of the larger laboratories before proceeding with overall cost of patient treatment episodes. It the evaluation of new instruments. Earlier should be viewed in the context of compli- European recommendations exist for proce- cated and potentially serious urinary tract dures in instrument evaluation [263]. When infections. replacing existing techniques with new sys- The following time limits are suggested for tems at least the following issues must be relevant uropathogens following receipt of the considered. specimen (Table XX). These time limits must incorporate a laboratory information system Analytical performance. that allows the clinician easy access to the Evaluation versus established reference or best result, either by telephone or, preferably, by comparison methods computer interface. The higher minimum per- Calibration and traceability of devices centage required re¯ects the clinical priority of Level of identi®cation (particles, microbes), antibiotic sensitivity over organism identi®ca- intended clinical use tion. It is more important to ensure correct A statistically signi®cant sample of patients choice of antimicrobial agent than it is to know and healthy reference individuals the name of the organism responsible for A detailed, pre-planned protocol infection. Analytical sensitivity and speci®city of results (false-positive and false-negative rates and pre- 9.5.4. Microscopy in the microbiology labora- dictive values at different prevalences) tory. All laboratories wishing to operate at Reliability of quantitation if applicable~ Level 2 must have facilities for microscopy. It linearity of measurement. is recognized that microscopy is labour-inten- sive and signi®cantly impacts on overall cost. Results from particle analysis (visual or auto- Practical performance. mated) and test-strip examination by general Ease of use and robustness or chemical laboratories or at point-of-care Ability to self-check and recognize and/or units may partially replace the need for diagnose faulty performance microscopy in the microbiology laboratory. Instrument throughput However, these results should only be used Ease of interfacing with laboratory computers. to assist in the microbiological evaluation of cultures. Only experienced evaluators, usually Clinical and economic impact. in microbiology laboratories, however, should Cost/bene®t assessment including all costs perform Gram staining. For detailed quality (reagents, manpower, running and mainte- speci®cations of microscopy, see Section 9.4. nance and indirect costs) Impact on patient care (if medical decision- making is altered, impact on outcomes). 9.6. Urinalysis device performance evaluation

Improvements in laboratory technology are Regulations. Any medical device intended for welcome and lead to better control of disease. in vitro diagnostics should be compatible with the European directive for such devices [32]. Generally, a minimum sensitivity of 80 ± 90%, TABLE XX. Acceptable turnaround times in bacter- depending on the patient population and appli- ial culture. cation, is considered adequate, while a speci®city Identi®cation to Level 2 90% within 24 h of 90 ± 95% should be maintained in diagnostic in pure culture reports. When the instrument or examination is Identi®cation to Level 2 90% within 48 h used for screening (exclusion of disease), a in mixed culture sensitivity of 95% with a speci®city of at least Available antibiotic sensitivity 98% within 48 h of relevant organism 90% should be the aims when trying to reduce the total cost of con®rmatory examinations. 48 ECLM ± European Urinalysis Group

APPENDIX (DETAIL OF Specimen details MEASUREMENTS AND PROCEDURES) Specimen identi®cation (ID) code (barcode, if ANNEX 10. TRANSMISSION OF INFOR- used) MATION AND DETAIL OF PRE- AND Date and time of voiding (®nal real time) POST-ANALYTICAL STAGE Collection method (mid-stream urine, single catheter urine, indwelling catheter urine, 10.1. Essential information in requests and suprapubic aspiration of urine, bag specimen of reports urine; other)

10.1.1. Specimen identi®cation and patient data. Success in patient preparation and collection The following boxes structure clinical back- (coded): ground information for laboratory processes. quali®ed specimen or defective collection (such as untimed collection, urgency, dif®culties in technique, etc.; classi®ed by health care 10.1.2. Requesting urinalysis examinations, siev- personnel when known) ing principle. Stepwise strategies (Section 8) should be translated into practical requesting Results from rapid examinations (if performed at point-of-care; such as test strip) routines. Considerable savings result if a sieve principle is applied for manual work, e.g. visual microscopy is performed for specimens positive with a sieving examination (usually de®ned of repeated lower urinary tract infection and no ®elds of a multiple strip or an automated par- background risk factor), a positive rapid ticle count) only. In selected cases, however, examination may be suf®cient to indicate sensitive protein measurements and even visual antimicrobial treatment without a need for microscopy should be requested independently culture. Patients with no clear-cut urinary tract of sieving examinations to detect the presence symptoms (medically justi®ed screening or of renal damage. Detailed requests help focus exclusion of urinary tract infection) may also laboratory activities correctly. be able to collect standardized morning urine, Bacterial culture should mostly be requested from which a rapid examination is able to independently of results from a rapid exami- exclude the presence of bacteriuria at 108 CFB/ nation (multiple strip or automated counter) for symptomatic high-risk patients to avoid false- L level with a sensitivity of 80 ± 90%. negative cases (this principle may change with The box below should be part of a request to improved sensitivity of sieving examinations communicate the need for additional proce- within a laboratory process). For low-risk dures if the sieving examination(s) is (are) symptomatic patients (females with symptoms performed in the laboratory. The preferred approach depends on the condition of each patient or patient group. Patient identi®cation Last name, ®rst name Personal ID code (recommended if available) Date of birth (if not included in the personal ID code) Requesting unit (where patient is being treated) Conditional request for visual microscopy ... Return address (to whom report should be sent) Detection of albumin/protein, leukocytes or Responsible physician/nurse (to be contacted if erythrocytes should be followed by visual consultation is needed) microscopy Conditional request for bacterial culture ... Concurrent antimicrobial therapy (bacterial Detection of leukocytes or bacteria by a rapid culture requested) examination should be followed by a bacterial culture (treatment of possible bacteriuria is Additional clinical information (signs, symptoms, anticipated) tentative diagnosis). No additional examinations with positive results ... If no information is given, a minimum level (Mark if applicable; may be used if the rapid of investigation should be applied as agreed examination is suf®cient for a clinical decision, locally on the basis of patient populations such as in low-risk urinary tract infection or served. acute cases) European Urinalysis Guidelines 49

10.1.3. Reports. Reporting results from ordinal particles, CFB/L). The species or genera scale examinations. Report format: Date, time, grown (depending on the methods of identi®- technique/instrument and operator ID should cation). Antimicrobial susceptibility results as be traceable. locally agreed. Test strip results: In order that the clinician is The bacterial culture report or laboratory aware of the uncertainty of the measurement, handbook should state the method of culture, ordinal scale reporting is recommended (nega- since the level of precision varies accordingly tive, 1z,2z or 3z). (from dipslide to accurate cultures with 1 ± It is recommended that the range of arbitrary 100 mL inocula; grown for 24 ± 48 h). concentrations corresponding to the assigned categories is always made available for the 10.2. Role of computers in transmission of clinical interpreter either with the report or, information more practically, in the laboratory handbook. A range is more informative than a single The need for increased ef®ciency goes in arbitrary concentration, e.g., for protein 1z parallel with automation and computerized data (0.2 ± 1.0 g/L) is preferred over 1z(0.3 g/L). To transmission. Speci®c needs of urinalysis usually avoid continual confusion, it would be of great create additional work to satisfy the transmis- help if all manufacturers were to adopt the same sion of medically relevant information. Details arbitrary categories for rapid examinations of of requests and reports are compiled in Section the same type. As a minimum, speci®ed 10.1. In addition, the following issues should be analytical sensitivity limits for positivity considered. should be identical. Standardization. National coding of laboratory Reporting results from particle analysis. The examinations is highly recommended. Addi- method should be stated in the report or in tional codes are needed to Supplement urina- the laboratory handbook (standardized sedi- lysis requests (Section 10.1). Morphological ment; standardized chamber counting; auto- and microbiological reports need special mated analysis). coding, but also space for free-form report- Morphological detail should be reported at a ing (occasional written notes on individual basic or advanced level (see Section 6.1). Units results). Units of quantities and formats of and quantities must be de®ned: particle counts/ reports must be standardized. high power ®eld (HPF, 6400 usually) are used for sediments examined under coverslips; count- Request speci®cations. Conditional requests ing chambers have de®ned volumes (e.g. 1 mL). (Section 10.1.2) and selections of correct pro- Even with HPF units, a calculation of particles/ cedures (such as shown in Table XIII for dif- litre (L) of original urine must be provided for ferent options for bacterial culture) need the clinician. Recommended unit for publica- decision rules in the on-line programmes when tions is particles/L. computerized; rapid results must be trans- Particle counts should be given as averages mitted to other working sites to organize per unit volume (HPF or L), not as ranges seen work-lists for microscopy or bacterial culture. in different view®elds. Only micro-organisms Algorithms applied in these decisions may be and clumps of cells are uncountable by visual simpli®ed from the given guidelines as agreed methods and may be reported in ordinal scale locally. from ``negative'' to ``3z''. Evaluation and reporting. Serial work-¯ow Reporting microbiological results Gram staining should be easy and accessible when working Oil immersion ®eld (OIF) is used for Gram with microscopes or reading culture plates. staining. The report may be formatted in Computer-assisted interpretation and graphic ordinal scale from ``negative'' to ``3z'' if display of daily results are encouraged where quantities are used. possible.

Bacterial cultures. Quantity of growth (recom- Internal quality control (IQC). Internal quality mended unit for particles: 10x colony-forming control data should be accessible for responsi- 50 ECLM ± European Urinalysis Group ble operators in both individual and cumula- means of understanding by individuals unfami- tive form; this means collection of separate liar with the native language (illustrations ®les. IQC data are expected to be stored for enclosed). Professional ``hands on'' assistance long periods (up to 15 years) in most quality is often needed for small children and elderly systems. people.

Connection to laboratory and hospital informa- 10.3.1. Collection of mid-stream urine specimen tion systems. A computerized urinalysis labora- (patient instructions) tory should have a standardized interface with Females: Wash your hands with soap and the general laboratory information system and water or a towelette. Dry-wipe them. Have hospital administrative system, including the clean collection container beside you. patients' medical records. Medical information Avoid touching the inside with your ®ngers. on patients will be increasingly available with While sitting on the toilet, wash your outer electronic patient records. Access to crucial genital organs, including the opening where parts of this information should be permitted the urine comes out, with a hand shower or based on laboratory user ID and patient IDs with lukewarm water and wet paper towels of investigated samples, as agreed locally. (or a sterile towelette) without using any dis- infectants. Dry-wipe. When urinating, let the Orientation towards a real time process. It is ®rst portion pass into the toilet (bedpan). Col- not possible to know the precise time or suc- lect the mid-portion in the container. Allow cess of quali®ed urine collection in advance; any excess urine to pass again into the toilet. sometimes even the method of collection may After urination, dry-wipe the outer surface of change. Thus, the possible preliminary compu- the container, secure the lid or transfer the urine to terized request must be completed after the tube(s) provided, and write or check your obtaining the specimen for such post- name and the date and time when you produced collection detail. the specimen on the label on the container. Then Computer transmissions should have a neg- proceed as advised locally. If any problems occur, ligible effect on turnaround times of emergency please consult the clinical attendant (see illustra- reports. Comparison with an earlier result of tions 1a and 2a, pages 91 and 93). the same patient should be possible. Males: Wash your hands with soap and water Safety of data: Back-up procedures must be or a towelette. Dry-wipe them. Have the clean organized to allow replacement when erroneous collection container beside you. Avoid touching transmissions occur or ®les become corrupted in the inside with your ®ngers. Uncover the repeated use. urethral opening by withdrawing the foreskin Quality control data should be easily acces- if necessary. Wash the end of your penis, sible for the reviewer of patients' results. including the opening where the urine comes out, with a hand shower or with lukewarm water and paper towels (or sterile towelette) 10.3. Collection of urine specimens without using any disinfectants. Dry-wipe. Nurses or laboratory personnel usually When urinating (either standing or sitting), let instruct patients how to obtain an adequate the ®rst portion pass into the toilet (bedpan). urine specimen. Health care personnel must ®rst Collect the mid-portion in the container. Allow understand the minimum requirements of any excess urine to pass again into the toilet. standardized specimen collection. Since the After urination, dry-wipe the outer surface of compliance of the patient or his/her parents is the container, secure the lid or transfer the urine usually needed to obtain an adequate specimen, to the tube(s) provided, and write or check your both oral and written guidance, often with name and the date and time when you produced illustrations, is necessary. Each institution is the specimen on the container label, then encouraged to modify the text given below to proceed as advised locally. If any problems make their local practice as ef®cient as possible. occur, please consult the clinical attendant (see Pictures showing the basic procedures for illustrations 1b and 2b, pages 92 and 94). females, males and children should be used Children (capable of controlled micturition): and can be freely copied; these may be the only After appropriate explanation, reasonably ade- European Urinalysis Guidelines 51 quate mid-stream specimens can be collected 7. Bend forward and hold the sterile specimen from children old enough to sit on a potty container (C) to catch the prostate secretion chair. This can be achieved by placing the while the physician massages the prostate. collection container in the potty chair (see Several drops are needed. illustration 3, page 95). Older children may 8. If no secretion is visible during massage, the follow the same advice as given to adults. physician collects a specimen with a 10 mL After producing the sample, dry-wipe the loop from urethral ori®ce for direct culture. outer surface of the container, secure the lid or 9. After prostatic massage, try to urinate an transfer the urine to the tube(s) provided, and additional 10 ± 15 mL into the container (D). write or check the child's name and the date The containers A ± D should be sent for and time when the specimen was produced on bacterial culture. If possible, particle analysis the container label, then proceed as advised is also of diagnostic value after inoculation of locally. If any problems occur, please consult the plates. the clinical attendant. 10.3.3. Collection of suprapubic aspiration spe- 10.3.2. Collection of sequential urine specimens cimen. From infants and toddlers able to con- for diagnosis of chronic bacterial prostatitis and trol their urination, a container inserted into a related disorders (Meares and Stamey pro- potty chair helps get a mid-stream specimen cedure). For diagnosis of prostatitis, sequen- (see illustration 3, page 95). If not, suprapubic tial collection of ®rst and middle portions of aspiration (or catheter specimen) should be a single-voided specimen is of diagnostic attempted when the diagnosis or exclusion of value, as well as drops expressed with prostate urinary tract infection is crucial and a bag massage and urine after prostatic massage specimen does not work (see illustration 4, [27]. The results are better if the patient has page 96). not ejaculated at least for 3 days before the collection of specimen. The given instructions 10.3.4. Timed collection of urine. A 24-h urine are to be followed with the assistance of the sample is the most common example of a physician performing the examination. timed collection. Patient instructions must be provided. Different preservatives to be used in Patient instructions timed collections are listed in Table XXI. The 1. Half an hour before specimen collection, table was compiled for analytes requested drink 400 mL of water (or juice). The from outpatients. Based on preservatives listed examination starts when you want to in the NCCLS Urinalysis Guidelines [3], new void. analytes have been added. Moreover, technical 2. Label four sterile collection vessels (A ± D) development in measurements and changes in and remove the closures from them. Avoid the body ¯uid have been taken into account touching the inside of the vessels or closures. (serum measurements replacing urine assays), 3. Wash your hands with soap and water or a as well as European experience in preservation towelette. Dry-wipe them. of the analytes [39, 264]. 4. Take the clean collection container with you. Uncover the urethral opening by withdraw- Patient instructions: 24-h collection of urine. As ing the foreskin. Wash the end of your penis, part of your medical examination you have including the opening where the urine comes been asked to collect a timed urine sample out, with a hand shower or with lukewarm because the doctor wants to know the exact water and paper towels (or sterile towelette) amount of (the examined substance) excreted without using any disinfectants. Dry-wipe. into your urine. You have been asked to col- 5. Urinate 10 ± 15 mL into the ®rst container lect for a 24-h period. If you are not in hospi- (A) in a standing position. tal, select a day when you expect to be able 6. Urinate 100 ± 200 mL into the toilet use the toilet where you keep the collection (bedpan). Without interrupting the stream, container throughout the continuous collection urinate 10 ± 15 mL into the second container period. READ THESE INSTRUCTIONS (B). Allow any excess urine to pass again CAREFULLY BEFORE YOU START THE into the toilet. COLLECTION. 52 ECLM ± European Urinalysis Group

TABLE XXI. Preservatives for single and timed urine collections (maximum documented stable time is expressed, when known, with the following abbreviations: h~hours, d~days, w~weeks, mo~months, y~years). The table assumes non-infected urine (bacteriuria may dramatically affect the preservation of some analytes). Usually, about 1% ®nal concentration is used.

Measurand Room Refrigerated Frozen HCl Boric Na2CO3 Notes temp (4 ± 6³C) -20³C 6 mol/L Acid (20³C)

Albumin 7 d 1 mo 6 mo* *Depends on the procedure Alpha-1 microglobulin 7 d 1 mo 6 mo* *Depends on the (Protein HC) procedure Alpha-2 macroglobulin 7 d 7 d Aluminium 3 d 7 d 1 y Special container Amino acids zz 5-Aminolevulinate, 1d 4d 1mo No z A (or delta-) Bacteria No 1 d 1 ± 2 d* *Combined preservatives also available Calcium 2 d 4 d 3 w z No Catecholamines 4 d 4 d 20 d z Chloride No Citrate 4 w* 1 d* * if pHv1.7 Collagen type I, z N-terminal telopeptide Cortisol z Creatinine 2 d 5 d 6 mo Cystine 3 mo* 1 y* 7 d* * add HCl Glucose v2h 2h 8h z Azide Glycosaminoglycans z Homovanillic acid zzz Human chorionic z gonadotropin (pregnancy examination) Hydroxyproline zz 5-Hydroxyindoleacetic 2h 2d* 2d* zz *add acid acid Immuno®xation and 7d 1mo 6mo electrophoresis of urine proteins Immunoglobulins 7 d 1 mo No* *cryoproteins may not (intact), quantitative redissolve Immunoglobulin, 7d 1mo 6mo kappa and lambda light chains Iron 3 d 7 d Years Lead L Lysergic acid 1m 1m 2m diethylamide Magnesium 3 d 3 d 1 y Metanephrines z Multiple strip 4 h*/1 d No *Erythrocyte and leukocyte detection most sensitive Myoglobin 12 d* 12 d* w12 d* *at pHw7 N-acetyl-beta- 1d 7d 1m glucosaminidase Oligosaccharides z Osmolality 3 h 7 d 3 mo Oxalate 4 mo* *z E, *if acidi®ed European Urinalysis Guidelines 53

TABLE XXI. (continued )

Measurand Room Refrigerated Frozen HCl Boric Na2CO3 Notes temp (4 ± 6³C) -20³C 6 mol/L Acid (20³C)

Particles except bacteria 1 ± 2 h 1 ± 8 h No No pH v 7 and Osmol (sediment) w300 mOsm/kgH20 favourable Phosphate (inorganic) 2 d* *if acidi®ed Porphyrins 4 d* 7 d* 1 mo* z * at pH 6-7, S Potassium 2 mo 2 mo 1 y Protein, total 1 d 7 d 1 mo z Depends on the procedure Pyridinium crosslinks 6 w 10 y Depends on (of collagen) the procedure Relative volumic mass z (relative density, old: speci®c gravity) Sodium 45 d 45 d 1 y Steroid pro®ling zz Test strip, see Multiple strip Tetrahydrocannabinol- 10 w 1 y 1 y Albumin 9-carboxylic acid stabilizes Urea 2 d 7 d 1 mo z Urate 4 d* No No *pHw8 Urobilinogen z S Vanillylmandelic acid 7 d* years* 7 d* *at pH 3 ± 5

A~acetic acid is not obligatory, E~EDTA addition in laboratory is helpful, L~lead-free container, S~protect from sunlight; No~not recommended. *~explained in Notes for each line.

Your collection may need preservatives for on to it. Check or write your name, personal reliable analysis. Your local advisor will tell you identi®cation number and detailed collection how to deal with these. Preservatives are usually date and times on the label. added to the collection container before the If any questions arise, please contact your start or immediately after the ®rst voided clinical attendant. portion. Write down the date and time when you start the collection (you can choose when to start). ANNEX 11. DETAIL OF Empty your bladder and discard that sample. CHEMISTRY EXAMINATIONS All voided urine after this start is to be collected in the container. 11.1. Detail of multiple test strips Exactly 24 hours after starting the collection, 11.1.1. Analytical principles. See Table XXII. empty your bladder and add this to the 11.1.2. Calculation of trueness in ordinal scale: collection container. Close the container tightly, grey zone. The trueness of ordinal scale mea- dry-wipe and place the label provided on the surements is expressed by de®ning detection container. Write or check the details of your limits (LD) and con®rmation limits (LC) from collection times and your personal identi®cation comparison measurements. They delineate a data. Store and transport the container to the grey zone. Below the detection limit, a strip laboratory as advised. examination should remain negative, while If only a portion of the 24-hour specimen was above the con®rmation limit, it should be requested, mix the complete collection thor- positive. At the grey zone, a gradual tran- oughly before pouring a small sample into the sition from negative to positive results should small container you have been given. Dry-wipe occur. The ratio between concentrations LC/ the small container and place the label provided LD is recommended to be ~5. 54 ECLM ± European Urinalysis Group

TABLE XXII. Detection principles and their limitations for multiple strips (modi®ed from references 14 and 15).

Measurand Measurement principle False-negative results False-positive results

Leukocytes Indoxyl esterase activity Vitamin C (intake Oxidizing detergents, (WBC) (granulocytes and grams/day), proteinw5 g/L, formaldehyde macrophages; not present glucosew20 g/L, mucous (0.4 g/L), sodium in lymphocytes) specimen, cephalosporins, azide, coloured nitrofurantoin; mercuric urine (beet ingestion, salts, trypsin inhibitor, bilirubinuria) oxalate, 1% boric acid Bacteria (nitrate Nitrite detected with No vegetables in diet, short Coloured urine, reductase positive)* Griess's test (azo dye) bladder incubation time, in vitro growth vitamin C, Gram-positive bacteria Erythrocytes (RBC) Pseudoperoxidase activity High nitrite concentration, Microbial peroxidases, by the haem moiety of delayed examination, high oxidizing detergents, haemoglobin density of urine, hydrochloric acid formaldehyde (0.5 g/L) Albumin (protein) Non-speci®c binding to Globulins, immunoglobulin Alkaline urine indicator dye light chains hardly detected; (pH 9), quaternary coloured urine ammonium detergents, chlorhexidine, polyvinylpyrrolidone (blood substitute) Glucose Glucose oxidase and Vitamin C, urinary tract Oxidizing detergents, peroxidase infection hydrochloric acid Ketone bodies Nitroprusside reaction Improper storage, Free sulphhydryl (acetoacetate; (Legal's test) beta-hydroxybutyrate not groups (e.g. captopril), acetone) detected coloured urines, L-dopa pH Two indicator dyes Formaldehyde lowers pH giving a pH range between 5 and 9 Relative volumic Ionic solutes of urine Falsely low: glucose, urea, Falsely high: mass (relative react with poly-electrolytes alkaline urine proteinw1 g/L, density; speci®c on the strip ketoacids gravity) Creatinine Oxidative reaction with EDTA Haemoglobin or copper complex myoglobin above 50 mg/L Urobilinogen Azo reaction with a Formaldehyde (2 g/L), Sulphonamide and diazonium salt; Ehrlich's exposure to light other drugs, aldehyde reaction coloured urine; porphobilinogen (Ehrlich) Bilirubin Azo reaction with a Vitamin C, high nitrite Coloured urine, diazonium salt concentration, exposure chlorpromazine to light metabolites Ascorbic acid Reduction reaction Not known Similar reducing with an indole dye agents

*Bacteria are detected on the basis of nitrate reductase present in most Gram-negative uropathogenic rods, such as E. coli (Griess's test). Nitrate reductase is lacking from some common uropathogens, i.e. Gram-positive bacteria, such as Staphylococcus saprophyticus and Enterococcus spp. European Urinalysis Guidelines 55

Example: Leukocyte detection by esterase leukocytes in diluted urines, in addition to activity. Acute urinary tract infections are random variation of the measurement. associated with urinary leukocyte counts §100 ± 200 WBC6106/L, while at the level of v10 WBC6106/L no association exists [180, 11.1.3. Detection and con®rmation limits for 265]. What is the performance of a test-strip different test-strip ®elds. The following detec- leukocyte esterase method detecting pyuria? tion limits (LD) and con®rmation limits (LC) Test-strip results are compared with chamber are proposed for general test strips counts of WBC from freshly voided (v2h) (Table XXV). They are created by multiplying urines with the following (imaginary) data the approximate healthy upper reference limits (Table XXIII): of concentrations in morning urine by a The following fractions describe the trueness of factor of 2, to avoid false-positive results due measurements: to intra-individual (biological) variation. To allow sensitive detection of albuminuria 1. The fraction of false positives at the detec- (microalbuminuria range), the following quality tion limit (LD)~FPD~b/(azb) (in the speci®cations are given to sensitive rapid example: 80/280~0.29 or 29%). albumin measurements (Table XXVI). Albumin 2. The fraction of false negatives at the grey concentrations (mg/L) are not expressed in

zone area ~FNG~c/(czd) (in the example: substance-based units (mol/L) to be comparable 25/125~0.20 or 20%). with total protein concentrations, such as in 3. The fraction of false negatives at the Table XXV.

con®rmation limit (LC)~FNC~e/(ezf) (in the example: 5/45~0.11 or 11%). 11.1.4. Concordance analysis (k coef®cient). Quality speci®cations for trueness: k coef®cients [249] should be calculated when If not otherwise clinically justi®ed, the comparing agreement between two or more trueness of a test-strip ®eld may be judged as ordinal scale categories, such as test-strip shown in Table XXIV (when used for an results obtained with two different measure- average patient population) ment procedures, or when comparing agree- In this example, a theoretical strip ®eld has a ment with four or more ordinal scale results poor performance in FPD, but an optimum to those measured with a quantitative compar- performance at FNG and a minimum perfor- ison method (procedure). mance of FNC (random variation, but also The principle is to subtract agreement possible problems with lysis of leukocytes on expected by chance alone. If formal signi®cance the reagent pad, etc.). In this case, the ``too testing is needed, the p value from McNemar's high'' sensitivity may, however, not be true, but test can be presented. Principles of the calcula- re¯ect delayed counting and disruption of tions of simple and weighted coef®cients have

TABLE XXIII. Example data for estimation of trueness of test strip examinations.

Comparison method (WBC 6 106/L) Negativev20 Grey zone 20 ± 99 Positive§100 Total

Test strip result Negative 200 (a) 25 (c) 5 (e) 230 Positive (1z or more) 80 (b) 100 (d) 40 (f) 220 TOTAL 280 125 45 450 Limits: LD LC

TABLE XXIV. Analytical quality speci®cations for trueness of test strip examinations.

Performance FPD~b/(azb) FNG~c/(czd) FNC~e/(ezf)

Optimum v10% v30% v5% Minimum v20% v50% v10% 56 ECLM ± European Urinalysis Group

TABLE XXV. Suggested detection and con®rmation limits for multiple test strips.

Property Comparison method Detection Con®rmation limit limit (LD) (LC)

Leukocytes (6106/L) Chamber countinga 20 100 Erythrocytes (6106/L) Chamber countinga 10 50 Albumin (protein) (g/L) Immunochemical 0.1 (alb), 0.5 (alb), (or dye binding for total protein) 0.2 (prot) 1 (prot) Nitrite (mg/L) Weighing out dry sodium nitrite; 0.5 2.5 applicable comparison method Glucose (mmol/L) Quantitative method (glucose 315 dehydrogenase or hexokinase method) Ketones (acetoacetate; mmol/L) Weighing out Li acetoacetate 1 5 pH pH meter (potentiometry) ¡1 unitb N/Ab Relative volumic mass Refractometry ¡0.005b N/Ab Creatinine (mmol/L) Enzymatic; (kinetic Jaffe no more ¡4d N/Ab recommended) Urobilinogen (mmol/L) Not commonly available 20c 100c Bilirubin (mmol/L) Bilirubin solution 10 50

a Microscopic chamber counting of fresh (less than 2 h) uncentrifuged specimens. b N/A~detection and con®rmation limits not applicable; an arbitrary class width is given. c Commonly available comparison methods are lacking. Manufacturers should document their evaluation. d Manufacturer reports an arbitrary scale of 0.9, 4.4, 8.8, 17.7 and 26.5 mmol/L.

TABLE XXVI. Analytical quality speci®cations suggested for sensitive albumin (rapid) examinations.

Property Comparison method Detection limit (LD) Con®rmation limit (LC)

Albumin (sensitive; mg/L) Immunochemical 10 50 Albumin (sensitive): Immunochemical, ratio to 315 Creatinine ratio (g/mol) quantitative creatinine method been published earlier [250]. A more detailed 11.1.5. Estimation of repeatability. Ordinal explanation is given elsewhere [266]. scale data are amenable to binomial statistics. Table XXVII gives selected 95% binomial k~(P(o)2P(e))/(12P(e)) con®dence limits to the observed proportion ~12(Q(o)/Q(e)) of results for those readers who do not have where P(o)~observed probability of agree- statistical tables or computer programmes at ment, P(e)~expected probability of agree- their bench. It shows that a specimen that lies ment by chance; Q(o)~observed within a certain concentration category (e.g. disagreement~1 ± P(o) and Q(e)~expected 1z) at a probability of 95% or more is a disagreement by chance~1 ± P(e). better control than another with a lower (even In a 262 table, a sensitivity of 90% and as low as 50%) original probability, because speci®city of 90% against the comparison the higher original probability has a narrower method result in a k~0.8. k~0.8 means that con®dence interval. the non-random agreement~P(o) ± P(e) was obtained with the examined method in 80% out of all expected disagreement by chance~ 11.1.6. Quali®ed procedures for test-strip Q(e)~1 ± P(e). reading A sensitivity of 80% and speci®city of 80% result in k~0.6. A zero value means no deviation from a random distribution (equiva- 11.1.6.1. Visual reading. Quali®ed test strip lent to sensitivity of 50% and speci®city of measurement is summarised in a tabular form 50%). k coef®cient varies between ± 1 (complete both for visual reading (11.1.6.1) and instru- disagreement) and z1 (complete agreement). mental reading (11.1.6.2). European Urinalysis Guidelines 57

Item Standard Method of checking

Identi®cation of specimen Label the specimen Compare label with the working list if analysing several specimens at once Homogeneous specimen Mix immediately before dipping Even colour Temperature of the specimen z20³C Allow to stand for 15 ± 30 min before analysis Quality of strips Date still acceptable Expiration date Environment Suf®cient light Arti®cial light is an adequate substitute for daylight to allow easy reading Calm space for working Allow no other activity during the procedure Dipping Follow manufacturer's guidance Observation by trainer Timing Use a timer showing seconds Not possible afterwards Reading Compare with the colours on the Train before actual patient analysis packing vial Internal quality control Control solutions measured daily Follow-up charts maintained if analysis is done daily External quality control Participation expected, organized Reports available with local supporting laboratory Storage of strips No physical problems associated Outlook of the strips (bent, wet, etc.), with storage closed vials Reporting Use the prede®ned units. Fill in Train before actual patient analysis the patient record or working list immediately

11.1.6.2. Re¯ectometric reading

Item Standard Method of checking

Identi®cation of specimen Label the specimen Compare label with the working list if analysing several specimens at once Homogeneous specimen Mix immediately before dipping Even colour Temperature of the specimen z20³C Allow to stand for 15 ± 30 min before analysis Quality of strips Date still acceptable Expiration date Protocol for instrumental Protocol written locally after Written protocol available measurement training Internal quality control Control solutions measured Follow-up charts maintained daily External quality control Participation expected, organized Reports available with national or foreign EQAS provider or within smaller groups Maintenance Instrument manual followed Documentation of service and repairs Calibration of the instrument Methods described mainly by the Documentation by the manufacturer, and methods, changes of reagents manufacturer. In vitro medical changes of strip lots recorded devices directive by the European Council 58 ECLM ± European Urinalysis Group

TABLE XXVII. Binomial 95% con®dence intervals of proportions.

Proportion of results falling into the same concentration categorya

No. of resultsb 50% 60% 70% 75% 80% 85% 90% 95%

20 27 ± 73% 36 ± 81% 46 ± 88% 51 ± 91% 56 ± 94% 62 ± 97% 68 ± 99% 75 ± 100% 50 36 ± 65% 45 ± 74% 55 ± 82% 62 ± 87% 66 ± 90% 73 ± 94% 78 ± 97% 86 ± 99% 100 40 ± 60% 50 ± 70% 60 ± 79% 65 ± 83% 71 ± 87% 77 ± 91% 82 ± 95% 89 ± 98% 1000 47 ± 53% 57 ± 63% 67 ± 73% 72 ± 78% 78 ± 83% 83 ± 87% 88 ± 92% 94 ± 96%

a Obtained proportion of results within the most frequent category when ®rst estimating the level of a quality control solution. b No. of test strip measurements from a patient or control specimen used for the estimate.

11.2. Detail of quantitative urine measurements 11.2.3. Detail of quantitative measurement pro- cedures 11.2.1. Health-associated reference intervals of Total protein urine proteins. The health-associated upper ref- Principles of measurement erence limits shown for excreted urine proteins derive from references 109 and 111 (Table Benzethonium chloride [272] and trichloroacetic XXVIII). These point estimates have wide con- acid precipitation [273], dye binding methods ®dence intervals owing to skewed distributions with Brome-phenol blue [274], Coomassie of values as shown in the latter reference. brilliant blue [275], Ponceau red [276] and pyrogallol-red [277], nephelometry and turbidi- metry have been suggested. All these methods 11.2.2. Diagnostic classi®cations based on pro- can be automated, in contrast to the biuret teinuria. Sensitive detection and differentiation examination [278]. Determination of total of prerenal, glomerular, tubular and postrenal protein is a compromise because no procedure proteinuria is now available with a low detec- detects all the proteins in urine. tion limit by means of speci®c protein mea- surements [267, 268]. Algorithms have also been developed to classify patients based on these measurements using two-dimensional Calibration. Calibration of total protein con- reference areas [269] (Fig. 5). centration can be performed with a reference Other diagnostic classi®cations have been sample of diluted serum. The use of albumin created from different protein to protein con- solution is preferable and more reproducible, centration ratios in urine [80, 123, 270, 271] however. CRM 470 standard is available as a (Table XXIX). primary calibrator [279].

TABLE XXVIII. Upper 95% reference limits (URL) for protein-creatinine ratios in urines from healthy indivi- duals. SI units are favoured over conventional units.

Protein Type of specimen Upper 95% reference limit Upper 95% reference limit (g/mol creatinine; SI unit) (mg/g creatinine; conventional unit)

Total protein Second morning 8a 70a Albumin First morning 3.0 27 Random 5.3 47 IgG First morning 0.7 6 Random 1.0 9 Protein HC First morning 0.5 4 (a1-microglobulin) Random 0.7 6 k-immuno-reactivity First morning 0.4 4 Random 0.7 6 l-immuno-reactivity First morning Below detection limit Below detection limit Random 0.7 6

a Turbidometric trichloroacetic acid precipitation method (see below). European Urinalysis Guidelines 59

FIG. 5. Differentiation of (1) primary glomerulopathies, (2) secondary glomerulopathies and (3) tubulo- interstitial nephropathies by speci®c protein measurements. The shaded area represents the health-associated values.

TABLE XXIX. Concentration ratios of proteins used for differentiation of proteinuria.

Concentration ratio Decision limit Suggested pathological condition a2-macroglobulin/albumin v0.02 Renal haematuria w0.02 Postrenal haematuria IgG/albumin v0.03 Selective glomerular proteinuria w0.03 Non-selective glomerular proteinuria a1-microglobulin/albumin w0.1 Mixed proteinuria v0.1 Glomerular proteinuria CRP (serum)/CRP (urine) w1.0 Bacterial infection v1.0 Rejection of renal transplant (AlbuminzIgGza1-microglobulin)/total protein) w0.6 Renal proteinuria v0.6 Suspicion of Bence-Jones proteinuria Immunoglobulin light chain k/l ratio v1 Monoclonal lambda chain 1 ± 5.2 Polyclonal light chains w5.2 Monoclonal kappa chain

Upper limits of reference intervals in healthy Albumin individuals. Because these are method-depen- Principles of measurement dent, several intervals are cited: Nephelometry, turbidimetry, RIA, ELISA and EMIT with monoclonal or polyclonal anti- bodies are commonly used. Pyrogallol red method: v180 mg/day [280], for pregnant women v260 mg/day [281] Calibration. CRM 470 standard is the primary Biuret method: v150 mg/day [278] reference material [279]. Professional societies Turbidimetric method: v75 mg/day [273] and major diagnostic companies have agreed Turbidimetric method: Protein/creatinine v8g/ to create secondary standards and join con- mol (second morning urine [109]); (or v70 mg/ sensus interim reference ranges based on the g in conventional units) CRM 470 [282]. 60 ECLM ± European Urinalysis Group

Prediction limit for incipient diabetic nephropa- Reference interval from healthy individuals thy. An albumin excretion rate of w20 mg/min and interpretation. Creatinine in urine: (traditionally cited unit), which corresponds to 7 ± 20 mmol/day (0.8 ± 2.2 g/day). Creatinine an approximate albumin/creatinine w4 g/mol excretion rate depends on muscle mass. Crea- (30 mg/g in conventional units), will predict tinine is almost entirely ®ltered by the glomer- incipient diabetic nephropathy [10]. uli and only traces are secreted by the tubules. The fraction of tubular secretion Interpretation. Albumin/creatinine ratios decrease increases, however, with reduced glomerular slightly with age [283]. The albumin/creatinine ®ltration rate. High protein meals and intense ratioisalsoslightlyhigherinwomen physical exercise lead to increased urinary than in men because of lower creatinine creatinine concentrations. excretion in women. Upper normal limit in pregnancy is 30 mg/day (24-h collection) [281]. Osmolality The average intra-individual coef®cient of varia- Principles of measurement tion of albumin excretion from day to day is approximately 20 ± 30% [109], and even larger Osmometry follows directly the de®nition of for diabetic patients [258]. Diagnostic decisions osmolality: it is based on either of the two should therefore not be based on a single mea- colligative properties, a decrease in freezing surement, especially if the result is equivocal. point or an increase in the evaporation point of solutions. a1-microglobulin (protein HC) Principles of measurement Interpretation. Urea, ammonia and mono- Nephelometry, turbidimetry, RIA and ELISA valent ions are mostly responsible for urine with polyclonal antibodies are commonly used. osmolality. With the maximum antidiuresis the urine reaches an osmolality of about

1200 mOsm/kg H2O. Maximal diuresis may Calibration. Measurement of a1-microglobulin or protein HC concentration has not been result in an osmolality as low as 50 mOsm/kg H O [127]. The concentrated morning urine standardized yet. An international calibrator 2 is highly desirable. after an overnight restriction of ¯uid intake reaches an osmolality of at least 700 mOsm/kg Interpretation. The within-subject coef®cient of H2O in healthy individuals. In chronic renal variation from day to day is 20% on average failure, the urine remains isotonic within the range 300 ± 350 mOsm/kg H2O. [109]. a1-microglobulin (30 ± 33 kDa) is pro- duced in the liver and lymphocytes. This glyco- protein appears in serum in free (50%), albumin Relative volumic mass (old terms: relative den- bound (v10%) and IgA-bound forms (40%). sity; speci®c gravity). Calibration. Relative Only the free form is ®ltered and reabsorbed in volumic mass measurement can be calibrated the proximal tubule to 99.8%. Increased concen- in practice by measuring volumic masses (den- trations in the urine are found in tubulo-intersti- sities) of pooled human urine, i.e., by weigh- tial dysfunction or in nephropathies. ing out accurately known volumes of pooled human urine. In this way, refractometers and Creatinine related instruments can be adjusted using a Principles of measurement calibrated balance. Methods based on the Jaffe reaction [284] should be replaced with more speci®c enzymatic Principles of measurement. Measuring princi- methods [285]. Reference measurement pro- ples include urinometry, refractometry, oscillo- cedure is based on isotope dilution mass metry and test strips. These have been spectrometry [286]. reviewed extensively [127]. It is to be noted that there are marked differences in the accu- Calibration. A known reference material for racy of these methods [287]. serum creatinine is SRM 914a [257]. Weighed out creatinine solution may serve as an Reference interval and interpretation. Relative approximate estimate. volumic mass is primarily a function of glu- European Urinalysis Guidelines 61 cose, phosphate and carbonate. For human Interpretation urine, the values are within the interval Ordinal scale reporting with two categories is 1.003 ± 1.035. Morning urine of healthy indivi- suf®cient for a rapid examination (either posi- duals has a relative density of 1.020 or more tive or negative result). A weak positive result after overnight restriction of ¯uid intake. Iso- indicates a slightly elevated hCG concentra- tonic range in chronic renal failure corre- tion but below 25 IU/L (or below 200 ± sponds to relative densities 1.010 ± 1.012 [127]. 500 IU/L if used in some ambulatory units). It is then good laboratory practice to ask for Conductivity. Conductometry of urine (meas- a new sample and retest it after an additional ured as a current ¯ow between two electrodes) 2 days. Pharmacies selling kits for home test- has recently also become readily available ing often apply the highest sensitivity of [128, 132]. Since the amount of charge in 25 IU/L. urine (the ionic strength) is related to urine concentration, the conductivity is also related to water excretion. Diet-dependent intake ANNEX 12. DETAIL OF PARTICLE affects the excretion of salt from healthy indi- ANALYSIS viduals as well as from patients. Data on 12.1. Visual microscopy methods reliability in clinical practice are accumulating. 12.1.1. Chamber counting of uncentrifuged urine for evaluation purposes (Level 3). Cham- 11.3. Detail of pregnancy examinations ber counting of uncentrifuged urine is recom- mended for comparisons of automated Specimen particle counting with visual methods, because The concentrated morning urine usually con- counts obtained from a chamber are more tains the highest concentrations of hCG and precise than those obtained under coverslip on is therefore the best specimen. If the urine a microscope slide. If counted without centri- specimen is too dilute, false-negative results fugation, the loss of particles during centrifu- may occur in early pregnancy. Stability and gation does not interfere with the results. In storage: If the examination is performed routine methods, cells and other particles in a within 48 h after collection, the specimen can volume of 1 mL are counted. For evaluation be kept in the refrigerator (2 ± 8³C). If exami- of particles other than WBC or RBC, this is nation is delayed for more than 48 h, the spe- not statistically suf®cient. For precise evalu- cimen should be frozen. ation, at least 100 cells should be counted to reach a CV~10%, and 400 cells should be Examinations counted to reach a CV~5%, based on the Different commercial systems for rapid and Poisson distribution (see Section 9.4.1). sensitive detection of human chorionic gona- Health-associated reference intervals of many dotropin (hCG) are available. Speci®c mono- urine particles are, however, below 2 particles/ clonal antibodies to the intact hCG molecule mL(v26106/L). (both alpha-chain and beta-chain) should be Variation of the counting volume should be used. less than 5% in a chamber used for reference Laboratories or point-of-care units should counting (compared to less than 10% in routine consider carefully what is the best sensitivity chambers). limit for their practice, and whether urine or Staining of specimens and/or phase contrast blood specimens should be used. Nominal microscopy is needed for accurate identi®cation sensitivity may be chosen between 25 and of the various urine particles. For evaluation of 500 IU/L in different clinical situations (emer- test-strip readers, leukocyte (granulocyte) and gency diagnostics of extrauterine pregnancies, erythrocyte counts are needed only. An example antenatal clinic detecting normal pregnancies, of a reliable comparison method to validate an X-ray department excluding pregnancies). The automated particle analyser was recently pub- rapid test selected should be evaluated against a lished which used both phase contrast optics quantitative hCG method with the help of local and supravital staining but only 1 mL volumes laboratory. for practical reasons [69]. Consequently, counts 62 ECLM ± European Urinalysis Group below 10 ( ± 30) particles6106/L were statisti- eosinophilic granulocytes, Hansel's stain is cally unreliable. recommended [145]. In the proper cytological examination of urine to detect cancer cells, 12.1.2. Standardized urine sediment exami- Papanicolaou staining with a ®xation procedure nation (Level 2). is a worldwide practice. This is beyond the Fixation and drying of the specimen, to scope of the general laboratory. permit careful cytological examination [41], is Quantitation. A high-quality binocular micro- generally considered too time-consuming for scope should be equipped with objectives for basic urinalysis. Wet sample preparation with low (106 to 166) and high (406) magni®ca- or without preservatives remains a pragmatic tion. Oculars are traditionally from 106 to compromise for routine laboratories. 12.5 6 magni®cation. The volume applied Centrifugation. The exact concentration under a coverslip de®nes the average height of factor is important to relate the ®ndings to the ¯uid layer under a coverslip of de®ned size. the original volume of urine. Optimum selection The diameter of the view®eld is equal to the of the factor depends on the size of the ocular view®eld number (usually from 18 to specimen tubes used and the suction equipment 22 mm; given by the manufacturer) divided by to remove the supernatant; the range is usually objective magni®cation. It can be measured 10 to 25-fold concentration. Centrifugation with special tools as well (metric scales are should be at 4006g for 5 min (adjust rpm available as microscope slides). values according to the radius of the rotor in Example: 10 mL of urine is concentrated into the centrifuge; see the equation below). A 0.5 mL~20-fold concentration; 0.05 mL stain refrigerated centrifuge is an advantage when is added. Then 13 mL is applied under working with larger series of non-preserved a18618 mm coverslip. Thus the ¯uid specimens or when undertaking parallel tasks. layer~13/(18618)~0.040 mm thick. If it is On the other hand, precipitation of crystals is investigated under 10640 magni®cation with increased at low temperatures. ocular view®eld number 22 (diameter of the view®eld ~22 mm/40~0.55 mm), the volume RCF~1.118610 ±56r6RPM2 of a high-power ®eld (HPF) is then: where r~ the radius of the centrifuge from 0.0406p6(0.55/2)2~0.00950 mm3 or mL the centre of the spindle to the bottom of the tube (cm) Because of the 20-fold concentration and 10% dilution because of the dye, one HPF corre- RPM~60006d 1/r, if RCF~4006g sponds to 0.173 mL of original urine volume. Staining and optics: Phase contrast micro- In this case, n RBC/HPF correspond to n scopy is recommended. Compared with bright- RBC/0.173 mL~5.8 6 n RBC6106/L. ®eld microscopy, the phase contrast technique E.g. 4 RBC/HPF equals to 23 RBC6 106/L. allows better detection, especially of bacteria, Reporting. An average particle count rather RBC and hyaline casts. With bright-®eld than a range should be reported. It is practical microscopy, supravital stains are obligatory to report it as particles/high power ®eld (HPF). for proper differentiation. In routine use, a It is, however, recommended that the number colour contrast with blue and red (e.g. Stern- concentration of particles be calculated also to heimer's Alcian blue ± pyronin B staining [171, the internationally comparable units of volume 288]) is better than the older violet-coloured (L), as shown above. Precise counting may have gentian violet ± saffranin O (Sternheimer- an upper limit for practical reasons, equalling Malbin) method [289]. Publications on the about 200 ± 300 particles 6 106/L. For numer- successful use of toluidine blue also exist ous microbes, ordinal scale reports, such as [172]. Stained preparations can also be routinely ``none'', ``few'', ``moderate'' and ``abundant'' reviewed by phase-contrast optics to bene®t are acceptable. from both techniques. Filters for polarizing The details for the standardized sediment light are helpful for correct identi®cation of examination are summarized in Table XXX. crystals and, in some cases, also of lipids. For special purposes, other staining pro- Visual microscopy procedure for routine. Cool cedures should be used. For detection of down the specimen (z4³C) to allow a delay European Urinalysis Guidelines 63 of up to 4 h if not preserved or speci®cally volume of sediment on to a microscope slide. investigated for crystals (precipitation of Add the coverslip horizontally to maximize urates will follow after cooling). At z20³C even distribution. Examples: 13 mLisan the delay of microscopy should not exceed 1 h appropriate volume for an 18618 mm cover- after voiding. slip, while 35 ± 50 mL may be used for a Always use a de®ned volume of urine (usually 24632 mm coverslip. The smaller volumes 5 ± 15 mL). Centrifuge the specimen at 4006g and coverslips have economic advantages in at z4³C (if not preserved) for 5 min. Remove bigger laboratories. the supernatant with an adjusted vacuum tool; Investigate the sample ®rst under low-power do not decant (an inaccurate method). Aim at (106 or 166) magni®cation to see the usually the decided concentration factor, e.g. 20 (12 mL uneven distribution of particles on the slide, as p 0.6 mL or 10 mL p 0.5 mL). well as to note the presence of rare elements, After suction, add 10% of the volume (60 mLor such as casts and epithelial cells. Then count 50 mL) of supravital stain if your laboratory is the number of different particles per high- using a stain. Mix gently. Use a cooling plate if power (406) objective ®eld, reporting the the specimens have to wait before investigation. average number observed in at least 10 ®elds After gentle resuspension, pipette a known selected from all areas of the coverslip (the

TABLE XXX. Summary of standardized urinary sediment examination.

Item Standard Method of checking

Delay Investigate within 4 h from micturition (if stored Documented times of at z4³C), or within 30 min at z20³C to allow collection evaluation of all cells; otherwise use preservatives after evaluation Original volume of urine 5 ± 12 mL Line marked on the tube Centrifugation 4006g for 5 min, preferably at z4³C if delays Check with the occur Supplier of the centrifuge Removal of supernatant Suction to a de®ned ®nal concentration factor Calibrate the ®nal volume by weighing pooled urine (buffer solutions have a different surface tension) Method of staining and Phase-contrast microscopy, or stainingzbright- Consult local supplier microscopy ®eld microscopy; polarized optics when needed; low and high-power magni®cation (6400) Volume investigated under De®ne and calculate Microscopic slide with a microscopic ®eld metric scale List of reported components De®ne the report format These guidelines Units of reporting Particles/L (particles/4006 magni®ed high-power Calculate the equivalence ®eld) Reproducible process Written operating procedures Training of personnel, blind peer reviews Internal quality control Training courses organized locally Two independent Double-check specimens weekly investigations for the same specimen External quality assessment Participation in an EQA programme Documents of results available Calibration Traceability of measured quantities Evaluation against uncentrifuged specimens 64 ECLM ± European Urinalysis Group total number of ®elds to be counted depends Their multilobular nucleus and cytoplasmic on the concentration of particles: lower num- granules make them readily identi®able. In bers require more ®elds to reach statistical dye-binding methods, living granulocytes do reliability). not always allow dyes to enter the cell. In Sternheimer staining, nuclei and inclusions 12.1.3. Morphologic criteria for urinary sedi- usually stain bright blue, while cytoplasm ment ®ndings (Level 2). Recent atlases show- remains reddish or brownish. Granulocytes ing detailed morphology of urinary elements may aggregate to form clumps. They also lyse for reference purposes must be consulted until easily when the osmolality is low or when the enough experience is accumulated [15, 41, 56, pH is high [89]. Eosinophils need special stains 288, 290 ± 292]. for their detection (Hansel's stain [145]). Macrophages (histiocytes) are often seen in association with in¯ammation. They show thin 12.1.3.1. Detailed characteristics. Erythrocytes. pink granular cytoplasm, often ®lled with red The size and haemoglobin content of erythro- blood cell remnants and other vacuoles, and cytes can vary according to the osmolality of bluish nuclei with unevenly distributed chroma- urine (the lower the osmolality the larger the tin on Sternheimer staining. cell and the lower the haemoglobin content). Lymphocytes have smooth nuclei that almost ``Ghost erythrocytes'' that have lost their hae- ®ll the cell. In Sternheimer staining, they are moglobin content can be missed if bright ®eld usually dark blue. Cytoplasm is scarce and microscopy is used alone. At low osmolality without granules. values, erythrocytes can even lyse. Erythro- Tubular epithelial cells. Different types of cytes usually have a diameter of 4 ± 7 mm. The tubular cells line the different segments of the appearance can vary according to the source renal tubules. As a consequence, several types of haematuria: isomorphic erythrocytes usually of tubular cells can be found in the urine, which indicate a post-renal bleeding, while dys- differ in size and shape. Tubular cells are morphic erythrocytes indicate glomerular mononucleated with granular cytoplasm and disease (see below). are larger than leukocytes. Most nuclei are Dysmorphic red cells. Phase-contrast micro- round to ovoid. They have an average diameter scopy is mandatory. Report the total number of of about 13 mm, and probably come from the erythrocytes/HPF and the percentage of dys- proximal segments of the tubules. Less fre- morphic red cells. Fassett et al. (1982) originally quently, rectangular, polygonal or even colum- proposed that a fraction of 80% dysmorphic RBC nar cells can be seen, which originate from or higher indicates renal bleeding [147]. A fraction distal tubules or collecting ducts. In Sternhei- of 80% isomorphic (normal) RBC indicates post- mer staining, tubular cells usually have dense, renal bleeding. In the middle, there are mixed granular red cytoplasm that may contain lipid cases. It is recommended that the fraction of droplets in patients with severe proteinuria, and acanthocytes be reported as well, because they are a blue or purple nucleus. A practical way to readily identi®ed by the blebs on the cell surface learn the morphology of tubular cells is to look and a presence of 5% or more acanthocytes is for these cells within casts: epithelial cells within indicative for renal bleeding (see also Section 6.1). casts are by de®nition tubular epithelial cells. The morphological appearances of dysmorphic Transitional epithelial (urothelial) cells. Divi- red cells, especially acanthocytes, can be studied sion of urothelial cells into super®cial and from the available publications. The correct deeper cells has been described recently [293]. evaluation of abnormal RBC morphology may Super®cial urothelial cells are usually round dependon pH andosmolality of the urine, which is to oval mononucleated cells with a mean why morning specimens are recommended if diameter of about 30 mm, and a pale halo possible [151]; acanthocyte formation does not around the nucleus. Occasionally they can be occur in vitro with the same frequency as other bi- or multinucleated. These cells are a frequent dysmorphic shapes [150]. Special training of the ®nding in patients with urinary tract infection investigator is needed. and urological disorders. Leukocytes. Polymorphonuclear granulocytes Deep urothelial cells are smaller than super- are the most frequent leukocytes found in urine. ®cial cells (mean diameter about 17 mm). They European Urinalysis Guidelines 65 may exhibit various shapes, but they have Owing to degenerative phenomena, it is often mostly a club-like or an ovoid appearance, a dif®cult to saywhether the cellswithin thecasts are central or peripheral nucleus and thin granular tubular rather than leukocytes. In such cases the cytoplasm. Usually they are found in associ- correct de®nition should be ``cellular cast'' only. ation with urothelial carcinoma, ureteric stones Haemoglobin and myoglobin casts. These are or hydronephrosis. These cells stain darker than brownish in colour with a granular surface. More super®cial urothelial cells. Atypical forms of frequently, haemoglobin casts derive from ery- urothelial cells can also be found in association throcyte casts. Therefore, they too indicate renal with the rapid techniques of cellular preparation parenchymal bleeding. However, haemoglobin used in routine urinalysis [172, 288]. Their casts may also be due to haemoglobinuria caused interpretation should be undertaken in specia- by intravascular haemolysis. Myoglobin casts lized cytological laboratories. may be seen in the urine of patients with renal Squamous epithelial cells. Squamous epithelial failure caused by crush syndrome. cells are the commonest cells of the urine. They Bilirubin casts stain yellow-brown due to have a polygonal shape, a central nucleus and water-soluble (conjugated) bilirubin excreted an average diameter of about 55 mm. Squamous into urine. Urinary bilirubin may be used in cells derive from urethra and vagina, and are differentiation of icteric patients if serum usually markers of contamination of urine measurements are lacking. during specimen collection. Bacterial and yeast casts. These are rare. Casts. Casts are elongated elements with a However, they may be seen in immuno- cylindrical shape that varies due to bending, compromised patients with bacterial or fungal wrinkling and irregular edges. The main types infection affecting the kidneys. of casts are described below. Lipids are usually seen as isolated or clumped Hyaline casts. These have a matrix with low droplets (either free or within cells and casts), or refractive index and are best identi®ed by phase as oval fat bodies (round-shaped particles contrast microscope. They are found in both containing packed lipid droplets), fatty casts renal parenchymal diseases and in normal or cholesterol crystals (see below). Lipids are subjects. identi®ed because of their refractility and their Granular casts. These can contain either ®ne ability to polarize light. or coarse granules. They are not usually found in normal subjects and when present suggest the Crystals. Uric acid. Lozenges, barrels or presence of renal disease. rosettes with a typical amber colour and bire- Waxy casts. These are usually large, with fringence under polarized light. clear-cut edges or indented borders, and are Calcium oxalate dihydrate. Typically bipyr- refractile. Waxy casts have a homogeneous amidal. They can appear also in aggregates. appearance, just like wax. They are found in Only large crystals show birefringence. patients with renal insuf®ciency or failure. Calcium oxalate monohydrate. Ovoid, dumb- Fatty casts. These contain lipid particles. bell or biconcave discs, always brightly bire- Fatty casts are typical of patients with heavy fringent. proteinuria associated with lipoproteinuria. Calcium phosphate. Prisms, needles or rosettes Cellular casts. According to the cells con- that polarize light. When occurring in plates, tained, cellular casts are classi®ed as: calcium phosphate is not birefringent. ± Erythrocyte casts (which always indicate Triple phosphate. Transparent birefringent bleeding from the renal parenchyma) prisms, usually with a ``cof®n lid'' appearance. ± Leukocyte (usually Granulocyte) casts Amorphous urates and phosphates. Granular (which may indicate acute pyelonephritis, particles, often in clumps. Urates are found in acute interstitial nephritis or proliferative glo- acid urine, phosphates in alkaline urine. Urates merulonephritides) polarize light while phosphates do not. ± Tubular epithelial cell casts (which are Cystine. Thin, hexagonal, birefringent plates found for example in patients with acute with irregular sides. They can be isolated, heaped tubular necrosis, acute interstitial nephritis, upon one another, or in clumps and rosettes. They acute cellular rejection of grafted kidney and can be seen at low pH (v6) and usually after an glomerular disorders). overnight incubation at z4³C. 66 ECLM ± European Urinalysis Group

Leucine. Oily-looking spheres with concentric establish the optimal intended diagnostic use striations. for a given instrument. Based on the technical Tyrosine. Thin needles, often aggregated in principles used, advice on specimen collection bundles or rosettes. and storage is essential to obtain reliable results Cholesterol. Transparent thin plates with and avoid artefacts. Lists of known interfer- sharp edges and corners. These are associated ences should be made generally available as with heavy proteinuria. soon as they are discovered during evaluations Crystals of drugs. Sulphadiazine (crystals with and in clinical practice. the appearance of ``sheaves of wheat''), triam- Customers should work out standard operat- terene, acyclovir (birefringent and needle- ing procedures with the help of manufacturers. shaped crystals), indinavir (plate or star-like These should include descriptions of regular crystals [294] and vitamin C [56, 163]. working, combinations of different analyses Microbes. Bacteria may be seen on routine from the same urine, quality assessment proto- microscopy; rods are particularly visible with cols and measures to be taken in the event of phase contrast microscopy. Cocci may be instrument alarms or error messages. Specimens confused with salt precipitation. not amenable to automated analysis should be Fungi. Cells of Candida spp. appear as ovoid listed, as well as the standardized alternative or roundish elements not absorbing stain. They manual methods. also appear as hyphi. Budding is the most typical morphological feature. On most occa- sions they are due to contamination from the 12.3. Health-associated upper reference limits vagina, although they may represent true of particles in morning urine infection in the chronically debilitated or immunosuppressed patient. Healthy reference intervals depend heavily on Protozoa. Trichomonas vaginalis, when alive, preanalytical as well as analytical standardiza- is easily identi®ed owing to the motility of the tion and delay of examination. Table XXXIII ¯agella and the rapid and irregular movements cites some of the 98% and 95% upper reference of the body. When dead it is similar to limits found in the earlier literature [170, 185, leukocytes. In most cases it is in the urine as 295 ± 297]. Comprehensive detail necessary to a consequence of genital contamination. evaluate the results has not been published. The Parasites. The diagnosis of parasitic infesta- ®gures are concentrations rather than excretion tion by Schistosoma haematobium relies on the rates, as originally proposed by Addis [298]; observation of the eggs in the urine. These lysis of cells unfortunately diminishes the measure about 140650 mm and are spindle- validity of excretion rates based on 12-h shaped with a round anterior and a conical collections. Adjustment to volume rate (diur- posterior end tapering into a delicate terminal esis) should be considered. spine. They may be seen to hatch if the urine is Rough estimates for arbitrary (non-para- dilute enough. metric) 95% upper reference limits in morning Details of urinary particles are tabulated for urine without centrifugation: leukocytes and easier reference in routine practice in Tables erythrocytes v10 x 106/L. XXXI and XXXII. NOTE 1. Owing to the procedural uncertain- ties these upper limits are given as arbitrary inequality ®gures only. Female specimens are more vulnerable than male specimens for pre- 12.2. Automated particle analysis analytical errors, as shown by the Stansfeld & Manufacturers of urine particle analysers Webb study. In addition, erythrocytes may be should describe in detail the differentiation dif®cult to see without phase contrast optics capability of their instrument, including sensi- and proper 6400 magni®cation [295]. For best tivity and speci®city data against a manual clinical comparability, concentrated ®rst (or comparison method, such as chamber counting, low-diuresis second) morning urine should be sediment morphology or bacterial culture. collected as a washed mid-stream collection and General as well as speci®c patient populations investigated according to an identical procedure should be targeted in the evaluations to that is applied to the patients as well. About European Urinalysis Guidelines 67

TABLE XXXI. Differentiation of urine cells by phase contrast optics or supravital staining.

Cell type Nucleus Cytoplasm Other

Granulocyte Multilobular or rod-shaped, granular, degenerates occurs often in does not stain always, bright easily clumps, round blue when stains shape Eosinophilic granulocytes Granular Use Hansel's staining for identi®cation Macrophage Bluish, chromatin uneven ``Thin'', granular, varying Phagocytosed pieces size of red cells inside of granules; round, tuberous and dendritic shapes occur Lymphocyte Fills the cell almost entirely, Smooth, scarce broken (dark) blue clear cell membrane thin rim Squamous epithelial cell Degenerated, small, round, Pale, large, slightly Often in clumps, (polygonal) central, even free granular adhering polygonal shapes, in urine bacteria may be folded Transitional epithelial cell, Oval or round, small, Large, with clear Round or oval shape, super®cial chromatin ®nely granular, perinuclear halo ®nely also in clumps nucleolus is often visible granular pink (catheter?) w1 nuclei may occur Transitional epithelial cell, Well-de®ned, large, evident Many granules may be Various shapes, deep nucleoli, central or eccentric dark red atypical forms occasionally atypical forms Renal tubular epithelial Homogeneous, clear, Granular, often dark Look at cells within cell blue/violet, nucleolus may red, ``thick'', degenerative the casts; round, or be evident granules or fat often oval, also in clumps inside

Prostatic epithelial cells cannot be differentiated from transitional epithelial cells by this method; prostatic particles may be seen occasionally. The intensity of staining varies and is dependent on the length of exposure to the stain as well as unknown factors related to the specimen. The ratio of blue to red is also affected by the batch of stain used in preparing the stain mixture.

TABLE XXXII. Differentiation of other urine particles (phase contrast or supravital staining).

Particle type Features

Microbes: Bacteria rods Dark, often in chains cocci Dark, isolated, in pairs, chains or clusters Yeasts Nucleus often visible, budding; do not stain well; also as branching hyphi (pseudomyceliae) Casts: Cellular casts Cells (erythrocytes, leukocytes, tubular cells) packed and/or plunged in the cast matrix Microbial casts Granular appearance; rods may appear clearly discernible Hyaline cast Low refractive index, may occur as darkly blue, compact or ®brillar, occasionally convoluted Granular cast Granules of various size and colour, usually red Waxy cast Refractile, with hard and indented edges; colour red rather than blue Fatty cast Lipid droplets can be isolated, in clumps, or packed; occasionally protruding cholesterol spicules Pigmented cast Haemoglobin and myoglobin (red-brown), bilirubin (yellow-orange-brown) 68 ECLM ± European Urinalysis Group

TABLE XXXIII. Health-associated upper reference limits (URL) of urine leukocyte (WBC) and erythrocyte (RBC) concentrations in uncentrifuged specimens.

Population Procedure details Sample size Estimated URL (particle: Reference percentile, value6106/L)

Boys, 0 ± 14 years, aMid-stream specimen 359 patients/ WBC: 98%, 10 295 in-patients 1142 specimens 95%, 5 Girls, 0 ± 14 years, aMid-stream specimen 273 patients/ WBC: 98%, 90 a 295 in-patients 1167 specimens 95%, 50 a aSingle-catheter specimen 100 WBC: 98%, 10 a 295 Boys and girls, Mid-stream specimen, 228 boys and WBC: 95%, 4 296 2 ± 16 years volumic mass¦1.005 160 girls RBC: 95%, 13 deleted Males, voluntary Mid-stream specimen, 75 WBC: 98%, 9 170 individuals Morning urine 95%, 5 ± 8 RBC: 98%, 4 95%, 3 Adults Not speci®ed 714 femalesz RBC: 95%, 8 297 293 males~1007 Adults bScanning electron 27 RBC: 95%, 2 185 microscopy

a Stansfeld and Webb (295) noticed the problem of collection of mid-stream urine specimens from girls. After collecting single in-and-out catheter urine specimens, upper reference limits comparable to those of boys were obtained. b Pollock et al. (185) used ®ltration of known volume of urine and scanning electron microscopy to establish the true value for RBC concentration in ®rst morning urine. GyoÈ ry et al. (299) have also published health-associated upper 95% reference limits from 1720 urine sediments counted in chamber: The limits were 1 RBC and 3 WBC6106/L (299). To avoid confusion, the ®gures obtained after centrifugation were not listed in the table. half of the erythrocytes and leukocytes are lost lower cell counts due to the higher cell numbers during centrifugation (when preparing urine counted. It seems likely that new more precise sediments). health-associated reference intervals will be NOTE 2. With statistically reliable total created based on re-assessed procedures and numbers of individuals, selection criteria of new technology [132, 175, 300]. reference individuals or calculations may have been defective, leading to non-standardized ANNEX 13. DETAIL OF MICROBIOLOGY results. Because of the skewed distributions, EXAMINATIONS parametric calculations may be misleading if the skewness is not corrected properly. Exact 13.1. Suggested comparison method for percentile estimates are in any case imprecise, bacterial culture of urine (Level 3) and a marked grey zone should be considered (expressed as a con®dence interval of the Substrates. percentile estimate). CLED (Cysteine-Lactose Electrolyte-De®cient) NOTE 3. Concentrations of erythrocytes and agar plate leukocytes in urines of newborn individuals are Blood agar most likely higher than those given above. Haematin agar NOTE 4. Diagnostic decision limits differ For more accurate identi®cation of yeasts, from the given upper healthy reference limits. special methods are required. For example, the probability of infection in a Control strains, see Table XXXV. non-standardized specimen increases when the amount of granulocytes (leukocytes) rises from Inoculation. Alternative A: 10 to 100 6 106/L [180, 265]. Ten microlitres of urine is inoculated with a Automated methods show higher precision at micropipette in the middle of a haematin agar European Urinalysis Guidelines 69

TABLE XXXIV. Quantitative interpretation of growth.

Volume of inoculum No. of colonies Colony count in urine Colony count in urine on plate (CFB/L) (CFU/mL)

1 mL1106 103 10 107 104 100 108 105 10 mL1105 102 10 106 103 100 107 104 100 mL1104 101 10 105 102 100 106 103 plate and a blood agar plate. In addition, 1 mL a neutral buffer; 100 mL of each dilution is is inoculated on to CLED agar with a plastic or inoculated by means of a Drygalski spatulum platinum loop. When performing cultures on and a rotating platform (1 or 10 mL may not be suprapubic aspiration urine, a large 100-mL precise enough volumes for normal pipettes). amount of inoculum is recommended on blood The colonies on the plates from at least three and haematin agar. Different amounts of dilutions are counted. The original concentra- inoculum lead to different interpretations of tion is calculated as the mean concentration quantitative growth for each colony count from the dilutions. (Scheme 1) observed (Table XXXIV). In order to raise the accuracy of the estimate The inoculum is spread with a platinum or for mean bacterial concentration the number of plastic loop by streaking back and forth replicates can be increased, e.g., for disinfectant through the drop a few times (Phase 1; Fig. examinations the same sample should be 6). Finally, several streaks are made very close analysed in 10 ± 20 replicates. The counts to each other and perpendicular to the ®rst line follow Poisson distribution. (Phase 2; Fig. 6). Quantitation is performed on plates inoculated by a micropipette. Plates not Incubation times. The following incubation used for quantitation can be divided for two temperatures, atmospheres and times are different specimens. recommended for best identi®cation of species Alternative B: and quantitative comparisons: When a more accurate quantitative result than that obtained by the method in alternative CLED agar 35 ± 37³C, aerobic 1z1 day A is necessary, the following procedure is Blood agar 35 ± 37³C, anaerobic 2 days recommended. The sample is diluted 1:10 with Haematin agar 35 ± 37³C, 5% CO2 1z1 day

FIG. 6. Inoculation of a culture plate. (limitations of computer graphics must be understood in the inter- pretation of these drawings). 70 ECLM ± European Urinalysis Group

13.2. Routine loop cultures Incubation and interpretation CLED agar 35 ± 37³C, aerobic 1z1 day The external work in quality assurance has Blood agar 35 ± 37³C, CO 1z1 day shown that there is great variation in the method- 2 ology of performing conventional urine cultures Clearly positive and negative results (108 CFB/L despite earlier attempts at standardization. of primary pathogens or no growth) can be Greater demands for faster and less expensive service also justify the description that a mini- reported after incubation for 1 day. Colony mum diagnostic performance be followed. The morphology develops signi®cantly after an lowest statistically reliable count to be reported additional day of growth (which can be carried with this method is 107 CFB/L (104 CFU/mL). out in a 5% CO2 atmosphere for CLED agar if blood agar is not used). This facilitates Substrate. identi®cation of mixed ¯ora that may contain CLED agar (for control strains see Table more resistant species or variants. An increase XXXV) of 8 ± 10% in the frequency of isolation has been Blood agar reported after 2 days compared to a 1-day culture [301, 302]. Inoculation. One microlitre is inoculated with As an alternative to CLED-agar, there are a plastic or platinum loop on blood and several new chromogenic media on the market. CLED agar. The inoculum is spread as These allow direct presumptive identi®cation described above. of E. coli, Proteus/Morganella/Providencia and

Scheme 1. Serial dilution for quantitative bacteria counting. European Urinalysis Guidelines 71

TABLE XXXV. Control strains for urine bacterial culture.

Medium Species ATCC no. Incubation Expected reaction

Blood agar E. coli 25922 35³C, aerobic 1 day Growth Blood agar S. aureus 25923 35³C, aerobic 1 day Growth Blood agar S. pneumoniae 6305 35³C, aerobic 1 day Growth, alpha haemolysis Blood agar S. pyogenes 19615 35³C, aerobic 1 day Growth, beta haemolysis Blood agar anaerobic C. perfringens 13124 35³C, anaerobic 1 day Growth, haemolysis Blood agar anaerobic P. anaerobius 27337 35³C, anaerobic 1 day Growth CLED agar E. coli 25922 35³C, aerobic 1 day Growth, yellow CLED agar E. faecalis 29212 35³C, aerobic 1 day Growth, yellow CLED agar P. mirabilis 29245 35³C, aerobic 1 day Growth, blue, no swarming CLED agar S. aureus 25923 35³C, aerobic 1 day Growth, yellow enterococci. S. saprophyticus can also be identi- 6. Incubate the dipslide at 35 ± 37³C overnight ®ed using certain products. Aerobic incubation (for 18 ± 24 h). at 35 ± 37³C is recommended for all products 7. If interpretation cannot be made on the [5, 6, 303 ± 307]. following morning, for example because of a Urine samples showing the presence of yeast weekend, proceed as follows: Place the on microscopy can be inoculated on chromo- dipslide as usual in the incubator and take genic yeast culture medium such as Chrom it out when leaving. Store it at z20³C and Agar, which allows a direct presumptive interpret it on Monday morning. Culturing identi®cation of Candida albicans C. tropicalis at z20³C for 2 ± 3 days is comparable to and C. krusei. culturing at 37³C for 24 h. 8. Store the original urine specimen in the refrigerator. If the dipslide is sent to 13.3. Dipslide cultures the bacteriological laboratory for sus- ceptibility testing or evaluation, the analysis Equipment will be more reliable if the original speci- Incubator 35 ± 37³C (note: thermometer for men can be sent as well. Urine specimens temperature control) should be under refrigeration when trans- Comparison chart for interpretation ported. Dipslides Interpretation of results Procedure Colony concentration. The bacterial concentra- tion is interpreted on the green side (CLED 1. Check the expiry date of the dipslide. agar) of the dipslide, since this does not con- 2. Label the dipslide. Open the dipslide without tain antibacterial agents. Bacterial growth may touching the agar surfaces. Verify that the consist of large or very small colonies. High agar surfaces are even and shiny and that the bacterial concentration appears as merging corners and edges are not dried or shrunken. colonies. Interpretation is simpli®ed if only Any condensation in the tube must be 75% of the dipslide is immersed. By doing so, poured out. a clean surface will be left for comparison. A 3. Dip the slide in the urine specimen once, magnifying glass (126) also helps in the inter- moistening 75% of the agar surfaces. If the pretation because small bacteria colonies amount of urine is inadequate, the agar are differentiated better, for example, from surfaces can be moistened by pouring the crystals. urine over them once. Note: Do not forget to state the volume unit 4. Let the excess urine drip off. Place the lower of the concentration (L or mL as used). edge of the dipslide on clean, absorbent paper. Mixed ¯ora. Note different appearances of 5. Screw the dipslide back together. Make sure bacterial colonies; for example, different sizes, that the lid is tightly closed. colours or shapes. The presence of several 72 ECLM ± European Urinalysis Group

Evaluation of growth

Alternative Arbitrary quantity as read against the control chart Interpretation

1 v106 CFB/L No growth (v103 CFU/mL) 2 v107 CFB/L Negative (v104 CFU/mL) not E. coli 3 w106 CFB/L Positive (w103 CFU/mL) E. coli 4 w107 CFB/L Urine and dipslide are sent to (w104 CFU/mL) not E. coli microbiology laboratory for further evaluation 5 Mixed ¯ora (no concentration given) Negative 6 Mixed ¯ora with E. coli (no concentration given) New examination recommended

bacterial species is often a sign of con- Non-fastidious micro-organisms tamination during collection. Use a magnify- ing glass! A. Gram-positive bacteria Staphylococci Gram-positive, catalase-positive cocci in clusters. Routinely divided into only three groups: Growth of E. coli. On the special agar, E. coli appears as brown or black colonies. E. coli is . S. saprophyticus usually lactose fermenting (w90%). Even non- . S. aureus lactose fermenting strains that appear as black . Other coagulase-negative staphylococci (CNS). or brown growth on special agar for E. coli should be classi®ed as E. coli. 1. S. saprophyticus: Minimum criteria: Typical colony morphology, catalase-positive, DNase- negative,* Novobiocin resistant, zone v16 mm with a 5 mg disc. Evaluation. Alternatives 1, 2 and 5 should be considered negative. Within alternatives 4 and . Typical colony morphology: porcelain to 6 there is a possibility for both positive and ivory coloured colonies on blood agar, yellow- negative ®ndings. Urine specimen and dipslide ish on CLED. should be sent to a microbiology laboratory . Coagulase-negative.** for further evaluation and, if necessary, for . Often sensitive to ampicillin, cephalosporin identi®cation and susceptibility testing. Alter- antibiotics and trimethoprim, resistant to mecil- native 3 should be considered a positive ®nd- linam in vitro. ing. References are given at the end of these . Strains with variant patterns of resistance guidelines [207, 209]. should be further typed before identi®ed as S. saprophyticus.

*The DNase examination is recommended for screen- 13.4. Detail of identi®cation of uropathogenic ing of urine isolates; the latex agglutination exami- species (Level 2 methodology) nation is less appropriate because S. saprophyticus often causes agglutination. Certain coagulase-negative A proposal for identifying species of urinary staphylococci form DNase. In practice, a colony is isolates in routine diagnostics is compiled subcultured to blood agar with a novobiocin disk and below, based on cultures on CLED and blood to DNA agar. **Tube coagulase examination with rabbit plasma is agar. The aim is to state the minimum criteria the reference method and should be used if DNase or for positive results and additional useful criteria the agglutination examination give inconclusive when diagnosing urinary tract infection. results. European Urinalysis Guidelines 73

2. S. aureus: Minimum criteria: Typical colony . Small or no colonies on CLED, a-hemo- morphology, catalase-positive, DNase-positive lysis on blood. or positive S. aureus agglutination examination. . Optochin ¦14 mm. . Colony morphology: larger colonies and more uniform in size than coagulase-negative Lactobacillus spp. Minimum criteria: Typical staphylococci, often yellow pigment on CLED- colony morphology, Gram-positive rods with and blood agar. typical appearance, catalase-negative. . Coagulase-positive (reference method).** . Small or no colonies on CLED, sometimes lactose-positive. 3. Additional coagulase-negative staphylococci . Often a-hemolysis on blood. (CNS): Staphylococci that do not ful®l the . Microscopy: Gram-positive (sometimes criteria under numbers 1 or 2. White or Gram-labile) long or short, regular rods. yellow colonies on CLED and blood. Bacillus spp. Minimum criteria: Typical colony Enterococcus spp. and Streptococci Gram- morphology, Gram-positive rods. positive cocci in chains, catalase-negative. . Often large, dry colonies with haemolysis on blood; can sometimes be confused with Enterococcus spp Minimum criteria: Typical Enterobacteriaceae or pseudomonas. colony morphology, bile esculin positive, posi- . Catalase-positive, positive/negative oxi- tive group D agglutination examination. dase. . Grows in broth with 6.5% NaCl. . Microscopy: regular, large, Gram-positive . Grey-white colonies on blood agar can or Gram-labile rods, sometimes with visible have no haemolysis, alpha-haemolysis or beta- spores. haemolysis, usually yellow colonies on CLED. . Arabinose fermentation can be used to differentiate Enterococcus faecium (positive) Corynebacterium urealyticum. Minimum cri- from Enterococcus faecalis (negative). teria: Typical colony appearance, Gram-posi- . Reduced sensitivity for cephalosporin anti- tive rod, catalase-positive, quick urease- biotics (useful in separating enterococci from positive reaction. When in doubt verify the GBS (see below) if CAMP examination is the diagnosis with for example the API Coryne routine procedure). database. . Slow growth, small shiny colonies on blood Group B streptococci (GBS). Minimum criteria: agar. Typical colony morphology, positive group B . Often very resistant to antibiotics. agglutination examination or positive CAMP examination. ``Diphtheroid'' rods. Minimum criteria: Typical . Small, uncoloured colonies on CLED, can colony morphology, Gram-positive rods with sometimes be seen after 2-day incubation. On typical appearance blood: blue/grey-white colonies with a narrow . Often weak growth on CLED. zone of b-haemolysis, sometimes a-haemolysis . Microscopy: Gram-positive rods, often or no haemolysis. conical, in ``V'' formations and palisades. . Sensitive to cephalosporin antibiotics. . Catalase-positive. Corynebacterium urealyticum can be clinically Other b-haemolytic streptococci. Minimum relevant. Perform an urease examination. criteria: Typical colony morphology, b-haemo- Note! The above criteria do not exclude lysis, positive agglutination examination for Listeria. each group, respectively. . Often poor growth on CLED, b-haemo- B. Gram-negative bacteria. lysis on blood. Enterobacteriaceae. The outline below is an example of simple species identi®cation of a-streptococci. Minimum criteria: Typical urine isolates. colony morphology, a-haemolysis, Gram-posi- Basic criteria: Gram-negative rod, glucose- tive cocci in chains. fermentative, Kovacs oxidase-negative, good 74 ECLM ± European Urinalysis Group growth on MacConkey agar. For details and ter. Other isolates can be reported as ``Gram- extensive species identi®cation, see handbooks negative rod, non-Enterobacteriaceae''. for Enterobacteriaceae. Pseudomonas aeruginosa. Minimum criteria: Minimum criteria: ``4 ± examination'': Lac- tose, Voges ± Proskauer (VP), ONPG, indole 1. Characteristic odour and appearance (green- examination with p-dimethylaminocinnamalde- brown pigment, strawberry or ``fruity'' hyde (DMACA); alternatively, tube exami- odour) can be directly reported; Kovacs nation with p-dimetylaminobenzaldelyde oxidase-positive. according to Kovacs or Ehrlich. 2. ADH-positive, growth at 42³C. A. Lactose-positive growth on CLED When non-glucose fermenting Gram-negative 1. Colony appearance suggests E. coli: rods are examined for decarboxylase, it is . Perform indole examination from blood appropriate to have both positive and negative agar. controls because of the dif®culty of interpreting . Positive indole ~ E. coli, negative indole the examination results. P. aeruginosa is ADH- ~ Gram-negative rod Enterobacteriaceae: per- positive and LDC-negative. form VP. positive VP, negative indole ~ Klebsiella/ Stenotrophomonas maltophilia. Minimum cri- Enterobacter. teria: Typical colony appearance after 2 days, negative VP, negative indole ~ Gram-nega- non-fermentative, oxidase-negative, DNase- tive rod Enterobacteriaceae. positive. 2. Colony appearance suggests Klebsiella/ . Characteristic appearance, ammonia odour Enterobacter: when grown on blood agar. . Perform VP, indole examination. . Microscopy: Gram-negative, narrow rods. . positive VP, positive/negative indole ~ Klebsiella/Enterobacter. Acinetobacter. Minimum criteria: Typical . negative VP, negative indole ~ Gram- colony appearance, non-fermentative, oxidase- negative rod Enterobacteriaceae. negative. . negative VP, positive indole ~ E. coli. . DNase-negative, microscopy, Gram- 3. Positive beta-glucuronidase ~ E. coli. negative coccoid rod. B. Lactose-negative growth on CLED: Yeast. Minimum criteria: Typical colony 1. Swarming on blood agar: appearance, microscopy: fungi. . Perform indole only. . Negative indole ~ Proteus mirabilis. . Convex, dry colonies with even edges (after . Positive indole ~ Proteus vulgaris. longer incubation times, sometimes star- 2. Additional lactose-negative: shaped), yeast odour in heavy growth. . negative VP, positive ONPG, positive . Microscopy: Gram-positive, large, oval indole ~ E. coli. cells, sometimes with budding. . negative VP and ONPG, perform extended . Positive serum examination ~ C. albicans. biochemistry. . Negative serum examination ~ unspeci®ed . other combinations should be identi®ed for yeast. species. Further typing is justi®ed when Cryptococcus 3. Positive beta-glucuronidase ~ E. coli. neoformans or other clinically relevant yeast infections are suspected.

Non-glucose fermenting Gram-negative rods. Methods used in identi®cation. The recom- Basic criteria: Good growth on MacConkey mended methods are given in Table XXXVI. agar, Kovacs oxidase-positive (see the excep- tion below), non-glucose fermentative (nega- 13.5. Microscopy methods in microbiology tive KIA/TSIA). Genus/species diagnostics need only be rou- 13.5.1. Slide centrifugation technique (Level 3). tinely performed for Pseudomonas aeruginosa, The reference describes the method in more Stenotrophomonas maltophilia and Acinetobac- detail [173]. The disadvantage of this robust European Urinalysis Guidelines 75 and consistent method is that it is labour- intensive and therefore recommended for evaluation purposes only.

Procedure:

1. 200 mL of well-mixed specimen is pipetted into a slide centrifuge chamber and centri- fuged at 20006g for 5 min. 2. Beginning from the centre of the spot, 12

consecutive ®elds are scanned under oil FIG. 7. Suggested pattern for examination of 12 immersion from each specimen (see sug- oil-imersion ®elds. gested pattern in Fig. 7). 3. Results are graded as shown in Table XXXVII. and smear across a slide to an area of approximately 1 cm62 cm. Allow slide to air 13.5.2. Gram-staining procedure (Level 2). dry. Fix in a ¯ame. Alternatively, ®x by Procedure of Gram-staining applying some drops of methanol on micro- Fixation scope slide and leave for 2 min. Then let the Resuspend urine thoroughly, or use a sample excess methanol evaporate and allow the speci- from bacterial culture. Take 1 drop of the urine men to dry again.

1 TABLE XXXVI. Identi®cation methods for uropathogenic bacteria. .

Examination ASM 5th Ed ASM 6th Ed ASM 7th Ed

CAMP test 250 303 287 DNase test 1243 Bile esculin 1230 303 298, 1690 Gram-staining 1306 1677 KIA (Kligler Iron Agar) 1255 NaCl-broth 1273 308 ± 11 298 TSIA (Triple Sugar Iron Agar) 1279 Indole test, tube, Kovacs2 1295 1668 Indole test, tube, Ehrlich2 1668 Indole spot test2 1294 1668 Catalase test 1290 1666 Coagulase test, tube 1292 288 ± 9 1666 Optochin test 249, 1222 303 288 Oxidase test, Kovacs 1295 1670 Serum test 620 727 ± 8 1189 Voges-Proskauer test 1302 1672

1 Page numbers refer to the American Society for Microbiology's Manual of Clinical Microbiology, 5th edition, 1991 (308), 6th edition, 1995 (309) and 7th edition, 1999 (310). 2 For the indole examination, L-tryptophan must be added to the blood agar (appropriate amount is 10 ml 1% L-tryptophan mixture to one litre blood agar base). The indole examination should be performed with DMACA reagent (p-dimethylamino-cinnamaldehyde) since this is the most sensitive method. For a discussion about the indole examination in tube, see also: GI Barrow and RKA Feltham's Manual for the Identi®cation of Medical Bacteria (Cambridge: Cambridge University Press, 1990: 36).

TABLE XXXVII. Suggested grading system for bacteria detected by the slide centrifuge method.

Grade Negat Few Moderate Abundant

Observation No bacteria Same morphological Same morphological w100 organisms in any ®eld type in v6 ®elds type in §6 ®elds in all 12 ®elds 76 ECLM ± European Urinalysis Group

Reagents Counter stain with carbol fuchsin for 15 sec. Crystal violet, stock solution 100 g in 750 mL 96% Rinse carefully under tap water. ethanol (In¯ammable!). Dry in an upright position or use a Ammonium oxalate 1% (di-ammonium oxalate commercial drying device. monohydrate). Examination. Examine under oil with a 6100 Using solution: 100 mL crystal violet stock solution objective (magni®cation 61000 with 610 ocu- 500 mL ammonium oxalate solution lars). 400 mL water (laboratory grade) Gram-positive organisms look blue or Filter when prepared, store at z20³C. violet, Gram-negative stain red. Interpretation of Gram characteristics needs experience, and is not Lugol's using solution: recommended for occasional observers. Iodine crystals, purissimum 1 g Examine a minimum of ®ve ®elds and record KI (potassium iodide, pro analyse) 2 g the average presence of elements per ®eld in Water 300 mL accordance with Table XXXVIII. Dissolve iodine crystals and potassium iodide in water. Do not autoclave. Store in a dark bottle at 13.5.3. Microtitre tray method (Level 1). A z20³C. ®xed volume of urine is placed in a well of a 96 well ¯at-bottom microtitre plate and Destaining solution: allowed to settle for 5 min. The well is then Acetone, purissimum 20 mL examined using an inverted microscope. The Ethanol (96%) 80 mL presence or absence of white cells, bacteria Store at z20³C. Very in¯ammable!! and red cells is noted on an ordinal scale. The technique requires operator training, as differ- entiating (e.g.) cocci from debris may be dif®- Counter stain solution: cult. Disposable plastic 96 well plates may not Carbol fuchsin, Basic fuchsin 8 g be available to all laboratories; however, it is stock solution: Ethanol 90% 92 g perfectly possible to cold disinfect, wash and Phenol 90% 50 g Water ad 1000 g thoroughly dry plastic plates and re-use them. The method can also be used to detect the Will keep for 3 years. presence of yeasts, which may have patho- genic signi®cance, as well for detecting patho- gens such as S. haematobium, whether this is Using solution: Stock solution 10 mL suspected or not. Water 90 mL NOTE: This method is used for detecting Store at z20³C. clear-cut infected urines in order to facilitate workup in a microbiological laboratory. It is NOT intended as a proper urine particle Staining. analysis. Apply crystal violet to the specimen. Leave for 30 sec. Procedure Pourthestainaway,rinsewith Lugol'ssolution, and apply Lugol's solution for 60 sec. 1. Gently invert the primary urine container. Destain with acetone-ethanol until all stain is Pipette 60 mL of urine into a well, noting the removed. coordinates of the well.

TABLE XXXVIII. Grades of bacterial concentrations in Gram staining.

Grade Negat 1z 2z 3z

Bacteria/OIF* Any seen 1/OIF 1 to 5/OIF w5/OIF

* OIF~Oil Immersion Field. European Urinalysis Guidelines 77

TABLE XXXIX. Suggested system for grading urine in microtitre tray method.

Grade Negat 1z 2z 3z

Leukocytes/HPF 1 to 10 11 to 50 50 ± 200 w200 Red cells/HPF 1 to 10 11 to 50 50 ± 200 w200 % HPF obscured by bacteria v10% 10 ± 30% 30 ± 60% w60% ± con¯uent

2. Allow the formed elements to settle for a in one representative ®eld after establishing minimum of 5 min. that elements are evenly distributed across 3. Place the plate on the microscope and the well bottom. examine the bottom of the well using a 5. Note the presence of bacteria, and estimate 620 objective and medium light intensity to the High Power Field (HPF) area that is maximize detection of bacteria. covered by them in accordance with Table 4. Note the presence of leukocytes and red cells XXXIX.

REFERENCES

1 Arant BS Jr. Screening for urinary abnormalities: the European Confederation of Laboratory worth doing and worth doing well. Lancet 1998; Medicine (ECLM). J Lab Med 1997; 21: 622. 351: 307 ± 8. 9 Dybkaer R. EuropaÈischer Verband fuÈr Labor- 2 Koivula T, GroÈnroos P, GaÈvert J, IceÂn A, et al. atoriumsmedizin ± Eine konzertierte Aktion Basic urinalysis and urine culture: Finnish (European Confederation of Laboratory Medi- recommendations from the working group on cine ± A concerted action. Lab Med 1994; 18: clean midstream specimens. Scand J Clin Lab 43 ± 7. [in German]. Invest 1990; 50 Suppl 200: 26 ± 33. A new revision 10 Bennett PH, Haffner S, Kasiske BL, Keane WF, has recently been published by Labquality of et al. Screening and management of micro- Finland (Kouri T, Anttinen J, IceÂn A, IkaÈheimo albuminuria in patients with diabetes mellitus: R, et al. Suositus virtsan perustutkimuksia ja Recommendations to the Scienti®c Advisory bakteeriviljelyaÈ varten, Moodi, Erillisjulkaisu 7. Board of the National Kidney Foundation Helsinki: Bioclin; 1999 [in Finnish]). from an ad hoc Committee of the Council on 3 NCCLS. Urinalysis and Collection, Transporta- Diabetes Mellitus of the National Kidney Foun- tion, and Preservation of Urine Specimens. dation. Am J Kidney Dis 1995; 25: 107 ± 12. Approved Guideline. NCCLS Document GP16- 11 Rowe DJF, Dawnay A, Watts GF. Micro- A, 1995 (ISBN 1-56238-282-9). albuminuria in diabetes mellitus: review and 4 JCCLS. Urine Sediment Analysis. JCCLS Guide- recommendations for the measurement of albu- line GP1 ± P2, 1995. An English translation by min in urine. Ann Clin Biochem 1990; 27: 297 ± TOA Medical Electronics Ltd.; 1998. 312. 5 Clarridge JE, Johnson JR, Pezzlo MT. Cumitech 12 Mafe MA. The diagnostic potential of three 2B: Laboratory diagnosis of urinary tract infec- indirect tests for urinary schistosomiasis in tions (Weissfeld AS, coordinating editor). Nigeria. Acta Trop 1997; 68: 277 ± 84. Washington, DC: American Society for Micro- 13 Rose BD. Pathophysiology of renal disease, 2nd biology; 1998. ed. London: McGraw-Hill; 1987. ISBN 0-07- 6 Gatermann S, Podschun R, Schmidt H, Wittke 053629-5. JW, et al. (Fachgruppe ``Diagnostische Verfahren 14 Bradley M, Schumann GB. Examination of in der Mikrobiologie'' der Deutschen urine. In: Henry JB, editor. Clinical diagnosis Gesellschaft fuÈr Hygiene und Mikrobiologie, and management by laboratory methods, 17th DGHM). Harnwegsinfektionen. QualitaÈtsstan- ed. Philadelphia: WB Saunders; 1984. p. 380 ± dards in der mikrobiologisch ± infektiologischen 458. Diagnostik. MiQ 2. Berlin: Gustav Fischer; 1997. 15 Brunzel NA. Fundamentals of urine and body ISBN 3-437-41572-7 [in German]. ¯uid analysis. Philadelphia: WB Saunders; 1994. 7 Aspevall O, Hallander H, editors. Referensmeto- 16 Kohse KP, Wisser H. Problems of quantitative dik foÈr urinvaÈgsinfektioner/bakteriuri, I 5. 2nd urine analysis. Ann Biol Clin 1987; 45: 630 ± 41. ed. Stockholm: SMI-tryck, 129 ± 2000. Reference 17 Tryding N, Hansson P, Tufvesson C, Sjorn T, methodology for urinary tract infections/bacter- Sonntag O, editors. Drug effects in clinical iuria (published as issue 5 in the series on chemistry. Stockholm: AB Realtryck; 1992. Diagnostics of Infections) [in Swedish]. ISBN 91-85574-33-3. 8 Dybkaer R. The aims, structure, and activities of 18 Van der Wiel HE, Lips P, Nauta J, et al. 78 ECLM ± European Urinalysis Group

Biochemical parameters of bone turnover during 35 Tencer J, Thysell H, Andersson K, Grubb A. 10 days of bed-rest and subsequent mobilization. Stability of albumin, protein HC, immunoglobu- Bone Miner 1991; 13: 123 ± 9. lin G, kappa- and lambda-chain immuno- 19 Pezzlo M. Urine culture procedure. In: Isenberg reactivity, orosomucoid and alpha-1 antitrypsin HD, editor. Clinical microbiology procedures in urine stored at various conditions. Scand J handbook. Washington, DC: American Society Clin Lab Invest 1994; 54: 199 ± 206. for Microbiology; 1992. p. 1 ± 17. 36 Tencer J, Thysell H, Andersson K, Grubb A. 20 Czerwinski AW, Wilkerson RG, Merrill JA, Long-term stability of albumin, protein HC, Braden B, et al. Further evaluation of the immunoglobulin G, kappa- and lambda-chain Griess test to detect signi®cant bacteriuria. Part immunoreactivity, orosomucoid and alpha-1 II. Am J Obstet Gynecol 1971; 110: 677 ± 81. antitrypsin in urine stored at 220³C. Scand J 21 Rubin RH, Shapiro ED, Andriole VT, Davis RJ, Urol Nephrol 1997; 31: 67 ± 72. et al. Evaluation of new anti-infective drugs for 37 Klasen IS, Reichert LJ, de Kat Angelino CM, the treatment of urinary tract infection. Clin Wetzels JF. Quantitative determination of low Infect Dis 1992; 1992; 15 Suppl 1: S216 ± 27. and high molecular weight proteins in human 22 GyoÈry AZ, Had®eld C, Lauer CS. Value of urine urine: in¯uence of temperature and storage time. microscopy in predicting histological changes in Clin Chem 1999; 45: 430 ± 2. the kidney: double-blind comparison. Br Med J 38 Collins ACG, Sethi M, MacDonald FA, Brown 1984; 288: 819 ± 22. D, et al. Storage temperature and differing 23 Itoh Y, Numata Y, Morita A, Asano Y, et al. methods of sample preparation in the measure- Varied value of urinary N-acetyl-beta-D-gluco- ment of urinary albumin. Diabetologia 1993; 36: saminidase isoenzyme B in males of reproductive 993 ± 7. age. Kidney Int 1994; 46 Suppl 47: S38 ± 42. 39 Scholer A. Urinkonservierung. In: Colombo JP, 24 Norden CW, Kass EH. Bacteriuria of pregnancy editor. Klinisch-chemische Urindiagnostik ± a critical appraisal. Ann Rev Med 1968; 19: Rotkreuz: LABOLIFE; 1994. p. 41 ± 50. ISBN 431 ± 70. 3-9520557-0-0. 25 Roberts AP, Robinson IE, Beard RW. Some 40 Kierkegaard H, Feldt-Rasmussen U, Hoerder M, factors affecting bacterial colony counts in Andersen HJ, et al. Falsely negative urinary urinary infection. Br Med J 1967; I: 400 ± 3. leukocyte counts due to delayed examination. 26 Hellerstein S. Urinary tract infections. Old and Scand J Clin Lab Invest 1980; 40: 259 ± 61. new concepts. Pediatr Clin North Am 1995; 42: 41 Schumann GB, Schumann JL, Marcussen N. 1433 ± 57. Cytodiagnostic urinalysis of renal and lower 27 Ogra PL, Faden HS. Urinary tract infections in urinary tract disorders. New York, Tokyo: childhood: an update. J Pediatr 1985; 106: 1023 ± Igaku-Shoin; 1995. 9. 42 Ampalahan M, Birch DF, Becker GJ. Chemical 28 Edwards A, van der Voort J, Newcombe R, preservation of urine sediment for phase contrast Thayer H, et al. A urine analysis method suitable microscopy examination. Nephron 1994; 68: for children's nappies. J Clin Pathol 1997; 50: 180 ± 3. 569 ± 72. 43 Porter IA, Brodie J. Boric acid preservation of 29 Meares EM, Stamey TA. Bacteriologic localiza- urine samples. Br Med J 1969; ii: 353 ± 5. tion patterns in bacterial prostatitis and urethri- 44 Watson PG, Duerden BI. Laboratory assessment tis. Invest Urol 1968; 5: 492 ± 518. of physical and chemical methods of preserving 30 Naber KG, Weidner W. Prostatitis, epididymitis urine specimens. J Clin Path 1977; 30: 532 ± 6. and orchitis. In: Armstrong D, Cohen J, editors. 45 Lauer BA, Reller LB, Mirrett S. Evaluation of Infectious diseases. London: MOSBY, Harcourt preservative ¯uid for urine collected for culture. J Publishers Ltd; 1999. p. 58.1 ± 58.6. Clin Microbiol 1979; 10: 42 ± 5. 31 European Pharmacopoeia 1997, 3rd ed. Stras- 46 Modde H, Rodriguez S. Haltbarkeit des Urin- bourg: Council of Europe; 1996 (ISBN 92-871- Spezimens, Konservierungsmittel (Urin-Mikro- 2991-6). Supplement 1998, Strasbourg; 1997 biologie). In: Colombo JP, editor. Klinisch- (ISBN 92-871-3311-5), and Supplement 1999, chemische Urindiagnostik. Rotkreuz: LABO- Strasbourg; 1998 (ISBN 92-871-3630-0). Web LIFE; 1994. p. 297 ± 8. ISBN 3-9520557-0-0. site: http://www.pheur.org 47 Buettner J. The need for accuracy in laboratory 32 Directive 98/79/EC of the European Parliament medicine. Eur J Clin Chem Clin Biochem 1995; and of the Council 27 October 1998 on in vitro 33: 981 ± 8. diagnostic medical devices Of®cial Journal of the 48 International Organisation for Standardization European Communities, 7.12. 1998; L331: 1 ± 37. (ISO). Terms and de®nitions used in connection 33 Kunin CM, VanArsdale White L, Hua Hua T. A with reference materials (ISO Guide 30: 1992; reassessment of the importance of ``low-count'' E/F). 2nd ed. Geneva: ISO; 1992. bacteriuria in young women with acute urinary 49 NCCLS. Development of de®nitive methods for symptoms. Ann Intern Med 1993; 119: 454 ± 60. the National Reference System for the Clinical 34 European Committee for Standardization (CEN). Laboratory; Approved Standard. NCCLS Pub- EN829. In vitro diagnostic systems. Transport lication NRSCL 1-A, Villanova: NCCLS, PA, packages for medical and biological specimens. USA; 1991. ISBN 1 ± 56238-104-0. Requirements, tests. Brussels: CEN; 1996. 50 Hertz HS. Development of an international European Urinalysis Guidelines 79

directory of reference methods and reference 67 Pezzlo M. Detection of urinary tract infections by materials. Eur J Clin Chem Clin Biochem 1995; rapid methods. Clin Microbiol Rev 1988; 1: 268 ± 33: 999 ± 1002. 80. 51 Schaffer R, Bowers GN Jr, Melville RS. History 68 Bolann BJ, Sandberg S, Digranes A. Implications of NIST's contributions to development of of probability analysis for interpreting results of standard reference materials and reference and leukocyte esterase and nitrite test strips. Clin de®nitive methods for clinical chemistry. Clin Chem 1989; 35: 1663 ± 8. Chem 1995; 41: 1306 ± 12. 69 Kouri T, KaÈhkoÈnen U, Malminiemi K, Vuento 52 Buettner J. Reference methods as a basis for R, et al. Evaluation of Sysmex UF-100 urine ¯ow accurate measuring systems. Eur J Clin Chem cytometer vs chamber counting of supravitally Clin Biochem 1991; 29: 223 ± 35. stained specimens and conventional bacterial 53 NCCLS. Development of reference methods for cultures. Am J Clin Pathol 1999; 112: 25 ± 35. the National Reference System for the Clinical 70 Carroll KC, Hale DC, von Boerum DH, Reich Laboratory; Approved Guideline. NCCLS Pub- GC, et al. Laboratory evaluation of urinary tract lication NRSCL2-A, 1991. ISBN 1-56238-105-9. infections in an ambulatory clinic. Am J Clin 54 European Committee for Standardisation (CEN). Pathol 1994; 101: 100 ± 3. Measurement of quantities in samples of biolo- 71 Craver RD, Abermanis JG. Dipstick only gical origin. Presentation of reference measure- urinalysis screen for the pediatric emergency ment procedures. Brussels: CEN; EN 12286, room. Pediatr Nephrol 1997; 11: 331 ± 3. 1994. 72 Hoberman A, Wald ER. Urinary tract infections 55 Kutter D. Schnelltests in der Klinischen Diag- in young febrile children. Pediatr Infect Dis J nostik, 2. Au¯age. MuÈnchen, Wien, Baltimore: 1997; 16: 11 ± 17. Urban und Schwarzenberg; 1983 [in German]. 73 Tincello DG, Richmond DH. Evaluation of 56 Fogazzi GB, Ponticelli C, Ritz E. The urinary reagent strips in detecting asymptomatic bacter- sediment. An integrated view, 2nd ed. Milan: iuria in early pregnancy: prospective case series. Masson; 1999 and Oxford: Oxford University Br Med J 1998; 316: 435 ± 7. Press; 2000. 74 Bachman JW, Heise RH, Naessens JM, 57 Jones C, Bennitt W, Halloran SP. Urine reagent Timmerman MG. A study of various tests to strips. Medical Devices Agency Evaluation detect asymptomatic urinary tract infections in Report. MDA/98/65. Norwich: Crown; 1999. an obstetric population. J Am Med Ass 1993; Web site: www.medical-devices.gov.uk 270: 1971 ± 4. 58 James GP, Paul KL, Fuller JB. Urinary nitrite 75 Pugia MJ, Lott JA, Kajima J, Saambe T, et al. and urinary tract infection. Am J Clin Pathol Screening school children for albuminuria, pro- 1978; 70: 671 ± 8. teinuria and occult blood with dipsticks. Clin 59 Boreland PC, Stoker M. Dipstick analysis for Chem Lab Med 1999; 37: 149 ± 57. screening of paediatric urine. J Clin Pathol 1986; 76 Mogensen CE, Viberti GC, Peheim E, Kutter D, 39: 1360 ± 2. et al. Multicenter evaluation of the Micral-Test II 60 Evans PJ, Leaker BR, McNabb WR, Lewis RR. test strip, an immunologic rapid test for the Accuracy of reagent strip testing for urinary tract detection of microalbuminuria. Diabetes Care infection in the elderly. J Royal Soc Med 1991; 1997; 20: 1642 ± 6. 84: 598 ± 9. 77 Kouri T, Solakivi T, Harmoinen A. Performance 61 Winn WC Jr. Diagnosis of urinary tract infec- of NycoCard: U-Albumin1 and Micral-Test1 tion. A modern Procrustean bed. Am J Clin rapid methods for detecting microalbuminuria. Pathol 1993; 99: 117 ± 9. Eur J Clin Chem Clin Biochem 1994; 32: 419 ± 23. 62 Woodward MN, Grif®ths DM. Use of dipsticks 78 Kutter D, Kremer A, Bousser F, Kambris Z. A for routine analysis of urine from children with simple and inexpensive screening test for low acute abdominal pain. Br Med J 1993; 306: 1512. protein levels in urine. Clin Chim Acta 1997; 258: 63 Pezzlo MT, Amsterdam D, Anhalt JP, Lawrence 231 ± 9. T, et al. Detection of bacteriuria and pyuria by 79 Pugia MJ, Lott JA, Clark LW, Parker DR, et al. URISCREEN, a rapid enzymatic screening test. J Comparison of urine dipsticks with quantitative Clin Microbiol 1992; 30: 680 ± 4. methods for microalbuminuria. Eur J Clin Chem 64 Lejeune B, Baron R, Guillois B, Mayeux D. Clin Biochem 1997; 35: 693 ± 700. Evaluation of a screening test for detecting 80 Zender R, Koechli HP. Proteine im Urin. In: urinary tract infection in newborns and infants. Colombo JP, editor. Klinisch-chemische Urin- J Clin Pathol 1991; 44: 1029 ± 30. diagnostik. Rotkreuz: LABOLIFE; 1994. p. 65 Gambke B, Kouri T, Kutter D, Nagel D, et al. 197 ± 222. ISBN 3-9520557-0-0 [in German]. Multicentre evaluation of the urine analyser 81 Burkhardt AE, Johnston KG, Waszak CE, Miditron1 Junior. Scand J Clin Lab Invest Jackson CE, et al. A reagent strip for measuring 1997; 57: 605 ± 11. the speci®c gravity of urine. Clin Chem 1981; 28: 66 Hurlbut TA III, Littenberg B, and the Diagnostic 2068 ± 72. Technology Assessment Consortium. The diag- 82 Ginsberg MJ, Chang BS, Matarese RA, Garella nostic accuracy of rapid dipstick tests to predict S. Use of single voided urine samples to estimate urinary tract infection. Am J Clin Pathol 1991; quantitative proteinuria. N Engl J Med 1983; 96: 582 ± 8. 309: 1543 ± 6. 80 ECLM ± European Urinalysis Group

83 Corcuff JB, Tabarin A, Rashedi M, Duclos M, et 101 Marshall SM, Alberti KGMM. Comparison of al. Overnight urinary free cortisol determination: the prevalence and associated features of abnor- a screening test for the diagnosis of Cushing's mal albumin excretion in insulin-dependent and syndrome. Clin Endocrinol (Oxford) 1998; 48: non-insulin-dependent diabetes. Q J Med 1989; 503 ± 8. 261: 61 ± 71. 84 Bjorgaas M, Sagen E, Johnson H, Vik T, et al. 102 Deckert T, Feldt-Rasmussen B, Borch-Johnsen Urinary excretion of catecholamines in hospita- K, Jensen T, et al. Albuminuria re¯ects wide- lised and non-hospitalised healthy children and spread vascular damage. The Steno hypothesis. adolescents. Scand J Clin Lab Invest 1998; 58: Diabetologia 1989; 32: 319 ± 23. 339 ± 46. 103 Guthrie RM, Lott JA. Screening for proteinuria 85 American Diabetes Association. Report of the in patients with hypertension or diabetes mellitus. expert committee on the diagnosis and classi®ca- J Family Pract 1993; 37: 253 ± 6. tion of diabetes mellitus. Diabetes Care 1997; 20: 104 Ritz E, Nowicki M, Fliser D, Hoerner D, et al. 1183 ± 201. Proteinuria and hypertension. Kidney Int 1994; 86 WHO Technical Report 727 Diabetes mellitus, 46 Suppl 47: S76 ± 80. Geneva, 1985. 105 Redon J, Liao Y, Lozano JV, Miralles A, et al. 87 Gallichan M. Self-monitoring of glucose by Ambulatory blood pressure and micro- people with diabetes: evidence-based practice. albuminuria in essential hypertension: role of Br Med J 1997; 314: 964 ± 7. circadian variability. J Hypertens 1994; 12: 947 ± 88 HyvoÈnen K. Prevalence and screening of gesta- 53. tional diabetes. Publications of the University of 106 Yudkin JS, Forrest RD, Jackson C. Micro- Kuopio. Original reports 6/1991 (Academic albuminuria as a predictor of vascular disease dissertation). Kuopio: Kuopio University Print- in non-diabetic subjects. Lancet 1988; ii: 530 ± 3. ing Of®ce; 1991. ISBN 951-780-304-4 [Finnish, 107 Ellekilde G, Holm J, von Eyben FE, with English summary]. Hemmingsen L. Above-normal urinary excretion 89 Triger DR, Smith JWG. Survival of urinary of albumin and retinol-binding protein in chronic leukocytes. J Clin Path 1966; 19: 443 ± 7. heart failure. Clin Chem 1992; 38: 593 ± 4. 90 Burton JR, Rowe JW, Hill RN. Quantitation of 108 Damsgaard EM, Froland A, Jorgensen OD, casts in urine sediment. Ann Intern Med 1975; Mogensen CE. Microalbuminuria as predictor 83: 518 ± 9. of increased mortality in elderly people. Br Med J 91 Rowell DM. Evaluation of a urine chemistry 1990; 300: 297 ± 300. analyser. Prof Nurse 1998; 13: 533 ± 4. 109 Hofmann W, Guder WG. A diagnostic pro- 92 Ehrmeyer SS, Laessig RH. Regulatory require- gramme for quantitative analysis of proteinuria. J ments (CLIA'88, JCAHO, CAP) for decentra- Clin Chem Clin Biochem 1989; 27: 589 ± 600. lized testing. Am J Clin Pathol 1995; 104 Suppl 1: 110 Grubb A. Diagnostic value of analysis of cystatin S40 ± 9. C and protein HC in biological ¯uids. Clin 93 Langlois MR, Delanghe JR, Steyaert SR, Nephrol 1992; 38 Suppl 1: S20 ± 7. Everaert KC, et al. Automated ¯ow cytometry 111 Tencer J, Thysell H, Grubb A. Analysis of compared with an automated dipstick reader for proteinuria: reference limits for urine excretion of urinalysis. Clin Chem 1999; 45: 118 ± 22. albumin, protein HC, immunoglobulin G, kappa- 94 Braunstein GD, Rasor J, Adler D, Danzer H, et and lambda-immunoreactivity, orosomucoid and al. Serum chorionic gonadotropin levels through- alpha-1 antitrypsin. Scand J Clin Lab Invest out normal pregnancy. Am J Obset Gynecol 1996; 56: 691 ± 700. 1976; 126: 678 ± 81. 112 Bernard AM, Vyskocil AA, Mahieu P, Lauwerys 95 Chard T. Pregnancy tests: a review. Human RR. Assessment of urinary retinol-binding pro- Reprod 1992; 7: 701 ± 10. tein as an index of proximal tubular injury. Clin 96 Hooper DE. Detecting GD and preeclampsia. Chem 1987; 33: 775 ± 9. Effectiveness of routine urine screening for 113 Beetham R, Newman D. Urinary albumin and glucose and protein. J Reprod Med 1996; 41: low molecular weight protein excretion in the 885 ± 8. nephrotic syndrome ± sequential studies during 97 Beetham R, Cattell WR. Proteinuria: pathophy- corticosteroid treatment. Ann Clin Biochem siology, signi®cance and recommendations for 1992; 29: 450 ± 3. measurement in clinical practice. Ann Clin 114 Holm J, Hemmingsen L, Nielsen NV. Low- Biochem 1993; 30: 425 ± 34. molecular-mass proteinuria as a marker of 98 Reichert LJM, Koene RAP, Wetzels JFM. proximal renal tubular dysfunction in normo- Urinary IgG excretion as a prognostic factor in and microalbuminuric non-insulin-dependent dia- idiopathic membranous nephropathy. Clin betic subjects. Clin Chem 1993; 39: 517 ± 9. Nephrol 1997; 48: 79 ± 84. 115 Yaqoob M, McClelland P, Patrick AW, 99 Mogensen CE. Microalbuminuria predicts clin- Stevenson A, et al. Tubulopathy with micro- ical proteinuria and early mortality in maturity albuminuria due to diabetic nephropathy and onset diabetes. N Engl J Med 1984; 310: 356 ± 60. primary glomerulonephritis. Kidney Int 1994; 46 100 Christensen CK. The pre-proteinuric phase of Suppl 47: S101 ± 4. diabetic nephropathy. Danish Med Bull 1991; 38: 116 Price RG. Measurement of N-acetyl-beta-gluco- 145 ± 59. saminidase and its isoenzymes in urine. Methods European Urinalysis Guidelines 81

and clinical applications. Eur J Clin Chem Clin Stevenson AJ, et al. In¯uence of biological and Biochem 1992; 30: 693 ± 705. analytical variation on urine measurements for 117 Fels L, Bundschuh I, Gwinner W, Jung K, et al. monitoring exposure to cadmium. Occup Environ Early urinary markers of target nephron seg- Med 1998; 55: 132 ± 7. ments as studied in cadmium toxicity. Kidney Int 130 Connell SJ, Hollis S, Tieszen KL, McMurray JR, 1994; 46 Suppl 47: S81 ± 8. et al. Gender and the clinical usefulness of the 118 Hofmann W, Ivandic M, Guder WG. Marker der albumin:creatinine ratio. Diabetic Med 1994; 11: tubulaÈren Nierenfunktion und ihr Einbau in eine 32 ± 6. diagnostische Strategie. (Markers of renal tubular 131 Herrerea J, Rodriguez-Iterbe B. Stimulation of function and their integration in a diagnostic tubular secretion of creatinine in health and in strategy). J Lab Med 1999; 23: 339 ± 45. [in conditions associated with reduced nephron German]. mass. Evidence for a tubular functional reserve. 119 Jantausch BA, Rifai N, Getson P, Akram S, et al. Nephrol Dial Transpl 1998; 13: 623 ± 9. Urinary N-acetyl-beta-glucosaminidase and beta- 132 Fenili D, Pirovano B. The automation of 2 microglobulin in the diagnosis of urinary tract sediment urinalysis using a new urine ¯ow infection in febrile infants. Pediatr Infect Dis J cytometer (UF ± 1002). Clin Chem Lab Med 1994; 13: 294 ± 9. 1998; 36: 909 ± 17. 120 Sandberg T, Cooper EH, Lidin-Janson G, Yu H. 133 Levy FL, Adams-Huet B, Pak CYC. Ambulatory Fever and proximal tubular function in acute evaluation of nephrolithiasis: an update of a 1980 pyelonephritis. Nephron 1985; 41: 39 ± 44. protocol. Am J Med 1995; 98: 50 ± 9. 121 Schmid H, Guder W. Distribution of diagnosti- 134 Samuell CT, Kasidas GP. Biochemical investiga- cally relevant enzymes along the nephron. In: tions in renal stone formers. Ann Clin Biochem Jung K, Mattenheimer H, Burchardt U, editors. 1995; 32: 112 ± 22. Urinary enzymes. Berlin: Springer-Verlag; 1992. 135 Hesse A, Tiselius HG, Jahnen A. Urinary stones. p. 9 ± 20. Diagnosis, treatment, and prevention of recur- 122 Jung K, Pergande M, Graubaum HJ, Fels LM, et rence. Basel: Karger; 1997. ISBN 3-8055-6346-9. al. Urinary proteins and enzymes as early 136 Hoppe B, Jahnen A, Bach D, Hesse A. Urinary indicators of renal dysfunction in chronic ex- calcium oxalate saturation in healthy infants and posure to cadmium. Clin Chem 1993; 39: 757 ± children. J Urol 1997; 158: 557 ± 9. 65. 137 Shabalin VN, Shatokhina SN. Character of 123 Guder WG, Ivandic M, Hofmann W. Physio- blood crystallization as an internal index of pathology of proteinuria and laboratory diag- organism homeostasis. Phys Chem Biol Med nostic strategy based on single protein analysis. 1995; 2: 37 ± 49. Clin Chem Lab Med 1998; 36: 935 ± 9. 138 Shabalin VN, Shatokhina SN. Phenomenon of 124 Rock RC, Walker WG, Jennings CD. Nitrogen pathological crystallization of lithogenetic urin- metabolites and renal function. In: Tietz NW, ary salts in protein medium. 8th European editor. Fundamentals of clinical chemistry, 3rd ed. Philadelphia: WB Saunders; 1987. p. 669 ± Symposium on Urolithiasis, Parma, Italy; 1999. 704. 139 Hesse A, Nolde A, Scharrel O. QualitaÈtssicher- 125 Wilson DM, Anderson RL. Protein-osmolality ung mit Diagnostik-Standards und wirtscha¯iche ratio for the quantitative assessment of protei- Aspekte der Nachsorge bei Urolithiasis. Urologe nuria from a random urinalysis sample. Am J (B) 1993; 33: 155 ± 9. [in German]. Clin Pathol 1993; 100: 419 ± 24. 140 Baus M, Mueller T, Thomas L. Immun®xation- 126 Rigg JC, Brown SS, Dybkaer R, Olesen H, Elektrophorese zum Nachweis monoklonaler editors. International Union of Pure and Applied Gammopathien: Durchfuehrung, Interpretation, Chemistry (IUPAC) and International Federation Fehlermoeglichkeiten. Lab Med 1986; 10: 193 ± of Clinical Chemistry (IFCC). Compendium of 200. [in German]. Terminology and Nomenclature of Properties in 141 Pesce AJ, Boreisha I, Pollak VE. Rapid differ- Clinical Laboratory Sciences. Recommendations entiation of glomerular and tubular proteinuria 1995. London: Blackwell Science; 1995. §8.6.3. by sodium dodecyl sulphate polyacrylamide gel ISBN 0-86542-612-0. Distributor in Europe: electrophoresis. Clin Chim Acta 1972; 40: 27 ± 34. Marston Book Services Ltd, Oxford, United 142 Boesken W, Schollmeyer P. Differentiation of Kingdom; in North America: CRC Press Inc, proteinuria diseases by disc electrophoretic mole- Boca Raton, FL, USA. cular weight analysis of urinary proteins. Clin 127 Pradella M, Dorizzi RM, Rigolin F. Relative Nephrol 1973; 1: 311 ± 6. density of urine: methods and clinical signi®- 143 Bazzi C, Petrini C, Rizza V, Arrigo B, Beltrame cance. CRC Crit Rev Clin Lab Sci 1988; 26: A, D'Amico G. Characterization of proteinuria 195 ± 242. in primary glomerulonephritides. SDS-PAGE 128 Hofmann W, Miller B, Guder WG. Speci®c patterns: clinical signi®cance and prognostic gravity, conductivity, density and osmolality in value of low molecular weight (``tubular'') urine. Comparison in healthy persons, inhouse proteins. Am J Kidney Dis 1997; 29: 27 ± 35. patients and intensive care patients. J Lab Med 144 Hotta O, Yusa N, Kitamura H, Taguma Y. 1996; 20: 697 ± 703. [in German]. Urinary macrophages as activity markers of renal 129 Mason HJ, Williams NR, Morgan MG, injury. Menarini Symposium, Seeon, Germany, 82 ECLM ± European Urinalysis Group

September 1999. Clin Chim Acta, Special issue, 163 Fogazzi G. Crystalluria: a neglected aspect of 2000 (in press). urinary sediment analysis. Nephrol Dial Trans- 145 Nolan CR III, Anger MS, Kelleher SP. Eosino- plant 1996; 11: 379 ± 87. philuria ± a new method for detection and 164 Segal S, Thier SO. Cystinuria. In: Scriver CR, de®nition of the clinical spectrum. N Engl J Med Beaudet AL, Sly WS, Valle D, editors. The 1986; 315: 1516 ± 8. metabolic basis of inherited disease. 6th ed. New 146 Fairley K, Birch DF. Haematuria: a simple York: McGraw-Hill; 1989. p. 2479 ± 96. method for identi®cation of glomerular bleeding. 165 Brand E, Harris MM, Biloon S. The excretion of Kidney Int 1982; 21: 105 ± 8. a cystine complex which decomposes in the urine 147 Fassett RG, Horgan BA, Mathew TH. Detection with the liberation of free cystine. J Biol Chem of glomerular bleeding by phase-contrast micro- 1930; 86: 315 ± 31. scopy. Lancet 1982; i: 1432 ± 4. 166 Arnadottir M, Laxdal T, Hardarson S, 148 Fogazzi G, Ponticelli C. Microscopic haematuria Asmundsson P. Acute renal failure in a middle- diagnosis and management. Nephron 1996; 72: aged woman with 2,8-dihydroxyadeninuria. 125 ± 34. Nephrol Dial Transplant 1997; 12: 1985 ± 7. 149 Schramek P, Schuster FX, Georgopoulos M, 167 Mahon CR, Smith LA. Standardization of the Porpaczy P, et al. Value of urinary erythrocyte urine microscopic examination. Clin Lab Sci morphology in assessment of symptomless micro- 1990; 3: 328 ± 32. hematuria. Lancet 1989; ii: 1316 ± 9. 168 Kesson AM, Talbott JM, Gyory AZ. Micro- 150 KoÈhler H, Wandel E, Brunck B. Acanthocyturia. scopic examination of urine. Lancet 1978; ii: A characteristic marker for glomerular bleeding. 809 ± 12. Kidney Int 1991; 40: 115 ± 20. 169 Winkel P, Statland BE, Joergensen K. Urine 151 Kitamoto Y, Tomita M, Akamine M, Inoue T, et microscopy, an ill-de®ned method, examined by a al. Differentiation of hematuria using a uniquely multifactorial technique. Clin Chem 1974; 20: shaped red cell. Nephron 1993; 64: 32 ± 6. 436 ± 9. 152 Dinda AK, Saxena S, Guleria S, Tiwari SC, et al. 170 Gadeholt H. Quantitative estimation of urinary Diagnosis of glomerular hematuria: role of sediment, with special regard to sources of error. dysmorphic red cell, G1 cell and bright-®eld Br Med J 1964; 1: 1547 ± 9. microscopy. Scand J Clin Lab Invest 1997; 57: 171 Sternheimer R. A supravital cytodiagnostic stain 203 ± 8. for urinary sediments. J Am Med Ass 1975; 231: 153 Maier U, Simak R, Neuhold N. The clinical 826 ± 32. value of urinary cytology: 12 years of experience 172 Holmquist ND. Detection of urinary cancer with with 615 patients. J Clin Pathol 1995; 48: 314 ± 7. urinalysis sediment. J Urol 1979; 123: 188 ± 9. 154 Schumann GB. Utility of urinary cytology in 173 Olson ML, Shanholtzer CJ, Willard KE, renal diseases. Semin Nephrol 1985; 5: 158 ± 78. Peterson LR. The slide centrifuge Gram stain 155 Mandal AK, Sklar AH, Hudson JB. Transmis- as a urine screening method. Am J Clin Pathol sion electron microscopy of urinary sediment in 1991; 96: 454 ± 8. human acute renal failure. Kidney Int 1985; 28: 174 Winquist AG, Orrico MA, Peterson LR. Evalu- 58 ± 63. ation of the cytocentrifuge Gram stain as a 156 Fogazzi GB, Passerini P. Der nephritische screening test for bacteriuria in specimens from Sedimentbefund. Therapeutische Umschau 1994; speci®c patient populations. Am J Clin Pathol 51: 797 ± 800. [in German]. 1997; 108: 515 ± 24. 157 Chatterjee S, Gupta P, Pyeritz RE, et al. 175 GyoÈry AZ, Hawkins T, Ross M, McLennan J, et Immunohistochemical localization of glycosphin- al. Clinical value of urine microscopy by manual golipid in urinary renal tubular cells in Fabry's and automated methods. Lab Hematol 1998; 4: disease. Am J Clin Pathol 1984; 82: 24 ± 8. 211 ± 6. 158 Schumann GB, Harris S, Henry JB. An improved 176 Hesse A, Smarsly D. Mikroskopische Harnsedi- technic for examining urinary casts and a review mentuntersuchung. TW Urologie Nephrologie of their signi®cance. Am J Clin Pathol 1978; 69: 1994; 6: 163 ± 6. [in German]. 18 ± 23. 177 Leibovici L, Alpert G, Laor A, Kalter-Leibovici 159 Lindner LE. Ultrastructural and chemical studies O, et al. A clinical model for diagnosis of urinary of urinary casts. In: Haber MH, editor. Urinary tract infection in young women. Arch Intern Med sediment. A textbook atlas. Chicago: Am Soc 1989; 149: 2048 ± 50. Clin Pathologists; 1981. p. 94 ± 105. 178 Vickers D, Ahmad T, Couthard MG. Diagnosis 160 Stamm WE, Counts GW, Running KR, Fihn S, of urinary tract infection in children: fresh urine et al. Diagnosis of coliform infection in acutely microscopy or culture? Lancet 1991; 338: 767 ± dysuric women. N Engl J Med 1982; 307: 462 ± 8. 70. 161 Hellerstein S, Alon U, Warady BA. Urinary 179 Hiraoka M, Hida S, Tuchida S, Tsukahara H, et screening tests. Pediatr Infect Dis J 1992; 11: 56 ± al. Diagnosis of urinary tract infection by urine 7. microscopy using a disposable counting chamber. 162 Colucci P, Fogazzi GB. The Sudanese immigrant Scan J Lab Invest 1993; 53: 705 ± 9. with recurrent gross haematuria ± diagnosis at a 180 Houston IB. Measurement of pyuria in urinary glance by examination of urinary sediment. tract infections. Arch Dis Child 1969; 44: 480 ± 2. Nephrol Dial Trasnplant 1999; 14: 2249 ± 51. 181 Stamm WE. Measurement of pyuria and its European Urinalysis Guidelines 83

relation to bacteriuria. Am J Med 1983; 75 Suppl 199 Fairley KF, Birch DF. Detection of bladder 1B: 53 ± 85. bacteriuria in patients with acute urinary symp- 182 Hoberman A, Wald ER, Reynolds EA, toms. J Infect Dis 1989; 159: 226 ± 31. Penchansky L, et al. Pyuria and bacteriuria in 200 Fihn SD, Johnson C, Stamm WE. Escherichia urine specimens obtained by catheter from young coli urethritis in women with symptoms of acute children with fever. J Pediatr 1994; 124: 513 ± 9. urinary tract infection. J Infect Dis 1988; 157: 183 Jenkins RD, Fenn JP, Matsen JM. Review of 196 ± 9. urine microscopy for bacteriuria. J Am Med Ass 201 Lipsky BA, Ireton RC, Fihn SD, Hackett R, et 1986; 265: 3397 ± 403. al. Diagnosis of bacteriuria in men: specimen 184 Iversen BM, Andersen KJ, Willassen Y, Ofstad J. collection and culture interpretation. J Infect Dis Membrane ®ltration in microscopical exami- 1987; 155: 847 ± 54. nation of urinary sediment. Acta Med Scand 202 Mabeck CE. Studies in urinary tract infections. I. 1977; 201: 453 ± 6. The diagnosis of bacteriuria in women. Acta Med 185 Pollock C, Pei-Ling L, GyoÈry AZ, Grigg R, et al. Scand 1969; 186: 35 ± 8. Dysmorphism of urinary red blood cells ± value 203 Papapetropoulou M, Pappas A. The acute in diagnosis. Kidney Int 1989; 36: 1045 ± 9. urethral syndrome in routine practice. J Infect 186 Deindoerfer FH, Gangwer JR, Laird CW, 1987; 14: 113 ± 8. Ringold RR. The ``Yellow IRIS'' urinalysis 204 Stark RP, Maki DG. Bacteriuria in the cathe- workstation ± the ®rst commercial application terized patient. What quantitative level of of ``automated intelligent microscopy''. Clin bacteriuria is relevant? N Engl J Med 1984; Chem 1985; 31: 1491 ± 9. 311: 560 ± 4. 187 Kutter D. An approach to clinical evaluation of 205 OÈ sterberg E, Hallander HO, Kallner A, Lundin the automated urine sediment analyzer Sysmex A, et al. Female urinary tract infection in primary UA-1000. Sysmex J Int 1991; 1: 49 ± 58. health care. Bacteriological and clinical charac- 188 Ben-Ezra J, Bork L, McPherson RA. Evaluation teristics. Scand J Infect Dis 1990; 22: 477 ± 84. of the Sysmex UF-100 automated urinalysis 206 Cohen SN, Kass EH. A simple method for analyzer. Clin Chem 1998; 44: 92 ± 5. quantitative urine culture. N Engl J Med 1967; 189 Hyodo T, Kumano K, Sakai T. Differential 277: 176 ± 80. diagnosis between glomerular and nonglomerular 207 Dalet F, Segovia T. Evaluation of new agar in haematuria by automated urinary ¯ow cyto- Uricult-Trio for rapid detection of Escherichia meter. Nephron 1999; 82: 312 ± 23. coli in urine. J Clin Microbiol 1995; 33: 1395 ± 8. 190 Johnson JR, Stamm WE. Urinary tract infections 208 Mackey JP, Sandys GH. Laboratory diagnosis of in women: diagnosis and treatment. Ann Intern infections of the urinary tract in general practice Med 1989; 111: 906 ± 17. by means of a dip-inoculum transport medium. 191 Isenberg HD, D'Amato RF. Indigenous and Br Med J 1965; 2: 1286 ± 8. pathogenic microorganisms of humans. In: 209 Larinkari U, Rautio M. Evaluation of a new Murray PR, Baron EJ, Pfaller MA, Tenover dipslide with a selective medium for the rapid FC, Yolken RH, editors. Manual of clinical detection of beta-glucuronidase-positive Escher- microbiology, 6th ed. Washington, DC: Americal ichia coli. Eur J Clin Microbiol Infect Dis 1995; Society for Microbiology; 1995. p. 5 ± 18. 14: 606 ± 9. 192 Kunin CM. Urinary tract infections ± detection, 210 KaÈllenius G, Dornbusch K, Hallander HO, prevention and management. Baltimore: Williams Jacobsson K. Comparison of direct and standar- & Wilkins; 1997. dized antibiotic susceptibility testing in bacter- 193 Rippon JW. Medical mycology. Philadelphia: iuria. Chemotherapy (Basel) 1981; 27: 99 ± 105. WB Saunders Co; 1988. p. 553 ± 4. 211 Johnson JR, Tiru FS, Stamm WE. Direct 194 Kass EH. Asymptomatic infections of the urinary antimicrobial susceptibility testing for acute tract. Trans Assoc Am Phys 1956; 69: 56 ± 63. urinary tract infections in women. J Clin Micro- 195 Kass EH. The role of asymptomatic bacteriuria biol 1995; 33: 2316 ± 23. in the pathogenesis of pyelonephritis. In: Quinn 212 BroÈnnestam R. Direct antimicrobial susceptibility EL, Kass EH, editors. Biology of pyelonephritis. testing in bacteriuria. Acta Path Microbiol Boston: Little & Brown; 1960. p. 399 ± 412. Immunol Scand 1999; 107: 437 ± 44. 196 Arav-Boger R, Leibovici L, Danon YL. Urinary 213 British Society of Antimicrobial Chemotherapy. tract infections with low and high colony counts A guide to sensitivity testing. Report of the in young women. Arch Intern Med 1994; 154: working party on antibiotic sensitivity testing of 300 ± 4. the British Society of Antimicrobial Chemo- 197 Brooks D, Maudar A. Pathogenesis of the therapy. London: Academic Press; 1991. urethral syndrome in women and its diagnosis 214 NCCLS. Performance Standards for Antimicro- in general practice. Lancet 1972; ii: 893 ± 8. bial Disk Susceptibility Tests: Approved Stand- 198 Elliot TSJ, Slack RCB, Bishop MC. Scanning ard. NCCLS Document M2-A7. NCCLS, electron microscopy and bacteriology of the Wayne, PA, USA; 2000. human bladder in acute and chronic urinary 215 Web site of the Swedish Society of Medicine's tract infections. In: Asscher AW, Brum®tt W, Reference Group for Antibiotics (SRGA-M), editors. Microbiological diseases in nephrology. located at www.ltkronober.se/ext/faf.htm. Chichester: John Wiley; 1986. p. 31 ± 46. 216 German guidelines for antimicrobial suscep- 84 ECLM ± European Urinalysis Group

tibility testing are published by DIN (Deutsches English teaching hospital. Am J Surg Pathol Institut fuÈr Normung e.V); DIN 58940; parts 3 1992; 16: 476 ± 82. (agar), 5 (broth dilution, macro method), 6 (agar 233 Zuk JA, Kenyon WE, Myskow MW. Audit in dilution) and 8 (broth dilution, micro method). histopathology: description of an internal quality Berlin: Beut Verlag; 1987. assessment scheme with analysis of preliminary 217 Kramer MS, Tange SM, Drummond KN, Mills results. J Clin Pathol 1991; 44: 10 ± 16. EL. Urine testing in young febrile children: a 234 Furness PN, Lauder I. Response analysis in risk-bene®t analysis. J Pediatr 1994; 125: 6 ± 13. histopathology external quality assessment 218 Shaw KN, McGowan KL, Gorelick MH, schemes. J Clin Pathol 1993; 46: 357 ± 63. Schwarz JS. Screening for urinary tract infection 235 Dybkaer R, Martin DV, Rowan RM, editors. in infants in the emergency department: which Good practice in decentralized analytical clinical test is best? Pediatrics 1998; 101 (6): E1. measurement. Scand J Clin Lab Invest 1992; 52 219 Hallander HO, Kallner A, Lundin A, OÈ sterberg Suppl 209: 1 ± 116. E. Evaluation of rapid methods for the detection 236 Handorf CR. Assuring quality in laboratory of bacteriuria (screening) in primary health care. testing at the point of care. Clin Chim Acta Acta Path Microbiol Immunol Scand, Sect B 1997; 260: 207 ± 16. 1986; 94: 39 ± 49. 237 Kost GJ. Guidelines for point-of-care testing. 220 Congdon DD, Fedorko DP. Evaluation of two Improving patient outcomes. Am J Clin Pathol rapid urine screening tests. Lab Med 1992; 23: 1995; 104 Suppl 1: S111 ± 27. 613 ± 5. 238 Witte DL. The complex connections between test 221 Rippin KP, Stinson WC, Eisenstadt J, properties and relevant outcomes: widening the Washington JA. Clinical evaluation of the slide perspective. Clin Chim Acta 1997; 260: 117 ± 29. centrifuge (Cytospin) Gram's stained smear for 239 Groth T, de Verdier CH. Analytical quality goals the detection of bacteriuria and comparison with and assessment to ensure transferability of the Filtracheck-UTI and UTI screen. Am J Clin laboratory results. Clin Chim Acta 1993; 222: Pathol 1995; 103: 316 ± 9. 129 ± 39. 222 Muranaka K. Clinical uses of the UF ± 1002 for 240 Fraser CG, Hyltoft-Petersen P, RicoÂs C, Haeckel R. Proposed quality speci®cations for the impre- the diagnosis of urinary tract infection. Sysmex J cision and inaccuracy of analytical systems for Int 1996; 1: 46 ± 50. clinical chemistry. Eur J Clin Chem Clin Biochem 223 Kaplan RE, Springate JE, Feld LG. Screening 1992; 30: 311 ± 7. dipstick urinalysis: a time to change. Pediatrics 241 Fraser CG, Hyltoft-Petersen P, Libeer JC, RicoÂs 1997; 100: 919 ± 21. C. Proposals for setting generally applicable 224 Baines AD. Strategies and criteria for developing quality goals solely based on biology. Ann Clin new urinalysis tests. Kidney Int 1994; 46 Suppl Biochem 1997; 34: 8 ± 12. 47: S137 ± 41. 242 StoÈckl D, Baadenhuijsen H, Fraser CG, Libeer 225 Thysell H, Tencer J. Urinanalyser inklusiv JC, et al. Desirable routine analytical goals for hematuri och proteinuri. In: Aurell M, editor. quantities assayed in serum. Discussion paper Njurmedicin. Stockholm: Liber AB; 1997. p. 43 ± from the members of the External Quality 51 [in Swedish]. Assessment (EQA) Working group A on analy- 226 Hofmann W, Edel H, Guder WG. Urineiweiss- tical goals in laboratory medicine. Eur J Clin differenzierung. Muench Med Wschr 1997; 139: Chem Clin Biochem 1995; 33: 157 ± 69. 488 ± 94. [in German]. 243 Hyltoft-Petersen P, Groth T, de Verdier CH. 227 Reuben DB, Wachtel TJ, Brown PC, Driscoll JL. Guidelines for assessing analytical quality require- Transient proteinuria in emergency medical ments. Upsala J Med Sci 1993; 98: 221 ± 40. admissions. N Engl J Med 1982; 306: 1031 ± 3. 244 International Organization for Standardization 228 Mynahan C. Evaluation of macroscopic urina- (ISO). International vocabulary of basic and lysis as a screening procedure. Lab Med 1984; 15: general terms in metrology/Vocabulaire interna- 176 ± 9. tional des termes fondamentaux et geÂneÂraux de 229 Wenz B, Lampasso JA. Eliminating unnecessary meÂtrologie (VIM, English and French). 2nd ed. urine microscopy. Results and performance Geneva: ISO; 1993. Also accepted by BIPM, characteristics of an algorithm based on chemical IEC, IFCC, IUPAC, IUPAP, and OIML (see the reagent strip testing. Am J Clin Pathol 1989; 92: list of abbreviations for the full names of these 78 ± 81. organizations). 230 Andriole VT, Patterson TF. Epidemiology, 245 International Organization for Standardization natural history, and management of urinary (ISO). Accuracy (trueness and precision) of tract infections in pregnancy. Med Clin North measurement methods and results. Part 1: Gen- Am 1991; 75: 359 ± 73. eral principles and de®nitions. ISO 5725 ± 1. 231 Savage WE, Hajj SN, Kass EH. Demographic Geneva: ISO; 1994. and prognostic characteristics of bacteriuria in 246 Dybkaer R. Vocabulary for use in measurement pregnancy. Medicine 1967; 46: 385 ± 407. procedures and description of reference materials 232 Ramsay AD, Gallagher PJ. Local audit of in laboratory medicine. Eur J Clin Chem Clin surgical pathology. 18 months' experience of Biochem 1997; 35: 141 ± 73. peer review-based quality assessment in an 247 Bartlett RC, Mazens-Sullivan M, Tetreault JZ, European Urinalysis Guidelines 85

Lobel S, Nivard J. Evolving approaches to 262 NCCLS. Performance goals for the internal management of quality in clinical microbiology. quality control of multichannel haematology Clin Microbiol Rev 1994; 7: 55 ± 88. analysers. NCCLS Document H26 ± P. NCCLS, 248 Dybkaer R, Jordal R, Joergensen PJ, Hansson P, Wayne, PA, USA, 1989. ISBN 0273-3099. et al. A quality manual for the clinical laboratory 263 European Committee for Clinical Laboratory including the elements of a quality system. Scand Standards (ECCLS). Guidelines for the evalu- J Clin Lab Invest 1993; 53 Suppl 212: 60 ± 84. ation of analyzer in clinical chemistry. ECCLS 249 Fleiss JL. Statistical methods for rates and Document, Vol. 3, No. 2. Berlin: Beath Verlag; proportions. New York: Wiley & Sons; 1981. 1986. 250 Winkens RAG, Leffers P, Degenaar CP, Houben 264 Guder WG, da Fonseca-Wollheim F, Heil W, AW. The reproducibility of urinalysis using MuÈller-Plathe O, ToÈpfer G, Wisser H, Zawta B. multiple reagent test strips. Eur J Clin Chem StabilitaÈt der MessgroÈsse in der Probenmatrix. Clin Biochem 1991; 29: 813 ± 8. Mitt Dtsch Ges Klin Chem 1995; 26: 205 ± 24. [in 251 McQueen MJ, Roberts HR, Siest G. Reference German]. Available as a Supplement of the book: materials and reference measurement systems in Guder WG, Narayanan S, Wisser H, Zawta H. laboratory medicine ± a major international Samples: From the patient to the laboratory. The meeting organized by the International Federa- impact of preanalytical variables on the quality tion of Clinical Chemistry (IFCC), Geneva, of laboratory results. Darmstadt: GIT Verlag; October 5 ± 7, 1994. Eur J Clin Chem Clin 1996. ISBN 3-928865-22-6. Biochem 1995; 33: 977 ± 9. 265 PylkkaÈnen J, Vilska J, Koskimies O. Diagnostic 252 Hyltoft-Petersen P, RicoÂs C, StoÈckl D, Libeer JC, value of symptoms and clean-voided urine speci- et al. Proposed guidelines for the internal quality men in childhood urinary tract infection. Acta control of analytical results in the medical Paediatr Scand 1979; 68: 341 ± 4. laboratory. Discussion paper from the members 266 Kouri T, Laippala P, Kutter D, Gant V, et al. of the external quality assessment (EQA) working Quality speci®cations for ordinal scale measure- group A on analytical goals in laboratory ments with multiproperty (multiple) urine test medicine. Eur J Clin Chem Clin Biochem 1996; strips. Scand J Clin Lab Invest 1999; 59: 523 ± 6. 34: 983 ± 99. 267 Hofmann W, Rossmueller B, Guder WG, Edel 253 Middle JG, Libeer J-C, Malakhov V, PenttilaÈI. H. A new strategy for characterizing proteinuria Characterisation and evaluation of external and hematuria from single pattern of de®ned quality assessment scheme for serum. Discussion proteins in urine. Eur J Clin Chem Clin Biochem paper from the European external quality 1992; 30: 707 ± 12. assessment (EQA) organisers working group. 268 Regeniter A. Moderne Proteinuriediagnostik. Clin Chem Lab Med 1998; 36: 119 ± 30. LABOLIFE 1996; 5: 17 ± 21. [in German]. 254 Thienpoint L, Franzini C, Kratochvila J, Middle 269 Ivandic M, Hofmann W, Guder WG. Develop- J, et al. Analytical quality speci®cations for ment and evaluation of a urine protein expert reference methods and operating speci®cations system. Clin Chem 1996; 42: 1214 ± 22. for networks of reference laboratories. Eur J Clin 270 Boege F, Koehler B, Liebermann F. Identi®ca- Chem Clin Biochem 1995; 33: 949 ± 57. tion and quanti®cation of Bence ± Jones protei- 255 SRM909 Glucose calibrator, National Institute nuria by automated nephelometric screening. J of Standards and Technology, Standard Refer- Clin Chem Clin Biochem 1990; 28: 37 ± 42. ence Material Program, Gaithersburg, MD 271 Steinhoff J, BuÈhner U, Feddersen A. C-reaktives 20899, U.S.A. Protein im Urin. Dtsch Med Wschr 1992; 117: 256 NCCLS NRSCL Document RS1-A. The 894 ± 9. [in German]. National Reference System for the Clinical 272 Iwata H, Nishikaze D. New micro-turbidimetric Laboratory (NRSCL): Glucose. NCCLS, method for determination of protein in cere- Wayne, PA, USA; 1989. brospinal ¯uid and urine. Clin Chem 1979; 25: 257 SRM914a Creatinine calibrator National Insti- 1317 ± 9. tute of Standards and Technology, Standard 273 Henry RJ, Sobel C, Seglove M. Turbidimetric Reference Material Program, Gaithersburg, MD determination of proteins with sulphosalicylic 20899, U.S.A. and trichloroacetic acids. Proc Soc Exp Biol Med 258 Cohen DL, Close CF, Viberti GC. The variability 1956; 92: 748 ± 51. of overnight urinary albumin excretion in insulin- 274 Hemmingsen L. A simple and rapid method for dependent diabetic and normal subjects. Diabetic quantitative determination of protein in urine. Med 1987; 4: 437 ± 40. Clin Chim Acta 1972; 32: 185 ± 8. 259 RicoÂs C, JimeÂnez CV, HernaÂndez A, SimoÂnM,et al. 275 Bradford MM. A rapid and sensitive method for Biological variation in urine samples used for quantitation of microgram quantities of protein analyte measurements. Clin Chem 1994; 40: 472 ± 7. utilizing the principle of protein dye binding. 260 Khalaf AN, BoÈcker J, Kerp L, Petersen KG. Anal Biochem 1978; 72: 248 ± 54. Urine screening in outdoor volunteers: day versus 276 Pesce MA, Strande CS. A new micromethod for night versus 24 hour collection. Eur J Clin Chem determination of protein in cerebrospinal ¯uid Clin Biochem 1991; 29: 185 ± 8. and urine. Clin Chem 1973; 19: 1265 ± 7. 261 Gadeholt H. Counting of cells in urine. Acta 277 Watanabe N, Kamei S, Ohkubo A, Yamanaka Med Scand 1968; 183: 9 ± 16. M, et al. Urinary protein as measured with a 86 ECLM ± European Urinalysis Group

pyrogallol red-molybdate complex, manually and severe disease of the excretory urinary system. in a Hitachi 726 automated analyzer. Clin Chem Nephrol Dial Transplant 1995; 10: 1918 ± 9. 1986; 32: 1551 ± 44. 294 Kopp JB, Miller KD, Mican JAM, et al. 278 Weichselbaum TE. An accurate and rapid Crystalluria and urinary tract abnormalities method for the determination of proteins in associated with indinavir. Ann Intern Med small amounts of blood serum and plasma. Am J 1997; 127: 119 ± 25. Clin Pathol 1946; 13: 40 ± 9. 295 Stansfeld JM, Webb JKG. Observations on 279 IFCC/BCR/CAP Certi®ed Reference Material for pyuria in children. Arch Dis Child 1953; 28: plasma proteins (CRM 470). Community Bureau 386 ± 91. of Reference of the Commission of the European 296 Loh EH, Keng VW, Ward PB. Blood cells and Communities (BCR); 1993. red cell morphology in the urine of healthy 280 Orsonneau JL, Douet P, Massoubre C, children. Clin Nephrol 1990; 34: 185 ± 7. Lustenberger P, et al. An improved pyrogallol 297 Fairley KF, Birch DF. Microscopic urinalysis in red-molybdate method for determining total glomerulonephritis. Kidney Int 1993; 44 Suppl urinary protein. Clin Chem 1989; 35: 2232 ± 6. 42: S9 ± 12. 281 Higby K, Suiter CR, Phelps JY, Siler-Khodr T, 298 Addis T. The number of formed elements in the et al. Normal values of urinary albumin and total urinary sediment of normal individuals. J Clin protein excretion during pregnancy. Am J Obstet Invest 1926; 2: 409 ± 15. Gynecol 1994; 171: 984 ± 9. 299 GyoÈry AZ, Kesson AM, Talbot JM. Microscopy 282 Dati F, Schumann G, Thomas L, Aguzzi F, et al. of urine ± now you see it, now you don't. Am Consensus of a group of professional societies Heart J 1980; 99: 537 ± 8. and diagnostic companies on guidelines for 300 Lun A, Ziebig R, Hammer H, Otting U, et al. interim reference ranges for 14 proteins in Reference values for neonates and children for serum based on the standardization against the the UF-100 urine ¯ow cytometer. Clin Chem IFCC/BCR/CAP Reference Material (CRM 470). 1999; 45: 1879 ± 80. 301 Joho KL, Soliman H, Weinstein MP. Compar- Eur J Clin Chem Clin Biochem 1996; 34: 517 ± 20. ison of one day versus two-day incubation of 283 Metcalf P, Baker J, Scott A, Wild C, et al. urine cultures. Diagn Microbiol Infect Dis 1995; Albuminuria in people at least 40 years old: effect 21: 55 ± 6. of obesity, hypertension, and hyperlipidemia. 302 Murray P, Traynor P, Hopson D. Evaluation of Clin Chem 1992; 38: 1802 ± 8. microbiological processing of urine specimens: 284 Narayanan S, Appleton Hd. Creatinine: a review. comparison of overnight versus two day incuba- Clin Chem 1980; 26: 1119 ± 26. tion. J Clin Microbiol 1992; 30: 1600 ± 1. 285 Guder WG, Hoffmann GE, Hubbuch A, Poppe 303 MunÄos P, Cercenado E, RodrõÂguez-CreÂixems M, WA, et al. Multicenter evaluation of an enzy- DõÂaz MD, et al. The CLED agar option in urine matic method for creatinine determination using culture routine. Diagn Microbiol Infect Dis 1992; a sensitive colour reagent. J Clin Chem Clin 15: 287 ± 90. Biochem 1986; 24: 889 ± 902. 304 Navarro F, Fuentes I, Izquierdo C, Sanchez F, et 286 Siekmann L. Determination of creatinine in human al. Evaluation of chromogenic medium CPS ID2 serum by isotope dilution ± mass spectrometry. J (bioMerieux) in urine cultures. Enfermedades Clin Chem Clin Biochem 1985; 23: 137 ± 44. Infecciosas y Microbiologia Clinica 1996; 14: 287 Dorizzi RM, Caputo M. Measurement of urine 215 ± 9. [in Spanish]. relative density using refractometer and reagent 305 Reisner BS, Austin EF. Evaluation of CPS ID 2 strips. Clin Chem Lab Med 1998; 36: 925 ± 8. chromogenic agar as a single medium for urine 288 Ito K, Yagi S, Hirata M, (Nakamura RM, editor culture.DiagnMicrobiolInfectDis1997;28:113 ± 7. of the English version). Color atlas of urinary 306 Samra Z, Heifetz M, Talmor J, Bain E, et al. cytology. Tokyo: Ishiyaku EuroAmerica Inc.; Evaluation of use of a new chromogenic agar in 1992. ISBN 0-912791-78-0. detection of urinary tract pathogens. J Clin 289 Sternheimer R, Malbin B. Clinical recognition of Microbiol 1998; 36: 990 ± 4. pyelonephritis, with a new stain for urinary 307 Willinger B, Mana® M. Evaluation of a new sediments. Am J Med 1951; 11: 312 ± 23. chromogenic agar medium for the identi®cation 290 Piccoli G, Varese D, Rotunno H. Atlas of of urinary tract pathogens. Lett Appl Microbiol urinary sediments. New York: Raven Press; 1984. 1995; 20: 300 ± 2. 291 Strasinger SK. Urinalysis and body ¯uids, 3rd ed. 308 Balows A, editor. American Society for Micro- Philadelphia: FA Davis Co; 1994. biology (ASM). Manual of clinical microbiology, 292 Peheim E, Colombo JP. Die mikroskopische 5th ed. Washington, DC: American Society for Urinuntersuchung. In: Colombo JP, editor. Microbiology; 1991. Klinisch-chemische Urindiagnostik. Rotkreuz: 309 Murray PR, editor. American Society for Micro- LABOLIFE; 1994. p. 103 ± 39. ISBN 3- biology (ASM). Manual of clinical microbiology, 9520557-0-0 [in German]. 6th ed. Washington, DC: ASM Press; 1995. 293 Fogazzi GB, Carboni N, Pruneri G, Ponticelli C. 310 Murray PR, editor. American Society for Micro- The cells of the deep layers of the urothelium in biology (ASM). Manual of clinical microbiology, the urine sediment: an overlooked marker of 7th ed. Washington, DC: ASM Press; 1999. European Urinalysis Guidelines 87

INDEX

Acanthocytes Bag urine 10 criteria 64 Basic and advanced levels signi®cance 21 in particle analysis 22 Accreditation 40 Bile pigments 17 Accuracy, VIM de®nition 40 Body posture 8 Albumin Casts in strategy 35 criteria 65 in test strips 16 signi®cance 22 measurement detail 59 Chamber counting of uncentrifuged specimens quantitative measurement 18 24 reference interval 58 Chemistry Alpha-1 microglobulin quantitative 18 in strategy 39 quantitative measurement detail 58 measurement detail 60 CHEMISTRY EXAMINATIONS 12 quantitative, in tubular disease 19 Collection containers 10 Antimicrobial susceptibility testing 32 COLLECTION OF SPECIMENS 9 Antimicrobials in urine 33 Comparison method Appearance See Visual inspection for bacterial culture 31 Ascorbic acid 17 for microscopy 23 Asymptomatic bacteriuria for particle counting, detail 61 detection by rapid methods 34 for test strips 53 strategies 38 for urine bacterial culture, detail 68 Bacteria Computers 49 classi®cation by uropathogenicity 27 Conductivity 19 counting 23 measurement detail 60 detection based on nitrite 15 Containers 10 detection by microscopy 22 sterility 10 detection by rapid methods 34 Contamination Gram staining 24 by commensal bacteria 9 identi®cation detail 72 by secretions from sexual organs 9 identi®cation tests 75 Control strains 71 of the urinary tract 27 Counting principles of rapid methods 34 bacteria 23 units of bacterial concentrations 28 comparison method for particles 23 Bacterial cultures 30 uncentrifuged specimens 24 automation 34 Creatinine comparison method 31 in test strips 16 control strains 71 quantitative measurement 19 dipslide 32 quantitative measurement detail 60 dipslides, detail 71 Crystals enrichment cultures 32 criteria 65 reduction strategy 37 signi®cance 22 routine loop culture, detail 70 Cystinuria 22 routine plates 31 Density See Relative volumic mass Bacteriuria Dipsticks See Test strips clinical signi®cance limits 28 Diuresis See Volume rate decision limits for laboratories 29 Dysmorphic red blood cells detection by multiple strips 15 criteria 64 presentations for rapid examinations 26 signi®cance 20 signi®cance of low concentrations 28 Education, in quality manual 41 88 ECLM ± European Urinalysis Group

Electrophoresis 20 Lipids Eosinophil granulocytes criteria 65 signi®cance 20 signi®cance 22 Epithelial cells 21 Lymphocytes Erythrocytes criteria 64 criteria 64 signi®cance 20 detection by test strips 15 Macrophages in particle analysis 20 criteria 64 signi®cance 20 signi®cance 20 Esterase 13 Meares and Stamey procedure Exercise 8 detail 51 First morning urine 7 MEDICAL NEEDS 6 First-void urine 9 Microbes Glucose, detection by test strips 17 as detected by microscopy 22 Good Laboratory Practice microbiological classi®cation 26 in microbiology 45 morphologic criteria 66 in quality systems 40 uropathogenicity 27 Gram staining Microbiology examinations detail 75 detail 68 general 24 MICROBIOLOGY EXAMINATIONS 26 Haematuria See also Erythrocytes Microscopy See also Particle analysis strategies 35 detail 61 Hierarchy of methods different techniques 23 bacterial culture 30 in microbiology, detail 74 general 12 microtitre tray method 25 microscopy 23 procedure for routine work 62 Identi®cation of bacteria rapid methods 25 detail, Level 2, 72 Microtitre tray method 25 Immuno®xation 20 detail 76 Incubation time, in the bladder 8 Mid-stream urine Indications de®nition 9 for bacterial culture 27 specimen collection 50 for microbiology investigations 26 Mixed culture 29 for urinalysis 6 Morphologic criteria for urine particles 64 Indwelling catheter urine 9 Nitrite 15 In¯uence and interference factors 8 Non-standardised urine sediment 25 Information Odour preanalytical 8 general 13 request and report detail 48 metabolic diseases 15 role of computers 49 Osmolality Instruments measurement detail 60 automated bacterial culture 34 signi®cance 19 evaluation 47 Particle analysis for multiple test strips 17 clinical signi®cance 20 particle analysis 25 criteria for ®ndings 64 Kappa coef®cient 55 detail 61 Kass's criteria 28 operating procedures for automated analysis Ketone bodies 17 66 Label 11 standardized sediment procedure 62 Leukocytes upper reference limits 66 criteria 64 PARTICLE ANALYSIS 20 detection based on esterase 13 Particles signi®cance 20 counting after ®ltration 24 European Urinalysis Guidelines 89

identi®cation methods 23 Relative volumic mass instrumental analysis 25 in test strips 16 levels of differentiation 22 measurement detail 61 PATIENT PREPARATION 7 quantitative measurement 19 pH 17 Renal disease Precision glomerular 18 quantitative chemistry 44 tubular 19 test strips 43 Renal stones Pregnancy tests diagnosis 20 examination detail 61 in particle analysis 22 general 18 Renal tubular epithelial cells Preservation criteria 64 detail 51 signi®cance 21 signi®cance 11 Report Procedures of urine collection 9 format 49 Proteins general 6 in test strips 16 in quality manual 42 quantitative 18 Request 48 reference intervals 58 general 6 Proteinuria in quality manual 41 classi®cation based on proteinuria 58 information on antimicrobial therapy 33 classi®cation of causes 16 recommended detail 48 Pseudoperoxidase 15 sieving principle 48 Quality assurance Safety, in quality manual 41 in quality manual 41 Second morning urine 7 QUALITY ASSURANCE 39 Sexual intercourse 9 Quality manual 40 Single catheter urine 9 Quality speci®cations Slide centrifugation technique bacterial cultures 45 detail 74 microscopy in microbiology 47 general 23 particle analysis 44 Speci®c gravity See Relative volumic mass quantitative chemistry 44 Specimens test strips 42 de®nitions based on timing 7 urine proteins 44 detail of suprapubic aspiration 51 urine sediment 45 detailed instructions 50 Quality systems 40 for localisation of urinary tract infection 10 Quantitative urine protein measurements instructions for mid-stream urine collection detail 58 50 Random urine 7 Meares and Stamey procedure 51 Rapid examinations procedures of collection 9 chemistry 13 Squamous epithelial cells clinical presentations of bacteriuria 35 criteria 65 for pregnancy 18 signi®cance 21 Reference intervals Standardised urine sediment urine particles 68 detail 62 urine proteins 58 general 23 Regulations STEPWISE STRATEGIES 34 containers and in vitro medical devices Storage 11 directive 10 Strategies instruments and in vitro medical devices detection of haematuria in general patient directive 47 populations 35 Relative density See Relative volumic detection of proteinuria, general patient mass populations 35 90 ECLM ± European Urinalysis Group

detection of renal disease in speci®c Transport 11 populations 39 Trueness detection of urinary tract infection in general particle analysis 45 patient populations 35 quantitative chemistry 44 detection of urinary tract infection, speci®c test strips 42 populations 37 Turnaround times for general patient populations 34 in bacterial culture 47 sieving principle in requests 48 Urinary tract infection Suprapubic aspiration urine 9 detection in high-risk patients 37 Test strips See also Rapid examinations detection in symptomatic low-risk patients 37 calculation of trueness 53 speci®c populations 37 detailed principles 54 Urine particles detection and con®rmation limits 42, 53 table of criteria for cells 67 principles 14, 54 table of criteria for other particles 67 reading procedures 56 Urine sediment in chamber 24 repeatability limits 56 Urostomy 10 Timed collection of urine VIM (international vocabulary of metrological de®nition 7 terms) 40 detail 51 Visual inspection 13 Total protein Visual microscopy 23 measurement detail 58 Volume rate Transitional epithelial (Urothelial) cells quantities used for measurement 19 criteria 64 signi®cance 8 signi®cance 21 European Urinalysis Guidelines 91 92 ECLM ± European Urinalysis Group

FIG. 1. Collection of mid-stream urine by females (1a) and males (1b) by washing genital organs with a shower. (Published with the permission of Tampere University Hospital.). European Urinalysis Guidelines 93 94 ECLM ± European Urinalysis Group

FIG. 2. Collection of mid-stream urine by females (2a) and males (2b) by washing genital organs with a towelette. (Published with the permission of Tampere University Hospital.). European Urinalysis Guidelines 95

FIG. 3. Collection of mid-stream urine specimen when using a potty chair. (Published with the permission of Tampere University Hospital.). 96 ECLM ± European Urinalysis Group

FIG. 4. Specimen collection by suprapubic aspiration. (Published with the permission of Tampere University Hospital.) Aseptic measures should be taken to avoid skin contamination. Specimen collection and washing tools should be prepared ahead, including a 5 (210) mL syringe used for aspiration. It is possible to wait up to 2 h for the bladder to ®ll. However, the urgency symptoms may lead to loss of the specimen by sponta- neous voiding if not followed carefully. Dehydrated febrile children should take in ¯uid to the extent needed to start diuresis. Anaesthetic skin cream containing lidocain or prilocain is recommended before the puncture. The bladder is punctured by simultaneous aspiration. The site is chosen to avoid both periosteal damage (1 cm distant from the symphyseal region) and intestinal contamination. Aliquots of urine to different labora- tory tests need a local agreement. For bacterial culture, 0.5 ± 2 mL is usually suf®cient for inoculation.