Simplified microscopy for rapid detection of significant bacteriuria in random urine specimens

(1)

JOURNALOFCLINICAL MICROBIOLOGY, Mar.1978,p.286-291 0095-1137/78/0007-0286$02.00/0

Vol. 7, No. 3

Printed in U.S.A.

Simplified Microscopy for Rapid Detection of

Significant

Bacteriuria

in

Random

Urine Specimens

GARY K. BARBIN,JEFFREY D. THORLEY,*ANDJAMES A. REINARZ Department of Medicine, University of Texas Medical Branch, Galveston, Texas 77550

Received forpublication1August1977

Simplified urine microscopy, nitrite testing, and dipstick culturewerecompared

with urine loop streak culturecolonycountsin219randomvoided specimensto

determine the accuracy of the threerapid screening techniques. Nitrite testing resultedin65% falsenegative results, which couldnotbesignificantly improved by incubation at37°C but whichcould be improved by adding nitrate substrate before incubation.Dipstick culturecouldnot bequantitated until after 18 h of incubation. A new, simplified microscopy technique, using unspun, unstained

urine, resulted in 4% false negative results and 4% false positive results in specimens containing over 105 organisms per ml and was the best method.

Centrifuges, Gram stainingreagents, andcounting chambers are not necessary

foraccuratemicroscopic screening of random urine specimens for thepresenceof

bacteriuria by this technique, and the resultsareimmediately available.

The human urinary tract isnormally sterile,

with the exception of the distalfewmillimeters

of theurethra. Bacteriuria is abnormal and has

been associated withacute urinarytract

infec-tions (24), chronic

pyelonephritis

(28), high

blood pressure (10), excess fetal prematurity,

andperinatal death (10). Since only 1 to 7% of

females and0.04 to0.05% of males(11, 15)have

bacteriuria, and since the majorityare

asymp-tomatic (9), screening methods must be rapid,

simple, and accurate. Detection andtreatment

ofpersons with bacteriuria may decrease

mor-bidity.

Manytestshave beenproposedtoscreen for

bacteriuriainvoidedurine(8), including

micro-scopic examinationofurine(centrifugedor

un-centrifuged) (3, 13, 16, 18, 22, 23, 26),chemical

testingof urine(by nitrite, catalase, tetrazolium,

glucose,orenzymes) (2,12, 17, 19,25),andrapid

culturetechniques

(dipstick culture,

agar cup,or

dipslide) (4, 7,20).Themostadvantageoustest

would be both sensitive and specific, utilize a

randomlyvoided urinespecimen, andgive rapid

results *hilethe patient was stillpresent. We

have prospectively studied and compared the

most promising rapid tests -for bacteriuria in

random urine specimens: simplified urine mi-croscopy, nitrite test, anddipstick cultures.

MATERIALS AND METHODS Specimens.Twohundred nineteen unselected ur-inesamples,which hadbeensubmittedtothe clinical

laboratory as clean catch, midstream, voided

speci-mensforculture,werestudied.Meanage ofpatients

was 36years (range 1 to85); 60weremale and 159

female. Noattempt wasmadetoidentifytreatment at

the time ofspecimencollection. Specimens were

re-frigeratedupon receipt, and all procedureswere

per-formed within24h.Except foranoccasional technical problem,everyspecimenwasquantitatively cultured, examined by simplified microscopy, tested for the

presenceofnitrite, and cultured bydipstick.

Quantitativeurineculture. Each urine specimen

wascultured inourlaboratory by the calibratedloop (0.01 ml) streak method using eosin-methylene blue

agarandsheep bloodagar.Colonycountsweredone from sheep blood agar andwere the standard with which other techniques were compared. Bacteria in numbersgreaterthan105/mlwereidentifiedtospecies by using APIstrips(Analytab Products, CarlePlace,

N.Y.) and other media andreagentsasnecessary.No

pourplate cultureswereperformed.

Simplified microscopyof urine. Because of the complexity of reported procedures,weuseda simpli-fiedtechniquesuggested by in vitro studies of Kunin (13). Adrop ofmanually agitated, uncentrifugedurine from eachspecimenwasplacedon aglass slide,

cov-ered withaglassslip,andexamined for bacteria under oil immersion (magnification x1,000, field diameter

190,um) (Fig. 1).Thefollowingguidelinesestablished a test aspositive: (i) identification of bacteria with

smooth surfaces without adherent material, and (ii)

demonstration ofatleast onebacterium per oil field

ineach of five fields. Carewastakennotto misinter-pretcrystalline materialasbacteria.

Nitritetest with incubation.A 2-ml sample of

urine from eachspecimenwaspipettedintoasterile tubeand testedbydipstickfor thepresence of nitrite. The urinewasthenincubatedat 37°C and retested

after 2, 4, 6, and 24 h. Two differentcommercially available nitrite indicator dipsticks were utilized si-multaneously (N-Uristix,AmesCo., Elkhart, Ind.,and

Bac-U-Dip, Warner-Chilcott Laboratories, Morris

Plains,N.J.).

286

on February 7, 2020 by guest

http://jcm.asm.org/

(2)

BACTERIURIA MICROSCOPY 287

Nitritetestwithsupplementalnitrate. Ten

ur-ine samples were selected because they contained more than 105 organisms per ml but tested nitrite

negative. Duplicate 1-ml samplesofurinewere

incu-batedat370C with and without addition of 0.04 ml of a sterile 5% solution ofpotassium nitrate. Samples were tested immediatelyfor the presenceofnitrite, thenincubatedat37°Cand retested after2,4,6,and 24h. These patientswerecomparabletothegeneral

testgroup. OrganismsculturedweresixEscherichia

coli andoneeach ofKlebsiellapneumoniae,Proteus

mirabilis, Pseudomonas aeruginosa, and

Staphylo-coccus aureus.

Urinenitrate concentration and culture. Urine

nitrate concentration wasmeasured in paired urine

specimens from nine healthy laboratory personnel,

four male and five female.Subjectsdidnotaltertheir normal activities. Onespecimenwascollectedatthe firstvoiding aftera night's sleep and the second at midafternoon of the same day. All specimens

con-tainedless than 103bacteriaper mlby quantitative

culture. Urine nitrate and nitrite were determined

simultaneouslyin allspecimens by utilizingthe

spec-trophotometricmethodofWegner(27).Asample

con-taining106 E.coli in sterilesaline, previouslyisolated

fromanitrite-positiveurinespecimen,wasadded to 1 ml ofeachspecimen,and the urinewasincubatedat

37°C.Nitritetestingwasperformed by dipstickat

15-minintervals untilpositive.

Dipstick culture. A commercially available

dip-stickwasused(Microstix, Ames).Astripwasdipped in each urinespecimen, placedinaplastic pouch,and observed during incubation at 37°C. After 18 h the

numberof bacteriawasestimatedbycomparing the padswithpicturesprovidedwith thestrips.

RESULTS

Quantitative urine culture. Two hundred

nineteen specimens were processed, and 52

(24%) contained more than 105 bacteriaperml.

Gram-negative bacteria, especially E. coli,

pre-dominated. Specimens containing greater than 105 bacteriapermlweretermedculturepositive;

those containing less than 105 bacteria per ml were termed culturenegative.

Simplified microscopy of urine. Two

hundred ten specimens were examined

micro-scopically. Microscopywas falsely positive in7

of 159 culture-negative specimens (4%) and falsely negativein2of51 culture-positive

spec-imens (4%) (Table 1). Cultures from theseven

false-positivespecimens revealedtwospecimens

with 104 to 105 E. coli, one with 104 to 105S. epidermidis, and four with less than 103

orga-nisms per ml. One falsely negative specimen contained E. coliand the otherdiphtheroids.

Nitritetestwithincubation.Two hundred eighteen specimensweretested for nitrite after

0,2, 4, 6, and24hofincubationat37°C (Table 2). Although chemical formulationwasdifferent,

results were similar with both nitrite dipstick

products. Of the 218, 51 specimens contained

over 105bacteria perml, but

only

18were

ini-tially nitritepositive (35%). Maximum positivity

occurred after4hofincubation, when23 (45%)

were positive. After incubation for 24 h, only

eight of thespecimensremained nitritepositive

(16%). There was no correlation between the

organismcultured and nitritepositivity.

Eighty-five urine specimens contained over

103but lessthan 105organismsperml;twowere

initially nitritepositive, but after incubation for

24h 18 werepositive.

Eighty-two urine specimens contained less

than103 organismsperml; three specimenswere

positiveinitiallyandremainedpositive

through-outthe entire incubation. Discoloration ofthese

specimensmade interpretation of thetest

diffi-cult. After incubation for 24 h, six additional

specimens converted from nitrite negative to

positive.

Nitritetestwithsupplemental nitrate. Of

the 10 culture-positive, nitrite-negative

speci-mens to which nitrate was added, 8 became

nitrite positive after incubationfor 6 hand the

other 2 (Proteusmirabilis and S. aureus) after

24 h. One control specimen without added

ni-trate became nitrite positive (E. coli) after in-cubation for 24 h.

Urine nitrateconcentration and culture.

Nitrateconcentrationwashigher(P < 0.025 by

pairedttest) infirstmorningthan inafternoon

urine specimens from nine healthy laboratory

personnel (Fig. 2).Nonitritewasdetectedin any

specimen. After addition of 106 E. coliperml,

five firstmorning and three afternoon

specimens

first became definitely nitrite positive after

in-cubation for2.25h; after incubation for 2.75h,

all

specimens

werenitritepositive.

Dipstick culture. Forty-five urine

specimens

containingmorethan105organismspermlwere

tested by dipstick culture; dipsticks from 39

(87%)wereinterpretedasindicating105 or more

bacteriapermlafter incubation for 18 h (Table 3). Onehundredsixty-seven specimens

contain-inglessthan 105 organismsper ml were tested;

dipsticks from three (2%) of these were

inter-pretedasindicating105 or morebacteriaper ml

afterincubation for 18 h.Although positive

dip-sticks could be identified after incubation for 4

to 6 h, estimation of the number of organisms was not possible before 18 h, so that dipstick culture was not useful forrapidscreening.

DISCUSSION

Quantitative urine cultures are valuable in

detectingbacteriuriaif properlycollected

speci-mensare studied (1, 9). Patients whose voided, cleancatch,midstream urine

specimens

contain over105 bacteria per ml have a highprobability

VOL. 7,1978

on February 7, 2020 by guest

http://jcm.asm.org/

(3)

*

..,

WP,

.

'

A,

.,

b

.,_4.

w

*

S{>

5 *

>

sx;

wP

?;-.,

a...

..

,

.

,

....

..

;

,8 /qf t

.,.

.

,

<

.s,

S::

.:. ':

; i;

*e'k'M

...,.~~ ~ ~ ~ ~ ~ ~ ~ ~ ~

:~ ~ ~ ~ ~4

* .U2...

:

X.

am~~~c

...,

< f4i:. s p

on February 7, 2020 by guest

http://jcm.asm.org/

(4)

VOL. 7, 1978

TABLE 1. Urine microscopy and urine culture

Bacterial No.of cul- Microscopyresults(%)

concn (orga- tures

exam-nism/ml)in ined Positive Negative samplestested

>105 51 49 (96) 2(4)

<105 159 7 (4) 152(96)

TABLE 2. Presenceofnitritein urinesamplesof varyingbacterial concentrations

No.(%)ofnitrite-positivesamples byconcn: Incubation(h)

>105/ml 103to105/ml <103/ml

0 18(35) 2(2) 3(4)

4 23 (45) 4(5) 3 (4)

24 8 (16) 18(21) 9(11)

ofhavingsignificant bacteriuria,whereas those

whosespecimens containfewer than 105bacteria

perml donot.However, since24hisrequiredto

obtain results from quantitative urine culture,

morerapidmethods have beensoughttoscreen

for the presence ofurinarybacteria.

The results of urine microscopy have been

reported using stained, unspun urine (3, 9),

stained sediment (13, 23), or both (16, 26, 28).

Robins etal. studied unstained urineutilizinga

counting chamber (22). None of these reports

states exactcriteria forpositive microscopy. Us-ing thesemethods,processingis

time-consuming

and may complicate interpretation, since it is

more difficult to separate bacteria from

amor-phous debris in sediment, and staining proce-dures introduce the

possibility

of false

negative

results, from lost bacteria if the specimen is

inadequately

fixedon the

slide,

and false

posi-tive

results,

from artifacts or contaminated

stains. Wehavedemonstrated thatanew

sim-plified urinemicroscopytechnique, omittingthe

useofstains, centrifuges,orcounting chambers,

is remarkably accurate forrapid screening for

bacteriuria in random urine specimens. This

method might also be useful for determining

responsetotherapyandcouldreplacethemore

expensivequantitative urine cultureforthis

pur-pose.

Nitrite is not normally detected in human

urine,

but bacteria have enzymes capable of reducing urinary nitrate tonitrite. Cruickshank

andMoyeswerethe firsttoreporturinary nitrite

as amarker forurinary bacteria, butfound

ni-BACTERIURIA MICROSCOPY 289

15

:~10 0

54

0

4, 4J

z

0.

FIG. 2. Urinary nitrate concentration in paired

first morning voidedurine(0)and afternoon voided

urine(0)from nine healthy individuals.

tritewas notpresentinallinfectedurine

speci-mens(5). This may be because notall bacteria

reduce nitrate, e.g., enterococci, acinetobacter,

oralcaligenes. Sleigh improvedthe accuracyof

the testby adding nitrate and incubating

sam-plesfor 6h beforetesting for nitrite (25). Czer-winski et al. first observed that nitrite testing

was more sensitive in infected first morning

voided specimensthan inrandom voided

speci-mens (6). This difference in nitrite sensitivity

depending upon timing ofspecimen

collection

has been confirmed by Craig et al. (4) and

ex-plains why some investigators have had good

results (21) and others poor results (20) when

using urinarynitriteas amarker for bacteriuria.

Inanefforttoincrease thesensitivityof nitrite testinginrandom voided urine

specimens,

which

are more practical for screening, to the level

reportedinfirst morning voided specimens, we

FIG. 1. Uncentrifuged, unstained urine (x1000) containing (A) E. coli and (B) Streptococcus faecalis.

Organisms aremarkedbyopen arrows, amorphous crystallinematerial by closedarrows.Notice multiple

planesoffocusinwetpreparationsand smoothnessofbacterial outline.Crystallinematerial ismorerefractile

thanbacteria.Note yeast in bottomphotograph.

on February 7, 2020 by guest

http://jcm.asm.org/

(5)

290 BARBIN, THORLEY, AND REINARZ

TABLE 3. Dipstickculturetechnique

Bacterial Interpreted dipstickresults

concn(orga- (:

nisms/ml) in

samples >105/ml <105/ml

tested

>105 39(87) 6(13)

<10-5 3(2) 164(98)

utilizedtwotechniques, incubationat37°Cand

additionofnitratesubstrate.Incubation of

spec-imensat 37°C failed to increase the sensitivity

ofthe nitrite test, suggesting that, contrary to

themanufacturers'suggestion(package inserts),

the difference insensitivity between random and first morning specimens isnot due to duration of incubation ofurine within thebladder. Sup-plementalnitrate convertedculture-positive, ni-trite-negative specimens to positive during in-cubation at 37°C, confirming Sleigh's results (25). Although nitrite testing may be useful in selected patients using first morning voided

specimens (14), addition ofnitrate followed by

incubation of random voidedspecimens is

cum-bersome and negates the usefulness forofficeor

hospital bacteriuria screening.

Craig et al. (4) and Moffat et al. (20) have reported good results from using the dipstick

culture test toscreenforbacteriuria. While our

results were similar, our purpose in using this

test was to determine whether it might be

adaptedforrapid screening. The manufacturer

recommendsinterpretationofthedipstickafter incubation for 12 to 18 h, but we hoped that

interpretation mightbe possible after ashorter incubation. This was not possible. The time required forincubation limits the usefulness of

the dipstick culture test for rapid bacteriuria screening.

In summary, we found neither the urinary

nitrite test nor the dipstick culture technique

useful forrapid detection ofsignificant bacteri-uria in random urinespecimens. However,

sim-plifiedmicroscopic detection ofbacteria in

un-stained, uncentrifugedurineaccuratelydetected bacteriuria. We suggest that when beginning microscopy, and periodically thereafter, a labo-ratory should compare the results ofurine mi-croscopy with simultaneous quantitative

cul-tures.Additionally, includingknownpositivesin

microscopy studies,suchasfrozendilutedbroth

cultures or quantities of positive urine, would serveas acontroltoensurethat bacteriacanbe

recognized. We suggest that both E. coli and enterococci beused,since cocciare moredifficult toidentifythanrods. Sincemicroscopyisquick

and inexpensive and canbe appliedto random,

clean catch, midstreamurine samples, thenew

technique of simplified urine microscopy ap-pears to be the best screening technique for bacteriuria currently available.

ACKNOWLEDGMENTS

The Clinical Microbiology Laboratory, supervised by Elaine C.Tacquard, helpfully provided urine specimens. Dip-sticks were kindlydonated by Warner-Chilcott Laboratories and the AmesCo. We express appreciation to NancyMillsaps for technicalassistance, to Susan Menotti, Valerie Williams, and Bonnie Rhew for manuscript preparation, and to Michael J.Megnafor expertphotographicassistance.

LITERATURE CITED

1. Boshell, B.R., and J. P.Sanford. 1958. Ascreening

method for the evaluation ofurinarytractinfections in female patients without catheterization. Ann. Intern. Med. 48:1040-1045.

2.Brenner,B.,and V. E.Gilbert.1963.Elevated levelsof

lacticdehydrogenase, glutamicoxalacetic transaminase,

and catalase in infected urine. Am. J. Med. Sci. 245:31-42.

3. Bulger,R.J.,and W. W.Kirby.1963.Simpletestsfor

significantbacteriuria. Arch.Intern.Med. 112:742-746. 4. Craig, W.A., C. M. Kunin, and J. DeGroot. 1973. Evaluation of new urinary tract infection screening

devices.Appl. Microbiol. 26:196-201.

5. Cruickshank, J.,and J. M.Moyes.1914.The presence andsignificance of nitrates inthe urine. Br. Med. J. 2:712-713.

6. Czerwinski,A.W.,R.G.Wilkerson,J. A.Merrill,B.

Braden,and J. P. Colmore. 1971. Further evaluation of the Griess testtodetectsignificantbacteriuria. Am. J. Obstet.Gynecol.110:677-681.

7. Edwards, B.,R.H.R.White,H.Maxted,I.Deverill, andP. A. White. 1975.Screeningmethods forcovert bacteriuria inschoolgirls.Br.Med. J. 1:463-467. 8. Gillenwater,J. Y. 1975.Diagnosisofurinarytract

infec-tion: appraisal ofdiagnostic procedures. Kidney Int. 8:S3-S1 1.

9. Kass,E.H. 1956.Asymptomaticinfections of theurinary

tract.Trans. Assoc. Am.Physicians69:56-63. 10. Kass,E. H. 1962. Prevention ofapparentlynon-infectious

diseasebydetection and treatment of infections of the

urinarytract.J. Chronic Dis. 15:665-673.

11. Kass, E. H. 1962. Pyelonephritis andbacteriuria. Ann. Intern. Med. 56:46-53.

12. Kincaid-Smith, P.,M.Bullen,J.Mills,V.Fussell,N.

Huston, andF.Goon.1964.Thereliabilityof screen-ingtestsfor bacteriuria in pregnancy. Lancet ii:61-62. 13. Kunin,C. M. 1961. Thequantitativesignificance of bac-teria visualized in the unstainedurinary sediment. N.

Engl.J.Med. 265:589-592.

14. Kunin,C.M.,andJ.DeGroot.1975.Self-screeningfor significantbacteriuria: evaluation ofadip-strip combi-nation nitrite/culture test. J. Am. Med. Assoc. 231:1349-1353.

15. Kunin,C.M.,E.Zacha,and A. J.Paquin.1962.Urinary

tractinfectionsin schoolchildren.I. Prevalence of bac-teriuria and associated urologic findings. N. Engl. J. Med. 266:1287-1296.

16. Lewis,J.F., and J. Alexander. 1976. Microscopyof stained urinesmears todetermine theneed for

quanti-tative culture. J.Clin. Microbiol. 4:372-374.

17. Lie,J. T. 1968. Evaluation ofanitrite test-kit(stat-test)

forthe detection ofsignificantbacteriuria. J. Clin. Pa-thol. 21:443-444.

18. McGeachie, J.,and A. C. Kennedy. 1963. Simplified quantitativemethods for bacteriuria andpyuria.J. Clin. Pathol. 16:32-38.

19. Matsaniotis, N.,C.Danelatou-Athanassiadou,C.

Ka-J. CLIN. MICROBIOL.

on February 7, 2020 by guest

http://jcm.asm.org/

(6)

BACTERIURIA MICROSCOPY 291

terelos, P. Hartokalis, and E. Apostolopoulou. 1971.Low urinary glucose concentration: reliable index. of urinary tractinfection. J. Pediatr. 78:851-858. 20. Moffat, C. M., M. R. Britt, and J. P. Burke. 1974.

Evaluation of miniaturetestforbacteriuria using

de-hydrated media and nitrite pads. Appl. Microbiol. 28:95-99.

21. Randolph, M. F., and K. Morris. 1974. Instant screening for bacteriuria in children: analysis ofadipstick. J. Pediatr. 84:246-248.

22. Robins, D.G., R. H. R. White, K. B. Rogers, and M. S. Osman. 1975. Urinemicroscopyas anaidto detec-tionof bacteriuria. Lancet i:476-478.

23. Sacks, T. G., and J. H. Abramson. 1967. Screeningtests forbacteriuria: avalidity study. J. Am. Med. Assoc.

201:1-4.

24. Savage, W. E., S. N. Hajj, and E. H. Kass. 1967.

Demographicandprognosticcharacteristics of bacteri-uria inpregnancy.Medicine 46:385-407.

25. Sleigh, J. D. 1965. Detection of bacteriuria bya modifi-cation of the nitritetest.Br. Med.J.1:765-767. 26.Thysell,H. 1969.Evaluationof chemical and

microscop-ical methods for massdetection ofbacteriuria. Acta Med.Scand. 186:393-400.

27. Wegner, T. N. 1972. Simple and sensitive procedure for determining nitrate and nitrite in mixtures of biological fluids. J.Dairy Sci. 55:642-644.

28. Zinner, S. H., and E. H. Kass. 1971. Long-term (10to14 years) followupofbacteriuria ofpregnancy.N.Engl. J. Med. 285:820-824.

VOL. 7, 1978