APpLnE MiconmooGy, July 1969, p. 57-63Copyright © 1969 Ameaikan Society for Micobiokwr

VoL 18, No. 1Pried in U.S.A.

Quantitative Urine Culture by Surface Drop MethodJOHANNA C. LORRIER AND HANS A. VALKENBURG

Department of Micdial Diseases, University Hospital, Leiden, The Netherlands

Received for publication 18 April 1969

A simple drop method for quantitative urine culture was developed and tested incomparison with methods for bacterial urinary counts. In a group of 452

urines all yielding Escherchia coil, 74 showed counts of more than 100,000 colonies,and 16 showed counts betwen 10,000 and 100,000 colonies per ml. Of these 90urines, 3 of the 16 in the doubtful group were false negative with the drop method.Another 7 urines in the total number of 452 showed discrepancies, but, because allwould have been repeated, the second urine sample would have corrected the pri-mary result The ease and clanlines of the method render it a suitable techniquefor n normal and patient populations. The method was applied on a popu-

lation sample of 1,330 pso from whom unwashed mid-stream urine was collectedand yielded figures comparable with results published in the literature. The methoddiscriminates between steps of 10-fold difference, whereas more accurate countmethods show a standard error of +25% and are reliable in a double dilutionseres.

Recently Guttman and Naylor (4) and Cohenand Kass (3) described a simple method forquantitative urine culture by the use of a micro-scopical slide covered with culture medium, whichis dipped into urine samples. This method provedto be reliable, easy, and practical, becuse theslide can be sent in by mail. One of the advantagesof the dip-slide method is that exact counting isunnecessary because the inoculated slide is com-pared with a standard photograph of a series ofdip slides prepared from 10-fold serial dilutions ofa culture of Escherichia colt. One other advantageis that both sides of the slide can be coated with adifferent culture medium, which enables theevaluation of contamination of the urine specimenwhen a nutrient agar is used in comparison witha selective culture medium. The dip-inoculummethod is ialy suitable for epidemiologicalinvestigations on urinary-tract infections and as ascreening procedure when delay in delivery ofurine specimens to the laboratory is likely, andtherefore particularly useful in general practiceand antenatal clinics (4).Based on the same principle, we developed a

method for laboratory use involving standardpetri dishes and selective media like Endo Agarand MacConkey Agar. The method consists ofdropping drops of urine of undetermined sizeonto the dried surface of a culture medium and,after incubating for 18 hr, comparing the overallgrowth pattern with standard growth pattern of10-fold serial dilutions of various microorganismson either Endo Agar or MacConkey Agar.

MATERIAIS AND METHODSUrine celia Urines were collected from 743

female patients attending the medical outpatientdepartment by the use of a slightly modified techniquefor collecting urines as described by Kass (7). Prepa-ration of the patients consists of washing the vulvafour times with sterile cotton ga sponges soakedin sterile buffered saline by an ing nurse. Next amid-stream specimen is obtained in a sterile, wide-mouthed, screw-cap container. The urine is deliveredto the laboratory within 1 hr of the time of voiding,refrigerated promptly, and plated within 2 hr of thetime of voiding. Sterile buffered saline was used be-cause the washing procedure proved to be less painfuland the yield of microorganisms would not be in-fluenced by the presence of soap or oter antisepticfluids (14).

Urines were also coleted during a random pop-ulation survey on r c dies from 1,330people, both ales and females, aged 45 years andover. Owing to local c uan cleaning of theexternal genitalia could not be performed, and only amid-stream speimen was collated after carefulwritten and oral instruction to the respondent. Urinesamples were collected in a l pared plasticcups covered with snap caps and cultured in generalwithin 1 hr after collection, occasiolly within 2 hr,and without refrigeration.

Culture media A comparison was made betweenbacterial counts on 5% sheep blood-agar plates,Endo Agar (Oxoid C.M.37), MacConkey Agar(Oxoid C.M.7), and Nutrient Agar.

Culture ntods. Nutrient Broth cultures (16 to18 hr) of Streptococcus faecalis, Staphylococcusaureus, and Micrococcus epidermidis and similarcultures of such gram-negative organisms as E. coli

57

on July 12, 2018 by guesthttp://aem

.asm.org/

Dow

nloaded from

LORRIER AND VALKENBURG

and Pseudomonas aeruginosa, and Klebsiella, Citro-bacter, and Proteus varieties in peptone broth wereused. Of these cultures serial 10-fold dilutions insterile buffered saline (pH 7.3 to 7.5) were made.

Duplicate determinations of both urines and bac-terial cultures were done by applying the pour-platemethod (8, 10, 13, 15) in Nutrient Agar and the sur-face viable count method on Endo Agar and Mac-Conkey Agar.Two methods of inoculation were employed.

(i) Seeding 0.1 ml of urine or culture dilution is per-formed on the well-dried surface (12, 16). The inocu-lum is distributed with a stainless-steel spreader.(ii) Inoculation with a calibrated bacteriological loop(2, 5, 6) is done on a similarly well-dried culture plate.

Technique of surface drop method. Culture platesto be inoculated should be well dried shortly beforeuse. By means of a noncalibrated sterile capillary pipetone or more drops of urine or bacterial culture aredropped onto the surface of the horizontally placedplate. The plates are left uncovered and exposed tothe air allowing the drops to dry (approximately 0.5hr). After overnight incubation for approximately18 hr, the results are read in comparison with a seriesof standard photographs obtained from 10-fold serialdilutions of known bacterial concentrations. Whensingle drops are applied, a 4-inch (10-cm) standardpetri dish can be inoculated with four different speci-mens.The aspect of bacterial growth is judged by the

edge and the central part (Fig. 1). Grading of the edge

FIG. 1. Aspect of bacterial growth on MacConkeyAgarfor various counts.

is done in five categories, in which grade 0 means nodetectable edge and in which grades 3 and 4 are com-patible with a bacterial count of 105 or more micro-organisms per ml. Occasionally a bacterial count of105 shows an edge grade 2.

In most instances an edge grade 2 is found inbacterial growth which yields 104 up to 105 micro-organisms per ml. Occasionally an edge grade 1 isfound for counts in the range between 104 and 105microorganisms per ml and could therefore erro-neously be judged as a count under 104 microorganismsper ml. Counts under 103 microorganisms per mlinvariably show an edge grade 0.The central growth pattern varies between loose,

separate colonies and massive growth without holes.In the range between 104 and 105 microorganisms perml, confluent growth of two or three colonies is seenwith large spaces between colonies. Increasing countsshow massive growth with small spaces in the rangeof 105 bacteria per ml. Counts over 106 bacteria perml are solid. The combination of the central growthpattern and the grade of the edge is more decisivefor a final count than either one alone.

Because Fig. 1 was obtained from cultures on Mac-Conkey Agar, the evaluation of bacterial counts onEndo Agar involved the preparation of another seriesof standard photographs, since spreading of the dropon Endo Agar is less pronounced, a relatively highercolony count would be found when Fig. 1 is used asa standard for cultures on Endo Agar.The size of the drop has hardly any effect on the

reading (Fig. 2). In this example one to four dropswere dropped onto one MacConkey Agar. This wastested for different bacterial concentrations, differentculture media, and different species of bacteria. Thegrowth pattern of different species of bacteria wasshown to be essentially the same with regard to thecounting method (Fig. 3).

RESULTS

The reliability of three count methods wasevaluated for three different technicians (Table 1).The mean of a duplicate count with the pour-plate method as performed by technician A wasconsidered to represent the exact number of bac-teria per ml. Three bacteriological loops made of1-mm platinum were standardized against themean counts of the pour plate method of tech-nician A. All experiments were performed withina period of 3 weeks, and therefore day-to-dayvariations are included. The standard error of thecounting procedures is expressed in percentages ofthe mean of duplicate counts.

Within methods, observer A showed the leasterror, but observer C was still within the limitsaccepted for a biological procedure. Betweenmethods, errors were in the same range and stillnot exceeding ±30% for any observer. As ex-pected the error between observers was leastwhen comparing observer A versus observer B,but again not exceeding ±25% for the other

58 APPL. MICROBIOL.

on July 12, 2018 by guesthttp://aem

.asm.org/

Dow

nloaded from

DROP METHOD URINE CULTURE

FIG. 2. Effect ofdrop size on bacterial growth pattern. (A) E. coli culture containing 5.6 X 10 microorganismsper ml. I, II, 111, and IV: 1, 2, 3, and 4 drops, respectively. (B) Streptococcus faecalis culture containing 1.2 X10; microorganisms per ml. I, II, III, and IV: 1, 2, 3, and 4 drops, respectively.

Bacteriat growth pattern of different species- n surface drop.methsd

minbero f . Strept.use/rn Ecol. Eterobacter faeci

0'10

'f3um0a g

5*@

FIG. 3. Bacterial growth pattern of different species

of bacteria in surface drop method.

compansons. This would mean that, for any

experiment applied here, a one tube or 100%cdifference in a series of double dilutions would besignificant at the 5 %- level. This is about as

accurate as can be anticipated for, e.g., a serologi-cal procedure.

To determine the influence of various culturemedia, 291 urines of outpatients were culturedon both blood-agar plates and MacConkey Agar(Table 2). The calibrated bacteriological loop(standard loop) method was used. All determina-tions were done in duplicate. The consecutiveurine samples were collected as described. Con-tamination was defined as the presence of morethan one species of bacteria per urine culture,whereas cultures of one species only were con-sidered to be pure cultures. In only 4.4% of thecultures on blood plates, no growth was seen incomparison with 55.0%,O negative cultures onMacConkey Agar. Nearly 40% of the urinesharbored one species of gram-positive micro-organisms when cultured on blood plates (staphy-lococci or diphtheroids) in the range between103 and 105 microorganisms per ml. For MacCon-key Agar, this figure was 4.8%. Concordance wasmet for pure cultures of gram-negative bacteriafor counts of 10 or more microorganisms perml. Including contaminations, 30.5%e was in therange between 104 and 105 microorganisms perml when cultured on blood plates, whereas thispercentage was 4.1 on MacConkey Agar. Afterexclusion of these obvious contaminations, thesepercentages were 9.6 and 2.4, respectively.Of 30 different cultures of E. colt, 10-fold serial

dilutions were prepared. Of each dilution, aduplicate count with a standard loop was per-formed, and duplicate drops were dropped onto

59VoL. 18, 1969

on July 12, 2018 by guesthttp://aem

.asm.org/

Dow

nloaded from

LORRIER AND VALKENBURG

TABLE 1. Standard error ofbacterial counting procedures (18-hr culture of E. coli)

Comparison No. of Observer A No. of Observer B No. of Observer Cspecimens specimens specimens

Within methodsPour plate 20 ± 8 NDa NDStandard loop 20 4:11 19 ±18 17 4300.1 ml on surface 20 4:13 19 =16 20 4-25

Between methodsPour plate vs. loop 20 4121 ND NDPour plate vs. 0.1 ml 20 ±10 ND NDLoop vs. 0.1 ml 20 :4:22 19 :4:18 17 :4:30

Between observers A vs. B A vs. C B vs. C

Loop vs. loop 19 413 17 4119 16 ±240.1 ml vs. 0.1 ml 19 415 17 ±22 16 ±23Loop vs. 0.1 ml 19 ±21 17 ±24 16 4220.1 ml vs. loop 19 ±19 17 ±25 16 ±23

a ND = not done, as mean of duplicate pour plate count of observer A was considered to representexact number of bacteria per ml.

TABLE 2. Distribution

Specimen

Culture of one speciesGram-negativeGram-positive

Contamination (morethan one species)

Both gram-negativeOne gram-positive/one gram-negative

Both gram-positive

Totals (contaminationsincluded)

Totals (contaminationsexcluded)

of bacterial counts in urine specimens of 291 outpatients cultured on blood-agarplates (BAP) and MacConkey Agar (McC)

Counts per milliliter

Negative 10O-9.104 104-9.104 10' or more Totals

BAP %

4.4

28.7

McC % BAP % McC %

55.0

59.1

6.232.3

1.0

2.4

41.9

38.5

11.04.8

0.3

2.1

18.2

15.8

BAP %

2.47.2

0.74.1

16.1

30.5

9.6

McC % BAP % McC % BAP % McC %

2.4

0.30.3

1.1

4.1

2.4

21.0 21.0 29.62.1 1.7 41.6

23.1

23.1

22.7

0.75.2

18.5

99.9

22.7 99.9

both Endo Agar and MacConkey Agar. Differentplates were used for duplicate determinations.Representative examples of bacterial growth inthe surface-drop method were selected for Fig. 1.The distribution of the various edges observed on

MacConkey Agar and Endo Agar is tabulatedversus bacterial counts in Table 3. In general, onEndo Agar, slightly higher grades were observedbecause the drop spreads less on the surface ofthis medium. Therefore the standard photographs(Fig. 1) cannot be applied as such to Endo Agar.

Urines (452) yielding E. coli obtained fromfemale outpatients were counted in duplicatewith both the standard loop method and thesurface drop method (Table 4). As judged fromthe standard loop method, yields of less than 104microorganisms per ml can be decided to betruly negative, counts between 104 and 9 X 104bacteria per ml to be doubtful positive, andcounts over 105 microorganisms per ml to betruly positive. In a similar way, the five edgegrades were divided, whereas screen grades 0 and

34.46.5

0.60.3

3.2

100.0

100.0

60 APPL. MICROBIOL.

on July 12, 2018 by guesthttp://aem

.asm.org/

Dow

nloaded from

DROP METHOD URINE CULTURE

1 were considered to be negative, grade 2 to bedoubtful positive, and grades 3 and 4 to be posi-tive. Four or 0.9% of the urines gave highervalues on screen in comparison with the standardloop count. On the other hand, 3 or 0.7% of theurines were negative on screen but yielded a bac-terial count between 10" and 105 mnicroorganismsper ml. These three specimens could therefore be

TABLE 3. Comparison of edge grading betweenMacConkey Agar and Endo Agar (30 different

cultures of E. coli)

Grades of edge as observed on0,Bacterial coum

Ma ey Aga En Agar

<103 0 010-9. 103 0-1-2 0-1-2-(3)1049. 104 1-2-(4) 3-4105.9.10 2-4 410'+ 4 4

£ Grades observed occasionally are in paren-theses.

i Expressed as microorganisms per milliliter.

considered to be false negative. As a consequence,a subsequent urine specimen would not have beencollected. Another three specimens were doubtfulon screen but positive on count. These urinesamples would anyhow have been repeated.Applying the surface drop method in total 10, or

2.3%, of the 452 specimens would have beenjudged as either false positive or false negative.This remarably small error was mainly due tothe fact that 10-fold dilutions were used, whereaspresunmbly any count method can reliably beperformed in a 2-fold dilution manner. Thisfivefold increase in dilution steps will thereforeconsiderably cut the error.During a population survey on rheumatic

diseas, urines were collected from 670 men and660 women, aged 45 years and over. Four differ-ent communites were involved. In three of these,all people over 45 years of age were included inthe survey; in the fourth community 33% of thepopulation 45 years and older was selected atrandom. In total, 4,000 people were examinedduring 1968. The results of the first 1,330 personsare given in Table 5.Of themen, only 0.8% showed

TABLE 4. Distribution of various bacterial counts among grades ofedge in suface drop method (452urines of outpatients yielding E. coli)

Conts per ml

Edge in grades'< l 103-0 i0'-0' | 10i-1i6 10' or more Desion onTotal

no.No. % NO. % NO. % NO. % NO. %

0 339 319 70.6 20 4.4 Negative1 23 20 4.4 3 0.7 Negative2 18 3 0.7 12 2.6 3 0.7 Doubtful3 12 1 0.2 11 2.4 Positive4 60 17 3.8 43 9.5 Positive

Totals 452 319 70.6 43 9.5 16 3.5 31 6.9 43 9.5

Decision True negative Doubtful True positiveon count

TABLE 5. Asymptomatic bacteriuria in a population sample (MacConkey Agar)

Counts (%o)Culture No.

Negative <106 io'-1i 10'-10' 106-10' >106

MenPure growth 658 93.4 2.1 1.2 0.7 0.5 0.3Contaminations 12 0.5 0.7 0.3 0.3 0.0

WomenPure growth 612 65.9 11.7 3.3 2.3 3.8 5.6Contaminations 48 2.6 1.7 0.9 1.1 1.1

61Vow. 18, 1969

on July 12, 2018 by guesthttp://aem

.asm.org/

Dow

nloaded from

LORRIER AND VALKENBURG

asymptomatic bacteriuria with counts of 105 orhigher and 0.7% had counts between 104 and 105.In the women, these percentages were 9.4 and2.3, respectively. Contamination was observed inurines of 0.6% of the men with counts of 104 orhigher and in 3.1% of the women.

DISCUSSION

Essentially the method described here has thesame advantages as the dip-inoculum method ofMackey and Sandys (11) and the dip-slidemethod described by Guttman and Naylor (4)and Cohen and Kass (3). However, it is some-what more "clean" because contamination ofcontainers for mailing and the runoff of the ex-cess urine on filter paper are avoided. Becausecalibration of the drops is not necessary, com-mercially available capillary loops can be used(even without sterilization, since air contamina-tion is very rare). The use of Fig. 1 as a compara-tive standard allows the inexperienced observerto make sufficiently accurate estimates of thebacterial count. Following the principle thatcounts of 104 or more should be repeated by ob-taining a second urine sample within 1 week afterthe first collection will eliminate false positiveresults. Of 452 urines, only 0.7% were negative"on screen") as judged by the edge of the dropand yielded counts between 104 and 105 micro-organisms per ml as determined by the standardloop method (Table 4). Controversially 90 urinespecimens yielded counts of 104 or more bacteriaper ml, and counts should have been repeated in asubsequent sample. When judged on screen only,3 of these 90 samples, or 3.3 %, would not havebeen repeated. It should be realized, however,that grading is done by both the edge and thecentral part of bacterial growth. This rendersmore reliable results.The use of selective media has a distinct in-

fluence on the number of "contaminations"(Table 2), expressed on one hand as counts be-tween 104 and 105 microorganisms per ml and onthe other hand as the low yield of negative cul-tures when blood-agar plates are used. Evenwhen obvious contaminations are excluded about10% of the urines yield counts between 104 and105 microorganisms per ml when cultured onblood-agar plates in comparison with less than3% on MacConkey Agar. The figures of 10 and3%0 are in good agreement with those of E. H.Kass (personal communication) and Brumfitt etal. (1). Obvious contaminations containing morethan one species of bacteria did not occur incounts of 105 or more when urines were collectedunder strict conditions. When only mid-streamurines were collected in elderly sometimes dis-

abled persons, 2.2% of the urines were obviouslycontaminated in counts over 105 microorganismsper ml in women and 0.3% in men (Table 5).These low figures are most likely the result of theapplication of MacConkey Agar and Endo Agarand are comparable with the results of Leighand Williams (9). Preliminary results in thisseries of random people show that of those hav-ing positive cultures (105 or more colonies) 80%were reproducible in a second urine sample col-lected within 1 week of the first urine sample.All of those being obviously contaminated andhaving counts of 104 or more colonies were simi-larly contaminated when repeated. Of the positiveurines, 80% from this population sample har-bored E. coli, and, between 10 and 14%, Entero-bacter species. It therefore seems likely that wash-ing of the external genitalia is not strictlynecessary, which facilitates the procedure consid-erably. Further details of this survey as well asa comparison with other culture media will bepublished later.

ACKNOWLEDGMENTS

We are indebted to R. J. Djajadiningrat for providing urinespecimens of outpatients, to L. de Bruin, H. C. M. Maas,G. Tromp, and M. A. Nagtegaal-'t Hart for performing the cul-tures and tests on reliability, and to the people of the villagesHaaksbergen, Oud-Vossemeer, Sluis, and Vleuten for their kindcooperation.

LITERATURE CITED

1. Brumfitt, W., and A. Percival. 1964. Pathogenesis and labora-tory diagnosis of non-tuberculous urinary tract infection: Areview. J. Clin. Path. (London) 17:482-491.

2. Cattell, W. R., and M. J. Lefford. 1963. Bacteriological exami-nation of urine. Brit. Med. J. 1:97-100.

3. Cohen, S. N., and E. H. Kass. 1967. A simple method forquantitative urine culture. N. Engl. J. Med. 277:176-180.

4. Guttmann, D., and G. R. E. Naylor. 1967. Dip-slide: An aid toquantitative urine culture in general practice. Brit. Med. J.3:343-345.

5. Guttmann, D., and E. J. Stokes. 1963. Diagnosis of urinaryinfection. Comparison of a pour-plate counting method witha routine method. Brit. Med. J. 1:1384-1387.

6. Hoeprich, P. D. 1960. Culture of the urine. J. Lab. Clin.Med. 56:899-907.

7. Kass, E. H. 1965. A cooperative study of bacteriuria in preg-nancy. Technique for performing quantitative bacterial cul-ture of urine. Technical Bulletin A.

8. Kass, E. H., and M. Finland. 1956. Asymptomatic infectionsof the urinary tract. Trans. Ass. Amer. Physicians Philadel-phia 69:56-64.

9. Leigh, D. A., and J. D. Williams. 1964. Method for the detec-tion of significant bacteriuria in large groups of patients.J. Clin. Pathol. (London) 17:498-503.

10. MacDonald, R. A., H. Levitin, G. Mallory, and E. H. Kass.1957. Relation between pyelonephritis and bacterial countsin the urine. An autopsy study. N. Engl. J. Med. 256:915-222.

11. Mackey, J. P., and G. H. Sandys. 1965. Laboratory diagnosisof infections of the urinary tract in general practice bymeans of a dip-inoculum transport medium. Brit. Med. J.2:1286-1288.

62 APPL. MICROBIOL.

on July 12, 2018 by guesthttp://aem

.asm.org/

Dow

nloaded from

DROP MEIHOD URINE CULTURE

12. Mamk R. X, and W. H. HaIL 1957. A apid mdbod forPe -u ra Lab.Cln. ed. 49819424

13. Maple, C. D. 191L The fiujency and dicr aof uimar

ulct infecdow xi an wnsmsed St ofwop Ann.

et M 972239.14 Robat,A.P, . .Roiwad R.W.B& 67.Sc

f-act bme kWcokar couma in urnary infecdkmBeL 1.400-403.

LS Sanford, L P., C. , andF. ]L Mao. 1956. E.valuadonfthe -paivewameockine. Am. M. I2 -3"

16. Trw. GQ C. 191. Ba1uria in prqmany. Lancet H:1062-104.

VoL 18, 1969 63

on July 12, 2018 by guesthttp://aem

.asm.org/

Dow

nloaded from