Detection of Bacteriuria in Pediatric Outpatients A New Culture Device Thomas J. Marr, MD, Howard S. Traisman, MD \s=b\ Urine


samples from 1,003 pediatric

examined for bacteriuria with a culture-nitrite strip test, along with microscopic assessment of pyuria and conventional chemical analysis. The latter two methods proved of little value in reliably detecting or excluding substantial bacteriuria. An overall prevalence of 5.2% was shown. Using Microstix Reagent Strips, 92 urine samples with substantial colony counts were detected. Subculture from the strips disclosed that 52 of these involved single pathogens, interpreted as true bacteriuria. When Microstix growth results were compared with the calibrated loop method on 108 randomly selected specimens, the strip gave only two false negatives and one false positive. Bacteriuria screening should be done in high-risk groups and in asymptomatic children as part of the health evaluation. The index of suspicion should be higher in girls, and reliable interpretation of culture results requires attention to proper specimen collection. (Am J Dis Child 129:940-943, 1975) were

urinary tract is often the focus of infection in the febrile child. Even without specific localizing signs, the normal laboratory and clin-


Received for

cepted Dec

publication May 31, 1974;

14. From the Department of western University Medical


Pediatrics, NorthSchool, and Chil-

Hospital, Chicago. Reprint requests to 1325 W Howard St, Evans-

dren's Memorial

ton, IL 60202 (Dr. Marr).

ical investigation will usually disclose the site of involvement. However, co¬ vert or silent bacteriuria is less often suspected, although Kunin et al have demonstrated its prevalence to be 1.1% in schoolgirls and 0.04% in

schoolboys.1 Although

such cases may be dis¬ closed in formal screening programs, their detection in the course of usual pediatrie practice is less likely, since the complexity or cost of urine cul¬ ture, or both precludes its routine use in most settings. Clues may be gained when routine urinalysis shows albu¬ min or abnormal microscopic find¬ ings, but sometimes these findings are not present in children with bac¬ teriuria. Simplified culture devices have long been available, but none seems to have achieved widespread use, possi¬ bly because of such factors as cost, in¬ convenience, or insufficient reliability. With the advent of a new product in¬ corporating both culture and chemical capability of detecting bacteriuria, it was considered desirable to test the capabilities of the device, while at the same time measuring the prevalence of substantial bacteriuria in several types of pediatrie outpatient popu¬ lations.

SUBJECTS AND METHODS Nicrostix Reagent Strips are sterilepackaged plastic strips with one chemical

Downloaded From: by a UQ Library User on 06/14/2015

two dehydrated culture medium pads attached near one end. The chemical reagents, based on the Griess re¬ action, detect the presence of nitrite pro¬ duced by the action of most pathogens on urinary nitrates. Development of a pink color 30 seconds after dipping is consistent with the presence of bacteria in concentra¬ tions of 105 or more per milliliter of urine. One culture medium pad supports the growth of both Gram-positive and Gramnegative organisms, while the other contains an inhibitor of Gram-positive bacteria, thus permitting Gram-negative growth only. The strip is dipped for five seconds into fresh, well-mixed urine, and after reading of the nitrite test at 30 sec¬ onds, it is placed in a plastic pouch and in¬ cubated at 37 C for 18 to 24 hours. Colonies present in the medium are manifest as pink spots resulting from reduction of a tetrazolium dye component in the pads. Semiquantitative counts are made by com¬ parison of growth patterns against a ref¬ erence density chart graded from 101 to 105 or more colonies per milliliter (Figure). Specimens for this study were obtained from 1,003 children (537 boys and 466 girls; age range, 1 month to 16 years) drawn from diverse clinical settings, as shown in the following list: Private pediatrie office practice 515 Routine visits (no illness) 89 Diabetes, periodic follow-up

reagent pad and

Hospital out-patient departments Nephrology clinic Urology clinic Acute care (nonlocalizing fever) Total All

203 162 34


patients (or supervising parents)


of test

strip configuration and color-density

instructed in proper genital cleans¬ ing and delivery of a midstream specimen. Each child had a complete urinalysis con¬ sisting of measurement of specific gravity, microscopic centrifuged sediment exami¬ nation, and a dipstick test for pH, glucose, ketones, occult blood, and protein as soon as possible after*collection, which was usu¬ ally within 15 minutes. All urine samples were also tested for bacteriuria by the Mi¬ were

crostlx method.

Those with substantial bacteriuria (10* colonies total of Gram-negative organisms, or both) were subcultured for bacterial identification in the manner sug¬ gested by the manufacturer—by pressing the culture test area against an agar plate and streaking through the inoculation with a loop. In addition, for purposes of compar¬ ing the validity of Microstix results with a reference method, 108 of the nephrologyurology clinic specimens were simulta¬ neously cultured on eosin-methylene blue (EMB) and blood agar plates for colony counting by the calibrated loop technique. or more


Of the 1,003 urine specimens tested with Microstix, 95 specimens were pos¬ itive, ie, greater than 105 colonies per milliliter on the culture areas, for a yield of 9.5%. When subcultures were taken from the positive growth areas, 40 instances of mixed growth were obtained (42% of the positive Micro¬ stix), 52 showed pure growth (55%),

charts used for


and 3 gave no growth (3%). In the lat¬ ter three cases, failure of transfer may have been related to the fact that the culture pads adhered to the

when the stick was removed from the pouch. The organisms isolated in the 52 in¬ stances of pure culture were Escherichia coli (32 isolates), Proteus spe¬ cies (13 isolates), Pseudomonas (4 isolates), Streptococcus faecalis (2 iso¬ lates), and KlebsielL· (1 isolate). In 11 of these (21%), a positive nitrite reac¬ tion was also obtained. A separate phase of the study was a comparison of the accuracy of the Microstix culture method with a stan¬ dard laboratory technique. Using the

plastic incubation pouch

nephrology-urology hospital out¬ patients, who would be expected to have a higher yield of substantial bacterial growth, 108 randomly se¬ lected urine specimens were simulta¬ neously inoculated with the strip and by calibrated loop on EMB and blood agar plates. There were 42 instances (39%) of substantial plate growth, and Microstix correctly identified 40 of these (sensitivity, 95%). No growth occurred in 66 urine specimens by the laboratory method; Microstix agreed in 65, giving one false positive (speci¬ ficity, 98%). In attempting to correlate bacte-

Downloaded From: by a UQ Library User on 06/14/2015

riuria with microscopie pyuria, a cri¬ terion of four white blood cells (WBCs) per high-power field of centrifuged urine was used, in accord¬ ance with our practice. Of the 95 specimens with bacteriuria by the Mi¬ crostix method, only 29 (30%) were found to have four or more WBCs per high-power field on microscopic exam¬ ination of the sediment. Conversely, there were 81 (71%) instances of pyuria not accompanied by substan¬ tial bacteriuria. As for positive re¬ sults on chemical urinalysis, there were occasional findings of protein or occult blood that did not correlate with presence or absence of bacte¬ riuria. The Table summarizes the principal findings and indicates the prevalence in each group of substantial bacte¬ riuria due to single etiologic agents. The 40 specimens with bacteriuria of mixed origin were attributed to con¬ tamination associated with imperfect specimen collection. Even though in¬ structed, the parent or child did not obtain a clean specimen. When these were repeated with better technique, in no instance was substantial bacte¬ rial growth found. All patients with "true" bacteriuria were followed up with serial cultures or intravenous urography, or both, where indicated.

Of the nine routine patients with pos¬ itive cultures, results of intravenous urography and cystography were nor¬ mal in all.

Summary of Findings in Five Patient Groups



Urinary tract infections are fre¬ quently seen in pediatrie practice and

often are manifestations of obstruc¬ tive uropathy. The signs and symptoms of urinary tract infection may be subtle, espe¬ cially in the infant. Unexplained fe¬ ver and failure to thrive were the most common presenting signs in one series of urinary tract infections.2 In another series of 66 infants with uri¬ nary tract infection in the first month of life, maternal infections, perinatal

anoxia, unsatisfactory weight gain, lethargy, anorexia, and birth weight inappropriate for length of gestation were all correlated with urinary tract infection.3 The prevalence of urinary tract in¬ fection and silent bacteriuria has been discussed by many authors. In the series reported by Edelmann et al, 0.7% of the full-term and 2.9% of the premature infants had bacte¬ riuria.4 Siegel and associates screened 600 infants in office practice and found that 2% of the boys and 7% of the girls had urinary tract infections.5 Savage et al evaluated urine speci¬ mens of schoolgirls entering the pri¬ mary grades and found 1.6% with bacteriuria. Of these, 70% had symp¬ toms, 35% had reflux, and 23% had roentgenographic changes of pyelo¬ nephritis.6 Lincoln and Winberg stud¬ ied 584 infants and found 27 boys and 5 girls with bacteriuria.7 Abbott screened 1,460 neonates, and 1% had bacteriuria.8 Kunin and co-workers have eval¬ uated urine specimens of thousands of schoolchildren since their survey in 1960, and several important facts have been discovered. The prevalence of bacteriuria in school-age children is 1.2% in schoolgirls, and the acquisi¬ tion rate for urinary tract infection is 0.7% per year.9 In studying 9,878 schoolchildren, 1.1% of the girls and 0.04% of the boys had covert bacte¬ riuria. Of 222 college girls, two had bacteriuria, and of 1,116 boys, none had bacteriuria.1 In a seven-year pe-

Routine Diabetes

Nephrology Urology

No. of Patients 515 89

Patients With Pure Growth on Subculture

No. With

No. With Positive



No. (%)



28 11 20 33

9.(1.8) 3 (3.4) 8 (3.9) 32 (19.8)



52 (5.2)

203 162




Fever without

localizing signs Totals





Four WBCs per high-power field. t >105 colonies of bacteria per milliliter of urine. *

riod, 2.2% of schoolgirls will have bac¬ a rate of 0.32% per year.10 Five percent of schoolgirls will ac¬


quire bacteriuria in grade school high school.11 A count of 100,000

or or

colonies per milliliter is the most useful criterion of an infected urine sample,9 but Rabinowitz and Saitz state that a value of 50,000 co¬ lonies per milliliter may also be use¬ ful.12 Investigators have noted the un¬ reliability of routine urinalysis for diagnosis of urinary tract infection. Kunin et al found that only 44.3% of patients with bacteriuria have more than four WBCs per high-power field.13 Only one quarter to two thirds of urine samples with pyuria will have substantial bacteriuria, accord¬ ing to Pryles and Lustik.14 In another series, only 19% of urine samples with 105 colonies per milliliter had pyuria, while 31% of those with 105 or more colonies per milliliter did not have pyuria.15 If a value of WBCs per highpower field is used as a criterion, 87% of these will be associated with bacte¬ riuria.16 The number of WBCs in the urine depends on many factors, in¬ cluding method of obtaining a urine specimen, urine flow rate, inflamma¬ tory response of the kidney, and vol¬ ume of urine in which the sediment is more


Earlier studies of bacteriuria in di¬ abetic patients noted that the preva¬ lence varies between 13% and 34%.18·20 However, two other studies showed no difference between diabetic and nondiabetic patients.21·22 In our series, the prevalence of 3.4% in diabetic pa¬ tients was twice that of the 1.8% in

Downloaded From: by a UQ Library User on 06/14/2015

nondiabetic patients. In view of these facts, it is indeed startling that a recent survey shows that only 50% of office pediatricians base their diagnosis of urinary tract infection on the results of a urine cul¬ ture.23


on a consensus



recommendations, together with


experience, groups at risk for urinary tract infection who should be screened by culture include prema¬ own

tures; full-term newborns who have

perinatal anoxia, unsatisfactory weight gain, lethargy, anorexia, or had

whose mothers have infection; smallor



borns; febrile children without local¬ izing signs; diabetic children every four to six months; and patients with findings suggestive of urinary tract involvement—abdominal pain, dys-

uria, enuresis, pyuria. Among asymptomatic children, it

has been suggested that girls be screened when starting school and again one year later.24 In office prac¬ tice, we favor including urine culture as part of the annual health eval¬ uation, especially in girls. The most recent recommendation of the Ameri¬ can Academy of Pediatrics is that a bacteriuria test be performed rou¬ tinely at age 3, before starting school, and at ages 8 to 9 and 11 to 12, and it points out that new, simple, and inex¬ pensive culture methods make this practical in office practice.25 In this series of 1,003 patients, the overall prevalence of true bacteriuria was 5.2%, with more than three times as many girls affected as boys. As seen from the Table, the incidence of

asymptomatic bacteriuria in well chil¬ dren was consistent with previous re¬ ports in the literature, but was twice this among our diabetic patients. The higher figures shown for those with

known structural or functional abnor¬ malities are not surprising, and con¬ firm the wisdom of frequent monitor¬ ing of such children by culture. The advantage of a properly obtained, clean-voided specimen is also appar¬ ent from this study. Brief discussion of the nitrite test is warranted. The key to obtaining a good yield has long been recognized as related to length of exposure of urinary nitrate to bacterial orga¬ nisms, preferably four hours or more. Thus, the first morning urine sample, after overnight "bladder incubation," is ideal for testing. Since it was not possible to impose these requirements

in our study, it is not surprising that positive nitrite results were obtained on only 11 of the 52 pure culture cases (21%). Kunin has characterized nitrite tests as having very high specificity and only fair sensitivity.24 Stated an¬ other way, positive results are quite important and negative findings are of little value. Seventy-nine percent of those with positive cultures had negative nitrite tests. However, when used in conjunction with culture, as in the Microstix configuration, comple¬ mentary information is provided. There were no false positives in our series. In our hands, the new strip test demonstrated a high degree of reli¬ ability. Personnel without previous experience were readily able to per¬ form the procedure and interpret re¬ sults. Compactness and stability at

temperature provide an obvious advantage. The test is economical (average cost per strip is approxi¬ mately 60 to 75 cents) when compared to a cost of $7 for a hospital labora¬ room

tory culture. A limitation of Microstix

is its inability to speciate organisms other than separating Gram-negative and Gram-positive growth. However, transfer of colonies for subculture and susceptibility testing, when nec¬ essary, is easily accomplished. Growth capability of the culture areas is in¬ hibited by presence of antimicrobials in urine; thus, the condition of pa¬ tients receiving such therapy is best evaluated no earlier than 48 hours af¬ ter use of such agents is discontinued.

This investigation was supported by a grant from the Ames Co, Elkhart, Ind (manufacturer of Microstix Reagent Strips), which also pro¬ vided materials and technical advice.

References 1. Kunin CM, Zadia E, Paguin AJ Jr: Prevalence of bacteriuria and urologic findings: Urinary tract infection in school children. N Engl J Med 266:1287-1296, 1962. 2. Kunin CM: Epidemiology and natural history of urinary tract infection in school age children. Pediatr Clin North Am 18:509-528, 1971. 3. Littlewood JM: Sixty-six infants with urinary tract infection in the first month of life. Arch Dis Child 47:218-226, 1972. 4. Edelmann CM, Ogwo JE, Fine BP, et al: Prevalence of bacteriuria in full term and premature newborn infants. J Pediatr 82:125-129, 1973. 5. Siegel SR, Sokoloff B, Siegel B: Asymptomatic and symptomatic urinary tract infection in infancy. Am J Dis Child 125:45-47, 1973. 6. Savage DCL, Wilson MI, McHeady D: Covert bacteriuria of childhood. Arch Dis Child 48:8\x=req-\ 20, 1973. 7. Lincoln K, Winberg J: Studies of urinary tract infections in infancy and childhood: II. Quantitative estimation of bacteriuria in unselected neonates with special reference to the occurrence of asymptomatic infections. Acta Paediatr Scand 53:307-316, 1964. 8. Abbott GD: A prospective study in 1,460 infants: Neonatal bacteriuria. Br Med J 1:267-269, 1972.

9. Kunin CM, Southall F, Paquin A: Epidemiology of urinary tract infection: A pilot study of 3,057 school children. N Engl J Med 263:817-823, 1960. 10. Kunin CM:

Emergence of bacteriuria and symptomatic urinary tract infection among a population of school girls followed for seven

years. Pediatrics 41:968-976, 1968. 11. Kunin CM: The natural history of recurrent bacteriuria in school girls. N Engl J Med 282:1443-1447, 1970. 12. Rabinowitz NK,Saitz EQ: Simplified screening for bacteriuria in children. Pediatrics 51:128\x=req-\ 132, 1973. 13. Kunin CM, Deutscher R, Paquin A: Urinary tract infection in school children: An epidemiologic, clinical and laboratory study. Medicine 41:91-130, 1962. 14. Pryles CD, Lustik B: Laboratory diagnosis of urinary tract infection. Pediatr Clin North Am 18:233-244, 1971. 15. Andriole VT: Diagnosis of urinary tract infection by urine cultures, in Kaye D (ed): Urinary Tract Infection and Its Management. St. Louis, CV Mosby Co, 1972, pp 42-47. 16. Lahy GR, Levin S: Diagnosis and therapy of urinary tract infection. Med Clin North Am 55:1439-1446, 1971. 17. Stamey TA: Office bacteriology. J Urol

Downloaded From: by a UQ Library User on 06/14/2015

97:926-934, 1967.

18. O'Sullivan DJ, Fitzgerald MG, Meynell MJ, et al: Urinary tract infection: A comparative

study in the diabetic and general populations. Br Med J 1:786-788, 1961. 19. Hunos A, Heonin R: Frequency of bacteriuria in patients with diabetes mellitus. N Engl J Med 261:1213-1216, 1959. 20. Rengarts RT: Asymptomatic bacilluria in 68 diabetic patients. Am J Med Sci 42:159-164, 1960. 21. Etzwiler DD: The incidence of urinary tract infection among juvenile diabetics. JAMA 191:81-83, 1965. 22. Pometta D, Rees SB, Younger D, et al: Asymptomatic bacteriuria in diabetes mellitus. N Engl J Med 276:1118-1121, 1967. 23. Dolan TF, Meyers A: A survey of office management of urinary tract infection in childhood. Pediatrics 52:21-26, 1973. 24. Kunin CM: Detection, Prevention and Management of Urinary Tract Infections. Philadelphia, Lea & Febiger, 1972, p 62. 25. Committee on Standards: Suggested Minimum Schedule for Preventive Health Care of Children and Youth. Evanston, Ill, American Academy of Pediatrics, 1974.

Detection of bacteriuria in pediatric outpatients. A new culture device.

Detection of Bacteriuria in Pediatric Outpatients A New Culture Device Thomas J. Marr, MD, Howard S. Traisman, MD \s=b\ Urine patients samples from...
2MB Sizes 0 Downloads 0 Views