QUANTITATIVE REAGENT JAMES
W. MORROW,
M.D.
M.D.
ROCKENMACHER,
M. L. GOLDMAN, KATHERINE
BY
STRIP INCUBATION
ALAN T. WEBB, MORRIS
URINE CULTURE
M.T.
ARLO’ITI,
PH.D. (A.S.C.P.)
M.T.
(A.S.C.P.)
From the Departments of Urology and Pathology, Rancho Los Amigos Hospita1 and University of Southern California School of Medicine, Downey, California
ABSTRACT - The $lter paper culture-tetraxolium test resuks have been compared with routine hospita1 streak plate cultures in 400 random urine specimens. Pseudomonas cultures tended to impart u color change to the test pads which rendered ínterpretation difficult. While a consistently accurute quantitatíve interpretation of these cultures díd emerge durtng the course of the trial studies, some instructions and degree of observer accommodation and bias, in terms of the manufacturer’s description, was necessary to achíeve an appropriate reading. When the hospita1 laboratory results und culture strip test results are compared in terms of colony counts greater than 1 O4 bacteria per míllíliter and less than 104 bacteria per milliliter, the method was found to be 93 per cent sensítive for the totul count and 83 per cent sensitive fw the gram-negative count with a specijìcity correlation of 96 per cent counts, respectively. fM- the total and g ram-negatíve
There are a number of opportunities in clinical practice in which a simple, inexpensive quantitative bacteriologie procedure could be of value in screening or monitoring for bacteriuria. In the last few years a number of systems have been introduced for this purpose. The improved quality and reliability of these methods suggest that they may play a major role in reducing demands on the bacteriologie resources of the clinical laboratory without sacrificing the reliability necessary in assessing the bacterial status of the urinary tract. These simplified testing procedures can be divided into the chemical screening tests and those based on bacteriologie culture methods. Chemical screening tests have been devised based on: (1) the reduction of nitrate to nitrite (the Griess test);’ (2) residual urinary glucose as related to bacterial action (Uriglox);2 (3) reduction
56
of tetrazolium by bacterial metabolites to the red precipitate triphenylformazan (uroscreen);” and (4) the urinary catalase test.4 In genera1 these tests have been insufficiently reliable for the clinician to accept the results with the required degree of certitude required when establishing the true state of bacteriuria. Most of the culture methods have greater reliability and in some instances permit subculturing, determination of species, and antibacterial sensitivity testing. It is perhaps a logica1 development that a combination of the features from both the chemical and culture methods should be incorporated into a single easily applied procedure. The purpose of this communication is to report the results of such a method and compare these results with those obtained by the routine bacteriologie processing of urine specimens in the laboratory at Rancho Los Amigos Hospital.
UROLOGY
/ JANCARY 1975
/ VOLUMEV.
UUMBER
1
Material and Methods The filter paper culture-tetrazolium test with combined nitrite indicator (Microstix)* is a clear plastic dip stick containing one chemical reagent pad and two culture media containing filter paper pads attached in series. The chemical reagent pad was designed to detect nitrite in the urine (Griess test). In the current study this portion of the test was not correlated with the culture data. The proximal media pad contains an inhibitor to the growth of gram-positive organisms and thus is representative of the total gram-negative count in the specimen. The distal pad supports both gram-negative and gram-positive growth and represents the total bacterial count present in the specimen. In addition the media pads contain colorless tetrazolium which, when reduced to the red precipitate triphenylformazan under bacterial action, produces punctate and hopefully discrete red spots which are indicative of the number of bacteria in the inoculum. In practice the reagent portion of the dip stick is saturated with urine and placed in a plastic envelope. The envelope is collapsed and sealed so as to exclude as much air as possible. The stick and the envelope are incubated overnight at 37“ C. Hospita1 laboratory
procedure
Each specimen of urine was mixed well, and a calibrated loopful was inoculated to each of three media plates; blood agar, phenylethanol agar, and McConkey’s medium agar. If the urine was clear, the loop volume used for streaking the plates was 0.01 ml. ; and if the urine was turbid, the loop volume calibration was 0.001 ml. After incubation for eighteen to twenty-four hours at 35” C. the plates were examined, and each colony type present was counted. The number of colonies per plate times the appropriate dilution factor produced the total bacterial count per milliliter. Bacterial plate count of each species is expressed and reported as (A) no growth, (B) less than 104 bacteria per milliliter, (C) less than 105 bacteria per milliliter, and (D) greater than 105 bacteria per milliliter. Species indentification and sensitivity studies were completed on only the latter two categories (C and D). Species identification of bacteria was made by subculturing each colony type onto differential media tubes in keeping with the modified WB Enteric Differential System1 with biochemical *AmesCompany,Elkhart, Indiana.
t Diagnostic Research Inc., Long Mand City, New York.
UROLOGY
/ JANUARY 1975 / VOLUMEV,
NUMBER 1
reactions being observed after overnight incubation. Antibiotic sensitivity testing for each bacterial species was performed by the Kirby-Bauer method. Urine specimens
The urine specimens for the study were obtained by random selection from among those received from al1 wards and services in the hospital. Any specimen which indicated that the patient donor was receiving antibacterial therapy was excluded from the study. The presence of most antibacterials in the urine is known to alter the filter paper culture test by inducing disproportionately low bacterial growth. Since the urine which provides the inoculum also hydrates the media, the media in turn contains a disproportionately high leve1 of antibacterial activity. Specimens were received at various times throughout the day and represented randomly collected specimens and not first morning specimens. While al1 reasonable effort was made to exclude specimens from patients who were on antibacterial therapy, it is probable that some antibacterial-containing specimens evaded the screening methods both at the ward leve1 and in the laboratory and were entered into the study. Since such unwanted inclusions obtain in the clinical setting wherein such tests are likely to be used, the influence of the occasional antibacterial-containing specimen appears justified in assessing the over-al1 reliability of the method. The dip stick was inoculated and sealed in a plastic envelope at the same time the agar plates were inoculated. The envelope was taped to one of the streak plate Petri dishes and incubated under the same conditions and in the same incubator as the agar plates. Results A total of 400 specimens were included in the study. Since the hospital method of reporting grouped results diHerently from the progressive exponential categories on the filter paper culture test, some restructuring of the data obtained by the dip stick method was necessary to compare ít with the hospita1 categories. Table 1 correlates al1 of the dip stick categories for total bacterial growth with the hospital reporting system. With this alignment of the data the degree of agreement or disagreement is difficult to assess; however, in the area counts greater than 105 the
57
TABLE
1.
obtained -
Comparison of total bacterial counts by streak plate method and filter paper culture-tetrazolium test* Streak Plate (Colonies per Milliliter) NO > 100,000 Growths < 10,000 < 100,000 _
Filter Paper (Colonies per Milliliter)
.J& 12 _7 7 4 3
17 4 8 26 5
2 7 7 6
in parentheses
indicate
Pseudomonas
Negative 10 10” 103 104 > los *Figures < 105.
3
4 11 10 7 15 109
(3) (6) (6) (1) (2) (5)
colonies
correlation does not appear to be optimum since the number of filter paper culture counts in this area is low when compared with the hospita1 group. The numbers in parentheses indicate cultures of Pseudomonas species. These are specifically noted, because during the course of the study the color change that this species tended to impart to the entire culture pad as wel1 as the suppression of definitive punctate spotting caused difficulty in interpreting the dip sticks in Pseudomonas specimens which were consistently more than 100,000 organisms by the streak plate method. As the study progressed the observer who interpreted the results accommodated to the altered appearance of the dip sticks as indicated by the fact that these inconsistencies declined as the series progressed. Despite the manufacturer’s comments on the enclosure that accompanied strains of Pseuthe strips, namely, “occasional domonas aeruginosa may produce gray-green or blue background colors which do not interfere with the reading of the test,” this coloration did in fact seriously interfere with the interpretation of test results.
TABLE
obtained
Comparison of gram-negative counts by streak plate method and the @er paper culture-tetrazolium test*
11.
Filter Paper (Colonies per Milliliter
Streak Plate (Colonies
NO Growths
< 10,000
per Milliliter) -
< 109,000
Negative 10 102
J4& 9 6
13 4 4
3
104 103 > 105
93 4
134 2
s3 2
indicate
Pseudomonas
*Figures < 105.
58
-
in parentheses
1
TABLE
-~ Number of Organisms < los > 10” > 104 Negative
TOTALS *Figures < 105.
15. (1) 102 cultures
Microstix-GramTotal negative
123 52 41 ---184 (16)
400 in parentheses
Hospita1 Laborator> Gram Total Cultures negative .~~.__~ _._
118 30 31 221 (20) .- _~~
158 25 80 137
156 19 ,42 lij1
400
400
~200
Pseudomonas
cultures
indicate
Table 11 correlates the hospita1 laboratory and the filter paper culture test results for gramnegative organisms. An overall pattern of results similar to that for the total counts is seen. By further consolidation of the dip stick results to correspond more closely to the hospita1 reporting system, filter paper culture test categories negative, lol, and 102, were considered as no growth (Table 111). The filter paper culture test category 103 was equated with hospita1 category less than lO*. Filter paper culture test category lO* was considered as hospita1 category less than 105, and the greater than the 105 filter paper culture category was aligned with the greater than 105 laboratory category. Using this convention the correlations of the data could be better appreciated. In the greater than 105 hospita1 culture category, however, there is significantly larger number of positive cultures as compared with the results obtained by the filter paper culture method, whereas in the less than 105 category the reverse is truc. If one further consolidates the data into two groups with colony counts less than lO* or greater than lO*, a significant correlation is seen for total bacterial counts and gram-negative bacterial counts (Table IV). If an additional adjustment is
TABLE IV. Correlation of laboratory and filter paper-culture test in terms of colony counts*
> 190,000 15 (7) 16 (12) 4
Correlation of filter paper culturcatest with hospita1 laboratory results*
111.
tetrazolium
II Number of Organisms
--.-
Microstix GramTotal negative
’ 104
175 225 (16)_
252 (20) _I_
TOTALS
400
490
< 104
*Figures < 105.
UROLOGY
in parentheses
i
148
indicate
Hospita1 Laboratory GrarnTotal negative Cultures .__~..~_ _ _____ ._
I-217
177 223 ----
400
400
Pseudomonas
cultures
183
JANC;ARY1975 / VOLUME~,NUMVIBERl
made for the misinterpretations of Pseudomonas colony counts by plating the appropriate group of Pseudomonas specimens in the greater than 104 categorization, an even more striking correlation between the filter paper culture and the hospita1 laboratory results is evident (Table V). Such a transfer of data cannot be sanctioned in recording the final result of the study; however, the development of a consistent and reliable bias on the part of the observer as the study progressed warrants comment and justifies a projection of what might be expected with the method in more extended use. Using the greater than 104 and less than 104 categories the sensitivity and specificity of the filter paper culture test in relation to the hospita1 laboratory results may be expressed. The sensitivity of the total bacterial growth portion of the test is 93 per cent, and the specificity is 97 per cent. The gram-negative portion of the test shows a sensitivity of 84 per cent and a specificity is 90 per cent. If the transfer of the Pseudomonas results to the appropriate growth category is made, the sensitivity for the total bacterial count portion of the test is 96 per cent and the specificity is also 96 per cent. The gram-negative portion of the test demonstrates the sensitivity is 96 per cent, and the specificity is 90 per cent. In a recent study of the filter paper culture method (total count only) compared with pour plate colony counts, a sensitivity of 90.7 per cent and a specificity of 99.1 per cent was found.5 It was also noted, in these 1,000 cases, that 30 specimens with colony counts of 105 or more by the pour plate method were accompanied by
UROLOGY / JANUARY1975 /
VOLUME V, NUMBER 1
TABLE V. Correlation of laboratory andfilter paperculture test in term of colony counts after correction for Pseudomonas readings
Number of Organisms < 104 >104 TOTALS
Microstix GramTotal negative
Hospita1 Laboratory Total GramCultures negative
191 gN_
168 ~232
183 217
177 223
400
400
400
400
lower counts on the filter paper test. In almost half of these specimens the counts on the filter study were less than 103. This discrepancy between the two methods, however, very rarely occurred with the commonly encountered enteric bacteria associated with urinary tract infection. 7601 E. Imperial Highway Downey, California 90242 (DR. MORROW) References 1. SMITH, L. T., THAYER, W. R., MALTA, E. M., and UTZ, J. P.: Relationship of the Griess nitrate test to bacterial culture in the diagnosis of urinary tract infection, Ann. Intern. Med. 54: 66 (1961). 2. COSGROVE, M. D., SPHALL, R. A., and MORROW, J. W.: A new office test for bacteriuria, J. Urol. 109: 868 (1973). 3. DOWNS, R. A.: A chnical evaluation of the uroscreen test: triphenyl-tetrazolium chloride, ibid. 94: 706 (1965). Detection of 4. BRAUDE, A. I., and BERKOWITZ, H.: urinary catalase by disk flotation, J. Lab. Clin. Med. 57: 499 (1961). 5. CRAIG, W. A., KUNIN, CALVIN M., and DEGROOT, J. : Evaluation of new urinary tract infection screening devices, Appl. Microbiol. 26: 196 (1973).
59
W. MORROW,
M.D.
M.D.
ROCKENMACHER,
M. L. GOLDMAN, KATHERINE
BY
STRIP INCUBATION
ALAN T. WEBB, MORRIS
URINE CULTURE
M.T.
ARLO’ITI,
PH.D. (A.S.C.P.)
M.T.
(A.S.C.P.)
From the Departments of Urology and Pathology, Rancho Los Amigos Hospita1 and University of Southern California School of Medicine, Downey, California
ABSTRACT - The $lter paper culture-tetraxolium test resuks have been compared with routine hospita1 streak plate cultures in 400 random urine specimens. Pseudomonas cultures tended to impart u color change to the test pads which rendered ínterpretation difficult. While a consistently accurute quantitatíve interpretation of these cultures díd emerge durtng the course of the trial studies, some instructions and degree of observer accommodation and bias, in terms of the manufacturer’s description, was necessary to achíeve an appropriate reading. When the hospita1 laboratory results und culture strip test results are compared in terms of colony counts greater than 1 O4 bacteria per míllíliter and less than 104 bacteria per milliliter, the method was found to be 93 per cent sensítive for the totul count and 83 per cent sensitive fw the gram-negative count with a specijìcity correlation of 96 per cent counts, respectively. fM- the total and g ram-negatíve
There are a number of opportunities in clinical practice in which a simple, inexpensive quantitative bacteriologie procedure could be of value in screening or monitoring for bacteriuria. In the last few years a number of systems have been introduced for this purpose. The improved quality and reliability of these methods suggest that they may play a major role in reducing demands on the bacteriologie resources of the clinical laboratory without sacrificing the reliability necessary in assessing the bacterial status of the urinary tract. These simplified testing procedures can be divided into the chemical screening tests and those based on bacteriologie culture methods. Chemical screening tests have been devised based on: (1) the reduction of nitrate to nitrite (the Griess test);’ (2) residual urinary glucose as related to bacterial action (Uriglox);2 (3) reduction
56
of tetrazolium by bacterial metabolites to the red precipitate triphenylformazan (uroscreen);” and (4) the urinary catalase test.4 In genera1 these tests have been insufficiently reliable for the clinician to accept the results with the required degree of certitude required when establishing the true state of bacteriuria. Most of the culture methods have greater reliability and in some instances permit subculturing, determination of species, and antibacterial sensitivity testing. It is perhaps a logica1 development that a combination of the features from both the chemical and culture methods should be incorporated into a single easily applied procedure. The purpose of this communication is to report the results of such a method and compare these results with those obtained by the routine bacteriologie processing of urine specimens in the laboratory at Rancho Los Amigos Hospital.
UROLOGY
/ JANCARY 1975
/ VOLUMEV.
UUMBER
1
Material and Methods The filter paper culture-tetrazolium test with combined nitrite indicator (Microstix)* is a clear plastic dip stick containing one chemical reagent pad and two culture media containing filter paper pads attached in series. The chemical reagent pad was designed to detect nitrite in the urine (Griess test). In the current study this portion of the test was not correlated with the culture data. The proximal media pad contains an inhibitor to the growth of gram-positive organisms and thus is representative of the total gram-negative count in the specimen. The distal pad supports both gram-negative and gram-positive growth and represents the total bacterial count present in the specimen. In addition the media pads contain colorless tetrazolium which, when reduced to the red precipitate triphenylformazan under bacterial action, produces punctate and hopefully discrete red spots which are indicative of the number of bacteria in the inoculum. In practice the reagent portion of the dip stick is saturated with urine and placed in a plastic envelope. The envelope is collapsed and sealed so as to exclude as much air as possible. The stick and the envelope are incubated overnight at 37“ C. Hospita1 laboratory
procedure
Each specimen of urine was mixed well, and a calibrated loopful was inoculated to each of three media plates; blood agar, phenylethanol agar, and McConkey’s medium agar. If the urine was clear, the loop volume used for streaking the plates was 0.01 ml. ; and if the urine was turbid, the loop volume calibration was 0.001 ml. After incubation for eighteen to twenty-four hours at 35” C. the plates were examined, and each colony type present was counted. The number of colonies per plate times the appropriate dilution factor produced the total bacterial count per milliliter. Bacterial plate count of each species is expressed and reported as (A) no growth, (B) less than 104 bacteria per milliliter, (C) less than 105 bacteria per milliliter, and (D) greater than 105 bacteria per milliliter. Species indentification and sensitivity studies were completed on only the latter two categories (C and D). Species identification of bacteria was made by subculturing each colony type onto differential media tubes in keeping with the modified WB Enteric Differential System1 with biochemical *AmesCompany,Elkhart, Indiana.
t Diagnostic Research Inc., Long Mand City, New York.
UROLOGY
/ JANUARY 1975 / VOLUMEV,
NUMBER 1
reactions being observed after overnight incubation. Antibiotic sensitivity testing for each bacterial species was performed by the Kirby-Bauer method. Urine specimens
The urine specimens for the study were obtained by random selection from among those received from al1 wards and services in the hospital. Any specimen which indicated that the patient donor was receiving antibacterial therapy was excluded from the study. The presence of most antibacterials in the urine is known to alter the filter paper culture test by inducing disproportionately low bacterial growth. Since the urine which provides the inoculum also hydrates the media, the media in turn contains a disproportionately high leve1 of antibacterial activity. Specimens were received at various times throughout the day and represented randomly collected specimens and not first morning specimens. While al1 reasonable effort was made to exclude specimens from patients who were on antibacterial therapy, it is probable that some antibacterial-containing specimens evaded the screening methods both at the ward leve1 and in the laboratory and were entered into the study. Since such unwanted inclusions obtain in the clinical setting wherein such tests are likely to be used, the influence of the occasional antibacterial-containing specimen appears justified in assessing the over-al1 reliability of the method. The dip stick was inoculated and sealed in a plastic envelope at the same time the agar plates were inoculated. The envelope was taped to one of the streak plate Petri dishes and incubated under the same conditions and in the same incubator as the agar plates. Results A total of 400 specimens were included in the study. Since the hospital method of reporting grouped results diHerently from the progressive exponential categories on the filter paper culture test, some restructuring of the data obtained by the dip stick method was necessary to compare ít with the hospita1 categories. Table 1 correlates al1 of the dip stick categories for total bacterial growth with the hospital reporting system. With this alignment of the data the degree of agreement or disagreement is difficult to assess; however, in the area counts greater than 105 the
57
TABLE
1.
obtained -
Comparison of total bacterial counts by streak plate method and filter paper culture-tetrazolium test* Streak Plate (Colonies per Milliliter) NO > 100,000 Growths < 10,000 < 100,000 _
Filter Paper (Colonies per Milliliter)
.J& 12 _7 7 4 3
17 4 8 26 5
2 7 7 6
in parentheses
indicate
Pseudomonas
Negative 10 10” 103 104 > los *Figures < 105.
3
4 11 10 7 15 109
(3) (6) (6) (1) (2) (5)
colonies
correlation does not appear to be optimum since the number of filter paper culture counts in this area is low when compared with the hospita1 group. The numbers in parentheses indicate cultures of Pseudomonas species. These are specifically noted, because during the course of the study the color change that this species tended to impart to the entire culture pad as wel1 as the suppression of definitive punctate spotting caused difficulty in interpreting the dip sticks in Pseudomonas specimens which were consistently more than 100,000 organisms by the streak plate method. As the study progressed the observer who interpreted the results accommodated to the altered appearance of the dip sticks as indicated by the fact that these inconsistencies declined as the series progressed. Despite the manufacturer’s comments on the enclosure that accompanied strains of Pseuthe strips, namely, “occasional domonas aeruginosa may produce gray-green or blue background colors which do not interfere with the reading of the test,” this coloration did in fact seriously interfere with the interpretation of test results.
TABLE
obtained
Comparison of gram-negative counts by streak plate method and the @er paper culture-tetrazolium test*
11.
Filter Paper (Colonies per Milliliter
Streak Plate (Colonies
NO Growths
< 10,000
per Milliliter) -
< 109,000
Negative 10 102
J4& 9 6
13 4 4
3
104 103 > 105
93 4
134 2
s3 2
indicate
Pseudomonas
*Figures < 105.
58
-
in parentheses
1
TABLE
-~ Number of Organisms < los > 10” > 104 Negative
TOTALS *Figures < 105.
15. (1) 102 cultures
Microstix-GramTotal negative
123 52 41 ---184 (16)
400 in parentheses
Hospita1 Laborator> Gram Total Cultures negative .~~.__~ _._
118 30 31 221 (20) .- _~~
158 25 80 137
156 19 ,42 lij1
400
400
~200
Pseudomonas
cultures
indicate
Table 11 correlates the hospita1 laboratory and the filter paper culture test results for gramnegative organisms. An overall pattern of results similar to that for the total counts is seen. By further consolidation of the dip stick results to correspond more closely to the hospita1 reporting system, filter paper culture test categories negative, lol, and 102, were considered as no growth (Table 111). The filter paper culture test category 103 was equated with hospita1 category less than lO*. Filter paper culture test category lO* was considered as hospita1 category less than 105, and the greater than the 105 filter paper culture category was aligned with the greater than 105 laboratory category. Using this convention the correlations of the data could be better appreciated. In the greater than 105 hospita1 culture category, however, there is significantly larger number of positive cultures as compared with the results obtained by the filter paper culture method, whereas in the less than 105 category the reverse is truc. If one further consolidates the data into two groups with colony counts less than lO* or greater than lO*, a significant correlation is seen for total bacterial counts and gram-negative bacterial counts (Table IV). If an additional adjustment is
TABLE IV. Correlation of laboratory and filter paper-culture test in terms of colony counts*
> 190,000 15 (7) 16 (12) 4
Correlation of filter paper culturcatest with hospita1 laboratory results*
111.
tetrazolium
II Number of Organisms
--.-
Microstix GramTotal negative
’ 104
175 225 (16)_
252 (20) _I_
TOTALS
400
490
< 104
*Figures < 105.
UROLOGY
in parentheses
i
148
indicate
Hospita1 Laboratory GrarnTotal negative Cultures .__~..~_ _ _____ ._
I-217
177 223 ----
400
400
Pseudomonas
cultures
183
JANC;ARY1975 / VOLUME~,NUMVIBERl
made for the misinterpretations of Pseudomonas colony counts by plating the appropriate group of Pseudomonas specimens in the greater than 104 categorization, an even more striking correlation between the filter paper culture and the hospita1 laboratory results is evident (Table V). Such a transfer of data cannot be sanctioned in recording the final result of the study; however, the development of a consistent and reliable bias on the part of the observer as the study progressed warrants comment and justifies a projection of what might be expected with the method in more extended use. Using the greater than 104 and less than 104 categories the sensitivity and specificity of the filter paper culture test in relation to the hospita1 laboratory results may be expressed. The sensitivity of the total bacterial growth portion of the test is 93 per cent, and the specificity is 97 per cent. The gram-negative portion of the test shows a sensitivity of 84 per cent and a specificity is 90 per cent. If the transfer of the Pseudomonas results to the appropriate growth category is made, the sensitivity for the total bacterial count portion of the test is 96 per cent and the specificity is also 96 per cent. The gram-negative portion of the test demonstrates the sensitivity is 96 per cent, and the specificity is 90 per cent. In a recent study of the filter paper culture method (total count only) compared with pour plate colony counts, a sensitivity of 90.7 per cent and a specificity of 99.1 per cent was found.5 It was also noted, in these 1,000 cases, that 30 specimens with colony counts of 105 or more by the pour plate method were accompanied by
UROLOGY / JANUARY1975 /
VOLUME V, NUMBER 1
TABLE V. Correlation of laboratory andfilter paperculture test in term of colony counts after correction for Pseudomonas readings
Number of Organisms < 104 >104 TOTALS
Microstix GramTotal negative
Hospita1 Laboratory Total GramCultures negative
191 gN_
168 ~232
183 217
177 223
400
400
400
400
lower counts on the filter paper test. In almost half of these specimens the counts on the filter study were less than 103. This discrepancy between the two methods, however, very rarely occurred with the commonly encountered enteric bacteria associated with urinary tract infection. 7601 E. Imperial Highway Downey, California 90242 (DR. MORROW) References 1. SMITH, L. T., THAYER, W. R., MALTA, E. M., and UTZ, J. P.: Relationship of the Griess nitrate test to bacterial culture in the diagnosis of urinary tract infection, Ann. Intern. Med. 54: 66 (1961). 2. COSGROVE, M. D., SPHALL, R. A., and MORROW, J. W.: A new office test for bacteriuria, J. Urol. 109: 868 (1973). 3. DOWNS, R. A.: A chnical evaluation of the uroscreen test: triphenyl-tetrazolium chloride, ibid. 94: 706 (1965). Detection of 4. BRAUDE, A. I., and BERKOWITZ, H.: urinary catalase by disk flotation, J. Lab. Clin. Med. 57: 499 (1961). 5. CRAIG, W. A., KUNIN, CALVIN M., and DEGROOT, J. : Evaluation of new urinary tract infection screening devices, Appl. Microbiol. 26: 196 (1973).
59
