- Email: [email protected]
Significance of Bacteriuria with Presumed Nonpathogenic Organisms
0022-5347/80/1241-0240$02.00/0 Vol. 124, August
THE JOURNAL OF UROLOGY
Printed in U.S.A.
Copyright© 1980 by The Williams & Wilkins Co.
SIGNIFICANCE OF BACTERIURIA WITH PRESUMED NONPATHOGENIC ORGANISMS MARYANNE B. MCGUCKIN, JEAN TOMASCO
AND
ROB ROY MACGREGOR
From the School of Allied Medical Professions, Department of Medical Technology and the Infectious Disease Section, School of Medicine, University of Pennsylvania, Philadelphia, Pennsylvania
ABSTRACT
The results of 2,793 clean catch midstream specimens submitted to the microbiology laboratory were examined to determine if the isolation of reputed non-pathogens such as Lactobacillus, Staphylococcus epidermidis, Micrococcus and diphtheroids represented poor collection technique, contaminant growth by failure to plate or refrigerate quickly or actual urinary tract infection. Eight per cent of the inpatient cultures and 18 per cent of the outpatient cultures had more than 105 colony-forming units per ml. of these non-pathogens. Of the outpatient cultures with more than 105 colony-forming units per ml. of non-pathogens 44 per cent had symptoms of urinary tract infection and more than 10 white blood cells per high power field in the urinary sediment, compared to 3 per cent of inpatients. In a prospective analysis of an outpatient population a second clean catch midstream specimen was collected within 2 hours of the first specimen with careful attention to sterile collection technique. Of 15 cases of significant bacteriuria with non-pathogens in the first specimen only 1 had significant bacteriuria with the same organism in the second specimen. In contrast, in the inpatient population 7 of 11 cases with pathogen significant bacteriuria in the first specimen had significant bacteriuria in the second specimen. These findings suggest that most episodes involving isolation of non-pathogens in greater than 105 colony-forming units per ml. from urine represent poor techniques of specimen collection rather than urinary tract infection. The diagnosis of urinary tract infection usually is suspected either on the basis of symptoms of cystitis or pyelonephritis, or by the demonstration of significant pyuria (> 10 white blood cells per high power field) in the urinary sediment1 and is confirmed by a culture yielding > 105 colony-forming units per ml. urine. 2 However, the diagnostic value of finding significant pyuria and bacteriuria in a specimen has been shown to be related to the quality of specimen submitted. In 1978 McGuckin and associates demonstrated that the criterion of > 10 white blood cells per high power field used to identify significant pyuria was appropriate only when carefully voided midstream specimens were being examined. Random urine specimens, voided without attention to sterile technique, often showed > 10 white blood cells per high power field without having > 105 colony-forming units per ml. 3 In addition, Sellin and associates showed that urine bacterial counts indicating significant bacteriuria can be caused by incorrect specimen collection, especially when the isolates include such seemingly non-pathogenic organisms as Staphylococcus epidermidis, diphtheroids or lactobacilli. 4 However, some investigators have presented data suggesting that Staphylococcus epidermidis is a significant cause of urinary tract infections. For example, in 1974 Maskell reported that 7 per cent of the urinary tract infections seen at his clinic were caused by Staphylococcus epidermidis and Micrococcus.5 Because these organisms can be isolated as contaminants and as pathogens proper collection of the specimen is critical; non-sterile collection can lead to misdiagnosis, overuse of antibiotics and increased patient cost. However, to our knowledge there has been no study to determine whether the instructions and collecting systems used for a clean catch midstream specimen can contribute to the contamination of the urine. The frequency of isolating non-pathogens in a clean catch midstream specimen was of concern in our institution and prompted this investigation. Therefore, the study was designed to 1) determine retrospectively the areas in our institution Accepted for publication September 28, 1979. Read at annual meeting of American Society for Medical Technology, Denver, Colorado, June 1978.
where there was a high incidence of non-pathogen bacteriuria, 2) determine prospectively whether such non-pathogen isolates represented poor technique of specimen collection or true infection and 3) evaluate the system and procedure used to collect the clean catch midstream specimen. MATERIALS AND METHODS
Patient population. A retrospective analysis of 2,793 clean catch midstream urine specimens submitted for routine culture was done between May and December 1977. Specimens came from inpatients and outpatients. A prospective analysis was done on 40 female outpatients from the hospital's walk-in clinic. Specimen collection. Specimens in the retrospective analysis were collected according to the standard procedure practiced in our institution, using a commercially available clean catch midstream specimen kit, including a sterile container, 3 prepackaged iodophor towelettes and printed instructions for collecting the specimen. Patients were given the kit and instructed to read the directions before collecting the specimen. During the prospective study 2 clean catch midstream specimens were collected. The first specimen was collected according to the procedure outlined previously. However, a second clean catch midstream specimen was collected from the same patients under the supervision of nursing personnel, who provided verbal and diagrammatic instructions to each patient. Microscopic and culturing procedures. Microscopic examination of the urinary sediment was done on all specimens. Urine was spun for 5 minutes at 3,000 revolutions per minute. Cultures were ordered if significant pyuria (> 10 white blood cells per high power field) was observed. Specimens for culture were inoculated with a 0.001 calibrated loop onto a colistin-nalidixic acid-MacConkey selective biplate. All specimens were cultured within 2 hours after collection, incubated at 37C and read at 24 and 48 hours. Identification of isolates was made by gram stain, morphology and biochemical reactions. 6 Criteria used to determine significance of non-pathogens in clean catch midstream specimen. A definite contaminant was defined as a second clean catch midstream specimen obtained
240
241
BACTERIURIA AND NON-PATHOGENS
within 2 hours of the first specimen showing no growth or a different organism. Definite infection was defined as > 10 5 colony-forming units per ml. of a non-pathogen plus symptoms and pyuria, or 2 cultures (first and second clean catch midstream specimens) yielding the same isolates plus either symptoms or pyuria. RESULTS
Table 1 shows the frequency in which >10 5 colony-forming units per ml. pathogens and non-pathogens were isolated in our population during the retrospective analysis. The finding that 8 per cent of inpatient urine cultures yielded > 10° colonyforming units per ml. non-pathogens versus 18 per cent for outpatients tended to support the idea that poor technique of specimen collection in the outpatient unit was causing contamination with non-pathogens. However, 44 per cent of the outpatients with significant bacteriuria with non-pathogens had symptoms of urinary tract infection and pyuria (> 10 white blood cells per high power field). Since some of the non-pathogen isolates from our outpatients occurred in the presence of symptoms and pyuria we were not able to assume that they were contaminants. Therefore, we sought direct evidence of contamination secondary to poor sterile technique by obtaining 2 cultures from patients coming to the clinic with symptoms of urinary tract infection. In contrast to the first specimen collection the second collection was preceded by careful technique instruction, including a diagram. The results of these sequential collections are shown in table 2. In 40 consecutive patients with symptoms of urinary tract infection we found 15 cases with non-pathogen significant bacteriuria in the first clean catch midstream specimen; only 1 of these 15 cases (6.7 per cent) had significant bacteriuria with the same isolate in the second specimen. In contrast, we found 11 cases that had significant bacteriuria with a pathogen in the first clean catch midstream specimen, and 7 cases (63.6 per cent) showed significant bacteriuria with the same isolate in the second clean catch midstream specimen. This difference in reproducibility of non-pathogen significant bacteriuria is significant at the 0.001 level by Chisquare analysis. DISCUSSION
The incidence of significant bacteriuria resulting from nonpathogens was higher in our outpa~ient population t~an among inpatients. However, our prospective study of 40 patients from whom 2 sequential specimens were collected showed that extra attention to sterile collection decreased the isolation of these non-pathogens from 15 to 1. Therefore, because improved sterile technique virtually abolished the isolation of > 105 colonyforming units per ml. of non-pathogens it appears that the first TABLE
1. Frequency of isolation of> 10" colony-forming units per
ml. pathogens and non-pathogens Inpatient No.(%)
Outpatient No.(%)
2,006 (71.8) 376 (18.7)
787 (28.2) 233 (29.6)
537
346 (92.0)
191 (82.0)
72
30 (8.0)
42 (18.0)
Total No.
All cultures No. cultures > 10·' colony-forming units per ml.
2,793 609
Pas. cultures > 10" colony-forming units per ml. pathogens > 10" colony-forming units per ml. non-pathogens Total No. pos. cultures TABLE 2.
609
Culture results when 2 urine specimens were obtained for culture from the same patient Specimen 1
No. pts. vnth >IO" colony-forming units non-pathogens No. pts. with> 10" colony-forming units pathogens
Specimen 2
6.7
15 11
% Reproducibility
7
63.6
isolates probably were contaminants. In contrast, 7 of 11 patients in our prospective population from whom pathogens were isolated in the first specimen continued to have the same isolates in the second clean catch midstream specimen. Significant bacteriuria with > 10 5 colony-forming units per ml. non-pathogens can result from 4 possible mechanisms: 1) non-sterile collection of the specimen, 2) contamination of the collection equipment, 3) failure to refrigerate the specimen promptly and process it expeditiously and 4) laboratory enor at the time of culture. Our data strongly suggested that improper collection of the specimen caused most high col~ny counts of non-pathogens. A review of our system for collectmg clean catch midstream specimens, along with the patients' responses to our inquiries showed 2 major causes of poor specimen collection in outpatients in our institution: 1) difficulty in opening the pre-packaged iodophor towelettes and 2) failure of the patients to read and comprehend the instructions for collecting the clean catch midstream specimen. Diagnostically, the clinical value of finding significant bacteriuria in each of the 3 diagnostic procedures has been well documented (clean catch· midstream specimen, catheter urine and suprapubic aspiration). Kaye showed that the clinical correlation between >10 5 colony-forming units per ml. and true bacteriuria is 80 per cent in the first specimen and increases to 95 per cent if the second and third specimens are similar. 7 On the other hand, urine obtained by catheterization has been shown to yield a lower incidence of single falsely positive urine cultures but can carry a 4 to 6 per cent risk of introducing infection. 8 Suprapubic aspiration probably is the more accurate method to determine the presence of true bacteriuria but this method does require a greater skill. It usually is performed when infection is suspected strongly and repeated voided specimens have not been clearly determinative. Therefore, the clean catch midstream specimen is the least invasive collection method of the 3 and can be of significant diagnostic value if performed properly. However, as our data confirm the collection system can determine the number of organisms isolated from a supposed clean catch urine specimen. Therefore, it appears that most episodes involving the isolation of non-pathogens in > 105 colony-forming units per ml. from a clean catch midstream specimen represent poor technique of specimen collection rather than true significant bacteriuria. REFERENCES 1. Center for Disease Control: Outline for Surveillance and Control of
Nosocomial Infections, p. 20, 1970. 2. Kass, E. H.: Bacteriuria and the diagnosis of infections of the urinary tract. With observation on the use of methionine as a urinary antiseptic. Arch. Intern. Med., HlO: 709, 1957. 3. McGuckin, M., Cohen, L. and MacGregor, R. R.: Significance of pyuria in urinary sediment. J. Urol., 120: 452, 1978. 4. Sellin, M., Cooke, D. I., Gillespie, W. A., Sylvester, D. G. H. and Anderson, J. D.: Micrococcal urinary tract infections in young women. Lancet, 2: 570, 1975. 5. Maskell, R.: Importance of coagulase-negative staphylococci as pathogens in the urinary tract. Lancet, 1: 1155, 1974. 6. Lennette, E. H., Spaulding, E. H. and Truant, J. P.: Manual of Clinical Microbiology, 2nd ed. Washington, D. C.: American Society of Microbiology, 1974. 7. Kaye, D.: Urinary Tract Infection and Its Management. St. Louis: The C. V. Mosby Co., 1972. 8. Brumfitt, W. and Percival, A.: Pathogenesis and laboratory diagnosis of non-tuberculous urinary tract infection: a review. J. Clin. Path., 17: 482, 1964. EDITORIAL COMMENT These authors emphasize the need to adhere to the classic definitions of significant bacteriuria when routine urine specimens are collected by the clean voided method. All too often disposable urine collection sets are purchased by a hospital or clinic without adequate evaluation by competent physicians. What is the point of using packaged iodophor towelettes in such kits if the patient finds them difficult to open?
C.M.K.
THE JOURNAL OF UROLOGY
Printed in U.S.A.
Copyright© 1980 by The Williams & Wilkins Co.
SIGNIFICANCE OF BACTERIURIA WITH PRESUMED NONPATHOGENIC ORGANISMS MARYANNE B. MCGUCKIN, JEAN TOMASCO
AND
ROB ROY MACGREGOR
From the School of Allied Medical Professions, Department of Medical Technology and the Infectious Disease Section, School of Medicine, University of Pennsylvania, Philadelphia, Pennsylvania
ABSTRACT
The results of 2,793 clean catch midstream specimens submitted to the microbiology laboratory were examined to determine if the isolation of reputed non-pathogens such as Lactobacillus, Staphylococcus epidermidis, Micrococcus and diphtheroids represented poor collection technique, contaminant growth by failure to plate or refrigerate quickly or actual urinary tract infection. Eight per cent of the inpatient cultures and 18 per cent of the outpatient cultures had more than 105 colony-forming units per ml. of these non-pathogens. Of the outpatient cultures with more than 105 colony-forming units per ml. of non-pathogens 44 per cent had symptoms of urinary tract infection and more than 10 white blood cells per high power field in the urinary sediment, compared to 3 per cent of inpatients. In a prospective analysis of an outpatient population a second clean catch midstream specimen was collected within 2 hours of the first specimen with careful attention to sterile collection technique. Of 15 cases of significant bacteriuria with non-pathogens in the first specimen only 1 had significant bacteriuria with the same organism in the second specimen. In contrast, in the inpatient population 7 of 11 cases with pathogen significant bacteriuria in the first specimen had significant bacteriuria in the second specimen. These findings suggest that most episodes involving isolation of non-pathogens in greater than 105 colony-forming units per ml. from urine represent poor techniques of specimen collection rather than urinary tract infection. The diagnosis of urinary tract infection usually is suspected either on the basis of symptoms of cystitis or pyelonephritis, or by the demonstration of significant pyuria (> 10 white blood cells per high power field) in the urinary sediment1 and is confirmed by a culture yielding > 105 colony-forming units per ml. urine. 2 However, the diagnostic value of finding significant pyuria and bacteriuria in a specimen has been shown to be related to the quality of specimen submitted. In 1978 McGuckin and associates demonstrated that the criterion of > 10 white blood cells per high power field used to identify significant pyuria was appropriate only when carefully voided midstream specimens were being examined. Random urine specimens, voided without attention to sterile technique, often showed > 10 white blood cells per high power field without having > 105 colony-forming units per ml. 3 In addition, Sellin and associates showed that urine bacterial counts indicating significant bacteriuria can be caused by incorrect specimen collection, especially when the isolates include such seemingly non-pathogenic organisms as Staphylococcus epidermidis, diphtheroids or lactobacilli. 4 However, some investigators have presented data suggesting that Staphylococcus epidermidis is a significant cause of urinary tract infections. For example, in 1974 Maskell reported that 7 per cent of the urinary tract infections seen at his clinic were caused by Staphylococcus epidermidis and Micrococcus.5 Because these organisms can be isolated as contaminants and as pathogens proper collection of the specimen is critical; non-sterile collection can lead to misdiagnosis, overuse of antibiotics and increased patient cost. However, to our knowledge there has been no study to determine whether the instructions and collecting systems used for a clean catch midstream specimen can contribute to the contamination of the urine. The frequency of isolating non-pathogens in a clean catch midstream specimen was of concern in our institution and prompted this investigation. Therefore, the study was designed to 1) determine retrospectively the areas in our institution Accepted for publication September 28, 1979. Read at annual meeting of American Society for Medical Technology, Denver, Colorado, June 1978.
where there was a high incidence of non-pathogen bacteriuria, 2) determine prospectively whether such non-pathogen isolates represented poor technique of specimen collection or true infection and 3) evaluate the system and procedure used to collect the clean catch midstream specimen. MATERIALS AND METHODS
Patient population. A retrospective analysis of 2,793 clean catch midstream urine specimens submitted for routine culture was done between May and December 1977. Specimens came from inpatients and outpatients. A prospective analysis was done on 40 female outpatients from the hospital's walk-in clinic. Specimen collection. Specimens in the retrospective analysis were collected according to the standard procedure practiced in our institution, using a commercially available clean catch midstream specimen kit, including a sterile container, 3 prepackaged iodophor towelettes and printed instructions for collecting the specimen. Patients were given the kit and instructed to read the directions before collecting the specimen. During the prospective study 2 clean catch midstream specimens were collected. The first specimen was collected according to the procedure outlined previously. However, a second clean catch midstream specimen was collected from the same patients under the supervision of nursing personnel, who provided verbal and diagrammatic instructions to each patient. Microscopic and culturing procedures. Microscopic examination of the urinary sediment was done on all specimens. Urine was spun for 5 minutes at 3,000 revolutions per minute. Cultures were ordered if significant pyuria (> 10 white blood cells per high power field) was observed. Specimens for culture were inoculated with a 0.001 calibrated loop onto a colistin-nalidixic acid-MacConkey selective biplate. All specimens were cultured within 2 hours after collection, incubated at 37C and read at 24 and 48 hours. Identification of isolates was made by gram stain, morphology and biochemical reactions. 6 Criteria used to determine significance of non-pathogens in clean catch midstream specimen. A definite contaminant was defined as a second clean catch midstream specimen obtained
240
241
BACTERIURIA AND NON-PATHOGENS
within 2 hours of the first specimen showing no growth or a different organism. Definite infection was defined as > 10 5 colony-forming units per ml. of a non-pathogen plus symptoms and pyuria, or 2 cultures (first and second clean catch midstream specimens) yielding the same isolates plus either symptoms or pyuria. RESULTS
Table 1 shows the frequency in which >10 5 colony-forming units per ml. pathogens and non-pathogens were isolated in our population during the retrospective analysis. The finding that 8 per cent of inpatient urine cultures yielded > 10° colonyforming units per ml. non-pathogens versus 18 per cent for outpatients tended to support the idea that poor technique of specimen collection in the outpatient unit was causing contamination with non-pathogens. However, 44 per cent of the outpatients with significant bacteriuria with non-pathogens had symptoms of urinary tract infection and pyuria (> 10 white blood cells per high power field). Since some of the non-pathogen isolates from our outpatients occurred in the presence of symptoms and pyuria we were not able to assume that they were contaminants. Therefore, we sought direct evidence of contamination secondary to poor sterile technique by obtaining 2 cultures from patients coming to the clinic with symptoms of urinary tract infection. In contrast to the first specimen collection the second collection was preceded by careful technique instruction, including a diagram. The results of these sequential collections are shown in table 2. In 40 consecutive patients with symptoms of urinary tract infection we found 15 cases with non-pathogen significant bacteriuria in the first clean catch midstream specimen; only 1 of these 15 cases (6.7 per cent) had significant bacteriuria with the same isolate in the second specimen. In contrast, we found 11 cases that had significant bacteriuria with a pathogen in the first clean catch midstream specimen, and 7 cases (63.6 per cent) showed significant bacteriuria with the same isolate in the second clean catch midstream specimen. This difference in reproducibility of non-pathogen significant bacteriuria is significant at the 0.001 level by Chisquare analysis. DISCUSSION
The incidence of significant bacteriuria resulting from nonpathogens was higher in our outpa~ient population t~an among inpatients. However, our prospective study of 40 patients from whom 2 sequential specimens were collected showed that extra attention to sterile collection decreased the isolation of these non-pathogens from 15 to 1. Therefore, because improved sterile technique virtually abolished the isolation of > 105 colonyforming units per ml. of non-pathogens it appears that the first TABLE
1. Frequency of isolation of> 10" colony-forming units per
ml. pathogens and non-pathogens Inpatient No.(%)
Outpatient No.(%)
2,006 (71.8) 376 (18.7)
787 (28.2) 233 (29.6)
537
346 (92.0)
191 (82.0)
72
30 (8.0)
42 (18.0)
Total No.
All cultures No. cultures > 10·' colony-forming units per ml.
2,793 609
Pas. cultures > 10" colony-forming units per ml. pathogens > 10" colony-forming units per ml. non-pathogens Total No. pos. cultures TABLE 2.
609
Culture results when 2 urine specimens were obtained for culture from the same patient Specimen 1
No. pts. vnth >IO" colony-forming units non-pathogens No. pts. with> 10" colony-forming units pathogens
Specimen 2
6.7
15 11
% Reproducibility
7
63.6
isolates probably were contaminants. In contrast, 7 of 11 patients in our prospective population from whom pathogens were isolated in the first specimen continued to have the same isolates in the second clean catch midstream specimen. Significant bacteriuria with > 10 5 colony-forming units per ml. non-pathogens can result from 4 possible mechanisms: 1) non-sterile collection of the specimen, 2) contamination of the collection equipment, 3) failure to refrigerate the specimen promptly and process it expeditiously and 4) laboratory enor at the time of culture. Our data strongly suggested that improper collection of the specimen caused most high col~ny counts of non-pathogens. A review of our system for collectmg clean catch midstream specimens, along with the patients' responses to our inquiries showed 2 major causes of poor specimen collection in outpatients in our institution: 1) difficulty in opening the pre-packaged iodophor towelettes and 2) failure of the patients to read and comprehend the instructions for collecting the clean catch midstream specimen. Diagnostically, the clinical value of finding significant bacteriuria in each of the 3 diagnostic procedures has been well documented (clean catch· midstream specimen, catheter urine and suprapubic aspiration). Kaye showed that the clinical correlation between >10 5 colony-forming units per ml. and true bacteriuria is 80 per cent in the first specimen and increases to 95 per cent if the second and third specimens are similar. 7 On the other hand, urine obtained by catheterization has been shown to yield a lower incidence of single falsely positive urine cultures but can carry a 4 to 6 per cent risk of introducing infection. 8 Suprapubic aspiration probably is the more accurate method to determine the presence of true bacteriuria but this method does require a greater skill. It usually is performed when infection is suspected strongly and repeated voided specimens have not been clearly determinative. Therefore, the clean catch midstream specimen is the least invasive collection method of the 3 and can be of significant diagnostic value if performed properly. However, as our data confirm the collection system can determine the number of organisms isolated from a supposed clean catch urine specimen. Therefore, it appears that most episodes involving the isolation of non-pathogens in > 105 colony-forming units per ml. from a clean catch midstream specimen represent poor technique of specimen collection rather than true significant bacteriuria. REFERENCES 1. Center for Disease Control: Outline for Surveillance and Control of
Nosocomial Infections, p. 20, 1970. 2. Kass, E. H.: Bacteriuria and the diagnosis of infections of the urinary tract. With observation on the use of methionine as a urinary antiseptic. Arch. Intern. Med., HlO: 709, 1957. 3. McGuckin, M., Cohen, L. and MacGregor, R. R.: Significance of pyuria in urinary sediment. J. Urol., 120: 452, 1978. 4. Sellin, M., Cooke, D. I., Gillespie, W. A., Sylvester, D. G. H. and Anderson, J. D.: Micrococcal urinary tract infections in young women. Lancet, 2: 570, 1975. 5. Maskell, R.: Importance of coagulase-negative staphylococci as pathogens in the urinary tract. Lancet, 1: 1155, 1974. 6. Lennette, E. H., Spaulding, E. H. and Truant, J. P.: Manual of Clinical Microbiology, 2nd ed. Washington, D. C.: American Society of Microbiology, 1974. 7. Kaye, D.: Urinary Tract Infection and Its Management. St. Louis: The C. V. Mosby Co., 1972. 8. Brumfitt, W. and Percival, A.: Pathogenesis and laboratory diagnosis of non-tuberculous urinary tract infection: a review. J. Clin. Path., 17: 482, 1964. EDITORIAL COMMENT These authors emphasize the need to adhere to the classic definitions of significant bacteriuria when routine urine specimens are collected by the clean voided method. All too often disposable urine collection sets are purchased by a hospital or clinic without adequate evaluation by competent physicians. What is the point of using packaged iodophor towelettes in such kits if the patient finds them difficult to open?
C.M.K.