JOURNAL OF CLINICAL MICROBIOLOGY, Mar. 2000, p. 1187–1190 0095-1137/00/$04.00⫹0 Copyright © 2000, American Society for Microbiology. All Rights Reserved.
Vol. 38, No. 3
Novel Screening Method for Urine Cultures Using a Filter Paper Dilution System CALVIN M. KUNIN1*
AND
WILLIAM J. BUESCHING2
1
Department of Internal Medicine and Division of Medical Microbiology, Department of Pathology,2 The Ohio State University College of Medicine and Public Health, Columbus, Ohio 43210 Received 10 September 1999/Returned for modification 17 October 1999/Accepted 5 December 1999
We have developed a novel method for urine culture for office practice based on the use of filter paper as a solid-phase dilution device. Filtration dilutes and spreads the inoculum onto a solid culture surface. Experiments were conducted to determine the optimum inoculum size, microbial permeability through filter papers, and ability to exclude vaginal epithelial cells. The filter paper dilution system was compared to the standard streak method to detect bacteriuria in specimens submitted to the diagnostic laboratory. The sensitivity and specificity of the filter paper dilution system for detection of high-count (>104 CFU/ml) gram-negative bacteriuria in 487 urine specimens were 98.2 and 97.4%, respectively. The sensitivity and specificity for gram-positive bacteriuria in 404 urine specimens were 91.2 and 99.2%, respectively. Low-count gram-negative bacteriuria (104 CFU/ml. The system has good storage properties and can be inoculated at the point of source without the need for refrigeration or preservatives. It should be a useful screening method for office practice, where members of the family Enterobacteriaceae and staphylococci cause most infections. Standard culture methods are preferred for hospital diagnostic microbiology laboratories, where there is a need to detect yeasts and fastidious microorganisms and to isolate individual colonies from mixed cultures. nants, detection of low-count bacteriuria, preliminary identification, and good storage properties. The method can be adapted for automated systems. In this report, we describe the design and characteristics of the filter paper culture system and its ability to detect significant bacteriuria compared to the agar streak method used in the diagnostic microbiology laboratory. (An application has been filed for a U.S. provisional patent for a solid-phase dilution system for urine culture.)
The quantitative culture is the “gold standard” for the diagnosis of urinary tract infections. Urine cultures account for a large proportion of the tests performed in the clinical microbiology laboratory (1, 2). The information must be clinically meaningful and reported to physicians in a timely manner. The quality of the work depends on how well the specimens are collected, transported, and processed. One of the major problems for both the laboratory and clinicians is distinguishing between contaminants and low counts of uropathogens in voided urine obtained from females (2–4, 7). The standard urine culture procedure is the surface streak plate method combined with isolation, identification, and antibiotic susceptibility testing. Several relatively simple and inexpensive screening methods are available for use in office practice (1, 2, 4). These include filter paper strips applied to the surface of small plastic plates; dip-slides; and tube, cup, and pad culture methods. A variety of automated tests have been developed to rapidly quantitate and identify microorganisms and determine antimicrobial susceptibility (1, 2). All of these methods involve inoculation of whole urine directly into the system. We have developed a novel method to perform quantitative urine cultures based on the use of filter paper as a solid-phase dilution device. Filtration dilutes and spreads the inoculum onto a solid culture surface. It removes yeasts and bacteria adherent to vaginal epithelial cells. The potential advantages include inexpensive construction, simplicity of use, inoculation at the site of collection, elimination of vaginal cell contami-
MATERIALS AND METHODS Description of the filter paper dilution system. The filter paper dilution system consists of a three-layer, circular filter paper sandwich applied to the surface of 60- by 15-mm-diameter petri dish (Fig. 1). The top layer filters the inoculum, the middle layer traps the microorganisms, and the bottom layer stores the media. The filtration component consists of a Whatman no. 3 47-mm-diameter filter fitted with a small tab (Whatman International, Ltd., Maidstone, England). The culture component consists an absorbent filter pad overlaid by a 47-mm-diameter, 0.45 M pore grid filter (Millipore Corporation, Bedford, Mass.) as used in water bacteriology. A white grid is used for gram-negative bacteria, and a black grid is used for gram-positive bacteria. The culture component is glued at several points to the bottom of the dish to prevent movement. The media consist of Endo broth MF (Difco, Detroit, Mich.) for gram-negative bacteria and Columbia broth containing 10 g of colistin sulfate per ml and 15 g of nalidixic acid per ml (Sigma, St. Louis, Mo.) for gram-positive bacteria. The liquid media are added to the culture system and stored in plastic bags at 4°C or lyophilized (Freezemobile 6; Virtis Co., Inc., Gardiner, N.Y.) and stored at room temperature until rehydrated with 2 ml of sterile water. The inoculum consists of 0.1 ml of urine applied to the center of the top layer. The top layer is removed at 15 to 30 min after inoculation. It can be left in place for at least 2 h before colonies begin to form. A forceps is applied to the tab to peel it off. The system may be kept at room temperature until incubated at 35°C. Colonies are counted after 24 and 48 h of incubation. Penetration of microorganisms through the filter papers. The microorganisms consisted of Escherichia coli ATCC 25922, Staphylococcus aureus ATCC 29213, Staphylococcus epidermidis ATCC 12228, Enterococcus faecalis ATCC 29212, Candida albicans ATCC 90028, and clinical isolates of Klebsiella pneumoniae and Enterobacter cloacae. Stock cultures were streaked on MacConkey agar for gramnegative bacteria and sheep blood agar plates for gram-positive bacteria and
* Corresponding author. Mailing address: Department of Internal Medicine, The Ohio State University College of Medicine and Public Health, Room M110, Starling Loving Hall, 320 West 10th Ave., Columbus, OH 43210. Phone: (614) 293-8976. Fax: (614) 293-5627. Email:
[email protected]. 1187
1188
KUNIN AND BUESCHING
J. CLIN. MICROBIOL.
FIG. 1. Illustration of the filter paper dilution system. The system is activated by adding 2 ml of selective medium or water to dehydrated medium on the pads. (A and E) Top and bottom, respectively, of a 60- by 15-mm petri dish. (B) Top layer. To the center of a Whatman no. 3 filter paper with a tab, 0.1 ml of urine is applied after activation. (C) Middle layer. Black and white 0.45 M pore grids trap the bacteria and provide the growth surface. (D) Bottom layer. An absorbent pad serves as a reservoir for the medium. (F) Papers C and D are fixed to the bottom of the petri dish at points H. Layer B is removed by the tab 15 to 60 min after application of the inoculum. (G) Bacterial colonies on the surface of the grid after incubation.
yeasts. A sample of a single colony was added to trypticase soy broth and incubated for several hours at 37°C until cloudy. The suspension was then diluted to a 0.5 McFarland standard and serially 10-fold diluted in Trypticase soy broth (Difco). Separate pipette tips were used for each dilution. A 0.1-ml sample of each of the serial dilutions was applied to the center of the surface of the system. An identical inoculum was spread on the surface of MacConkey or Trypticase soy agar plates used for controls. The plates were incubated for 1 h at room temperature. The filter papers were then removed, the plates were incubated overnight at 37°C, and the colonies were counted. The ratio of the colony counts was determined for the filter paper systems versus the controls at dilutions at which discrete colonies could be readily distinguished (about 30 to 100 colonies). The optimum time for penetration of bacteria through the surface filter paper was determined by using E. coli ATCC 25922 as the test microorganism. A 0.1-ml sample of each of the serial dilutions (as described above) was added to the surface of the system containing Endo broth. The Whatman filters were removed at intervals, the plates were incubated overnight at 37°C, and the colonies were counted. Penetration of vaginal epithelial cells through filter paper. Vaginal cells were collected from three adult women. One woman had normal menstrual cycles, and two were postmenopausal and receiving hormone replacement therapy. A cotton swab was gently rotated over the mucosal surface of the vaginal vault. The swab was then swirled in 30 ml of Dulbecco’s phosphate-buffered saline (Gibco BRL, Grand Island, N.Y.). The cell suspension was vortexed, and 5 ml was passed through a Whatman no. 3 filter paper cone. A 0.5-ml sample of the cell suspension and the filtrate was added to PreservCyt solution (Cytyc Corporation, Boxborough, Mass.). An even layer of cells was deposited on a microscope slide. The slides were processed by the Papanicolaou method in the Cytology Laboratory of the Ohio State University Hospital. Cell counts were performed on 10 100⫻ or ⫻400 fields depending on the cell density. Comparison of the filter paper system with standard urine cultures performed in the clinical diagnostic microbiology laboratory. Urine specimens were obtained from 487 adult patients (females, 57.4%; males, 42.5%) from the clinic and inpatient services at the Ohio State University Medical Center. The specimens were transported to the clinical diagnostic microbiology laboratory by using the Vacutainer Urine Collection kit (Becton Dickinson, Franklin Lakes, N.J.). Urine specimens, without preservative, were obtained from an additional 54
clinic patients. The specimens were cultured by the streak plate method (4) with 0.01-ml quantitative loops on MacConkey and Columbia colistin-nalidixic acid agar (Remel, Lenexa, Kans.). The inoculum was evenly distributed by a rotating device. The urine specimens were stored at 4°C and cultured by the filter paper system within 24 h. The findings were compared with laboratory reports obtained from the hospital computer system. In a parallel study, freshly obtained urine, without preservative, was tested directly in the filter paper culture system before the sample was sent in the Vacutainer system to the laboratory. End points and statistical analysis. Calculations of sensitivity, specificity, and related values were performed as described by Sackett et al. (6) with the following assumptions. For the clinical diagnostic microbiology laboratory, a negative urine culture was considered to be ⬍10,000 CFU/ml. A positive culture was considered to be ⱖ10,000 CFU/ml. For the filter paper dilution system, a negative culture was considered to be no growth or ⬍10 CFU. A positive culture was considered to be from ⬎10 CFU to too numerous to count.
RESULTS Penetration of microorganisms through filter papers. In preliminary experiments, an 0.1-ml inoculum of E. coli cells applied to Whatman no. 3 filter paper was found to provide optimal dilution and spread compared to smaller or larger inocula, more porous Whatman no. 1 and 2 filters, and the thinner no. 5 and 6 filters. The growth zone occupied almost the full diameter of the HA filters when 0.1-ml inocula of ⱖ105 CFU/ml were applied to the filtration component. Penetration was virtually complete within 15 min (Table 1). In this experiment, the filter paper diluted the inoculum about 1:74. In other experiments, the dilution factor for E. coli was 1:8, 1:47, and 1:63. Other dilution factors for microorganisms tested with the filter paper system were as follows: S. aureus, 1:8 and 1:13; S. epidermidis, 1:24; E. faecalis, 1:14; E. cloacae,
VOL. 38, 2000
NOVEL URINE CULTURE SYSTEM
TABLE 1. Effect of time at room temperature before removal of the Whatman no. 3 filtration component from the filter paper dilution system on the colony counts of E. coli ATCC 25922 Time (min)
15 30 45 60 120 MacConkey agar controls
TABLE 2. Comparison of 487 urine cultures for gram-negative bacteria performed with the filter paper dilution system versus the streak plate method used in the diagnostic microbiology laboratorya
No. of CFU at dilutiona: 10⫺2
10⫺3
⬃138 ⬃112 ⬃190 ⬃150 ⬃160
18 28 19 15 22 b
TNTC
TNTC
10⫺4
10⫺5
10⫺6
1 2 0 5 5
2 0 1 0 0
0 0 0 0 0
150
3
4
Mean count of the culture system at a 10⫺3 dilution ⫽ 20.4. Ratio to counts on MacConkey agar plates ⫽ 1:73.5. b TNTC, too numerous to count. a
1:6.2; and K. pneumoniae, 1:127. C. albicans penetrated very poorly through Whatman no. 3 paper (dilution factor of about 1:600). Up to about 100 colonies could be counted on the HA filters. E. coli produced small, relatively discrete, red colonies. Other members of the family Enterobacteriaceae tended to form larger colonies. S. aureus and S. epidermidis produced characteristic orange colonies and white colonies, respectively. Streptococci formed tiny “salt- and pepper-like” colonies. Penetration of vaginal epithelial cells through Whatman no. 3 filter paper. The counts of squamous epithelial vaginal cells of the unfiltered suspensions were 69.2, 62.6, and 32.2 per ⫻100 field for the three women, respectively. Some of the preparations contained 0 to 4 leukocytes per field and occasional colonies of bacteria. In contrast, there were virtually no intact vaginal squamous epithelial cells, leukocytes, or bacterial colonies in the Whatman no. 3 filtrates. Only a rare cell was seen on the entire slide. The filtrates contained small numbers of red-stained nonnucleated cell fragments. These were about a quarter of the size of the intact cells and numbered 0.8, 4.0, and 4.7 per ⫻100 field for the three women, respectively. Thus, filtration was shown to remove 94% of the epithelial cells. Comparison of the filter paper dilution system and streak plate method. An inoculum of 0.1 ml was selected to achieve optimum spread, and Whatman no. 3 filter paper was used to achieve maximum dilution. The filtration component was removed 15 to 60 min after incubation at room temperature. About half of the systems contained fresh media. The other half consisted of dehydrated plates stored at room temperature for up to 6 weeks. Two milliliters of sterile water was added to the dehydrated plates prior to use. The data from the fresh and dehydrated systems were pooled for analysis, since identical results were obtained. The results for gram-negative and grampositive bacteria are shown in Tables 2 and 3, respectively. The sensitivity and specificity of the filter paper dilution system correlated well with the clinical diagnostic microbiology laboratory at a positive cutoff point of ⱖ104 CFU/ml. Low-count gram-negative bacteriuria (⬍104 CFU/ml) determined by the streak method was detected in five of nine specimens (55.6%) by the filter paper dilution system (as one or more colonies). Mixed cultures (⬍104 CFU/ml) by the streak method were sterile in the filter paper dilution system in 29 of 41 specimens (70.7%). The remaining 12 of 41 (29.3%) contained one or more colonies of gram-negative bacteria. Lactobacillus and Gardnerella at counts of 104 to 105 CFU/ml by the streak method were not detected by the filter paper dilution system. Only 3 of 8 Candida sp. isolates at counts of ⬎104 CFU/ml were detected by the filter paper system. These ob-
1189
Streak plate category
No. of specimens
No. (%) of cultures with result by filter paper dilution system Negativeb
Negative cultures (⬍10,000 CFU/ml)
Positivec
431
420 (97.4)
11 (2.6)
Positive cultures ⱖ10,000–50,000 CFU/ml ⱖ50,000–100,000 CFU/ml ⱖ100,000 CFU/ml
18 15 23
1 (5.6) 0 (0.0) 0 (0.0)
17 (94.5) 15 (100.0) 23 (100.0)
Total positive
56
1 (1.8)
55 (98.2)
d
a
The sensitivity and specificity were 98.2% and 97.4%, respectively. The positive and negative predictive values were 83.3% and 99.8%, respectively. b No growth or ⬍10 CFU. c Growth of ⱖ10 CFU to too numerous to count; 54 of 55 cultures (98.2%) were too numerous to count. d E. coli (n ⫽ 27), Klebsiella and Enterobacter spp. (n ⫽ 14), Proteus, Morganella, and Providencia spp. (n ⫽ 7), Pseudomonas aeruginosa (n ⫽ 7), Serratia marcescens (n ⫽ 1), Citrobacter sp. (n ⫽ 1), Acinetobacter baumanii (n ⫽ 1), or unspecified gram-negative rod (n ⫽ 5).
servations suggest that the filter removed cell-associated bacteria and yeasts usually did not penetrate the filter well. The filter paper system appeared to be able to isolate gram-negative bacteria from mixed cultures. Although the gram-negative and gram-positive filter paper sets tended to be selective for their respective groups, several species of gram-negative bacteria were isolated in both systems when used at the same time. These included Klebsiella and Enterobacter spp. (6 of 14); Pseudomonas sp. (2 of 7); and Proteus, Providencia, or Morganella (6 of 7). This was presumed to be due to resistance to the concentrations of nalidixic acid and colistin contained in the Columbia broth. A positive test in both filter paper sets suggests the presence of these microorganisms. Fifty-four fresh urine specimens, without preservative, were cultured directly on the filter paper system. The tests were
TABLE 3. Comparison of 404 urine cultures for gram-positive bacteria performed with the filter paper dilution system versus the streak plate method in the diagnostic microbiology laboratorya
Streak plate category
Negative cultures (⬍10,000 CFU/ml)
No. of specimens
370
No. (%) of cultures with result by filter paper dilution system Negativeb
Positivec
367 (99.2)
3 (0.8)
d
Positive cultures ⱖ10,000–50,000 CFU/ml ⱖ50,000–100,000 CFU/ml ⱖ100,000 CFU/ml
16 13 5
1e (6.3) 15 (93.7) 2e (15.4) 11 (84.6) 0 (0.0) 5 (100.0)
Total positive
34
3e (8.8)
a
31 (91.2)
The sensitivity and specificity were 91.2% and 99.2%, respectively. The positive and negative predictive values were 91.2% and 99.2%, respectively. b No growth or ⬍10 CFU. c Growth at ⱖ10 CFU to too numerous to count; 25 of 31 cultures (80.6%) were too numerous to count. d Includes 24 isolates of E. faecalis, E. faecium, Enterococcus-like and Streptococcus agalactiae; 4 isolates of S. aureus; and 6 isolates of coagulase-negative staphylococci. e Streptococci.
1190
KUNIN AND BUESCHING
performed and interpreted as positive or negative by a medical secretary who had no prior microbiology training. The system detected five of five E. coli isolates and one of one each of Proteus mirabilis and S. aureus isolates, as well as a coagulasenegative Staphylococcus isolate at counts of ⱖ104 CFU/ml. E. faecalis was not detected in one specimen. DISCUSSION The novel features of the filter paper dilution urine culture system are its ability to dilute and spread the inoculum and its ability to remove cellular elements. It overcomes the problem of contamination of female-voided urine by cell-associated, commensal microorganisms (3, 4). The magnitude of dilution is not great—about 8- to 75-fold with E. coli with a Whatman no. 3 filter. A 10-fold-lower final concentration is achieved by using an inoculum of 0.1 ml. Exact quantitation is not possible at counts of ⱖ104 CFU/ml. Nevertheless, the system provides clinically meaningful results for rapidly growing bacteria equivalent to those achieved by the standard streak method. The system has several advantages over current urine culture screening methods (4). These include (i) the ability to detect low-count bacteriuria, even in the presence of mixed cultures; (ii) inexpensive construction; (iii) ease of use; and (iv) excellent storage characteristics. The system should reduce the need for repeat cultures by eliminating low-count contaminants and the costs of inappropriate therapy by eliminating false positives. The differential medium allows preliminary identification. Colonies can be isolated from cultures containing single microorganisms for formal identification and susceptibility tests. The system can be inoculated at the point of collection. Point-of-source collection is controversial (5), but has several advantages. These include elimination of refrigeration or preservatives, decreased workload for laboratory personnel, and faster turnaround times. The intact units can be transported to the laboratory from the point of source in hospitals and clinics several hours after inoculation. If there is a long delay in transport, the filtration device can be removed at the site of collection. Preliminary results can be obtained after suitable incubation in clinics and the field. It should be possible to discard the negatives and transport only the positives to the diagnostic microbiology laboratory for further identification and susceptibility testing. The units can also be used for selftesting by patients with recurrent urinary tract infections and for screening. The drawbacks for use in the field include the need to inoculate urine specimens soon after collection unless refrigerated and the need to remove the filtration component if there is a long delay prior to transport, as well as issues of quality control. The disadvantages of the filter paper system for use in hospital diagnostic laboratories are (i) the difficulty in isolating individual colonies from mixed cultures, (ii) the in-
J. CLIN. MICROBIOL.
ability to provide quantitative end points, (iii) the poor sensitivity for yeasts, and (iv) uncertainty about its ability to detect fastidious microorganisms. The prototype device described in this report needs to be improved before it is suitable for use in the field or can replace standard methods in hospital diagnostic laboratories. Additional features might include a more selective filtration component; improved media; heavier plastic construction; a domeshaped, twistoff screw top; an inoculation port at the top of the dome; supporting threads in the filtration component to prevent tears; and a 0.1-ml inoculator. We are currently exploring the use of filter papers of different sizes and porosities to inoculate differential agar culture plates. The system might also be adapted for automation. Multiple syringes could be used to aspirate urine from trays of collection tubes and apply a fixed inoculum to the culture systems. After penetration is complete, the filtration component could be mechanically removed at the tab. It might be possible to automate direct susceptibility testing by devising a system with multiple syringes to inoculate urine on multiple small filter papers impregnated with antimicrobial drugs. The end point would be growth or no growth under the papers on agar plates. In summary, we have described a novel system for urine cultures based on filtration methods. The system has the potential to reduce costs of screening, avoid contamination, and provide more rapid results. It can be adapted for point-ofsource and automated cultures. ACKNOWLEDGMENTS This study was supported by the Leadership Council of the Ohio State University Medical Center. We are grateful to Hua Hua Tong and Karen Blackburn for technical assistance. REFERENCES 1. Clarridge, J. E., J. R. Johnson, and M. T. Pezzlo. 1998. Cumitech 2B, Laboratory diagnosis of urinary tract infections. Coordinating ed., A.S. Weissfeld, American Society for Microbiology, Washington, D.C. 2. Eisenstadt, J., and J. A. Washington. 1996. Diagnostic microbiology for bacteria and yeasts causing urinary tract infections, p. 29–66. In H. L. T. Mobley and J. W. Warren (ed.), Urinary tract infections: molecular pathogenesis and clinical management. American Society for Microbiology, Washington, D.C. 3. Kunin, C. M., L. V. White, and H. H. Tong. 1993. A reassessment of the importance of “low-count” bacteriuria in young women with acute urinary symptoms. Ann. Intern. Med. 119:454–460. 4. Kunin, C. M. 1997. Urinary tract infections. Detection, prevention and management, 5th ed. Williams & Wilkins, Baltimore, Md. 5. Robinson, A., M. Marcon, J. E. Mortensen, Y. S. McCarter, M. LaRocco, L. R. Peterson, and R. B. Thomson, Jr. 1999. Controversies affecting the future practice of clinical microbiology. J. Clin. Microbiol. 37:883–889. 6. Sackett, D. L., R. B. Haynes, and P. Tugwell. 1985. Clinical epidemiology. A basic science for clinical medicine, p. 71–73. Little Brown & Co., Boston, Mass. 7. Stamm, W. E., G. W. Counts, K. R. Running, S. Fihn, M. Turck, and K. K. Holmes. 1982. Diagnosis of coliform infection in acutely dysuric women. N. Engl. J. Med. 307:463–468.
