DIAGN MICROBIOLINFECTDIS 1991;14:191-194
Latex Agglutination Detection of Group-B Streptococcal Inoculum in Urine Terrance J. Zuerlein, Barbara Christensen, and Robert T. Hall
It is not uncommon for a newborn to be blood culture negative, but urine latex particle agglutination (LPA) positive for group-B streptococcus (GBS). We hypothesized that whole-cell GBS contamination of urine, which is possible in bag-collected specimens, could cause LPA positivity. Serial tenfold dilutions through 10~ of an 18-hr broth suspension of GBS serotype III were prepared in sterile urine. Directigen (Becton-Dickenson Microbiologic Systems, Cockeyville, MD) LPA testing was performed on each sample at inoculation and, if negative, every 2 hr until becoming LPA positive. All tubes were quantitatively cultured when becoming LPA positive. The assay was performed four times to evaluate independently the effects of urine sample centrifugation, volume, glucose content, and
incubation temperature. An 18-hr GBS suspension dilution of 1:10,000 in urine and as few as 5.7 x 103 organisms/ml caused LPA positivity at time O. Urine supported sufficient replication to enable log 0 concentrations (10~ dilution) to become LPA positive within 8 hr at 35°C, a routine infant incubator setting. Alteration by centrifugation, volume, or glucose content had no effect on LPA positivity or time needed to become positive. We conclude that Directigen LPA can detect a relatively small concentration of GBS organisms in urine. Due to potential contamination from skin bacteria, positive LPA resuits from bag urine should be interpreted with caution in the absence of positive cultures.
Numerous rapid diagnostic tests have been developed in an effort to diagnose invasive GBS disease more rapidly. These include counterimmunoelectrophoresis, staphylococcal protein-A coagglutination, and group antigen-specific latex particle agglutination (LPA). LPA has been determined to be the most rapid and sensitive of these diagnostic tests, with a sensitivity of 90%-100% and a specificity 81%-98.7% (Friedman et al. 1984, Rabalais et al. 1987; Rench et al., 1984). Anti-GBS antibodies have been shown to detect as little as 31.25 ng/ml of purified GBS polysaccharide antigen in serial dilution assays (Rench et al. 1984). Urine has proven to be the most sensitive and dependable source of GBS antigen in numerous studies evaluating various body fluids of GBS-infected infants (Rench et al., 1984; Bromberger et al., 1980; Hamoudi et al., 1983; Ingram et al., 1982). Many infants have mucocutaneous colonization with GBS, and urine specimens are frequently obtained by bag collection. Conflicting opinions have
Group-B streptococci (GBS) are significant bacterial pathogens in the neonatal population. The incidence of genital tract or rectal GBS colonization in pregnant women has been reported to be 5%-30% (Baker, 1979), and 65%-72% of infants born to these mothers will acquire mucocutaneous colonization with the organism (Baker, 1979). Approximately 1% of colonized infants will develop early onset GBS disease and many of these die in spite of antibiotic therapy (Baker, 1979). From the Department of Neonatology(T.J.Z., R.T.H.) and the MicrobiologyLaboratory(B.C.), Children'sMercyHospital, University of Missouri--Kansas City School of Medicine, Kansas City, Missouri, USA. Address reprint requests to Dr. T.J. Zuerlein, Children'sMercy Hospital, 24th and Gfllham,Kansas City, MO 64108, USA. Received 5 March 1990; revised and accepted 27 July 1990. © 1991 ElsevierScience PublishingCo., Inc. 655 Avenue of the Americas, New York, NY 10010 0732-8893/91/$3.50
been reported (Friedman et al., 1984; Rabalais et al., 1987; Hamoudi et al., 1983) regarding the ability of whole-cell GBS contamination in urine to cause LPA positivity. LPA-kit instruction manuals do not make any recommendation concerning the method of urine collection for LPA testing or comment on the issue of specimen contamination. The purpose of the present study was to evaluate the sensitivity of a LPA system (Directigen, Becton-Dickenson Microbiologic Systems, Cockeyville, MD) to detect a GBS organism inoculum in urine under varying sample conditions.
METHODS GBS serotype III was kindly supplied by Richard Facklam (Centers for Disease Control, Atlanta, GA). The lyophilized bacteria were rehydrated with SeptiChek tryptic soy broth (TSB; Roche Diagnostic Systems, Nutley, NJ), plated on 5% sheep blood agar, and incubated overnight at 35°C. The following morning B-hemolytic colonies with typical GBS morphology were present and confirmed as GBS by Phadebact coagglutination (Pharmacia, Piscataway, NJ). Three colony-forming-units (CFU) were then inoculated into 50 ml of tryptic soy broth and allowed to incubate at 35°C overnight, yielding a high-integer log7/ml to low-integer logs/ml concentration. The following morning, at time 0, serial tenfold dilutions through 104 of the original suspension were prepared by inoculating 0.4 ml of organism suspension into 3.6 ml of sterile donor urine. The anticipated organism concentration was therefore 107-100 in a volume of 4 ml. Donor urine was analyzed by urinalysis and culture to document sterility. To legitimize the use of donor urine for this study, donor urinalyses were compared with those of 10 randomly selected newborn urine specimens by unpaired t test for continuous variables and the Fisher test for categorical variables. For determination of the number of CFU in broth dilutions at time 0, the 10-4 and 10-5 dilutions were quantitatively plated in triplicate on 5% sheep blood agar. Colony counts on each plate were performed and averaged to obtain the number of CFU/ml present. Directigen LPA testing was performed on all urine-organism dilutions at time 0. Dilution tubes not agglutinating at time 0 were allowed to incubate and were retested every 2 hr until positive or until 24 hr had passed. When the LPA became positive, the dilution tube suspension was cultured quantitatively in triplicate. This enabled determination of the number of CFU/ml necessary to obtain a positive LPA. All LPA testing was performed by one of the investigators (T.J.Z.) to ensure uniformity of proce-
T.J. Zuerlein et al.
dure and interpretation. A second investigator (B.C.) blinded to the dilutional status of the specimen confirmed LPA and quantitative culture results. Any degree of macroscopic agglutination was considered positive, as per kit instructions. Positive and negative GBS antigen controls were performed each day of the study to verify LPA kit sensitivity and specificity. In addition, latex controls were performed to detect the presence of nonspecific agglutination. All specimens were heated according to LPA kit instructions to eliminate nonspecific agglutination. The study procedure was repeated in four groups as shown in Table 1 in order to evaluate independently the effects of centrifugation (I), urine sample volume (II), glucose content (III), and incubation temperature (IV). Group II-A consisted of serial dilutions beginning with 0.2 rnl of overnight GBS growth into 1.8 ml donor urine, whereas group II-B consisted of 0.4 ml into 3.6 ml donor urine. Donor urine glucose content in group III-A was 10 mg/dl by Kodak Ektachem 700 Analyzer. Urine glucose concentration in group III-B was modified to 100 mg/dl by addition of D10W to assimilate the occasionally encountered Dextrostick-positive infant urine specimen. Specimens in group IV-A were incubated at 35°C, a routine temperature setting for neonates on open bed warmers or in isolettes. Group IV-B incubated at 27°C (room temperature) to evaluate the effect of lack of refrigeration on urine specimens with delay in LPA testing. Group IV-C "incubated" in refrigeration at 5°C.
RESULTS Donor urine used for serial dilutions was found to be negative for blood, ketones, glucose, and protein. Mean specific gravity was 1.010 ( _+0.002 SD) and pH was 7.0 (0 SD). Sheep blood agar and EMB cultures were negative in all specimens at 24 and 48 hr. Comparison of the four donor urinalyses with 10 infant urinalyses revealed no significant differences. Concordance between unblinded and blinded investigators regarding LPA interpretation and quantitative culture results was 100%. The 10-1-10~ (containing 107-104 organisms) dilution tubes were LPA positive at time 0 under all study conditions. The mean lowest concentration yielding a positive LPA at time 0 was 2.1 x 104 CFU/ml. The mean highest concentration yielding a negative LPA at time 0 was 2.5 x 103 CFU/ml. The threshold for LPA positivity for whole-GBS organisms using Directigen, therefore, lies between these two values. In support of this, the mean number of CFU in dilution tubes 10-5-10~ at the time of becoming LPA positive during incubation was 1.0 x 10 4 .
The effect of varying processing and incubation
Latex Agglutination of Group-B Steptococcus
TABLE 1 Nu m ber of Hours Required for Serial Dilutions to Become LPA Positive Under Varying Processing and Incubation Conditions Procedure Variables I. Centrifugation A. Group Centrifugation (-+) Volume (ml) Glucose (rag%) Temperature (°C) B. Dilution 10-4
104 10 3
102 101 10o
III. Glucose Content
IV. Incubation Temperature
+ 4 10 35
4 10 35
+ 2 10 35
+ + 4 4 10 100 35 35
+ 4 10 35
+ 4 10 27
+ 4 10 5
II. Urine Volume
+ 4 10 35
C. Hours Required to Become LPA Positive 0 6 8 8 8
0 6 8 8 8
0 6 8 8 8
0 6 8 8 8
0 4 6 6 8
0 4 6 6 8
0 6 8 10 10
0 10 12 16 18
0 b b b b
~Per milliliter. bRemained LPA negative at 24 hr.
conditions on the num be r of hours required to become LPA positive is shown in Table 1.
Centrifugation There was no difference in LPA results between unspun and spun supernatant urine samples. All urine-GBS suspensions became LPA positive within 6-8 hr of incubation.
Urine Volume When GBS where inoculated into 2 versus 4 ml at the same concentration, incubation time required to become LPA positive was not different. Thus, no volume-independent effect was observed, and all samples were LPA positive within 6-8 hr of incubation.
Urine Glucose Content The supply of additional glucose substrate did not enable whole-organism suspensions to replicate and attain a "critical concentration" for LPA positivity any sooner. All urine samples were LPA positive within 4-8 hr.
Incubation Temperature Aliquots incubated at 35°C became LPA positive within 6-10 hr. Urine samples incubated at 27°C had sufficient GBS replication to become LPA positive at the times shown in Table 1. Incubation at 5°C ar-
rested replication and did not allow any of the 10-s10-a dilutions to become positive within 24 hr.
DISCUSSION Previously published work addressing the issue of whole-GBS organism detection by LPA has been marked by conflicting statements. Several authors reported that LPA testing may be limited because of inability to differentiate between colonization and infection (Friedman et al., 1984; H o p p e et al., 1986). Others indicate that LPA detects antigen rather than living organisms (Hamoudi et al., 1983), and is sensitive only in infants w ho are diseased and have sufficient antigen in their body fluids for detection (Lim et al., 1986). Rabalais et al. (1987) state that the explanation for occurrence of detectable antigen in the absence of proven infection is unclear. In the present study using anti-GBS antibodies and urine-GBS suspensions, LPA positivity occurred without acid or enzymatic extraction of carbohydrate antigen from the bacterial cell wall. Routine heating of the urine specimen to 100°C for 5 min as recom m ended by the manufacturer to avoid nonspecific agglutination apparently liberated or unmasked sufficient group-specific antigen for LPA interaction. Furthermore, urine appeared to be an excellent culture media for GBS. Donor urine supported sufficient replication to allow log0 concentrations (1-9 organisms) to become LPA positive within 8 hr at routine infant incubator or warmer settings. Numerous clinical trials have evaluated LPA performance characteristics in infants with culture-pos-
itive GBS infection, infections with other bacteria, or no infection (Friedman et al., 1984; Rabalais et al., 1987; Rench et al., 1984; Bromberger et al., 1980; Hamoudi et al., 1983; Ingram et al., 1982; Baker and Rench, 1983). Explanations for false-positive LPAs include GBS colonization, but this was not documented (Hamoudi et al., 1983). Friedman et al. (1984) identified a number of infants with urine LPA false positives who had GBS isolated from rectal, gastric, or skin cultures. None of these clinical trials listed above included a description of how urine specimens were obtained except for one author noting that all three of the false positives in her series were collected by bag (Hamoudi et al., 1983).
T.J. Zuerlein et al.
positive within a surprisingly short period of time. Bacterial contamination of bag-obtained urine samples is a well-recognized risk (Cole and Cloherty, 1986). Urine collection bags are often left in place for prolonged periods of time while waiting for the infant to urinate a "sufficient" volume. Presence of urine in the bag is not always immediately noted, especially in the prone, nursed infant. This may allow contaminant GBS to incubate and replicate prior to LPA performance, potentially leading to LPA positivity. The possibility of LPA positivity secondary to GBS contamination may limit the utility of LPA to discriminate invasive GBS disease in the neonatal sepsis evaluation.
IMPLICATIONS As we have shown, a very small GBS inoculum is capable of causing urine samples to become LPA
This work was supported by a grant from the William Randolph Hearst Foundation.
REFERENCES Baker CJ (1979) Group B streptococcal infections in neonates. Pediatr Rev 1:5-14. Friedman CA, Wender DF, Rawson JE (1984) Rapid diagnosis of group B streptococcal infection utilizing a commercially available latex agglutination assay. Pediatrics 78:27-30. Rabalais GP, Bronfin DR, Daum RS (1987) Evaluation of a commerciallyavailable latex agglutination test for rapid diagnosis of group B streptococcal infection. Pediatr Infect Dis 6:177-181. Rench MA, Metzger TG, Baker CJ (1984) Detection of Group B streptococcal antigen in body fluids by a latex-coupled monoclonal antibody assay. J Clin Microbio120:852854. Bromberger PI, Chandler B, Gezon H, Haddow JE (1980) Rapid detection of neonatal group B streptococcal infections by latex agglutination. ] Pediatr 96:104-106. Hamoudi AC, Marcon MJ, Cannon HJ, McClead RE (1983) Comparison of three major antigen detection methods
for the diagnosis of group B streptococcal sepsis in neonates. Pediatr Infect Dis 2:432-435. Ingram DL, Suggs DM, Pearson AW (1982) Detection of group B streptococcal antigen in early-onset and lateonset group B streptococcal disease with the Wellcogen Strep B latex agglutination test. J Clin Microbiol 16:656658. Hoppe JE, Grieshaber J, Hofler W (1986) Colonization of Nigerian neonates with group B streptococci and its rapid detection. Infection 14:30-34. Lira DV, Morales WJ, Walsh AF, Kazanis D (1986) Reduction of morbidity and mortality rates for neonatal group B streptococcal disease through early diagnosis and chemoprophylaxis. J Clin Microbiol 23:489-492. Baker CJ, Rench MA (1983) Commercial latex agglutination for detection of group B streptococcal antigen in body fluids. J Pediatr 102:393-3952. Cole FS, Cloherty JP (1986) Infection: prevention and treatment. In Manual of Neonatal Care, vol 8. Ed, JP Cloherty. Boston: Little, Brown and Company, p 147.