- Email: [email protected]
Inadequacy of rapid immunoassays for intrapartum detection of group B streptococcal carriers
Inadequacy of Rapid Immunoassays for Intrapartum Detection of Group B Streptococcal Carriers CAROL 1. BAKER, MD Objective: To determine the accuracy of two currently used immunoassays and a newly developed optical immunoassay for rapid intrapartum detection of group B streptococcal colonization compared with culture methods. Methods: Rayon-tipped swabs were used to collect specimens from the distal vagina of 502 women at admission for labor or rupture of membranes. Four tests were performed on specimens from the first 197 patients: culture in selective broth medium, semiquantitative culture on blood agar medium, and ICON Strep B and Quidel Group B Strep Test enzyme immunoassays. For the remaining 305 women, a fifth test, Strep B OIA, a newly developed optical immunoassay, was also performed. Results: The prevalence of group B streptococcal vaginal colonization was 25.1% by selective broth medium and 17.3% when swabs were plated directly onto blood agar medium, giving the latter method a sensitivity of 69%. When compared with selective broth medium results, the sensitivities of the rapid immunoassays were 12% (Quidel), 15% (ICON), and 37% (Strep B OIA). These values rose to 16% (Quidel), 21% (ICON), and 53% (Strep B OIA) when compared with nonselective blood agar medium results. For women with heavy group B streptococcal colonization (more than lo6 colony forming units/ml), the sensitivities were 36% (Quidel), 46% (ICON), and 100% (Strep B OIA). Specificities for all assays were high (98-lOO%), but variability was found in positive (79-100%) and negative (77-85%) predictive values. Conclusion: Although Strep B OIA reliably detects women with heavy group B streptococcal colonization and is more sensitive than either the ICON or Quidel enzyme immunoassays, none of these rapid assays is sufficiently accurate for
From the Departments of Pediatrics and Microbiology and Immunology, Baylor College of Medicine, Houston, Texas. Supported by U.S. Public Health Service contract uo. AI 25152 from the National lnstitufe of Allergy and Infectious Diseases. Biostar, Inc., furnished the Strep B OlA devices and the training of personnel. The content of this publication does not necessarily reflect the viezos or policies the Department Health and Human Services, nor does mention trade names, commercial products, or organizations imply endorsement by the U.S. governmeut.
of of
VOL.
88, NO.
of
1, JULY
1996
routine use in the intrapartum detection of women colonized with group B streptococcus. (Obstet Gynecol 2996;88:52-51
Group B streptococci are a prominent causeof perinatal infection.’ Active surveillance in a population of ten million indicates that neonatal sepsis due to group B streptococcus occurs in 1.8 per 1000 live births in the United States, with a casefatality ratio of 5%.* Among survivors who have meningitis, at least one-third have permanent neurodevelopmental deficits. In pregnant women, group B streptococcus causes a spectrum of infections that range from asymptomatic bacteriuria to acute and fatal endocarditis, but the most typical clinical picture is that of chorioamnionitis or early postpartum febrile morbidity, often associated with bacteremia.3 More than a decade ago, the economic burden of perinatal group B streptococcus was estimated at $727 million4 Both ACOG and the American Academy of Pediatrics5 agree that group B streptococcal diseaseprevention is desirable and that intrapartum chemoprophylaxis of women at risk for delivery of group B streptococcus-infected infants significantly reduces maternal febrile morbidity and early-onset group B streptococcal neonatal sepsis. To date, all trials evaluating the efficacy of maternal prophylaxis in preventing early onset group B streptococcal sepsis in neonates have been based on antenatal detection of maternal carriers.6 However, antenatal group B streptococcal screening highlights three practical concerns: the accuracy of antenatal cultures in detecting women carrying group B streptococci at delivery, the logistics of providing cul-
Financial
Disclosure
Dr. Baker in 1994.
was a consultant
with
Biostar
once in 1993 and once
0029.7844/96/$15.00 PII SOO29-7844(96)00111-l
51
ture results to the delivery site, and the management of women with no or incomplete prenatal care.7 These concerns would be obviated if a rapid and reliable method were available to identify group B streptococcal carriers at hospital admission for delivery. An ideal test for the detection of group B streptococcus at hospital admission would be rapid, accurate, and inexpensive. Many candidate methods have been reviewed by Yancy et a1.8 Although the most accurate of these is the enzyme immunoassay for detection of group B streptococcal antigen in vaginal swab specimens, most studies have compared enzyme immunoassay sensitivity with culture methods insufficient to detect women with a low inoculum of group B streptococci.’ Because these parturients give birth to approximately 15% of the neonates with early-onset group B streptococcal infection,6 their identification is important in developing intrapartum prevention strategies that will have the greatest impact. Recently, a new method for rapid group B streptococcal antigen detection in vaginal swabs has been developed and is being marketed as Strep B OIA test (Biostar Inc., Boulder, CO). The principle of this optical immunoassay is that the reflection of light changes when an antigen binds to an antibody on a reflecting surface, and this can be visualized directly. The optical immunoassay was devised to be more sensitive than enzyme immunoassays, and as such has the potential of meeting the criteria for rapid group B streptococcal detection mentioned above. The purpose of this study was to compare three commercially available assays for the rapid detection of group B streptococcus in vaginal swab specimens with results of cultures performed by two different methods to determine their potential usefulness for detection of group B streptococcal carriers at hospital admission for labor or rupture of membranes. To do so, we determined the sensitivity, specificity, and positive and negative predictive values for each assay. We compared these values with a commonly used culture method, direct plating of swabs onto blood agar medium, and the method advocated by the American Academy of Pediatric@ and the Centers for Disease Control and Prevention,7 in which swabs are incubated in selective broth medium before subculture onto blood agar medium.
Materials and Methods Between November 1, 1993, and September 28, 1994, 502 women evaluated for admission to Ben Taub General Hospital in Houston, Texas, for labor or rupture of membranes who had not received antibiotics or undergone pelvic examination were eligible for study. After
52
Baker
Antip
Assays
verbal consent, women were enrolled consecutively between 8 AM and 3 PM on Mondays, Tuesdays, and Wednesdays. Demographic data and gravidity, parity, gestation, time of membrane rupture (if present), and time vaginal swabs were collected were obtained from the medical records by a research nurse on a standard form. For the first 197 women, three rayon swabs (Becton Dickinson Microbiology Systems, Cockeysville, MD) were rotated 360” in the distal vagina to obtain specimens. When the Strep B OIA became available for study, an additional swab was obtained from the remaining 305 women. Three swabs were placed into modified Stuart’s transport media (Becton Dickinson) and the remaining one was placed in Amies transport medium without charcoal (Cultureswab Transport System, Difco, Surrey, UK). Swabs were delivered to the research laboratory and processed within 6 hours of collection. The swab in Amies transport medium was used to inoculate a 5% sheep blood agar (trypticase soy agar base) plate using a semiquantitative, four-zone streaking method.’ Growth of beta-hemolytic colonies was recorded as l+ to 4+ (l+ = 102-103, 2+ = >103-105, 3+ = >105-109 4+ = >106 colony forming units [cfu] per milliliter). One to 2+ indicated light and 3-4+ heavy group B streptococcal colonization. This swab was then placed into selective Todd-Hewitt broth (BBL; Becton Dickinson) containing 8 pg / mL gentamicin and 15 pg/mL nalidixic acid, incubated at 35C overnight, and subcultured onto a 5% sheep blood agar plate. Plates were incubated overnight at 35C, removed, and inspected for beta-hemolytic and nonhemolytic colonies with morphology typical of group B streptococci. The identification of group B streptococcus was confirmed by a positive reaction with the Streptex grouping kit reagent (Murex Diagnostics Limited, Dartford, UK). Each swab transported in modified Stuart’s medium was used for individual immunoassay testing. Kits for the two enzyme immunoassays were purchased from the manufacturers. All tests were performed at room temperature as directed by the manufacturers’ instructions. The first enzyme immunoassay, Quidel Group B Strep Test (San Diego, CA), consists of a Quidel test cartridge that contains a filtration membrane coated with rabbit polyclonal antibodies to group B streptococci. Group B streptococcal antigen was extracted from the swab with reagent provided by the manufacturer; the extracted fluid sample was then transferred to the test cartridge and gravity-filtered through the membrane. If group B streptococcal antigen was present, it bound to antibodies on the surface of the membrane. Reagent containing polyclonal anti-group B streptococcal antibody enzyme conjugate was then added to the cartridge and allowed to incubate for 1 minute. After
Obstetrics 6 Gymcology
rinsing the membrane to remove unbound antibodyenzyme conjugate, an enzyme substrate solution was added to the cartridge and allowed to react for 2 minutes. If group B streptococcal antigen was present, a solid blue circle appeared on the cartridge membrane, indicating a positive result. A positive group B streptococcal control was performed on each assay day. The entire procedure required less than 10 minutes. The second enzyme immunoassay, ICON Group B Strep (Hybritech, San Diego, CA), provided a plastic vial in which a swab was placed and extraction reagents were added. After incubation for 1 minute, five drops of a second extraction reagent and one drop of group B streptococcal polyclonal antibody-conjugate were added and mixed thoroughly. The swab was then pressed against the wall of the vial to express as much fluid as possible, and the sample was passed through a pre-filter onto the center of the immunoassay filter cylinder after 2 minutes. The membrane of the cylinder was washed and incubated for 2 minutes. After another wash, the immunoassay was read as positive if a purple circle appeared in the center of the membrane. Positive and negative controls were impregnated on each immunoassay filter membrane. The assay usually required 10 minutes for completion. For the Strep B OIA, a swab was placed into a tube with the first extraction reagent, then a second extraction reagent for sequential incubation for up 5 minutes. After neutralization, the swab was pressed against the tube to extract as much fluid as possible, then discarded. Group B streptococcal antibody conjugate was added to the solution, the solution was incubated for 3-5 minutes, and a few drops were pipetted onto the surface of the Strep B OIA cartridge device. After 10 minutes, the surface was washed and blotted, and substrate was added. After an additional 10 minutes of incubation, the surface was washed again, blotted, and inspected for a large blue to purple circle on the gold surface, which indicated that group B streptococcal antigen was present in the specimen. A group B streptococcal positive antigen control was provided and performed for each assay day. The Strep B OIA procedure usually took 35 minutes to complete. Both culture methods and the two enzyme immunoassays were performed on vaginal swabs from each of the 502 patients; the final 305 women also had swabs tested by the optical immunoassay method, Strep B OIA. The sensitivity, specificity, and positive and negative predictive values for the immunoassays were calculated as compared with each of two culture methods. A further analysis for patients heavily colonized with group B streptococcus also was performed. Comparison between groups was analyzed by uncorrected 2 test; P < .05 was considered significant.
VOL.
88, NO.
1, JULY
1996
Table
1. Sensitivity, Specificity, and Predictive Values of Three Rapid Immunoassays for Group B Streptococcus Assessed by Direct Blood Agar Plate Culture Method Assav method Quidel
Sensitivity
16.1%
EIA
[8.5-23.71
ICON 20.7%
Positive predictive value Negative predictive value
98.8% (4101415) 73.7%
[53.9-93.51
[81.8-881
(4101483)
53.2%
(18187) 94.7%
[84.7-1001
86.2%
W/19) 85.9%
[82.8-891
B OIA* [38.9-67.51
(25147) 98.4% (254/258)
99.8% (4141415)
(14/19) 84.9%
Strep
[12.3-29.11
(14187) Specificity
EIA
[73.7-98.71 (25129)
92.0%
(415/483)
[88.9-95.11 (2541276)
EIA = enzyme immunoassay; OIA = optical immunoassay. * The final 305 of the 502 patients were tested (see Materials and Methods). Data are presented as % [95% confidence interval], with the proportion underneath.
Red ts Among the 502 study participants, there were 22 (4.4%) whites, 86 (17.1%) African-Americans, 385 (76.7%) Hispanics, and 9 (1.8%) Asians. Ninety (17.9%) were less than 20 years of age. One hundred twenty-six (25.1%) had group B streptococci isolated from vaginal swabs processed in selective broth medium, and 87 (17.3%) from swabs plated directly onto 5% blood agar medium. The latter method failed to identify 39 intraparturn group B streptococcal carriers or 7.8% of this population of pregnant women, giving this culture method a sensitivity of 69%. Vaginal group B streptococcal colonization was more prevalent in AfricanAmericans (43 of 86,50%) and less prevalent in Hispanics (74 of 385, 19.2%) when compared with whites (seven of 22, 31.8%), but these differences were not statistically significant. The sensitivity, specificity, and positive and negative predictive values of the three immunoassays compared with culture results from direct plating of swabs onto 5% sheep blood agar media are shown in Table 1. The Strep B OIA was two- to threefold more sensitive (53%) than either of the enzyme immunoassays (Quidel, 16%; ICON, 21%) in detecting group B streptococcal carriers. Each assay had excellent specificity, and the ICON and Strep B OIA tests were more accurate than the Quidel enzyme immunoassay, as indicated by their positive predictive values of 95 and 86%, respectively. When degree of colonization or vaginal group B streptococcal inoculum was considered, all three assays had improved sensitivity for detection of heavily colonized women who had a group B streptococcal inoculum exceeding lo” cfu/mL. Among the 33 heavily
Baker
Antigen
Assnys
53
Table 2. Effect of Bacterial Inoculum” on Sensitivity of the Strep B OIA for Vaginal Swabs Growth
on
No. of
NO.
specimens
plate*
0 1+ 2+ 3+ 4t
OIA-positive
258 14 18 14 1
OIA = opticalimmunoassay. * Swab inoculateddirectlv on 5%
semiquantitative technique.’
Sensitivity
4 1 9 14 1
7.1% 50% 1OO’Y 100;
sheepbloodagarplateusinga
colonized study patients, 12 (36%) and 15 (46%) had positive Quidel or ICON enzyme immunoassays, respectively. Previously, inoculum has been shown to have a strong direct relationship on the sensitivity of these assays,9-1’ and we also found this to be true of the Strep B OIA (Table 2). Each of the heavily colonized women tested by OIA was positive. However, the sensitivity of Strep B OIA assay fell substantially in women with lower group B streptococcal inocula. Next, the results of the three immunoassays were compared with cultures processed in selective broth medium, the recommended method for antenatal group B streptococcal screening.5,6Each method was limited in its ability to identify lightly colonized group B streptococcal carriers (Tables 2 and 3). These women with lower vaginal group B streptococcal inocula have been estimated to deliver 15% of the neonates who develop early-onset group B streptococcal sepsis.1°
Table
3.
Sensitivity, Specificity, and Predictive Valuesof Three Immunoassays for Group B Streptococcus Assessed by Vaginal Cultures Processed in Selective Broth Medium
Assaymethod Quidel Sensitivity Specificity Positive predictive value Negative predictive value
EIA
ICON
EIA
Strep B OIA*
11.9% [6.2-17.61 (15/126) 98.9% (372 / 376) 78.9% [60.5-97.31 (15/19)
15.1% [8.8-21.41 (191126) 100% (375/376) 100% (19/19)
36.8% [25.4-48.23 (25168) 98.3% (235 / 239) 86.2% [73.7-98.71 (25129)
77.0% [73.3-80.71 (3721483)
77.8%) [74.1-81.5) (376 / 483)
84.5% [80.2-88.81 (235/278)
EIA = enzyme immunoassay; OIA = optical immunoassay. * The final 305 of the 502 patients were tested (see Materials and Methods). Data are presented as % [confidence interval], with the proportion underneath.
54
Baker
A~+zyz Assnys
Discussion The findings of this study repeat a theme in the search for rapid intrapartum detection of group B streptococcal carriers. Although relatively sensitive for the identification of pregnant women with moderate or high numbers of vaginal group B streptococci, immunoassays were unable to identify those with light colonization.‘O This shortcoming is most apparent when a vaginal swab is processed in antibiotic-containing or selective broth medium. Armer et all’ reported that the ICON Strep B assayhad an overall sensitivity of 11% for vaginal carriers of group B streptococcus identified using selective broth medium for culture. We also found ICON Strep B to be insensitive (15%), as was the Quidel Group B Strep test (12%). When compared with the commonly used, but lesssensitive,12culture method of directly plating swabs onto 5% sheep blood agar medium, the sensitivity of the immunoassays rose substantially. However, only women with moderate or heavy colonization will be identified by this method.13 Because approximately 15% of neonates with invasive group B streptococcal infection are born to women with lower quantities of group B streptococci, intrapartum identification of women with any level of group B streptococcal colonization is desirable.5,6 The new optical immunoassay, Strep B OIA, appears promising as a rapid diagnostic technique for identification of group B streptococcal antigen in vaginal swab specimens.When an assay based on the sameprinciple asthe Strep B OIA was tested for the detection of group A streptococci in children with pharyngitis, the results were encouraging14,i5: sensitivity and specificity exceeded 90%.14In our patients, the Strep B OIA had an overall sensitivity of 37%, but this rose to 100% for women with heavy group B streptococcal colonization. The disparity in the sensitivity of the Strep A and Strep B OIA is explained by the clinical populations targeted for screening. In children with low inocula of group A streptococci in throat cultures, the sensitivity of the Strep A OIA is poor (13%), but only 7% of children with acute pharyngitis fall into this category.r4 By contrast, nearly 60% of our study patients had low numbers of group B streptococcus in their vaginal cultures. If the clinical goal of a rapid antigen test was detection of heavily group B streptococcus-colonized women only, then Strep B OIA would appear to be an excellent test given its high sensitivity and specificity as well as its reasonably high positive and negative predictive values. Cumulative evidence supports the effectiveness of intrapartum ampicillin or penicillin G prophylaxis in preventing early-onset group B streptococcal sepsisin neonates and early postpartum febrile morbidity in
Ohsfetrics Er Gynecology
their mothers.5,h,‘“,‘h Although detection of colonization by culture during the third trimester is one method by which candidates for prophylaxis may be identified,5,‘6 there will continue to be a need for intrapartum detection in certain populations. Both rapid immunoassays for detection of group B streptococcal antigen in vaginal swab specimens that we studied had poor sensitivity (16-21%) in colonized women at hospital admission. Although the new optical immunoassay, Strep B OIA, improved sensitivity two- to threefold, this was not adequate in women who had light or low inoculum colonization (31%). Their detection would be critical to the design of an intrapartum strategy target that would increase the number of infant cases prevented. However, its 100% sensitivity in detecting heavily colonized women, combined with its reasonable predictive values for all group B streptococcal colonization, suggest that Strep B OIA has potential use in select obstetrical populations when antenatal detection of group B streptococcal carriers by culture is not feasible and a laboratory facility could offer 24 hour a day rapid testing at admission. None of the currently available rapid antigen methods for detection of group B streptococcal carriers appears to be sufficiently sensitive for routine use in the selection of women for intrapartum antibiotic prophylaxis. Because rapid detection of carriers at hospital admission continues to be germane to strategies that would prevent perinatal group B streptococcal infections through intrapartum chemoprophylaxis, improved antigen detection methods should be devised. Meanwhile, detection of carriers for intrapartum chemoprophylaxis must depend on antenatal vaginal and rectal swab specimens cultured in selective broth medium.
References 1. Baker CJ, Edwards MS. Group B streptococcal infections. In: Remington JS, Klein JO, eds. Infectious diseases of the fetus and newborn infant. 4th ed. Philadelphia: WB Saunders, 1995:9801054. 2. Zangwill disease
KM,
surveillance 3. American streptococcal
4. Institute
88, NO.
A, States,
Wenger 1990:
JD. Report
Group from
of
DC:
American
College
1, JULY
New
1996
vaccine
Group bulletin
of Obstetricians
development:
1: Diseases National
5. American Academy group B streptococcal 1992;90:775-8.
of importance in the United Academy Press, 1985:424-39.
States.
of Pediatrics. infection
prevention Pediatrics
Guidelines for by chemoprophylaxis.
Washof
6. Centers for Disease Control and Prevention. Prevention of group B streptococcal disease: A public health perspective. Federal Register 1994;(Dec 15);59(24O)d:647-764-773. 7. Larsen JW, Dooley SL. Group obstetrical 8. Yancey
viewpoint. MK, Armer
B streptococcal
Pediatrics T, Clark
infections:
1993;91:148-9. I’, Duff I’. Assessment
An
of
rapid
identification tests for genital carriage of group B streptococcal. Obstet Gynecol 1992;80:1038-47. 9. Gentry Y, Hillier SL, Eschenbach DA. Evaluation of a rapid enzyme Obstet
immunoassay test for detection Gynecol 1991;78:397-401.
10. Tuppurainen streptococcal
N, Hallman M. disease: Intrapartum
laxis of heavily 583-7. 11. Armer T, Clark
colonized I’, Duff
I’, Saravanos
Goroff DK, Alpert of bacteriological
streptococcus 13. Granato PA,
from Petosa
Prevention detection
parturients.
tion of group B streptococcal noassay. Am J Obstet Gynecol 12. Baker CJ, Comparison
of group of
B streptococcus.
neonatal group and chemoprophy-
Obstet K. Rapid
Gynecol
1989;73:
intrapartum
colonization with 1993;168:39-43.
an enzyme
B
detecimmu-
SL, Hayes C, McCormack WM. methods for the isolation of group B
vaginal cultures. MT. Evaluation
J Clin Microbial 1976;4:46-8. of a rapid screening test
for
detecting group B streptococci in pregnant women. J Clin Microbiol 1991;29:1536-8. 14. Roddey OF Jr, Clegg HW, Martin ES, Swetenburg RL, Koonce EW. Comparison ture methods 15.
of an optical immunoassay for the detection of
group
technique with A streptococci
pediatric office. J Pediatr 1995;126:931-3. Fries SM. Diagnosis of group A streptococcal private method
clinic: and
pharyngitis
Comparative evaluation of an optical culture. J Pediatr 1995;126:933-6.
16. Allen UD, Navas cillin prophylaxis infection: Results
L, King SM. Effectiveness in preventing early-onset of a meta-analysis. Can
two
culin a in
a
immunoassay
of intrapartum penigroup B streptococcal Med Assoc J 1993;149:
1659-65.
Address reprint requests to: Carol J Baker, MD Deyartment qf Pediatrics Baylor School of Medicine One Baylor Plaza, Room 307E Houstou, TX 77030
active
1992. Medicine.
DC:
B streptococcal a multistate
system. MMWR 1992;41:25-32. College of Obstetricians and Gynecologists. infections in pregnancy. ACOG technical
170. Washington, Gynecologists,
VOL.
Schuchat
in the United
priorities, ington,
Establishing
B no. and
Rmived Noven~her 29, 1995. Rrckoed in revised fom ~anunry Accepted February 14, 1996. Copyright Gynecologists.
0
1996 by Published
The American by Elsevier
22, 1996.
College of Science Inc.
Baker
Obstetricians
Antigen Assays
and
55
From the Departments of Pediatrics and Microbiology and Immunology, Baylor College of Medicine, Houston, Texas. Supported by U.S. Public Health Service contract uo. AI 25152 from the National lnstitufe of Allergy and Infectious Diseases. Biostar, Inc., furnished the Strep B OlA devices and the training of personnel. The content of this publication does not necessarily reflect the viezos or policies the Department Health and Human Services, nor does mention trade names, commercial products, or organizations imply endorsement by the U.S. governmeut.
of of
VOL.
88, NO.
of
1, JULY
1996
routine use in the intrapartum detection of women colonized with group B streptococcus. (Obstet Gynecol 2996;88:52-51
Group B streptococci are a prominent causeof perinatal infection.’ Active surveillance in a population of ten million indicates that neonatal sepsis due to group B streptococcus occurs in 1.8 per 1000 live births in the United States, with a casefatality ratio of 5%.* Among survivors who have meningitis, at least one-third have permanent neurodevelopmental deficits. In pregnant women, group B streptococcus causes a spectrum of infections that range from asymptomatic bacteriuria to acute and fatal endocarditis, but the most typical clinical picture is that of chorioamnionitis or early postpartum febrile morbidity, often associated with bacteremia.3 More than a decade ago, the economic burden of perinatal group B streptococcus was estimated at $727 million4 Both ACOG and the American Academy of Pediatrics5 agree that group B streptococcal diseaseprevention is desirable and that intrapartum chemoprophylaxis of women at risk for delivery of group B streptococcus-infected infants significantly reduces maternal febrile morbidity and early-onset group B streptococcal neonatal sepsis. To date, all trials evaluating the efficacy of maternal prophylaxis in preventing early onset group B streptococcal sepsis in neonates have been based on antenatal detection of maternal carriers.6 However, antenatal group B streptococcal screening highlights three practical concerns: the accuracy of antenatal cultures in detecting women carrying group B streptococci at delivery, the logistics of providing cul-
Financial
Disclosure
Dr. Baker in 1994.
was a consultant
with
Biostar
once in 1993 and once
0029.7844/96/$15.00 PII SOO29-7844(96)00111-l
51
ture results to the delivery site, and the management of women with no or incomplete prenatal care.7 These concerns would be obviated if a rapid and reliable method were available to identify group B streptococcal carriers at hospital admission for delivery. An ideal test for the detection of group B streptococcus at hospital admission would be rapid, accurate, and inexpensive. Many candidate methods have been reviewed by Yancy et a1.8 Although the most accurate of these is the enzyme immunoassay for detection of group B streptococcal antigen in vaginal swab specimens, most studies have compared enzyme immunoassay sensitivity with culture methods insufficient to detect women with a low inoculum of group B streptococci.’ Because these parturients give birth to approximately 15% of the neonates with early-onset group B streptococcal infection,6 their identification is important in developing intrapartum prevention strategies that will have the greatest impact. Recently, a new method for rapid group B streptococcal antigen detection in vaginal swabs has been developed and is being marketed as Strep B OIA test (Biostar Inc., Boulder, CO). The principle of this optical immunoassay is that the reflection of light changes when an antigen binds to an antibody on a reflecting surface, and this can be visualized directly. The optical immunoassay was devised to be more sensitive than enzyme immunoassays, and as such has the potential of meeting the criteria for rapid group B streptococcal detection mentioned above. The purpose of this study was to compare three commercially available assays for the rapid detection of group B streptococcus in vaginal swab specimens with results of cultures performed by two different methods to determine their potential usefulness for detection of group B streptococcal carriers at hospital admission for labor or rupture of membranes. To do so, we determined the sensitivity, specificity, and positive and negative predictive values for each assay. We compared these values with a commonly used culture method, direct plating of swabs onto blood agar medium, and the method advocated by the American Academy of Pediatric@ and the Centers for Disease Control and Prevention,7 in which swabs are incubated in selective broth medium before subculture onto blood agar medium.
Materials and Methods Between November 1, 1993, and September 28, 1994, 502 women evaluated for admission to Ben Taub General Hospital in Houston, Texas, for labor or rupture of membranes who had not received antibiotics or undergone pelvic examination were eligible for study. After
52
Baker
Antip
Assays
verbal consent, women were enrolled consecutively between 8 AM and 3 PM on Mondays, Tuesdays, and Wednesdays. Demographic data and gravidity, parity, gestation, time of membrane rupture (if present), and time vaginal swabs were collected were obtained from the medical records by a research nurse on a standard form. For the first 197 women, three rayon swabs (Becton Dickinson Microbiology Systems, Cockeysville, MD) were rotated 360” in the distal vagina to obtain specimens. When the Strep B OIA became available for study, an additional swab was obtained from the remaining 305 women. Three swabs were placed into modified Stuart’s transport media (Becton Dickinson) and the remaining one was placed in Amies transport medium without charcoal (Cultureswab Transport System, Difco, Surrey, UK). Swabs were delivered to the research laboratory and processed within 6 hours of collection. The swab in Amies transport medium was used to inoculate a 5% sheep blood agar (trypticase soy agar base) plate using a semiquantitative, four-zone streaking method.’ Growth of beta-hemolytic colonies was recorded as l+ to 4+ (l+ = 102-103, 2+ = >103-105, 3+ = >105-109 4+ = >106 colony forming units [cfu] per milliliter). One to 2+ indicated light and 3-4+ heavy group B streptococcal colonization. This swab was then placed into selective Todd-Hewitt broth (BBL; Becton Dickinson) containing 8 pg / mL gentamicin and 15 pg/mL nalidixic acid, incubated at 35C overnight, and subcultured onto a 5% sheep blood agar plate. Plates were incubated overnight at 35C, removed, and inspected for beta-hemolytic and nonhemolytic colonies with morphology typical of group B streptococci. The identification of group B streptococcus was confirmed by a positive reaction with the Streptex grouping kit reagent (Murex Diagnostics Limited, Dartford, UK). Each swab transported in modified Stuart’s medium was used for individual immunoassay testing. Kits for the two enzyme immunoassays were purchased from the manufacturers. All tests were performed at room temperature as directed by the manufacturers’ instructions. The first enzyme immunoassay, Quidel Group B Strep Test (San Diego, CA), consists of a Quidel test cartridge that contains a filtration membrane coated with rabbit polyclonal antibodies to group B streptococci. Group B streptococcal antigen was extracted from the swab with reagent provided by the manufacturer; the extracted fluid sample was then transferred to the test cartridge and gravity-filtered through the membrane. If group B streptococcal antigen was present, it bound to antibodies on the surface of the membrane. Reagent containing polyclonal anti-group B streptococcal antibody enzyme conjugate was then added to the cartridge and allowed to incubate for 1 minute. After
Obstetrics 6 Gymcology
rinsing the membrane to remove unbound antibodyenzyme conjugate, an enzyme substrate solution was added to the cartridge and allowed to react for 2 minutes. If group B streptococcal antigen was present, a solid blue circle appeared on the cartridge membrane, indicating a positive result. A positive group B streptococcal control was performed on each assay day. The entire procedure required less than 10 minutes. The second enzyme immunoassay, ICON Group B Strep (Hybritech, San Diego, CA), provided a plastic vial in which a swab was placed and extraction reagents were added. After incubation for 1 minute, five drops of a second extraction reagent and one drop of group B streptococcal polyclonal antibody-conjugate were added and mixed thoroughly. The swab was then pressed against the wall of the vial to express as much fluid as possible, and the sample was passed through a pre-filter onto the center of the immunoassay filter cylinder after 2 minutes. The membrane of the cylinder was washed and incubated for 2 minutes. After another wash, the immunoassay was read as positive if a purple circle appeared in the center of the membrane. Positive and negative controls were impregnated on each immunoassay filter membrane. The assay usually required 10 minutes for completion. For the Strep B OIA, a swab was placed into a tube with the first extraction reagent, then a second extraction reagent for sequential incubation for up 5 minutes. After neutralization, the swab was pressed against the tube to extract as much fluid as possible, then discarded. Group B streptococcal antibody conjugate was added to the solution, the solution was incubated for 3-5 minutes, and a few drops were pipetted onto the surface of the Strep B OIA cartridge device. After 10 minutes, the surface was washed and blotted, and substrate was added. After an additional 10 minutes of incubation, the surface was washed again, blotted, and inspected for a large blue to purple circle on the gold surface, which indicated that group B streptococcal antigen was present in the specimen. A group B streptococcal positive antigen control was provided and performed for each assay day. The Strep B OIA procedure usually took 35 minutes to complete. Both culture methods and the two enzyme immunoassays were performed on vaginal swabs from each of the 502 patients; the final 305 women also had swabs tested by the optical immunoassay method, Strep B OIA. The sensitivity, specificity, and positive and negative predictive values for the immunoassays were calculated as compared with each of two culture methods. A further analysis for patients heavily colonized with group B streptococcus also was performed. Comparison between groups was analyzed by uncorrected 2 test; P < .05 was considered significant.
VOL.
88, NO.
1, JULY
1996
Table
1. Sensitivity, Specificity, and Predictive Values of Three Rapid Immunoassays for Group B Streptococcus Assessed by Direct Blood Agar Plate Culture Method Assav method Quidel
Sensitivity
16.1%
EIA
[8.5-23.71
ICON 20.7%
Positive predictive value Negative predictive value
98.8% (4101415) 73.7%
[53.9-93.51
[81.8-881
(4101483)
53.2%
(18187) 94.7%
[84.7-1001
86.2%
W/19) 85.9%
[82.8-891
B OIA* [38.9-67.51
(25147) 98.4% (254/258)
99.8% (4141415)
(14/19) 84.9%
Strep
[12.3-29.11
(14187) Specificity
EIA
[73.7-98.71 (25129)
92.0%
(415/483)
[88.9-95.11 (2541276)
EIA = enzyme immunoassay; OIA = optical immunoassay. * The final 305 of the 502 patients were tested (see Materials and Methods). Data are presented as % [95% confidence interval], with the proportion underneath.
Red ts Among the 502 study participants, there were 22 (4.4%) whites, 86 (17.1%) African-Americans, 385 (76.7%) Hispanics, and 9 (1.8%) Asians. Ninety (17.9%) were less than 20 years of age. One hundred twenty-six (25.1%) had group B streptococci isolated from vaginal swabs processed in selective broth medium, and 87 (17.3%) from swabs plated directly onto 5% blood agar medium. The latter method failed to identify 39 intraparturn group B streptococcal carriers or 7.8% of this population of pregnant women, giving this culture method a sensitivity of 69%. Vaginal group B streptococcal colonization was more prevalent in AfricanAmericans (43 of 86,50%) and less prevalent in Hispanics (74 of 385, 19.2%) when compared with whites (seven of 22, 31.8%), but these differences were not statistically significant. The sensitivity, specificity, and positive and negative predictive values of the three immunoassays compared with culture results from direct plating of swabs onto 5% sheep blood agar media are shown in Table 1. The Strep B OIA was two- to threefold more sensitive (53%) than either of the enzyme immunoassays (Quidel, 16%; ICON, 21%) in detecting group B streptococcal carriers. Each assay had excellent specificity, and the ICON and Strep B OIA tests were more accurate than the Quidel enzyme immunoassay, as indicated by their positive predictive values of 95 and 86%, respectively. When degree of colonization or vaginal group B streptococcal inoculum was considered, all three assays had improved sensitivity for detection of heavily colonized women who had a group B streptococcal inoculum exceeding lo” cfu/mL. Among the 33 heavily
Baker
Antigen
Assnys
53
Table 2. Effect of Bacterial Inoculum” on Sensitivity of the Strep B OIA for Vaginal Swabs Growth
on
No. of
NO.
specimens
plate*
0 1+ 2+ 3+ 4t
OIA-positive
258 14 18 14 1
OIA = opticalimmunoassay. * Swab inoculateddirectlv on 5%
semiquantitative technique.’
Sensitivity
4 1 9 14 1
7.1% 50% 1OO’Y 100;
sheepbloodagarplateusinga
colonized study patients, 12 (36%) and 15 (46%) had positive Quidel or ICON enzyme immunoassays, respectively. Previously, inoculum has been shown to have a strong direct relationship on the sensitivity of these assays,9-1’ and we also found this to be true of the Strep B OIA (Table 2). Each of the heavily colonized women tested by OIA was positive. However, the sensitivity of Strep B OIA assay fell substantially in women with lower group B streptococcal inocula. Next, the results of the three immunoassays were compared with cultures processed in selective broth medium, the recommended method for antenatal group B streptococcal screening.5,6Each method was limited in its ability to identify lightly colonized group B streptococcal carriers (Tables 2 and 3). These women with lower vaginal group B streptococcal inocula have been estimated to deliver 15% of the neonates who develop early-onset group B streptococcal sepsis.1°
Table
3.
Sensitivity, Specificity, and Predictive Valuesof Three Immunoassays for Group B Streptococcus Assessed by Vaginal Cultures Processed in Selective Broth Medium
Assaymethod Quidel Sensitivity Specificity Positive predictive value Negative predictive value
EIA
ICON
EIA
Strep B OIA*
11.9% [6.2-17.61 (15/126) 98.9% (372 / 376) 78.9% [60.5-97.31 (15/19)
15.1% [8.8-21.41 (191126) 100% (375/376) 100% (19/19)
36.8% [25.4-48.23 (25168) 98.3% (235 / 239) 86.2% [73.7-98.71 (25129)
77.0% [73.3-80.71 (3721483)
77.8%) [74.1-81.5) (376 / 483)
84.5% [80.2-88.81 (235/278)
EIA = enzyme immunoassay; OIA = optical immunoassay. * The final 305 of the 502 patients were tested (see Materials and Methods). Data are presented as % [confidence interval], with the proportion underneath.
54
Baker
A~+zyz Assnys
Discussion The findings of this study repeat a theme in the search for rapid intrapartum detection of group B streptococcal carriers. Although relatively sensitive for the identification of pregnant women with moderate or high numbers of vaginal group B streptococci, immunoassays were unable to identify those with light colonization.‘O This shortcoming is most apparent when a vaginal swab is processed in antibiotic-containing or selective broth medium. Armer et all’ reported that the ICON Strep B assayhad an overall sensitivity of 11% for vaginal carriers of group B streptococcus identified using selective broth medium for culture. We also found ICON Strep B to be insensitive (15%), as was the Quidel Group B Strep test (12%). When compared with the commonly used, but lesssensitive,12culture method of directly plating swabs onto 5% sheep blood agar medium, the sensitivity of the immunoassays rose substantially. However, only women with moderate or heavy colonization will be identified by this method.13 Because approximately 15% of neonates with invasive group B streptococcal infection are born to women with lower quantities of group B streptococci, intrapartum identification of women with any level of group B streptococcal colonization is desirable.5,6 The new optical immunoassay, Strep B OIA, appears promising as a rapid diagnostic technique for identification of group B streptococcal antigen in vaginal swab specimens.When an assay based on the sameprinciple asthe Strep B OIA was tested for the detection of group A streptococci in children with pharyngitis, the results were encouraging14,i5: sensitivity and specificity exceeded 90%.14In our patients, the Strep B OIA had an overall sensitivity of 37%, but this rose to 100% for women with heavy group B streptococcal colonization. The disparity in the sensitivity of the Strep A and Strep B OIA is explained by the clinical populations targeted for screening. In children with low inocula of group A streptococci in throat cultures, the sensitivity of the Strep A OIA is poor (13%), but only 7% of children with acute pharyngitis fall into this category.r4 By contrast, nearly 60% of our study patients had low numbers of group B streptococcus in their vaginal cultures. If the clinical goal of a rapid antigen test was detection of heavily group B streptococcus-colonized women only, then Strep B OIA would appear to be an excellent test given its high sensitivity and specificity as well as its reasonably high positive and negative predictive values. Cumulative evidence supports the effectiveness of intrapartum ampicillin or penicillin G prophylaxis in preventing early-onset group B streptococcal sepsisin neonates and early postpartum febrile morbidity in
Ohsfetrics Er Gynecology
their mothers.5,h,‘“,‘h Although detection of colonization by culture during the third trimester is one method by which candidates for prophylaxis may be identified,5,‘6 there will continue to be a need for intrapartum detection in certain populations. Both rapid immunoassays for detection of group B streptococcal antigen in vaginal swab specimens that we studied had poor sensitivity (16-21%) in colonized women at hospital admission. Although the new optical immunoassay, Strep B OIA, improved sensitivity two- to threefold, this was not adequate in women who had light or low inoculum colonization (31%). Their detection would be critical to the design of an intrapartum strategy target that would increase the number of infant cases prevented. However, its 100% sensitivity in detecting heavily colonized women, combined with its reasonable predictive values for all group B streptococcal colonization, suggest that Strep B OIA has potential use in select obstetrical populations when antenatal detection of group B streptococcal carriers by culture is not feasible and a laboratory facility could offer 24 hour a day rapid testing at admission. None of the currently available rapid antigen methods for detection of group B streptococcal carriers appears to be sufficiently sensitive for routine use in the selection of women for intrapartum antibiotic prophylaxis. Because rapid detection of carriers at hospital admission continues to be germane to strategies that would prevent perinatal group B streptococcal infections through intrapartum chemoprophylaxis, improved antigen detection methods should be devised. Meanwhile, detection of carriers for intrapartum chemoprophylaxis must depend on antenatal vaginal and rectal swab specimens cultured in selective broth medium.
References 1. Baker CJ, Edwards MS. Group B streptococcal infections. In: Remington JS, Klein JO, eds. Infectious diseases of the fetus and newborn infant. 4th ed. Philadelphia: WB Saunders, 1995:9801054. 2. Zangwill disease
KM,
surveillance 3. American streptococcal
4. Institute
88, NO.
A, States,
Wenger 1990:
JD. Report
Group from
of
DC:
American
College
1, JULY
New
1996
vaccine
Group bulletin
of Obstetricians
development:
1: Diseases National
5. American Academy group B streptococcal 1992;90:775-8.
of importance in the United Academy Press, 1985:424-39.
States.
of Pediatrics. infection
prevention Pediatrics
Guidelines for by chemoprophylaxis.
Washof
6. Centers for Disease Control and Prevention. Prevention of group B streptococcal disease: A public health perspective. Federal Register 1994;(Dec 15);59(24O)d:647-764-773. 7. Larsen JW, Dooley SL. Group obstetrical 8. Yancey
viewpoint. MK, Armer
B streptococcal
Pediatrics T, Clark
infections:
1993;91:148-9. I’, Duff I’. Assessment
An
of
rapid
identification tests for genital carriage of group B streptococcal. Obstet Gynecol 1992;80:1038-47. 9. Gentry Y, Hillier SL, Eschenbach DA. Evaluation of a rapid enzyme Obstet
immunoassay test for detection Gynecol 1991;78:397-401.
10. Tuppurainen streptococcal
N, Hallman M. disease: Intrapartum
laxis of heavily 583-7. 11. Armer T, Clark
colonized I’, Duff
I’, Saravanos
Goroff DK, Alpert of bacteriological
streptococcus 13. Granato PA,
from Petosa
Prevention detection
parturients.
tion of group B streptococcal noassay. Am J Obstet Gynecol 12. Baker CJ, Comparison
of group of
B streptococcus.
neonatal group and chemoprophy-
Obstet K. Rapid
Gynecol
1989;73:
intrapartum
colonization with 1993;168:39-43.
an enzyme
B
detecimmu-
SL, Hayes C, McCormack WM. methods for the isolation of group B
vaginal cultures. MT. Evaluation
J Clin Microbial 1976;4:46-8. of a rapid screening test
for
detecting group B streptococci in pregnant women. J Clin Microbiol 1991;29:1536-8. 14. Roddey OF Jr, Clegg HW, Martin ES, Swetenburg RL, Koonce EW. Comparison ture methods 15.
of an optical immunoassay for the detection of
group
technique with A streptococci
pediatric office. J Pediatr 1995;126:931-3. Fries SM. Diagnosis of group A streptococcal private method
clinic: and
pharyngitis
Comparative evaluation of an optical culture. J Pediatr 1995;126:933-6.
16. Allen UD, Navas cillin prophylaxis infection: Results
L, King SM. Effectiveness in preventing early-onset of a meta-analysis. Can
two
culin a in
a
immunoassay
of intrapartum penigroup B streptococcal Med Assoc J 1993;149:
1659-65.
Address reprint requests to: Carol J Baker, MD Deyartment qf Pediatrics Baylor School of Medicine One Baylor Plaza, Room 307E Houstou, TX 77030
active
1992. Medicine.
DC:
B streptococcal a multistate
system. MMWR 1992;41:25-32. College of Obstetricians and Gynecologists. infections in pregnancy. ACOG technical
170. Washington, Gynecologists,
VOL.
Schuchat
in the United
priorities, ington,
Establishing
B no. and
Rmived Noven~her 29, 1995. Rrckoed in revised fom ~anunry Accepted February 14, 1996. Copyright Gynecologists.
0
1996 by Published
The American by Elsevier
22, 1996.
College of Science Inc.
Baker
Obstetricians
Antigen Assays
and
55