- Email: [email protected]
Rapid intrapartum detection of group B streptococcal colonization with an enzyme immunoassay
Rapid intrapartum detection of group B streptococcal colonization with an enzyme immunoassay Traci Armer, MD, Penny Clark, PhD, Patrick Duff, MD, and Karen Saravanos, MD
Gainesville, Florida OBJECTIVE: The purpose of this study was to determine the accuracy of a rapid immunoenzyme assay for detecting intrapartum colonization with group B streptococci. STUDY DESIGN: Three rayon-tipped swabs were used to collect specimens from the posterior vaginal wall of 424 preterm and term patients in labor. Three tests were performed on specimens obtained from the first 182 patients: semiquantitative culture on blood agar, culture in selective Todd-Hewitt broth, and ICON Strep B (Hybritech, San Diego) immunoconcentration assay. For the remaining 242 patients, the ICON test was performed only when the Todd-Hewitt broth culture was positive. RESULTS: The prevalence of positive cultures was 23%. For the first 182 patients, the immunoassay had a sensitivity of 11 %, a specificity of 100%, a positive predictive value of 100%, and a negative predictive value of 78%. The overall sensitivity for all 424 patients was 9%. In eight women with heavy colonization, defined as an inoculum of > 105 cfu/ml the sensitivity of the assay was 100%. In the study population there were three cases of group B streptococcal sepsis in infants whose mothers were only lightly colonized. None of these cases of colonization were detected by the assay. CONCLUSION: The ICON immunoconcentration assay is very sensitive in identifying heavy group B streptococcal colonization of > 105 cfu/ml but quite insensitive in detecting lower levels of colonizaton. Thus it is not a suitable test for general screening. (AM J OBSTET GYNECO~ 1993;168:39-43.)
Key words: Group B streptococci, rapid detection, immunoassay
Up to 15,000 cases of neonatal sepsis each year are caused by group B streptococci. The vast majority of neonatal infections result from vertical transmission during labor. Of colonized women, 17% to 65% transmit this organism to their infants, and subsequently sepsis develops in I % to 2% of these infants.' The mortality rate among these newborns is < 5% in term infants but can be as high as 50% in premature infants. 2 Because 60% of the pregnant women who harbor this organism are only intermittent carriers, antenatal cultures do not consistently identity the women who will actually be colonized at the time of delivery.' In addition, microbiologic cultures for group B streptococci require 24 to 48 hours of incubation, which delays identification of the patients in labor who are colonized. Because intrapartum antibiotic therapy can decrease the frequency of vertical transmission, there has been great interest recently in the development of a rapid diagnostic test to select colonized patients for antibiotic prophylaxis.··6 One new rapid test, the ICON immunoassay (Hy-
britech, San Diego), provides for acid extraction of group B streptococcal antigen directly from vaginal swabs, followed by solid enzyme immunoassay. The test takes < 10 minutes to perform. A positive test is denoted by the presence of color development at the center of a filter membrane. Recently, Gentry et aJ.' evaluated the performance of the ICON assay compared with standard blood agar cultures and found a sensitivity of 33%, a specificity of 95%, and positive and negative predictive values of 43% and 93%, respectively. The sensitivity of the test improved to 100% in patients with heavy group B streptococcal colonization, which corresponded to a bacterial inoculum of > 10' cfulml. The purpose of the current study was to determine the accuracy of the ICON immunoassay in detecting both heavy and light maternal group B streptococcal colonization, with Todd-Hewitt broth culture as the gold standard for identification of the organism. ToddHewitt broth is a selective enrichment medium for the growth of group B streptococci and enhances the likelihood of identifYing even lightly colonized patients.
From the Division of Maternal-Fetal Medicine, Department of Obstetnes and Gynecology, University of Florida. Received for publicahon February 26, 1992; revised June 5, 1992; accepted June 11, 1992. Repnnt requests: Patnek Duff, MD, P.O. Box 100294, University of Florida, College of Medicine, Gainesville, FL 32610-0294. 6/1/40250
This study was performed from June through September 1991 at Shands Hospital, University of Florida, which primarily serves indigent patients insured under the Medicaid program. Women evaluated in the labor and delivery ward for preterm or term labor were
0002·9378/93 $1.00
+ 0.20
Material and methods
39
40
Armer et al.
included in the study if an investigator was available to collect specimens. Patients were excluded if they had received antibiotics within the previous 2 weeks. Informed written consent was obtained from all qualified patients in accordance with institutional review board guidelines. Vaginal specimens were collected by sweeping a sterile rayon-tipped swab from the Culturette Collection and Transport System (Becton-Dickinson, Cockeysville, Md .) along the lateral vaginal walls down to the perineum. This procedure was repeated with three individual swabs, which were then reinserted into separate transport systems. Ampules containing 0.5 ml of modified Stuart's bacterial transport medium were crushed in two swabs; one swab was kept dry in the transport system. The swabs were delivered to the research microbiology laboratory and processed within 2 hours of collection. Three tests were performed on these specimens. One of the moistened swabs was transferred to a selective Todd-Hewitt broth (BBL, Becton-Dickinson) containing 5% sheep blood, 8 J..Lglml gentamicin, and 15 J..Lg/ml nalidixic acid and incubated at 37° C in a 5% carbon dioxide atmosphere. At the end of 24 hours of incubation any organisms present in the broth culture were subcultured onto a trypticase soy blood agar plate and incubated for another 24 hours. Group B streptococci were confirmed by the presence of ~-hemolysis and a positive reaction with the PathoDx Latex Strep grouping kit (Diagnostic Products, Los Angeles). The other moistened swab was used for semiquantitative culture on neomycin-blood agar after a prescribed streaking method. 8 The plate was incubated at 37° C in an atmosphere of 5% carbon dioxide and screened for the presence of ~-hemolytic colonies after 24 hours of incubation. Negative plates were further incubated and reexamined after 48 hours. Colony growth on the first and second quadrants was scored as 1 + and 2 +, respectively. Growth on the third quadrant was scored as 3 + and then further subdivided into 3 + with < 10 ~-hemolytic colonies and 3 + with ~ lO [3-hemolytic colonies. Any growth on the fourth quadrant was scored as 4 +. A positive culture in selective broth accompanied by no growth on the semiquantitative plate was scored as O. The dry swab was used to perform the ICON Strep B immunoassay.9 Test kits were purchased from the manufacturer. The swabs were either processed immediately or were refrigerated and tested within 72 hours, in accordance with the manufacturer's guidelines. Refrigeration of the bacteria does not affect test results because the assay identifies bacterial antigens regardless of the presence of live organisms. The test was performed at room temperature, and the procedure prescribed by the manufacturer was strictly followed for
January 1993 Am J Obslel Gynecol
each swab processed. Briefly, the vaginal swab was subjected to nitrous acid extraction of bacterial antigen for 1 minute. Antibody conjugate, an alkaline phosphatase-conjugated rabbit anti-group B streptococci, was added to the extracted antigen. The extract was then squeezed through a prefilter onto the center membrane filter of an ICON cylinder and allowed to react for 2 minutes. The patient zone on the filter membrane contained immobilized rabbit anti-group B streptococci antibody, and the positive control zone contained group B streptococci antigen. Binding of the immobilized antibody with the antibody-antigen complex resulted in the formation of an enzyme-labeled antibody-group B streptococci antigen-immobilized rabbit antibody sandwich. After any unbound labeled antibody on the filter membrane was washed off, enzyme substrate was added. The presence of group B streptococcal antigen was determined by the development of any shade of purple at the central patient zone at the end of 2 minutes of incubation. The three tests described were performed on the first 182 patients. We calculated the sensitivity, specificity, and positive and negative predictive values for the immunoassay compared with those of Todd-Hewitt broth culture. As noted subsequently, we observed perfect specificity and positive predictive value for the initial 182 assays. Therefore in the succeeding group of 242 patients only those with positive Todd-Hewitt broth cultures were tested with the ICON immunoassay. As a further refinement of our analysis, the limit of sensitivity of the ICON immunoassay was determined on known titers of group B streptococci. A tenfold serial dilution was made from an overnight culture of Streptococcus agalactiae (American Type Culture Collection 12386, Rockville, Md.). A volume of 0.1 ml from each of the dilutions was absorbed onto duplicate Culturette swabs. The ICON immunoassay was performed on one swab, and a semiquantitative culture as described was performed on the duplicate swab. Heavy growth, defined as ~ 105 cfulml, correlated with semiquantitative scores of 4 + and 3 + with ~ 10 ~-hemolytic colonies. Light growth, defined as ~ 10' cfulml, correlated with semiquantitative scores of 0, 1 + , 2 +, and 3 + with < lO ~-hemolytic colonies. These definitions of heavy and light growth were then applied to all of the semiquantitative cultures from clinical specimens. At the end of the microbiologic study we sought to review the infant charts of all 98 mothers with positive cultures. A random sample of neonatal charts from 35 mothers who had negative cultures was also reviewed . Neonatal diagnoses of suspected sepsis, proved sepsis, pneumonia, and meningitis were recorded. We compared the performance of the rapid test in patients with light versus heavy colonization and in those with ruptured versus intact membranes. We also
Rapid detection of group B streptococci 41
Volume 168 Number 1, Part 1
RESULTS OF CULTURE IN TODD-HEWITT BROTH
+ ~>-
0< + s< t/)t/) f-ot/) t/)Z
~~
5
0
5
39
138
177
44
138
182
Sensitivity =5/44 (11%, 95% CJ. 6-16%) Specificity = 138/138 (100%) Positive predictive value =5/5 (100%) Negative predictive value =138/177 (78%, 95% C.I. 72-84%) Fig. 1. Sensitivity, specificity, and predictive values of the ICON immunoassay for detection of group B streptococci. C.I., 95% Confidence interval.
Table I. Effect of bacterial inoculum on sensitivity of ICON immunoassay Inoculum size * (cfulml) Light
1 X 10' 6 X 10' 1 X 10' Heavy 1 X 105 > 10'
Culture result
No. of specimens
No. with positive ICON
2+ 3+t
37 20 23 10
0 0 1 0
0 0 4 0
3H 4+
5 3
5 3
100 100
0
1+
SensItivity
(%)
·See Material and methods section.
t < 10 group B streptococci colonies on third quadrant. t> 10 group B streptococci colonies on third quadrant.
assessed the frequency of infection in neonates born to patients with light versus heavy colonization. The uncorrected X· test was used to determine statistical significance. A p value < 0.05 was considered significant. Results
Ninety-eight of the 424 cultures (23%, 95% confidence interval 19% to 27%) were positive for group B streptococci. Fig. 1 shows the overall performance of the ICON immunoassay in the first group of 182 patients, with Todd-Hewitt broth culture as the gold standard. For the last group of 242 patients the sensitivity of the immunoassay was 7%. We cannot report ~pecificity or predictive values for this final sample of patients because we performed the ICON test only on patients who were known to have a positive culture. For the total 424 patients the test had an overall sensitivity of 9%. In patients with ruptured membranes at the time
of specimen collection the prevalence of colonization and the assay sensitivity, specificity, and positive and negative predictive values did not differ significantly from those in patients with intact membranes. Ninety of the 98 positive cultures showed light colonization on semiquantitative plates. As shown in Table I, only one of these lightly colonized patients was identified by the ICON immunoassay. All eight of the heavily colonized patients (100%) tested positive with the immunoassay. Thus this diagnostic assay had a limit of sensitivity that was equivalent to an inoculum size of 5 ~ 1 x 10 cfulml. Of the infants born to the 90 women with light colonization, 79 had charts available for review. Nineteen (24%) of these infants were diagnosed with suspected sepsis on the basis of maternal history and neonatal leukocytosis with a left shift. Three of these infants had proved group B streptococcal sepsis, two by
42 Armer et al.
positive blood cultures and one by positive urine latex agglutination test. The three term infants were treated with intravenous ampicillin and were discharged from the hospital within 14 days. None of the 19 mothers had positive ICON assays. All eight infants born to the heavily colonized women had charts available for review. Four (50%) of these infants were diagnosed with suspected sepsis, all of whom were born to mothers with a positive immunoassay. Two of the mothers had previously been identified as carriers of group B streptococci and had received intrapartum prophylaxis with ampicillin. None of infants born to heavily colonized women had proved group B streptococcal sepsis. Of the infants born to mothers with negative cultures, 35 charts were reviewed for comparison. Four (11 %) infants had suspected sepsis, and none had proved group B streptococcal sepsis. The percentage of infants who were diagnosed with suspected sepsis from any cause was not significantly different between heavily and lightly colonized mothers or between colonized and noncolonized mothers.
Comment The results of this study suggest that the ICON immunoassay reliably detects the presence of heavy group B streptococcal colonization of ~ 10' cfulml and is unlikely to yield a false-positive result. The assay failed to identify women with light colonization. The overall test sensitivity of 9% was considerably lower than the 33% sensitivity reported by Gentry et al. 7 This discrepancy is most likely explained by our use of Todd-Hewitt broth culture as the standard for group B streptococci identification instead of direct plating on selective blood agar. The broth culture improves detection of low inoculum levels. The high sensitivity of the assay when testing heavily colonized patients was consistent with the findings of Gentry et aI. 7 In the instructions enclosed with the ICON test kit the manufacturers state that the majority of other vaginal microorganisms, including non-group B streptococci, were tested with the assay at concentrations of 10 8 organisms per test and cross reactivity was not present. We likewise found no evidence of cross reactivity with other organisms or interference of the assay as a result of blood, meconium, or mucus on the swab. The interpetation of the assay was straightforward, and the end point was easy to read. Other rapid detection tests evaluated in mixed highand low-risk obstetric populations have also demonstrated low overall sensitivities, which improved when analysis was restricted to heavily colonized patients. 7 • 10-1 6 These rapid tests have used latex particle agglutination and enzyme immunoassay techniques, the performance of which has been summarized in our
J anuary 1993 Am J Obstet Gyneco l
previous report. 14 Clearly, heavily colonized women are at higher risk of transmitting group B streptococci to their newborns. B• 17 However, we cannot ignore the fact that sepsis can develop in infants born to lightly colonized mothers, as was demonstrated in this study and others. 6 • 18. 19 Because this assay and the other rapid tests previously studied fail to reliably identify lightly colonized patients, the search for another test or method with improved sensitivity must continue. REFERENCES I . Boyer KM, Gadzala CA, Kelly PO, Burd LI, Gotoff SP. Selective intraprtum chemoprophylaxis of neonatal group B streptococcal early-onset disease. n. Predictive value of prenatal cultures. 1 Infect Dis 1983; 148:802-9. 2. Minkhoff H, Mead P. An obstetric approach to the prevention of early-onset group B (3-hemolytic streptococcal sepsis. AM 1 OBSTET GYNECOL 1986;154:973-7. 3. Mean PB. Group B streptococcal sepsis. In: Charles D, Glover DD, eds. Current therapy in obstetrics. Philadelphia: BC Decker, 1988; 189-95. 4. Boyer KM, Gadzala CA, Kelly PD, Burd LI, Gotoff SP. Selective intrapartum chemoprophylaxis of neonatal group B streptococcal early-onset disease. Ill. Interruption of mother-to-infant transmission. 1 Infect Dis 1983; 148:810-6. 5. Boyer KM, Gotoff SP. Prevention of early-onset neonatal group B streptococcal disease with selective intrapartum chemoprophylaxis. N Engll Med 1986;314:1665-9. 6. Lim DV, Morales Wl, Walsh AF, Kazanis D. Reduction of morbidity rates and mortality rates for neonatal group B streptococcal disease through early diagnosis and chemoprophylaxis. 1 Clin Microbiol 1986;23:489-92 . 7. Gentry YM, Hillier SL, Eschenbach DA. Evaluation of a rapid enzyme immunoassay for the detection of group B streptococcus. Obstet Gynecol 1991;78:397-401. 8. Baron lB, Finegold SB. Cultivation and isolation of viable pathogens. In: Manning S, ed. Bailey and Scou's diagnostic microbiology. 8th ed. St. Louis: CV Mosby, 1990:94-5. 9. Donatelli 1, Macone A, Goldmann DA, et al. Rapid detection of group A streptococci: comparative performance by nurses and laboratory technologists in pediatric satellite laboratories using three test kits. J Clin Microbiol 1992; 30:138-42. 10. Granato PA, Petosa MT. Evaluation of a rapid screening test for detecting group B streptococci in pregnant women. 1 Clin Microbiol 1991;29:1536-8. II. Wald ER, Dashefsky B, Green M, et al. Rapid detection of group B streptococci directly from vaginal swabs. 1 Clin Microbiol 1987;25:573-4. 12. Kontnick CM, Edberg SC. Direct detection of group B streptococci from vaginal specimens compared with quantitative culture. 1 Clin Microbiol 1990;28:336-9. 13. Skoll MA, Mercer BM, Baselski V, Gray lP, Ryan G, Sibai BM . Evaluation of two rapid group B streptococcal antigen test in labor and delivery patients. Obstet Gynecol 1991;77;322-6. 14. Brady K, Duff P, Schilhab lC, Herd M. Reliability of a rapid latex fixation test for detecting group B streptococci in the genital tract of parturients at term. Obstet Gynecol 1989;73:678-81 . 15. Clark P, Armer T, Duff P, Davidson K. Assessment of a rapid latex agglutination test for group B streptococcal colonization of the genital tract. Obstet Gynecol 1992;79: 358-63. 16. Lotz-Nolan L, Amato T, litis J, Wallen W, Packer B. Evaluation of a rapid latex agglutination test for detection
Volume 168 Number 1. Part 1
Rapid detection of group B streptococci
of group B streptococci in vaginal specimens. Eur J Clin Microbiol Infect Dis 1989;8:289-93. 17. Jones DE. Kanarek KS, Lim DV. Group B streptococcal colonization patterns in mothers and their infants. J Clin Microbiol 1984;20:438-40. 18. Morales WJ, Lim D. Reduction of group B streptococcal maternal and neonatal infections in preterm pregnancies
with premature rupture of membranes through a rapid identification test. AM J OBSTET GYNECOL 1987;157:13-6. 19. Towers CV, Garite TJ, Friedman WW, Pircon RA, Nageotte MP. Comparison of a rapid enzyme-linked immunoasorbent assay test and the Gram stain for detection of group B streptococcus in high-risk antepartum patients. AM J OBSTET GYNECOL 1990;163:965-7.
The development of abnormal heart rate patterns after absent end-diastolic velocity in umbilical artery: Analysis of risk factors Domenico Arduini, MD," Giuseppe Rizzo, MD," and Carlo Romanini, MDb Rome and Ancona, Italy OBJECTIVES: Our objectives were to evaluate the time interval elapsing between the occurrence of absent end-diastolic velocity in the umbilical artery and either the development abnormal fetal heart rate patterns or delivery and to establish the maternal and fetal factors that may affect this interval. STUDY DESIGN: Thirty-seven fetuses free of structural and chromosomal abnormalities in which the development of absent end-diastolic velocity in umbilical artery was evidenced by serial Doppler recordings were studied. At the first occurrence the following factors were considered: gestational age, presence of hypertension or preeclampsia, amniotic fluid index, severity of growth retardation, and 10 different Doppler indices calculated from extracardiac and intracardiac vascular districts. Actuarial statistical methods were applied, with the occurrence of antepartum late heart rate deceleration as the censoring variable. RESULTS: The interval between the first occurrence of absent end-diastolic velocity in umbilical artery and delivery ranged from 1 to 26 days. Indications for delivery were the development of antepartum late heart rate decelerations in 23 fetuses (62.1 %) and different maternal or fetal complications in the remaining 14 fetuses. Multivariate analysiS revealed that gestational age and the presence of hypertension and pulsations in umbilical vein were the dominant factors in determining the length of this time interval. CONCLUSION: The duration of the time interval between the occurrence of absent end-diastolic velocity in umbilical artery and abnormal heart rate pattern differs considerably among fetuses, and it is mainly determined by gestational age and presence of maternal hypertension and pulsations in umbilical vein. (AM J OBSTET GYNECOL 1993;168:43-50.)
Key words: Doppler ultrasonography, abnormal heart rate patterns, fetal distress, fetal circulation Doppler ultrasonography has been widely applied in obstetrics to study umbilical artery velocity waveforms From the Laboratory of Fetal PhYSIOlogy, Department of Obstetrus and Gynecology, Universttii Cattolzca S. Cuoreo" and the Department of Obstetncs and Gynecology, Unzversitii dl Ancona. b Supported by Grant No. CNR 920004 PF41 from the Italzan National Counczl of Research. Received for publication March 26, 1992; reVISed June 23, 1992; accepted June 25, 1992. Reprint requests: Carlo Romanmi, MD, Universltd dl Roma "Torvergata ",1st Clznzca Ostetnea e Ginecologica PollClznico S. EugenIO, P. Ie Dell 'Umanesimo, 10, 00144 Roma, Italy. 6/1/40640
0002-9378/93 $1.00
+ 0.20
as a reflection of placental vascular resistance. Several associations have been reported between qualitative changes in indices of flow and perinatal complications. I. 2 However, the absence of end-diastolic velocity seems to be the finding with the greatest clinical interest,"' 3 and it is frequently associated with fetal hypoxemia, acidemia, and poor perinatal outcome.4-7 There is no account, however, on the time interval elapsing between the occurrence of absent end-diastolic velocity and evidence of abnormal heart rate patterns requiring early delivery. Previous studies have shown
43
Gainesville, Florida OBJECTIVE: The purpose of this study was to determine the accuracy of a rapid immunoenzyme assay for detecting intrapartum colonization with group B streptococci. STUDY DESIGN: Three rayon-tipped swabs were used to collect specimens from the posterior vaginal wall of 424 preterm and term patients in labor. Three tests were performed on specimens obtained from the first 182 patients: semiquantitative culture on blood agar, culture in selective Todd-Hewitt broth, and ICON Strep B (Hybritech, San Diego) immunoconcentration assay. For the remaining 242 patients, the ICON test was performed only when the Todd-Hewitt broth culture was positive. RESULTS: The prevalence of positive cultures was 23%. For the first 182 patients, the immunoassay had a sensitivity of 11 %, a specificity of 100%, a positive predictive value of 100%, and a negative predictive value of 78%. The overall sensitivity for all 424 patients was 9%. In eight women with heavy colonization, defined as an inoculum of > 105 cfu/ml the sensitivity of the assay was 100%. In the study population there were three cases of group B streptococcal sepsis in infants whose mothers were only lightly colonized. None of these cases of colonization were detected by the assay. CONCLUSION: The ICON immunoconcentration assay is very sensitive in identifying heavy group B streptococcal colonization of > 105 cfu/ml but quite insensitive in detecting lower levels of colonizaton. Thus it is not a suitable test for general screening. (AM J OBSTET GYNECO~ 1993;168:39-43.)
Key words: Group B streptococci, rapid detection, immunoassay
Up to 15,000 cases of neonatal sepsis each year are caused by group B streptococci. The vast majority of neonatal infections result from vertical transmission during labor. Of colonized women, 17% to 65% transmit this organism to their infants, and subsequently sepsis develops in I % to 2% of these infants.' The mortality rate among these newborns is < 5% in term infants but can be as high as 50% in premature infants. 2 Because 60% of the pregnant women who harbor this organism are only intermittent carriers, antenatal cultures do not consistently identity the women who will actually be colonized at the time of delivery.' In addition, microbiologic cultures for group B streptococci require 24 to 48 hours of incubation, which delays identification of the patients in labor who are colonized. Because intrapartum antibiotic therapy can decrease the frequency of vertical transmission, there has been great interest recently in the development of a rapid diagnostic test to select colonized patients for antibiotic prophylaxis.··6 One new rapid test, the ICON immunoassay (Hy-
britech, San Diego), provides for acid extraction of group B streptococcal antigen directly from vaginal swabs, followed by solid enzyme immunoassay. The test takes < 10 minutes to perform. A positive test is denoted by the presence of color development at the center of a filter membrane. Recently, Gentry et aJ.' evaluated the performance of the ICON assay compared with standard blood agar cultures and found a sensitivity of 33%, a specificity of 95%, and positive and negative predictive values of 43% and 93%, respectively. The sensitivity of the test improved to 100% in patients with heavy group B streptococcal colonization, which corresponded to a bacterial inoculum of > 10' cfulml. The purpose of the current study was to determine the accuracy of the ICON immunoassay in detecting both heavy and light maternal group B streptococcal colonization, with Todd-Hewitt broth culture as the gold standard for identification of the organism. ToddHewitt broth is a selective enrichment medium for the growth of group B streptococci and enhances the likelihood of identifYing even lightly colonized patients.
From the Division of Maternal-Fetal Medicine, Department of Obstetnes and Gynecology, University of Florida. Received for publicahon February 26, 1992; revised June 5, 1992; accepted June 11, 1992. Repnnt requests: Patnek Duff, MD, P.O. Box 100294, University of Florida, College of Medicine, Gainesville, FL 32610-0294. 6/1/40250
This study was performed from June through September 1991 at Shands Hospital, University of Florida, which primarily serves indigent patients insured under the Medicaid program. Women evaluated in the labor and delivery ward for preterm or term labor were
0002·9378/93 $1.00
+ 0.20
Material and methods
39
40
Armer et al.
included in the study if an investigator was available to collect specimens. Patients were excluded if they had received antibiotics within the previous 2 weeks. Informed written consent was obtained from all qualified patients in accordance with institutional review board guidelines. Vaginal specimens were collected by sweeping a sterile rayon-tipped swab from the Culturette Collection and Transport System (Becton-Dickinson, Cockeysville, Md .) along the lateral vaginal walls down to the perineum. This procedure was repeated with three individual swabs, which were then reinserted into separate transport systems. Ampules containing 0.5 ml of modified Stuart's bacterial transport medium were crushed in two swabs; one swab was kept dry in the transport system. The swabs were delivered to the research microbiology laboratory and processed within 2 hours of collection. Three tests were performed on these specimens. One of the moistened swabs was transferred to a selective Todd-Hewitt broth (BBL, Becton-Dickinson) containing 5% sheep blood, 8 J..Lglml gentamicin, and 15 J..Lg/ml nalidixic acid and incubated at 37° C in a 5% carbon dioxide atmosphere. At the end of 24 hours of incubation any organisms present in the broth culture were subcultured onto a trypticase soy blood agar plate and incubated for another 24 hours. Group B streptococci were confirmed by the presence of ~-hemolysis and a positive reaction with the PathoDx Latex Strep grouping kit (Diagnostic Products, Los Angeles). The other moistened swab was used for semiquantitative culture on neomycin-blood agar after a prescribed streaking method. 8 The plate was incubated at 37° C in an atmosphere of 5% carbon dioxide and screened for the presence of ~-hemolytic colonies after 24 hours of incubation. Negative plates were further incubated and reexamined after 48 hours. Colony growth on the first and second quadrants was scored as 1 + and 2 +, respectively. Growth on the third quadrant was scored as 3 + and then further subdivided into 3 + with < 10 ~-hemolytic colonies and 3 + with ~ lO [3-hemolytic colonies. Any growth on the fourth quadrant was scored as 4 +. A positive culture in selective broth accompanied by no growth on the semiquantitative plate was scored as O. The dry swab was used to perform the ICON Strep B immunoassay.9 Test kits were purchased from the manufacturer. The swabs were either processed immediately or were refrigerated and tested within 72 hours, in accordance with the manufacturer's guidelines. Refrigeration of the bacteria does not affect test results because the assay identifies bacterial antigens regardless of the presence of live organisms. The test was performed at room temperature, and the procedure prescribed by the manufacturer was strictly followed for
January 1993 Am J Obslel Gynecol
each swab processed. Briefly, the vaginal swab was subjected to nitrous acid extraction of bacterial antigen for 1 minute. Antibody conjugate, an alkaline phosphatase-conjugated rabbit anti-group B streptococci, was added to the extracted antigen. The extract was then squeezed through a prefilter onto the center membrane filter of an ICON cylinder and allowed to react for 2 minutes. The patient zone on the filter membrane contained immobilized rabbit anti-group B streptococci antibody, and the positive control zone contained group B streptococci antigen. Binding of the immobilized antibody with the antibody-antigen complex resulted in the formation of an enzyme-labeled antibody-group B streptococci antigen-immobilized rabbit antibody sandwich. After any unbound labeled antibody on the filter membrane was washed off, enzyme substrate was added. The presence of group B streptococcal antigen was determined by the development of any shade of purple at the central patient zone at the end of 2 minutes of incubation. The three tests described were performed on the first 182 patients. We calculated the sensitivity, specificity, and positive and negative predictive values for the immunoassay compared with those of Todd-Hewitt broth culture. As noted subsequently, we observed perfect specificity and positive predictive value for the initial 182 assays. Therefore in the succeeding group of 242 patients only those with positive Todd-Hewitt broth cultures were tested with the ICON immunoassay. As a further refinement of our analysis, the limit of sensitivity of the ICON immunoassay was determined on known titers of group B streptococci. A tenfold serial dilution was made from an overnight culture of Streptococcus agalactiae (American Type Culture Collection 12386, Rockville, Md.). A volume of 0.1 ml from each of the dilutions was absorbed onto duplicate Culturette swabs. The ICON immunoassay was performed on one swab, and a semiquantitative culture as described was performed on the duplicate swab. Heavy growth, defined as ~ 105 cfulml, correlated with semiquantitative scores of 4 + and 3 + with ~ 10 ~-hemolytic colonies. Light growth, defined as ~ 10' cfulml, correlated with semiquantitative scores of 0, 1 + , 2 +, and 3 + with < lO ~-hemolytic colonies. These definitions of heavy and light growth were then applied to all of the semiquantitative cultures from clinical specimens. At the end of the microbiologic study we sought to review the infant charts of all 98 mothers with positive cultures. A random sample of neonatal charts from 35 mothers who had negative cultures was also reviewed . Neonatal diagnoses of suspected sepsis, proved sepsis, pneumonia, and meningitis were recorded. We compared the performance of the rapid test in patients with light versus heavy colonization and in those with ruptured versus intact membranes. We also
Rapid detection of group B streptococci 41
Volume 168 Number 1, Part 1
RESULTS OF CULTURE IN TODD-HEWITT BROTH
+ ~>-
0< + s< t/)t/) f-ot/) t/)Z
~~
5
0
5
39
138
177
44
138
182
Sensitivity =5/44 (11%, 95% CJ. 6-16%) Specificity = 138/138 (100%) Positive predictive value =5/5 (100%) Negative predictive value =138/177 (78%, 95% C.I. 72-84%) Fig. 1. Sensitivity, specificity, and predictive values of the ICON immunoassay for detection of group B streptococci. C.I., 95% Confidence interval.
Table I. Effect of bacterial inoculum on sensitivity of ICON immunoassay Inoculum size * (cfulml) Light
1 X 10' 6 X 10' 1 X 10' Heavy 1 X 105 > 10'
Culture result
No. of specimens
No. with positive ICON
2+ 3+t
37 20 23 10
0 0 1 0
0 0 4 0
3H 4+
5 3
5 3
100 100
0
1+
SensItivity
(%)
·See Material and methods section.
t < 10 group B streptococci colonies on third quadrant. t> 10 group B streptococci colonies on third quadrant.
assessed the frequency of infection in neonates born to patients with light versus heavy colonization. The uncorrected X· test was used to determine statistical significance. A p value < 0.05 was considered significant. Results
Ninety-eight of the 424 cultures (23%, 95% confidence interval 19% to 27%) were positive for group B streptococci. Fig. 1 shows the overall performance of the ICON immunoassay in the first group of 182 patients, with Todd-Hewitt broth culture as the gold standard. For the last group of 242 patients the sensitivity of the immunoassay was 7%. We cannot report ~pecificity or predictive values for this final sample of patients because we performed the ICON test only on patients who were known to have a positive culture. For the total 424 patients the test had an overall sensitivity of 9%. In patients with ruptured membranes at the time
of specimen collection the prevalence of colonization and the assay sensitivity, specificity, and positive and negative predictive values did not differ significantly from those in patients with intact membranes. Ninety of the 98 positive cultures showed light colonization on semiquantitative plates. As shown in Table I, only one of these lightly colonized patients was identified by the ICON immunoassay. All eight of the heavily colonized patients (100%) tested positive with the immunoassay. Thus this diagnostic assay had a limit of sensitivity that was equivalent to an inoculum size of 5 ~ 1 x 10 cfulml. Of the infants born to the 90 women with light colonization, 79 had charts available for review. Nineteen (24%) of these infants were diagnosed with suspected sepsis on the basis of maternal history and neonatal leukocytosis with a left shift. Three of these infants had proved group B streptococcal sepsis, two by
42 Armer et al.
positive blood cultures and one by positive urine latex agglutination test. The three term infants were treated with intravenous ampicillin and were discharged from the hospital within 14 days. None of the 19 mothers had positive ICON assays. All eight infants born to the heavily colonized women had charts available for review. Four (50%) of these infants were diagnosed with suspected sepsis, all of whom were born to mothers with a positive immunoassay. Two of the mothers had previously been identified as carriers of group B streptococci and had received intrapartum prophylaxis with ampicillin. None of infants born to heavily colonized women had proved group B streptococcal sepsis. Of the infants born to mothers with negative cultures, 35 charts were reviewed for comparison. Four (11 %) infants had suspected sepsis, and none had proved group B streptococcal sepsis. The percentage of infants who were diagnosed with suspected sepsis from any cause was not significantly different between heavily and lightly colonized mothers or between colonized and noncolonized mothers.
Comment The results of this study suggest that the ICON immunoassay reliably detects the presence of heavy group B streptococcal colonization of ~ 10' cfulml and is unlikely to yield a false-positive result. The assay failed to identify women with light colonization. The overall test sensitivity of 9% was considerably lower than the 33% sensitivity reported by Gentry et al. 7 This discrepancy is most likely explained by our use of Todd-Hewitt broth culture as the standard for group B streptococci identification instead of direct plating on selective blood agar. The broth culture improves detection of low inoculum levels. The high sensitivity of the assay when testing heavily colonized patients was consistent with the findings of Gentry et aI. 7 In the instructions enclosed with the ICON test kit the manufacturers state that the majority of other vaginal microorganisms, including non-group B streptococci, were tested with the assay at concentrations of 10 8 organisms per test and cross reactivity was not present. We likewise found no evidence of cross reactivity with other organisms or interference of the assay as a result of blood, meconium, or mucus on the swab. The interpetation of the assay was straightforward, and the end point was easy to read. Other rapid detection tests evaluated in mixed highand low-risk obstetric populations have also demonstrated low overall sensitivities, which improved when analysis was restricted to heavily colonized patients. 7 • 10-1 6 These rapid tests have used latex particle agglutination and enzyme immunoassay techniques, the performance of which has been summarized in our
J anuary 1993 Am J Obstet Gyneco l
previous report. 14 Clearly, heavily colonized women are at higher risk of transmitting group B streptococci to their newborns. B• 17 However, we cannot ignore the fact that sepsis can develop in infants born to lightly colonized mothers, as was demonstrated in this study and others. 6 • 18. 19 Because this assay and the other rapid tests previously studied fail to reliably identify lightly colonized patients, the search for another test or method with improved sensitivity must continue. REFERENCES I . Boyer KM, Gadzala CA, Kelly PO, Burd LI, Gotoff SP. Selective intraprtum chemoprophylaxis of neonatal group B streptococcal early-onset disease. n. Predictive value of prenatal cultures. 1 Infect Dis 1983; 148:802-9. 2. Minkhoff H, Mead P. An obstetric approach to the prevention of early-onset group B (3-hemolytic streptococcal sepsis. AM 1 OBSTET GYNECOL 1986;154:973-7. 3. Mean PB. Group B streptococcal sepsis. In: Charles D, Glover DD, eds. Current therapy in obstetrics. Philadelphia: BC Decker, 1988; 189-95. 4. Boyer KM, Gadzala CA, Kelly PD, Burd LI, Gotoff SP. Selective intrapartum chemoprophylaxis of neonatal group B streptococcal early-onset disease. Ill. Interruption of mother-to-infant transmission. 1 Infect Dis 1983; 148:810-6. 5. Boyer KM, Gotoff SP. Prevention of early-onset neonatal group B streptococcal disease with selective intrapartum chemoprophylaxis. N Engll Med 1986;314:1665-9. 6. Lim DV, Morales Wl, Walsh AF, Kazanis D. Reduction of morbidity rates and mortality rates for neonatal group B streptococcal disease through early diagnosis and chemoprophylaxis. 1 Clin Microbiol 1986;23:489-92 . 7. Gentry YM, Hillier SL, Eschenbach DA. Evaluation of a rapid enzyme immunoassay for the detection of group B streptococcus. Obstet Gynecol 1991;78:397-401. 8. Baron lB, Finegold SB. Cultivation and isolation of viable pathogens. In: Manning S, ed. Bailey and Scou's diagnostic microbiology. 8th ed. St. Louis: CV Mosby, 1990:94-5. 9. Donatelli 1, Macone A, Goldmann DA, et al. Rapid detection of group A streptococci: comparative performance by nurses and laboratory technologists in pediatric satellite laboratories using three test kits. J Clin Microbiol 1992; 30:138-42. 10. Granato PA, Petosa MT. Evaluation of a rapid screening test for detecting group B streptococci in pregnant women. 1 Clin Microbiol 1991;29:1536-8. II. Wald ER, Dashefsky B, Green M, et al. Rapid detection of group B streptococci directly from vaginal swabs. 1 Clin Microbiol 1987;25:573-4. 12. Kontnick CM, Edberg SC. Direct detection of group B streptococci from vaginal specimens compared with quantitative culture. 1 Clin Microbiol 1990;28:336-9. 13. Skoll MA, Mercer BM, Baselski V, Gray lP, Ryan G, Sibai BM . Evaluation of two rapid group B streptococcal antigen test in labor and delivery patients. Obstet Gynecol 1991;77;322-6. 14. Brady K, Duff P, Schilhab lC, Herd M. Reliability of a rapid latex fixation test for detecting group B streptococci in the genital tract of parturients at term. Obstet Gynecol 1989;73:678-81 . 15. Clark P, Armer T, Duff P, Davidson K. Assessment of a rapid latex agglutination test for group B streptococcal colonization of the genital tract. Obstet Gynecol 1992;79: 358-63. 16. Lotz-Nolan L, Amato T, litis J, Wallen W, Packer B. Evaluation of a rapid latex agglutination test for detection
Volume 168 Number 1. Part 1
Rapid detection of group B streptococci
of group B streptococci in vaginal specimens. Eur J Clin Microbiol Infect Dis 1989;8:289-93. 17. Jones DE. Kanarek KS, Lim DV. Group B streptococcal colonization patterns in mothers and their infants. J Clin Microbiol 1984;20:438-40. 18. Morales WJ, Lim D. Reduction of group B streptococcal maternal and neonatal infections in preterm pregnancies
with premature rupture of membranes through a rapid identification test. AM J OBSTET GYNECOL 1987;157:13-6. 19. Towers CV, Garite TJ, Friedman WW, Pircon RA, Nageotte MP. Comparison of a rapid enzyme-linked immunoasorbent assay test and the Gram stain for detection of group B streptococcus in high-risk antepartum patients. AM J OBSTET GYNECOL 1990;163:965-7.
The development of abnormal heart rate patterns after absent end-diastolic velocity in umbilical artery: Analysis of risk factors Domenico Arduini, MD," Giuseppe Rizzo, MD," and Carlo Romanini, MDb Rome and Ancona, Italy OBJECTIVES: Our objectives were to evaluate the time interval elapsing between the occurrence of absent end-diastolic velocity in the umbilical artery and either the development abnormal fetal heart rate patterns or delivery and to establish the maternal and fetal factors that may affect this interval. STUDY DESIGN: Thirty-seven fetuses free of structural and chromosomal abnormalities in which the development of absent end-diastolic velocity in umbilical artery was evidenced by serial Doppler recordings were studied. At the first occurrence the following factors were considered: gestational age, presence of hypertension or preeclampsia, amniotic fluid index, severity of growth retardation, and 10 different Doppler indices calculated from extracardiac and intracardiac vascular districts. Actuarial statistical methods were applied, with the occurrence of antepartum late heart rate deceleration as the censoring variable. RESULTS: The interval between the first occurrence of absent end-diastolic velocity in umbilical artery and delivery ranged from 1 to 26 days. Indications for delivery were the development of antepartum late heart rate decelerations in 23 fetuses (62.1 %) and different maternal or fetal complications in the remaining 14 fetuses. Multivariate analysiS revealed that gestational age and the presence of hypertension and pulsations in umbilical vein were the dominant factors in determining the length of this time interval. CONCLUSION: The duration of the time interval between the occurrence of absent end-diastolic velocity in umbilical artery and abnormal heart rate pattern differs considerably among fetuses, and it is mainly determined by gestational age and presence of maternal hypertension and pulsations in umbilical vein. (AM J OBSTET GYNECOL 1993;168:43-50.)
Key words: Doppler ultrasonography, abnormal heart rate patterns, fetal distress, fetal circulation Doppler ultrasonography has been widely applied in obstetrics to study umbilical artery velocity waveforms From the Laboratory of Fetal PhYSIOlogy, Department of Obstetrus and Gynecology, Universttii Cattolzca S. Cuoreo" and the Department of Obstetncs and Gynecology, Unzversitii dl Ancona. b Supported by Grant No. CNR 920004 PF41 from the Italzan National Counczl of Research. Received for publication March 26, 1992; reVISed June 23, 1992; accepted June 25, 1992. Reprint requests: Carlo Romanmi, MD, Universltd dl Roma "Torvergata ",1st Clznzca Ostetnea e Ginecologica PollClznico S. EugenIO, P. Ie Dell 'Umanesimo, 10, 00144 Roma, Italy. 6/1/40640
0002-9378/93 $1.00
+ 0.20
as a reflection of placental vascular resistance. Several associations have been reported between qualitative changes in indices of flow and perinatal complications. I. 2 However, the absence of end-diastolic velocity seems to be the finding with the greatest clinical interest,"' 3 and it is frequently associated with fetal hypoxemia, acidemia, and poor perinatal outcome.4-7 There is no account, however, on the time interval elapsing between the occurrence of absent end-diastolic velocity and evidence of abnormal heart rate patterns requiring early delivery. Previous studies have shown
43