January 1992

FETAL MONITORING IN PERINATAL SEPSIS Dorine Day, M.D., Jay H. Ugol, M.D., Janice I. French, C.N.M., Albert Haverkamp, M.D., Robert E. Wall, M.D., and James A. McGregor, M.D., CM.

To evaluate the utility of conventional electronic fetal monitoring in detection of established perinatal sepsis, we conducted a case-controlled study of fetal monitor results in 18 patients delivering newborns with sepsis. Eleven of the 18 newborns (61%) demonstrated clinically reassuring fetal heart rate patterns, not significantly different from controls (p = 0.80). No pattern predictive of presumed perinatal sepsis could be identified. Obstetric factors traditionally associated with increased risk of perinatal infection, such as long duration of labor, use of internal fetal monitoring, and increased number of vaginal examinations were not significantly different from controls. Preterm rupture of membranes greater than 24 hours occurred in less than 50% of the cases. Clinical diagnosis of chorioamnionitis was established in only one third of the patients who delivered newborns with sepsis. In this study, use of conventional electronic fetal monitoring did not accurately identify newborns with presumed intrauterine infection.

Perinatal infection present at birth is now recognized as a common cause of neonatal mortality and morbidity.1 Interpretation of electronic fetal heart rate monitoring (EFM) is widely used to assess fetal well-being during labor. Freeman and Garite2 list perinatal infection present at birth as being associated with low Apgar scores despite normal intrapartum fetal heart rate monitoring results. They suggest that chorioamnionitis, various congenital infections, and presumed perinatal bacterial sepsis may compromise the fetus in the presence of normal EFM results and in the absence of fetal hypoxemia. In a small observational study of EFM in newborns with presumed in utero bacterial infection, Youchah et al3 noted that each of their cases demonstrated some EFM abnormality, but no identifiable pattern was associated with sepsis nor were patterns suggestive of fetal compromise specifically discussed. If interpretation of fetal heart rate patterns could identify presumed intrapartum perinatal sepsis, then appropriate treatment could be initiated before birth and improve perinatal outcome.4 Recognition of such patterns could also allow for further understanding of the pathogenesis of intrapartum fetal distress. To examine these questions, we undertook a retrospective, case-controlled, blinded analysis of EFM interpretation to determine the utility of EFM in detect-

ing presumed intrapartum perinatal sepsis manifested soon after birth (blood culture positive and obtained less than 72 hours after birth).5 Relevant antepartum and intrapartum complications were also assessed. A subset of neonates infected with early onset group B streptococcal sepsis was further evaluated. METHODS

Maternal and newborn records over a 21-month period between January 1984 and September 1985 were reviewed at the University of Colorado Health Sciences Center and Denver General Hospital. The majority of patients at the two hospitals required financial assistance for care. Criteria for inclusion in this study included: a positive blood culture drawn from the infant within 72 hours of birth, no other recognized cause of infection, and availability of intrapartum fetal monitor tracings for review. The next consecutive, uninfected newborn delivered at the same hospital who was matched for gestational age, birthweight, and gender, and for the presence of ruptured membranes became the control group. The fetal monitor tracings for the study cases and controls were interpreted blindly by one of

University of Colorado Health Sciences Center, Denver, Colorado, and Denver General Hospital, Denver, Colorado Reprint requests: Dr. Day, Denver General Hospital, 750 Cherokee St, Denver, CO 80204


Copyright © 1992 by Thieme Medical Publishers, Inc., 381 Park Avenue South, New York, NY 10016. All rights reserved.

Downloaded by: National University of Singapore. Copyrighted material.



Maternal Age (yrs) (mean/range) Race White Black Hispanic Unknown Marital status Single Married Divorced Unknown Gravidity (mean/range) Parity (mean/range) Elective abortion (mean/range) Spontaneous abortion (mean/range)

Study Cases (n = 18)

Controls (n = 18)



7 5 4 2

12 3 3 0

8 8 1 1


6 12 0 0 2.2/1-5







*AII women received public support for pregnancy-related

Table 2. Selected Outcomes of Case and Control Pregnancies


Eighteen cases of neonatal sepsis acquired in utero met the inclusion criteria. During the study period, 9249 patients were delivered, representing an incidence of neonatal sepsis of 1.95 per 1000. Eighteen control newborns were also identified. All newborns were the products of singleton gestations. No statistical difference existed between cases and controls for maternal age, race, marital status, gravidity, parity, spontaneous abortions, or elective abortions (Table 1). In addition, no difference was found for maternal weight, number of clinic visits, or estimated gestational age at time of first clinic visit. Table 2 compares the matched parameters for cases and controls. Average estimated gestational ages in study cases and controls were 35.2 versus 35.8 weeks, respectively (p = 1.0). Average birthweight was 2507 gm for study cases versus 2516 gm in controls (p = 1.0). Newborn gender (11 male and 7 female) were similarly distributed in each group (p = 1.0). Ten of 18 newborns from each group delivered preterm (less than 37 weeks) (p = 1.0). Causative organisms are listed in Table 3. Sixteen of the 18 (89%) positive blood cultures were drawn within 12 hours of birth, and the remaining two were within 60 hours of birth. Group B Streptococcus was cultured from ten of the 18 cases (GBS) and 3 of the 18 cases had Escherichia coli. Control cases had no laboratory or clinical evidence of infection, although routine cultures were not obtained. Because of the large number of GBS sepsis in the

Selected Demographic Characteristics (Case Versus Control Mothers)*

Downloaded by: National University of Singapore. Copyrighted material.

the authors (A.H.), a board-certified Maternal-Fetal medicine specialist. Tracings from latent phase labor, active phase labor, and second stage were classified as reassuring, nonreassuring, or ominous.6 Ominous patterns were defined as late decelerations, severe variable decelerations, fetal tachycardia with loss of short-term variability, marked bradycardia (less than 80 beats/minute), and absent short- or long-term variability. Tracings with diminished shortor long-term variability, mild or moderate variables, fetal tachycardia with good variability, and moderate bradycardia (80 to 100 beats/min) were classified as nonreassuring. The remaining tracings were considered reassuring. The presence of accelerations on each tracing were also evaluated. An overall subjective impression for each labor pattern was reported. Clinical chorioamnionitis was based on the presence of two or more of the following: maternal fever (temperature more than 37.8°C), maternal tachycardia (more than 100 beats/min), fetal tachycardia (more than 160 beats/min), uterine tenderness, maternal leukocytosis, or purulent or foul amniotic fluid, similar to criteria listed in Gibbs et al's work.4 In addition to descriptive statistics, analysis of continuous data was performed by Student Mest, with the chi-square or Fischer's exact test for nonparametric data. These data were processed with the SPSS computer program on an IBM XT.

Gestational age at delivery (weeks) Average birthweight Newborn gender Male Female Duration of ruptured membranes prior to labor None 0-12 hours 12-24 hours >24 hours

Table 3.

Study Cases (n = 18)

Controls (n = 18)

P Value







11 7

11 7


0 0 2 8

0 0 2 8



Organisms Group B Streptococcus Escherichia coli Group D Enterococcus Streptococcus pneumoniae Gardnerella vaginalis Haemophilus influenzae (nontypable) Staphylococcus aureus "Recovered from neonatal blood cultures.

Cases (n = 18) 10 3 1 1 1 1 1



Table 4.

Interpretations of Fetal Monitor Tracings31 Study Cases (n = 18)

Latent-phase labor Reassuring Nonreassuring Ominous Unknown or NA + Active-phase labor Reassuring Nonreassuring Ominous Unknown or NA Second stage Reassuring Nonreassuring Ominous Unknown or NA Overall impression Reassuring Nonreassuring Ominous Unknown or NA


Controls (n = 18)

GBS+ (n = 10)

13 3 1 1

(72.2%) (16.7%) (5.6%) (5.6%)

3 1 2 2

(72.2%) (5.6%) (11.1%) (11.1%)

9 (90%) 1 (10%) 0 0

9 6 2 1

(50.0%) (33.3%) (11.1%) (5.6%)

8 5 2 3

(44.4%) (27.8%) (11.1%) (16.7%)

6 (60%) 3 (30%) 0 1 (10%)

4 6 2 6

(22.2%) (33.3%) (11.1%) (33.3%)

7 2 3 6

(38.9%) (11.1%) (16.7%) (33.3%)

2 4 1 3

11 (61.1%) 5 (27.8%) 2 (11.1%) 0

13 (72.2%) 2 (11.1%) 3 (16.7%) 0

(20%) (40%) (10%) (30%)

8 (80%) 2 (20%) 0 0

*Electronic fetal monitor tracings interpreted according to criteria in Haverkamp et al. 6 No significant differences in overall or in any category over the groups existed. +GBS: group B Streptococcus; NA: not available.

Table 5.

Anemia (hematocrit < 30%) Urinary tract infection Vaginitis, any Cere I age Surgery Preterm labor

Antepartum Complications Study Cases (n = 18)

Controls* (n = 18)

GBS* (n = 10)

1 (3 unknown) 1

6 (1 unknown) 2

1 (1 unknown) 1

4 1 1 6

0 0 0 2

3 0 0 1

*No statistical significance was present in comparing study or group B Streptococcus (CBS) cases to controls.

18 controls (p = 0.11). Statistically significant differences were found in comparing some intrapartum complications. Ten of 18 study cases required oxytocin augmentation versus 3 of 18 control patients (p = 0.04). None of the GBS-infected cases labored spontaneously compared with 8 of 18 control patients (p = 0.04). None of the other intrapartum complication comparisons demonstrated statistical significance although clinically recognized chorioamnionitis approached it, occurring in 7 of 18 study cases versus 2 of 18 control mothers (p = 0.12). The three indications for cesarean delivery in the study group were placenta previa with chorioamnionitis, failure to progress, and fetal distress. In each case, GBS was cultured from the neonate. The only cesarean delivery performed in the control group was an elective repeat cesarean. No statistically significant difference was found in mode of delivery. Other labor complications in the study cases included one patient each with previa with preterm rupture of membranes, fetal distress, prolonged second stage, failure to progress, and preeclampsia. In the control group, one case of preeclampsia was noted. None of these complications occurred with enough frequency to allow statistical evaluation. Because of the high frequency of chorioamnionitis in the study group, maternal signs suggestive of intrapartum infection were statistically analyzed (Table 7). Only the study case and GBS postpartum temperature were statistically different from those of the control group (p = 0.03). None of the patients with chorioamnionitis received antipyretic agents. Four of the seven study cases with chorioamnionitis received antibiotics, whereas both control cases with chorioamnionitis received antibiotics prior to delivery (p = 0.24). Table 8 lists neonatal outcomes. Apgar scores were not statistically different at 1 minute between any group. A trend was noted for decreased 5-minute Apgar scores among study cases versus control newborns (p = 0.08). However, 13 of the 18 newborns with sepsis (72%) had 5-minute Apgar scores greater than or equal to 7, compared to 15 of the 18 controls (83%). Nine often cases with GBS sepsis had 5-min-

Downloaded by: National University of Singapore. Copyrighted material.

study group, subsequent tables will include these patients as a subset for comparison. Analysis of the fetal monitor tracings are tabulated in Table 4. No statistically significant differences were found between cases and controls for each stage of labor or the overall impression of labor (p = 0.80). The baseline fetal heart rate, presence of late or variable decelerations, presence of accelerations, or decreased variability did not differ significantly between study and control cases. Overall, 11 of the 18 tracings from cases with sepsis (61%) were judged to be reassuring. Similarly, analysis of fetal monitor tracings from GBS-infected children did not demonstrate any difference from control patients (p = 1.0). Overall, eight often GBS-infected cases had reassuring tracings. Fetal tachycardia was noted in only four of the 18 newborns with sepsis (22%) versus one of the 18 control newborns (6%) (p = 0.16). Two of the four cases with sepsis and fetal tachycardia occurred in association with chorioamnionitis. Only one of these had associated maternal fever. In the three cases of fetal tachycardia not associated with maternal fever, no other maternal etiology was identified. Cases and controls were evaluated for antepartum (Table 5) and intrapartum complications (Table 6). There were no statistically significant differences between study cases and controls or GBS-infected cases and controls in antepartum complications. Although more preterm labor was present in the study group, this failed to reach statistical significance (p = 0.23). A trend was noted in symptomatic vaginitis as it was present in 4 of 18 study cases and none of the

January 1992

FETAL MONITORING IN SEPSIS/Day, et al. Intrapartum Complications

Labor Spontaneous

Induced1 Augmented* No labor Incidence of meconium No. vaginal examinations [mean (range)] Duration of labor [mean (range)] First stage Second stage Third stage Use internal monitors Cardiac Tocodynameter Duration (hours) of internal monitors [mean (range)] Cardiac Tocodynameter Diagnosis of chorioamnionitis Labor Other Mode of delivery Unassisted vaginal Assisted vaginal Cesarean section

Study Cases (n = 18)

Controls (n = 18)

3 4


10 6 *

3 1 3


(n = 10) 0 4 5 I 1


1 2

6.0 (1-15)

6.2 (2-11)

8.8 (2-25) 0.6 (0.1-2.5) 0.1 (0-0.3)

8.3 (1-28.6) 0.7 (0.2-3.3) 0.1 (0-0.3)

10.1 (0-24) 0.7 (0-2.5) 0.1 (0-0.3)

14 10

13 7


3.2 (0-14) 2.7 (0-14)

6.9 (1-15)

3.3 (0-8.1) 2.1 (0-7.9)


6 I



2 1

12 3 3

10 7 1

6 1 3

*p = 0.04 for group B Streptococcus (GBS) cases versus controls Labor induced with either artificial rupture of membranes, oxytocin infusion, or prostaglandin gel application. •Continuous intravenous oxytocin infusion. § p = 0.04 for study cases versus controls. f

ute Apgar scores greater than or equal to 7. Infant resuscitation techniques were similar in all groups. All five newborns with sepsis under 2000 gm died. The other neonatal death occurred in a newborn with overwhelming GBS sepsis. This infant weighed 2760 gm at birth and had Apgar scores of 8 and 8 at 1 and 5 minutes. No neonatal deaths occurred in the control group, although 5 of the 18 newborns weighed less than 2000 gm. DISCUSSION

In this retrospective, case-controlled, blinded analysis, interpretation of EFM did not differ from

Table 7.

Maternal temperature Intrapartum (°C) Postpartum (°C) Maternal pulse Predelivery* Postdelivery Maternal white blood cells Predelivery Postdelivery

*p = 0.03.

Maternal Signs of Infection*

Study Cases (n = 18)

Controls (n = 18)

37.4 37.5

37.0 37.1

37.4 37.7



102 94


14,600 (5 unknown) 15,200 (11 unknown)

13,200 (6 unknown) 17,100 (14 unknown)

11,800 (4 unknown) 15,600 (4 unknown)

*AII values listed are averaged values. GBS: group B Streptococcus.


controls, and no patterns indicative of presumed intrapartum perinatal infection were identified. Sixty-one percent of case newborns had reassuring tracings, not statistically different from the control group (72%) (p = 0.80). Similar to Youchah et al,3 no fetal heart rate pattern characteristic of intrapartum fetal sepsis was consistently identified, although fetal tachycardia was reported by Youchah et al in 60% of cases. In contrast, fetal tachycardia was present in only 22% of our study cases. In a controlled, prospective study, Hager and Pauly7 investigated whether intrapartum fetal tachycardia was a reliable indicator of fetal infectious morbidity. Of their 30 neonates with tachycardia, only one had a positive culture, indicating bacte-

GBS+ (n=



Downloaded by: National University of Singapore. Copyrighted material.

Table 6.

Table 8.

Apgar scores 1 minute average 5 minute average Neonatal deaths >2000 gm

Fetal monitoring in perinatal sepsis.

To evaluate the utility of conventional electronic fetal monitoring in detection of established perinatal sepsis, we conducted a case-controlled study...
1MB Sizes 0 Downloads 0 Views