Milwidsky, Finci-Yeheskel, and Mayer
11. 12. 13.
14. 15.
16. 17.
U. Ripening of the human uterine cervix related to changes in collagen, glycosaminoglycans, and collagenolytic activity. AM J OBSTET GVNECOL 1983; 14 7:662-6. Shotton DM. Elastase. Methods EnzymoI1970;19: 113-40. Cawston Te, Mercer E, Tyler JA. The activation oflatent pig synovial collagenase. Biochim Biophys Acta 1981;657:73-83. Rajabi MR, Dean DD, Beydoun SN, Woessner JF. Elevated tissue levels of collagenase during dilation of uterine cervix in human parturition. AM J OBSTET GV:>iECOL 1988;159:971-6. Mayer M. Biochemical and biological aspects of the plasminogen activation system. Clin Biochem 1990;23:197211. Junqueira LC, Zugaib M, Montes GS, Toledo OM, Krisztan RM, Shigihara KM. Morphologic and histochemical evidence for the occurrence of collagenolysis and for the role of neutrophilic polymorphonuclear leukocytes during cervical dilation. AM J OBSTET GVNECOL 1980; 138:273-81. McGregor JA, Schoonmaker IN, Lunt BD, Lawellin DW. Antibiotic inhibition of bacterially induced fetal membrane weakening. Obstet Gynecol 1990;76: 124-8. Higazi AAR, Mayer M. In vitro inhibition of urokinase by penicillins. Thromb Haemost 1988;60:305-7.
February 1992 Am J Obstet Gynecol
18. Higazi A, Mayer M. Inhibition of tissue type plasminogen activator by penicillins. J Clin Chern Clin Biochem 1988;26:380. 19. Higazi AAR, Mayer M. Effects of penicillins and 6-amino hexanoic acid on the kinetics of human plasmin. Biochern J 1989;260:609-12. 20. Deutsch DG, Mertz ET. Plasminogen: purification from human plasma by affinity chromatography. Science 1970; 170: 1095-6. 21. Finci Z, Nachshon I, Sharoni Y, Mayer M. Functional assay of plasminogen activator by hydrolysis of HC-globin, Exp Hematol 1986; 14:293-7. 22. Morales WJ, AngelJL, O'Brien WF, Knuppel RA, Finazo MA. Randomized study of antibiotic therapy in idiopathic preterm labor. Obstet Gynecol 1988;72:829-33. 23. Segal IH. Biochemical calculations. 2nd ed. New York: John Wiley, 1976:208-323. 24. Harris ED, Vater CA. Vertebrate collagenase. Methods Enzymol 1982;82:423-52. 25. Newton ER, Dinsmoor M, Gibbs RS. A randomized, blinded, placebo-controlled trial of antibiotics in idiopathic pre term labor. Obstet Gynecol 1989;74:562-6.
The effect of fetal sepsis on umbilical cord blood gases Bruce A. Meyer, MD, Jan E. Dickinson, MD, Charles Chambers, MD, and Valerie M. Parisi, MD, MPH Houston, Texas The relationship between fetal sepsis and acid-base status is unknown. We hypothesized that in utero sepsis would result in fetal metabolic acidemia. In a retrospective study during a 38-month period, the acid-base status at birth of neonates with in utero sepsis, documented by positive blood CUltures, was reviewed. Compared with term neonates, preterm neonates had a 22-fold increase in the risk of bacteremia at birth. In spite of this increased risk of sepsis, there was no significant alteration in arterial pH in preterm septic neonates when compared with preterm controls. Fetal sepsis at term was accompanied by a statistically significant reduction in arterial pH (7.21 ± 0.07) compared with controls (7.26 ± 0.06, P < 0.05). When controlled for other variables, the decrease in arterial pH at term was correlated with an increased duration of labor (7.3 ± 0.7 in controls vs 10.8 ± 0.9 hours in neonates with sepsis, p < 0.05). The classic predictors of chorioamnionitis were found to be poor prognostiC indicators of fetal bacteremia. Fetal sepsis at term is associated with a deterioration in the fetal acid-base status and a prolongation of labor. (AM J OBSTET GVNECOL 1992;166:612-7.)
Key words: Fetal sepsis, fetal acid-base balance, umbilical cord blood gases
From the Departments of Obstetrics, Gynecology and Reproductive Sciences and Pediatrics, University of Texas Medical School at Houston. Received for publication Ma,V 23,1990; revised May 8,1991; acceptedJune 26,1991. Reprint requests: Bruce A. Me,ver, MD. St. Luke's Perinatal Center, 4400 Womall Road, Kansas City, MO 64111. 611 132025
612
Controversy exists regarding fetal infection in utero and its effect on acid-base status. In the adult, septicemia has been associated with progressive metabolic acidemia. 1 This has lead to the inference that in utero sepsis may have a similar deleterious effect on fetal acidbase balance. Fetal in utero infection has been implicated as a cause of birth asphyxia.2.3 In contrast, in the
Fetal sepsis 613
Volume 166 Number 2
presence of premature rupture of the membranes, chorioamnionitis does not alter umbilical cord pH.' Additionally, in the presence of chorioamnionitis, immediate delivery affords no maternal' or neonatal 6 advantage when appropriate antibiotic therapy is provided. Thus the exact relationship between fetal sepsis and umbilical cord blood gas parameters is unknown. The purpose of this study was to establish the effect of infection acquired in utero on fetal acid-base status at birth. We hypothesized that in utero sepsis would result in a fetal metabolic acidemia. Material and methods
The neonatal records of all infants with posItIve blood cultures within the first 24 hours of life who were admitted to the nursery at Hermann Hospital (Houston) from January 1986 through March 1989 were reviewed. Neonates with positive blood cultures within 24 hours of birth were considered to have acquired their infection in utero. Because of concerns with contamination, infants with cultures that revealed nonaureus staphylococcus were included only if two blood cultures were positive. When mothers with premature rupture of the membranes were treated with antepartum antibiotics, their neonates were excluded from analysis. Umbilical cord arterial and venous blood gas parameters were used to determine the fetal acid-base status at birth. Sixty-six neonates met these criteria. In five newborns with sepsis blood gases could not be determined from the umbilical cord, and these values were unavailable for analysis. Only those neonates with an arterial specimen were included in the study. The maternal obstetric records of the 61 neonates who met the study criteria were then analyzed. Controls were obtained by manually reviewing the maternal and neonatal medical record of the next consecutive delivery; they were specifically matched for maternal age and parity, socioeconomic status, gestational age, birth weight, and mode of delivery. Health care delivery service was used as a marker for socioeconomic status. Patients cared for by the University of Texas Women's Clinic, by a private health maintenance organization, or by private physicians were eligible for inclusion in the study. The following maternal criteria were evaluated: chronic medical conditions, complications of pregnancy, duration of ruptured membranes, number of vaginal examinations, duration of labor, requirement for oxytocin augmentation, fetal heart rate (FHR) abnormalities, presence of meconium-stained amniotic fluid, mode of delivery, type of analgesia, febrile morbidity, and antibiotic administration. All nulliparous women were managed by a modified active management of labor protocol, as previously described.' Oxytocin was administered to multiparous women at the discretion of the attending physician. Continuous e1ec-
tronic FHR monitoring was performed in all patients. FHR abnormalities were defined according to the criteria of Freeman and Garite. 8 The diagnosis of chorioamnionitis was based on the presence of maternal temperature :::::38 C and one or more of the following: maternal tachycardia, fetal tachycardia, uterine tenderness, or malodorous amniotic fluid. Umbilical cord arterial and venous blood gases were obtained at all births at Hermann Hospital during the study period. Immediately after delivery, a segment of umbilical cord was clamped, divided, and handed to the delivery room nurse. Preheparinized 3 ml syringes (Marquest Medical Products, Englewood, Col.) were used to obtain a consistent dose and preparation of heparin. All samples were obtained in the delivery room and placed on ice. Analysis was performed within 15 minutes of delivery by a laboratory technician on a blood gas analyzer (IL1312, Allied Co., Lexington, Mass.) located in the satellite laboratory adjacent to the labor and delivery area. Statistical analysis was performed by paired t test, multivariate Hotelling [2 test, X", and Fisher's exact test as appropriate. This was accomplished with statistical software packages (SPSS-PC, SPSS Inc., Chicago, and Abstat, Anderson-Bell, Parker, Col.). A p value of 0.20). The preterm birth rate at our institution during the study time period was 5.0%. Thus 4.3% of preterm and 0.2% of term deliveries met the criteria for sepsis acquired in utero. The incidence of in uteroacquired infection was significantly different between these groups (p < 0.05, odds ratio 22.4, 95% confidence limits 15.9 and 31.7). Analysis of the arterial umbilical cord gases revealed a significant decrease in pH in the septic neonates when compared with controls (p < 0.05). In addition, the bicarbonate and base excess were significantly lower in the septic neonates. There
614 Meyer et al.
February 1992 Am J Obstet Gynecol
Table I. Demographic characteristics Term
Overall Septic (n = 61)
Control (n = 29)
25 ± 0.7 1.2 ± 0.2 34.4 ± 0.6
26 ± 0.7 1.3 ± 0.3 34.3 ± 0.6
2418 ± 133
2369 ± 142
Control (n = 61)
Age (yr) Parity Gestational age (wk) Birth weight (gm)
I
Preterm Septic (n = 29)
Control (n = 32)
26 ± 1.0 1.2 ± 0.2 39.3 ± 0.2
26 ± 1.0 1.5 ± 0.5 39.1 ± 0.2
25 ± 0.9 l.l ± 0.3 30.5 ± 0.6
26 ± 0.9 1.1 ± 0.2 30.0 ± 0.6
3422 ± 73
3449 ± 99
1489 ± 98
1456 ± 103
I
I
Septic (n = 32)
Table II. Umbilical arterial cord gas values Overall Control (n = 61) Septic (n = 61) Term Control (n = 29) Septic (n = 29) Preterm Control (n = 32) Septic (n = 32) *p
pH
P02
Peo2
Bicarbonate
Base excess
7.26 ± 0.06 7.23 ± 0.07*
18.0 ± 5.8 18.1 ± 6.8
53.1 ± 9.1 54.6 ± 8.8
24.1 ± 2.3 22.5 ± 3.1*
-2.9 ± 2.3 -4.5 ± 3.3*
7.26 ± 0.06 7.21 ± 0.07*
18.1 ± 4.8 17.2 ± 5.4
55.2 ± 8.7 57.6 ± 9.2
24.7 ± 2.0 22.1 ± 3.4*
-2.7 ± 2.2 -5.0 ± 3.1*
7.27 ± 0.07 7.25 ± 0.07
18.0 ± 6.7 19.0 ± 8.0
51.1 ± 9.1 51.7 ± 7.5
23.5 ± 2.6 22.9 ± 2.7
-3.3 ± 2.4 -3.9 ± 3.3
< 0.05, versus control.
was no alteration in arterial respiratory gas values when septic and control groups were compared. These data are summarized in Table II. Venous umbilical cord blood gases analysis also showed a significant decrease in pH in the septic cases when compared with controls (P < 0.05). Bicarbonate and base excess were also significantly lower in the septic neonates. Venous respiratory parameters were not altered. These data are summarized in Table III. The data were divided into preterm «37 completed weeks of gestation) and term subsets to evaluate the effect of gestational age on cord blood gases in this population. There were no differences in umbilical arterial or venous blood gas parameters when preterm septic neonates were compared with preterm controls (Tables II and III). However, within the term subset there was a significant decrease in the arterial and venous pH (p < 0.05). In both arterial and venous specimens, the bicarbonate and base excess were significantly lower in the septic neonates. There were no alterations in arterial P0 2 or Peo 2 when term septic neonates were compared with controls. The respiratory component of venous gases were altered, as venous P02 was significantly lower while Peo 2 was significantly higher in the septic group (Tables II and III). FHR tracings were evaluated by the attending physician. The occurrence of any abnormality on FHR tracing was noted in 36% (n = 22) of septic and 33% (n = 20) of control neonates (p > 0.20). Two thirds of these abnormalities consisted offetal tachycardia (> 160
and 0.20). There was a significant increase in the duration of labor in the overall septic versus control group (p < 0.05). When preterm septic and control groups were compared, there was no significant difference in the length of labor (p > 0.20). However, there was a significant prolongation of labor when the term septic subset was compared with term controls (p < 0.05). The observed difference in the duration of labor in the overall group was accounted for by the 45% increase in the length of labor among term septic neonates. We have previously defined fetal acidemia as an arterial cord pH 24 hr Labor> 12 hr Meconium Chorioamnionitis Endometritis
Term
(7) (5) (1) (9)
I
(19)* (15)* (18)* (17)*
17% 17% 0% 7%
(5) (5) (0) (2)
I
(14)* (9) (5)* (6)
(2) (0) (1) (7)
Septic (n = 32)
58.7 ± 8.1*
(5) (6)* (13)* (11)
*P < 0.05, versus control. crease in febrile morbidity among the mothers of septic neonates. Intrapartum fever (>38 0 C) occurred more frequently in mothers of septic newborns when compared with mothers of controls (26% vs 2%, respectively; p < 0.05). This febrile morbidity was evident in both the term and preterm septic subgroups when compared with controls (24% vs 0% and 28% vs 3%, respectively). The clinical diagnosis of chorioamnionitis was made significantly more often in the overall term and in the preterm septic groups when compared with controls (p < 0.05). In addition, there was a significant increase in the incidence of postpartum endomyometritis when mothers of septic newborns were compared with controls (p < 0.05) (Table IV). The duration of ruptured membranes was significantly longer in the overall septic group when compared with controls (p < 0.05). The preterm septic subset accounted for this difference (Table IV). In the overall group, there was a significant increase in the incidence of prolonged rupture of membranes (>24 hours) when septic neonates were compared with controls (p < 0.05). This difference was also attributable to the preterm subset (Table IV). There was a significant increase in the incidence of meconium-stained amniotic fluid in the over-
all septic group when compared with controls (p < 0.05). There was no statistically significant difference in the requirement for oxytocin augmentation in any of the populations studied (overall, 28% (17) vs 39% (24); term, 38% (11) vs 48% (14); preterm, 19% (6) vs 31 % (10); P > 0.20). Similarly, there were no significant differences in regional analgesia usage (overall, 31% (20) vs 39% (24); term, 59% (17) vs 69% (20); preterm, 9% (3) vs 13% (4); P > 0.20). There were also no significant differences in the assignment of a 5-minute Apgar score 0.20). The bacterial pathogens isolated included a variety of microorganisms, two thirds of which were group B streptococcus, Escherichia coli and staphylococcus species. Analysis of the isolates demonstrated a significant increase in the incidence of group B streptococcal infection in term septic when compared with preterm septic neonates (19/29 vs 11/32, respectively, p < 0.05). There were no differences in arterial or venous umbilical cord blood gas parameters between neonates with group B streptococcal infection compared with septic neonates with
616 Meyer et al.
other pathogens. One neonate in each subset had more than one pathogen identified. There were eight neonatal deaths among the septic neonates (three term, five preterm), whereas there was one neonatal death in the control group, in the preterm subgroup.
Comment Delivery of the fetus when intrauterine infection is diagnosed or strongly suspected is currently accepted obstetric practice. 9 . 10 However, no maternal' or neonatal6 benefit to immediate delivery has been demonstrated with intrapartum antibiotic administration. Little is known about the relationship between in utero infection and fetal acid-base status. In this study, we have demonstrated that fetal sepsis is associated with a decrease in the umbilical arterial pH, accompanied by metabolic alterations in the arterial blood gases. These metabolic changes are evidenced by decreases in pH, bicarbonate, and base excess without alterations in arterial Po 2 , or Pea•. Assessment of the influence of gestational age on the arterial blood gas values indicates that the differences are accounted for by the term septic subgroup. Term septic neonates display a significant decrease in arterial pH, bicarbonate, and base excess when compared with matched nonseptic controls. Similar changes are observed in the venous umbilical cord blood gases. In the term septic subgroup, both metabolic and respiratory alterations are observed, as venous P0 2 is significantly decreased while Peo z was significantly elevated. This may be due to an impairment in placental function due to local infection or to a reduction in maternal perfusion in the presence of sepsis. The latter speculation is unlikely to be of significance, as only four of these women had blood culture-proved sepsis and only 30% had a clinical diagnosis of chorioamnionitis. These factors may compound the metabolic changes associated with sepsis in fetal arterial blood gases. Curiously, cord blood gases from the preterm septic newborns are not significantly different from their nonseptic controls, in spite of the greater incidence of chorioamnionitis and prolonged rupture of the membranes in the mothers of this group. We postulate that the lack of alteration in acid-base status is likely due to the significantly shorter duration of labor in this subpopulation. This suggests that fetal sepsis with a short duration of labor does not cause significant changes in fetal acid-base status. We speculate that fetal sepsis plus prolonged labor may cause accelerated deterioration in acid-base status, as demonstrated in the term septic subgroup. To evaluate the etiology of the observed metabolic changes in the term septic neonates, several obstetric factors were assessed. There were no significant differences found in the duration of ruptured mem-
February 1992 Am J Obstet Gynecol
branes, incidence of prolonged ruptured membranes, oxytocin usage for augmentation of labor, fetal meconium passage, or regional analgesia use in the term septic versus nonseptic groups. The only obstetric factor in which term septic neonates differed from controls was in the duration of labor. A significant prolongation in the mean duration of labor was present when term septic neonates were compared with their nonseptic term controls. This significant prolongation of labor held true even when term septic neonates were compared with the preterm septic and preterm control groups. In addition, the incidence of labor duration > 12 hours was significantly higher only in the term septic subset. Eight of the nine neonates with an arterial pH of 12 hours. This lends further credence to the detrimental effect of prolonged labor on acid-base balance in the presence of fetal sepsis. It is unclear from this data whether the in utero sepsis per se had a direct effect on the fetus (i.e., through lactic or other fixed acids) or whether this was a physiologic response to the increased duration of labor in the face of an already stressed septic fetus. Deterioration in fetal acid-base status with increasing duration of labor has been demonstrated previously by other authors. I I However, as there were no differences in the length of labor when preterm septic neonates were compared with preterm controls, it is unlikely that sepsis directly causes an increase in the length of labor. Thus the observed metabolic alterations in arterial blood gas parameters appears to be attributable to the combination of fetal sepsis and increased duration of labor. To ensure that our study population was representative of true fetal infection, we chose to assess only neonates with bacteremia documented by positive blood culture within the first 24 hours of life. These selection criteria may not include all infected neonates but do provide a study population with clearly defined sepsis. The intrapartum maternal administration of antibiotics may have rendered neonatal cultures negative in a bacteremic newborn; the exclusion of these patients from our study may have confounded the interpretation of our results. The spectrum of infectious microorganisms in this study was comparable to those of previous investigators. 12 . 13 Group B streptococcal infection did not disproportionately contribute to the observed differences in arterial blood gas values. The classic maternal markers of chorioamnionitis were poor prognostic indicators of fetal sepsis. Clinical chorioamnionitis was diagnosed in only 27% of mothers of septic neonates. Features often associated with maternal antenatal sepsis, such as prolonged rupture of the membranes, duration of labor> 12 hours, and postpartum endomyometritis, were present significantly more often in mothers of septic neonates than in those of nonseptic
Fetal sepsis
Volume 166 Number 2
controls. Because the majority of mothers ofbacteremic neonates do not have clinical findings of chorioamnionitis, the clinical diagnosis of fetal intrauterine sepsis is extremely difficult. Additionally, the occurrence of fetal sepsis may be independent of clinical maternal infection. Previous studies" 6 assessing the effect of in utero sepsis on umbilical cord blood gas parameters have focused on the diagnosis of chorioamnionitis and its impact on the fetus. Laboratory investigation revealed that the majority of these neonates were not septic. Therefore the umbilical cord blood gas values do not necessarily reflect the effect of sepsis on fetal acid-base balance. The incidence of early neonatal bacterial sepsis in our study was four in 1000 live births, in keeping with previous reports of 1 to 5 in 1000 live births. 12 Early identification of the fetus with in utero sepsis is critical for instituting appropriate antibiotic therapy. However, the diagnosis of fetal sepsis is difficult, with recognizable chorioamnionitis being present in only the minority of cases. A high index of suspicion remains the major diagnostic criterion, with prolonged rupture of membranes and increased duration of labor being major risk factors. The 22-fold increase in the risk of bacteremia at birth in preterm versus term neonates may be due to a number of factors. These include ascending infection as the cause of preterm premature rupture of the membranes and the preterm delivery and the significant increase in the duration of membrane rupture in preterm neonates. In our study fetal sepsis at term was accompanied by a statistically significant reduction in arterial and venous pH. The metabolic alterations in the arterial blood gases are likely due to lactic acidemia secondary to the infection. Although of statistical significance, these differences remain well within the previously reported normal range at our institution 11 and may not be clinically significant. Our data suggest that acidemia at birth is a poor predictor of fetal sepsis, as >85% of septic neonates had a normal cord pH. Additionally, these data confirm the findings of Ramin et al. 15 that neither fetal nor maternal sepsis is an indication for cesarean section. The clinical significance of these findings lies in the increased incidence of fetal acidemia in septic neonates and in the association between increased duration of labor and fetal sepsis. From our study it is unclear whether fetal sepsis can contribute to a prolongation of labor or whether prolonged labor leads to fetal sepsis. However, our findings lend support to the concepts of active management of labor, which have been dem-
617
onstrated to reduce the duration of labor.16 In our institution, application of these principles has decreased the duration of labor without altering fetal acid-base status. 7 There are four major conclusions that can be drawn from our study: (1) Fetal sepsis with short duration of labor does not cause significant changes in acid-base status; (2) fetal sepsis alone does not cause an increase in the duration of labor; (3) there is, however, an association between the duration of labor, fetal sepsis and fetal acidemia at term; (4) preterm neonates have a significantly increased risk of fetal sepsis when compared with term neonates. Further study of the potential clinical effects of fetal sepsis is indicated. REFERENCES I. Andriole VT. Bacterial infection. In: Burrow GN, Ferris
2. 3. 4.
5. 6. 7.
8. 9. 10. 11. 12. 13. 14.
15. 16.
TF, eds. Medical complications during pregnancy. Philadelphia: WB Saunders, 1987:302-32. Peevey KJ, Chalhub EG. Occult group B streptococcal infection: an important cause of intrauterine asphyxia. AM J OBSTET GYNECOL 1983;146:989-90. Naeye RL, Peters Ee. Amniotic fluid infections with intact membranes leading to perinatal death: A prospective study. Pediatrics 1978;61: 171. Vintzileos AM, Campbell WA, Nochimson DJ, et al. The fetal biophysical profile in patients with premature rupture of the membranes-an early predictor of fetal infection. AM J OBSTET GYNECOL 1985; 152:510-6. Gilstrap LC, Leveno KJ, Cox SM, et al. Intrapartum treatment of acute chorioamnionitis: impact on neonatal sepsis. AM J OBSTET GYNECOL 1988; 159:579-83. Hauth JC, Gilstrap LC, Hankins GDV, et al. Term maternal and neonatal complications of acute chorioamnionitis. Obstet Gynecol 1985;66:59-62. Thorp .lA, Boylan PC, Parisi VM, et al. Effects of highdose oxytocin augmentation on umbilical cord blood gas values in primigravid women. AM J OBSTF.T GYNECOL 1988; 159:670-5. Freeman RK, Garite TJ. Fetal heart rate monitoring. ed I. Baltimore: Williams & Wilkins, 1981 :63-83. Schwartz RH, Fruiterman JP. Life-threatening infections in pregnancy. Clin Obstet Gynecol 1976; 19:561. MacVicar J. In: Charles D, Finland M, eds. Chorioamnionitis, obstetric and perinatal infections. Philadelphia: Lea & Febiger, 1973:491-5. Kubli FW. Influence of labor on fetal acid-base balance. Clin Obstet Gynecol 1968; 11: 168-91. St Geme JW, Polin RA. Neonatal sepsis. Progress in diagnosis and management. Drugs 1988;36:784-800. Siegel JD, McCracken GH. Sepsis neonatorum. N Engl J Med 1981;304:642-7. Boylan PC, Parisi VM. Fetal acid-base balance. In: Creasy RK, Resnick R, eds. Maternal-fetal medicine: principles and practice. ed 2. Philadelphia: WB Saunders, 1989:36274. Ramin S, Maberry M, Gilstrap L, et al. Fetal acidemia in pregnancies with chorioamnionitis. Obstet Gynecol 1990;76:351-4. O'Driscoll K, Foley M, MacDonald D. Active management of labor as an alternative to cesarean section for dystocia. Obstet Gynecol 1984;63:485-90.
11. 12. 13.
14. 15.
16. 17.
U. Ripening of the human uterine cervix related to changes in collagen, glycosaminoglycans, and collagenolytic activity. AM J OBSTET GVNECOL 1983; 14 7:662-6. Shotton DM. Elastase. Methods EnzymoI1970;19: 113-40. Cawston Te, Mercer E, Tyler JA. The activation oflatent pig synovial collagenase. Biochim Biophys Acta 1981;657:73-83. Rajabi MR, Dean DD, Beydoun SN, Woessner JF. Elevated tissue levels of collagenase during dilation of uterine cervix in human parturition. AM J OBSTET GV:>iECOL 1988;159:971-6. Mayer M. Biochemical and biological aspects of the plasminogen activation system. Clin Biochem 1990;23:197211. Junqueira LC, Zugaib M, Montes GS, Toledo OM, Krisztan RM, Shigihara KM. Morphologic and histochemical evidence for the occurrence of collagenolysis and for the role of neutrophilic polymorphonuclear leukocytes during cervical dilation. AM J OBSTET GVNECOL 1980; 138:273-81. McGregor JA, Schoonmaker IN, Lunt BD, Lawellin DW. Antibiotic inhibition of bacterially induced fetal membrane weakening. Obstet Gynecol 1990;76: 124-8. Higazi AAR, Mayer M. In vitro inhibition of urokinase by penicillins. Thromb Haemost 1988;60:305-7.
February 1992 Am J Obstet Gynecol
18. Higazi A, Mayer M. Inhibition of tissue type plasminogen activator by penicillins. J Clin Chern Clin Biochem 1988;26:380. 19. Higazi AAR, Mayer M. Effects of penicillins and 6-amino hexanoic acid on the kinetics of human plasmin. Biochern J 1989;260:609-12. 20. Deutsch DG, Mertz ET. Plasminogen: purification from human plasma by affinity chromatography. Science 1970; 170: 1095-6. 21. Finci Z, Nachshon I, Sharoni Y, Mayer M. Functional assay of plasminogen activator by hydrolysis of HC-globin, Exp Hematol 1986; 14:293-7. 22. Morales WJ, AngelJL, O'Brien WF, Knuppel RA, Finazo MA. Randomized study of antibiotic therapy in idiopathic preterm labor. Obstet Gynecol 1988;72:829-33. 23. Segal IH. Biochemical calculations. 2nd ed. New York: John Wiley, 1976:208-323. 24. Harris ED, Vater CA. Vertebrate collagenase. Methods Enzymol 1982;82:423-52. 25. Newton ER, Dinsmoor M, Gibbs RS. A randomized, blinded, placebo-controlled trial of antibiotics in idiopathic pre term labor. Obstet Gynecol 1989;74:562-6.
The effect of fetal sepsis on umbilical cord blood gases Bruce A. Meyer, MD, Jan E. Dickinson, MD, Charles Chambers, MD, and Valerie M. Parisi, MD, MPH Houston, Texas The relationship between fetal sepsis and acid-base status is unknown. We hypothesized that in utero sepsis would result in fetal metabolic acidemia. In a retrospective study during a 38-month period, the acid-base status at birth of neonates with in utero sepsis, documented by positive blood CUltures, was reviewed. Compared with term neonates, preterm neonates had a 22-fold increase in the risk of bacteremia at birth. In spite of this increased risk of sepsis, there was no significant alteration in arterial pH in preterm septic neonates when compared with preterm controls. Fetal sepsis at term was accompanied by a statistically significant reduction in arterial pH (7.21 ± 0.07) compared with controls (7.26 ± 0.06, P < 0.05). When controlled for other variables, the decrease in arterial pH at term was correlated with an increased duration of labor (7.3 ± 0.7 in controls vs 10.8 ± 0.9 hours in neonates with sepsis, p < 0.05). The classic predictors of chorioamnionitis were found to be poor prognostiC indicators of fetal bacteremia. Fetal sepsis at term is associated with a deterioration in the fetal acid-base status and a prolongation of labor. (AM J OBSTET GVNECOL 1992;166:612-7.)
Key words: Fetal sepsis, fetal acid-base balance, umbilical cord blood gases
From the Departments of Obstetrics, Gynecology and Reproductive Sciences and Pediatrics, University of Texas Medical School at Houston. Received for publication Ma,V 23,1990; revised May 8,1991; acceptedJune 26,1991. Reprint requests: Bruce A. Me,ver, MD. St. Luke's Perinatal Center, 4400 Womall Road, Kansas City, MO 64111. 611 132025
612
Controversy exists regarding fetal infection in utero and its effect on acid-base status. In the adult, septicemia has been associated with progressive metabolic acidemia. 1 This has lead to the inference that in utero sepsis may have a similar deleterious effect on fetal acidbase balance. Fetal in utero infection has been implicated as a cause of birth asphyxia.2.3 In contrast, in the
Fetal sepsis 613
Volume 166 Number 2
presence of premature rupture of the membranes, chorioamnionitis does not alter umbilical cord pH.' Additionally, in the presence of chorioamnionitis, immediate delivery affords no maternal' or neonatal 6 advantage when appropriate antibiotic therapy is provided. Thus the exact relationship between fetal sepsis and umbilical cord blood gas parameters is unknown. The purpose of this study was to establish the effect of infection acquired in utero on fetal acid-base status at birth. We hypothesized that in utero sepsis would result in a fetal metabolic acidemia. Material and methods
The neonatal records of all infants with posItIve blood cultures within the first 24 hours of life who were admitted to the nursery at Hermann Hospital (Houston) from January 1986 through March 1989 were reviewed. Neonates with positive blood cultures within 24 hours of birth were considered to have acquired their infection in utero. Because of concerns with contamination, infants with cultures that revealed nonaureus staphylococcus were included only if two blood cultures were positive. When mothers with premature rupture of the membranes were treated with antepartum antibiotics, their neonates were excluded from analysis. Umbilical cord arterial and venous blood gas parameters were used to determine the fetal acid-base status at birth. Sixty-six neonates met these criteria. In five newborns with sepsis blood gases could not be determined from the umbilical cord, and these values were unavailable for analysis. Only those neonates with an arterial specimen were included in the study. The maternal obstetric records of the 61 neonates who met the study criteria were then analyzed. Controls were obtained by manually reviewing the maternal and neonatal medical record of the next consecutive delivery; they were specifically matched for maternal age and parity, socioeconomic status, gestational age, birth weight, and mode of delivery. Health care delivery service was used as a marker for socioeconomic status. Patients cared for by the University of Texas Women's Clinic, by a private health maintenance organization, or by private physicians were eligible for inclusion in the study. The following maternal criteria were evaluated: chronic medical conditions, complications of pregnancy, duration of ruptured membranes, number of vaginal examinations, duration of labor, requirement for oxytocin augmentation, fetal heart rate (FHR) abnormalities, presence of meconium-stained amniotic fluid, mode of delivery, type of analgesia, febrile morbidity, and antibiotic administration. All nulliparous women were managed by a modified active management of labor protocol, as previously described.' Oxytocin was administered to multiparous women at the discretion of the attending physician. Continuous e1ec-
tronic FHR monitoring was performed in all patients. FHR abnormalities were defined according to the criteria of Freeman and Garite. 8 The diagnosis of chorioamnionitis was based on the presence of maternal temperature :::::38 C and one or more of the following: maternal tachycardia, fetal tachycardia, uterine tenderness, or malodorous amniotic fluid. Umbilical cord arterial and venous blood gases were obtained at all births at Hermann Hospital during the study period. Immediately after delivery, a segment of umbilical cord was clamped, divided, and handed to the delivery room nurse. Preheparinized 3 ml syringes (Marquest Medical Products, Englewood, Col.) were used to obtain a consistent dose and preparation of heparin. All samples were obtained in the delivery room and placed on ice. Analysis was performed within 15 minutes of delivery by a laboratory technician on a blood gas analyzer (IL1312, Allied Co., Lexington, Mass.) located in the satellite laboratory adjacent to the labor and delivery area. Statistical analysis was performed by paired t test, multivariate Hotelling [2 test, X", and Fisher's exact test as appropriate. This was accomplished with statistical software packages (SPSS-PC, SPSS Inc., Chicago, and Abstat, Anderson-Bell, Parker, Col.). A p value of 0.20). The preterm birth rate at our institution during the study time period was 5.0%. Thus 4.3% of preterm and 0.2% of term deliveries met the criteria for sepsis acquired in utero. The incidence of in uteroacquired infection was significantly different between these groups (p < 0.05, odds ratio 22.4, 95% confidence limits 15.9 and 31.7). Analysis of the arterial umbilical cord gases revealed a significant decrease in pH in the septic neonates when compared with controls (p < 0.05). In addition, the bicarbonate and base excess were significantly lower in the septic neonates. There
614 Meyer et al.
February 1992 Am J Obstet Gynecol
Table I. Demographic characteristics Term
Overall Septic (n = 61)
Control (n = 29)
25 ± 0.7 1.2 ± 0.2 34.4 ± 0.6
26 ± 0.7 1.3 ± 0.3 34.3 ± 0.6
2418 ± 133
2369 ± 142
Control (n = 61)
Age (yr) Parity Gestational age (wk) Birth weight (gm)
I
Preterm Septic (n = 29)
Control (n = 32)
26 ± 1.0 1.2 ± 0.2 39.3 ± 0.2
26 ± 1.0 1.5 ± 0.5 39.1 ± 0.2
25 ± 0.9 l.l ± 0.3 30.5 ± 0.6
26 ± 0.9 1.1 ± 0.2 30.0 ± 0.6
3422 ± 73
3449 ± 99
1489 ± 98
1456 ± 103
I
I
Septic (n = 32)
Table II. Umbilical arterial cord gas values Overall Control (n = 61) Septic (n = 61) Term Control (n = 29) Septic (n = 29) Preterm Control (n = 32) Septic (n = 32) *p
pH
P02
Peo2
Bicarbonate
Base excess
7.26 ± 0.06 7.23 ± 0.07*
18.0 ± 5.8 18.1 ± 6.8
53.1 ± 9.1 54.6 ± 8.8
24.1 ± 2.3 22.5 ± 3.1*
-2.9 ± 2.3 -4.5 ± 3.3*
7.26 ± 0.06 7.21 ± 0.07*
18.1 ± 4.8 17.2 ± 5.4
55.2 ± 8.7 57.6 ± 9.2
24.7 ± 2.0 22.1 ± 3.4*
-2.7 ± 2.2 -5.0 ± 3.1*
7.27 ± 0.07 7.25 ± 0.07
18.0 ± 6.7 19.0 ± 8.0
51.1 ± 9.1 51.7 ± 7.5
23.5 ± 2.6 22.9 ± 2.7
-3.3 ± 2.4 -3.9 ± 3.3
< 0.05, versus control.
was no alteration in arterial respiratory gas values when septic and control groups were compared. These data are summarized in Table II. Venous umbilical cord blood gases analysis also showed a significant decrease in pH in the septic cases when compared with controls (P < 0.05). Bicarbonate and base excess were also significantly lower in the septic neonates. Venous respiratory parameters were not altered. These data are summarized in Table III. The data were divided into preterm «37 completed weeks of gestation) and term subsets to evaluate the effect of gestational age on cord blood gases in this population. There were no differences in umbilical arterial or venous blood gas parameters when preterm septic neonates were compared with preterm controls (Tables II and III). However, within the term subset there was a significant decrease in the arterial and venous pH (p < 0.05). In both arterial and venous specimens, the bicarbonate and base excess were significantly lower in the septic neonates. There were no alterations in arterial P0 2 or Peo 2 when term septic neonates were compared with controls. The respiratory component of venous gases were altered, as venous P02 was significantly lower while Peo 2 was significantly higher in the septic group (Tables II and III). FHR tracings were evaluated by the attending physician. The occurrence of any abnormality on FHR tracing was noted in 36% (n = 22) of septic and 33% (n = 20) of control neonates (p > 0.20). Two thirds of these abnormalities consisted offetal tachycardia (> 160
and 0.20). There was a significant increase in the duration of labor in the overall septic versus control group (p < 0.05). When preterm septic and control groups were compared, there was no significant difference in the length of labor (p > 0.20). However, there was a significant prolongation of labor when the term septic subset was compared with term controls (p < 0.05). The observed difference in the duration of labor in the overall group was accounted for by the 45% increase in the length of labor among term septic neonates. We have previously defined fetal acidemia as an arterial cord pH 24 hr Labor> 12 hr Meconium Chorioamnionitis Endometritis
Term
(7) (5) (1) (9)
I
(19)* (15)* (18)* (17)*
17% 17% 0% 7%
(5) (5) (0) (2)
I
(14)* (9) (5)* (6)
(2) (0) (1) (7)
Septic (n = 32)
58.7 ± 8.1*
(5) (6)* (13)* (11)
*P < 0.05, versus control. crease in febrile morbidity among the mothers of septic neonates. Intrapartum fever (>38 0 C) occurred more frequently in mothers of septic newborns when compared with mothers of controls (26% vs 2%, respectively; p < 0.05). This febrile morbidity was evident in both the term and preterm septic subgroups when compared with controls (24% vs 0% and 28% vs 3%, respectively). The clinical diagnosis of chorioamnionitis was made significantly more often in the overall term and in the preterm septic groups when compared with controls (p < 0.05). In addition, there was a significant increase in the incidence of postpartum endomyometritis when mothers of septic newborns were compared with controls (p < 0.05) (Table IV). The duration of ruptured membranes was significantly longer in the overall septic group when compared with controls (p < 0.05). The preterm septic subset accounted for this difference (Table IV). In the overall group, there was a significant increase in the incidence of prolonged rupture of membranes (>24 hours) when septic neonates were compared with controls (p < 0.05). This difference was also attributable to the preterm subset (Table IV). There was a significant increase in the incidence of meconium-stained amniotic fluid in the over-
all septic group when compared with controls (p < 0.05). There was no statistically significant difference in the requirement for oxytocin augmentation in any of the populations studied (overall, 28% (17) vs 39% (24); term, 38% (11) vs 48% (14); preterm, 19% (6) vs 31 % (10); P > 0.20). Similarly, there were no significant differences in regional analgesia usage (overall, 31% (20) vs 39% (24); term, 59% (17) vs 69% (20); preterm, 9% (3) vs 13% (4); P > 0.20). There were also no significant differences in the assignment of a 5-minute Apgar score 0.20). The bacterial pathogens isolated included a variety of microorganisms, two thirds of which were group B streptococcus, Escherichia coli and staphylococcus species. Analysis of the isolates demonstrated a significant increase in the incidence of group B streptococcal infection in term septic when compared with preterm septic neonates (19/29 vs 11/32, respectively, p < 0.05). There were no differences in arterial or venous umbilical cord blood gas parameters between neonates with group B streptococcal infection compared with septic neonates with
616 Meyer et al.
other pathogens. One neonate in each subset had more than one pathogen identified. There were eight neonatal deaths among the septic neonates (three term, five preterm), whereas there was one neonatal death in the control group, in the preterm subgroup.
Comment Delivery of the fetus when intrauterine infection is diagnosed or strongly suspected is currently accepted obstetric practice. 9 . 10 However, no maternal' or neonatal6 benefit to immediate delivery has been demonstrated with intrapartum antibiotic administration. Little is known about the relationship between in utero infection and fetal acid-base status. In this study, we have demonstrated that fetal sepsis is associated with a decrease in the umbilical arterial pH, accompanied by metabolic alterations in the arterial blood gases. These metabolic changes are evidenced by decreases in pH, bicarbonate, and base excess without alterations in arterial Po 2 , or Pea•. Assessment of the influence of gestational age on the arterial blood gas values indicates that the differences are accounted for by the term septic subgroup. Term septic neonates display a significant decrease in arterial pH, bicarbonate, and base excess when compared with matched nonseptic controls. Similar changes are observed in the venous umbilical cord blood gases. In the term septic subgroup, both metabolic and respiratory alterations are observed, as venous P0 2 is significantly decreased while Peo z was significantly elevated. This may be due to an impairment in placental function due to local infection or to a reduction in maternal perfusion in the presence of sepsis. The latter speculation is unlikely to be of significance, as only four of these women had blood culture-proved sepsis and only 30% had a clinical diagnosis of chorioamnionitis. These factors may compound the metabolic changes associated with sepsis in fetal arterial blood gases. Curiously, cord blood gases from the preterm septic newborns are not significantly different from their nonseptic controls, in spite of the greater incidence of chorioamnionitis and prolonged rupture of the membranes in the mothers of this group. We postulate that the lack of alteration in acid-base status is likely due to the significantly shorter duration of labor in this subpopulation. This suggests that fetal sepsis with a short duration of labor does not cause significant changes in fetal acid-base status. We speculate that fetal sepsis plus prolonged labor may cause accelerated deterioration in acid-base status, as demonstrated in the term septic subgroup. To evaluate the etiology of the observed metabolic changes in the term septic neonates, several obstetric factors were assessed. There were no significant differences found in the duration of ruptured mem-
February 1992 Am J Obstet Gynecol
branes, incidence of prolonged ruptured membranes, oxytocin usage for augmentation of labor, fetal meconium passage, or regional analgesia use in the term septic versus nonseptic groups. The only obstetric factor in which term septic neonates differed from controls was in the duration of labor. A significant prolongation in the mean duration of labor was present when term septic neonates were compared with their nonseptic term controls. This significant prolongation of labor held true even when term septic neonates were compared with the preterm septic and preterm control groups. In addition, the incidence of labor duration > 12 hours was significantly higher only in the term septic subset. Eight of the nine neonates with an arterial pH of 12 hours. This lends further credence to the detrimental effect of prolonged labor on acid-base balance in the presence of fetal sepsis. It is unclear from this data whether the in utero sepsis per se had a direct effect on the fetus (i.e., through lactic or other fixed acids) or whether this was a physiologic response to the increased duration of labor in the face of an already stressed septic fetus. Deterioration in fetal acid-base status with increasing duration of labor has been demonstrated previously by other authors. I I However, as there were no differences in the length of labor when preterm septic neonates were compared with preterm controls, it is unlikely that sepsis directly causes an increase in the length of labor. Thus the observed metabolic alterations in arterial blood gas parameters appears to be attributable to the combination of fetal sepsis and increased duration of labor. To ensure that our study population was representative of true fetal infection, we chose to assess only neonates with bacteremia documented by positive blood culture within the first 24 hours of life. These selection criteria may not include all infected neonates but do provide a study population with clearly defined sepsis. The intrapartum maternal administration of antibiotics may have rendered neonatal cultures negative in a bacteremic newborn; the exclusion of these patients from our study may have confounded the interpretation of our results. The spectrum of infectious microorganisms in this study was comparable to those of previous investigators. 12 . 13 Group B streptococcal infection did not disproportionately contribute to the observed differences in arterial blood gas values. The classic maternal markers of chorioamnionitis were poor prognostic indicators of fetal sepsis. Clinical chorioamnionitis was diagnosed in only 27% of mothers of septic neonates. Features often associated with maternal antenatal sepsis, such as prolonged rupture of the membranes, duration of labor> 12 hours, and postpartum endomyometritis, were present significantly more often in mothers of septic neonates than in those of nonseptic
Fetal sepsis
Volume 166 Number 2
controls. Because the majority of mothers ofbacteremic neonates do not have clinical findings of chorioamnionitis, the clinical diagnosis of fetal intrauterine sepsis is extremely difficult. Additionally, the occurrence of fetal sepsis may be independent of clinical maternal infection. Previous studies" 6 assessing the effect of in utero sepsis on umbilical cord blood gas parameters have focused on the diagnosis of chorioamnionitis and its impact on the fetus. Laboratory investigation revealed that the majority of these neonates were not septic. Therefore the umbilical cord blood gas values do not necessarily reflect the effect of sepsis on fetal acid-base balance. The incidence of early neonatal bacterial sepsis in our study was four in 1000 live births, in keeping with previous reports of 1 to 5 in 1000 live births. 12 Early identification of the fetus with in utero sepsis is critical for instituting appropriate antibiotic therapy. However, the diagnosis of fetal sepsis is difficult, with recognizable chorioamnionitis being present in only the minority of cases. A high index of suspicion remains the major diagnostic criterion, with prolonged rupture of membranes and increased duration of labor being major risk factors. The 22-fold increase in the risk of bacteremia at birth in preterm versus term neonates may be due to a number of factors. These include ascending infection as the cause of preterm premature rupture of the membranes and the preterm delivery and the significant increase in the duration of membrane rupture in preterm neonates. In our study fetal sepsis at term was accompanied by a statistically significant reduction in arterial and venous pH. The metabolic alterations in the arterial blood gases are likely due to lactic acidemia secondary to the infection. Although of statistical significance, these differences remain well within the previously reported normal range at our institution 11 and may not be clinically significant. Our data suggest that acidemia at birth is a poor predictor of fetal sepsis, as >85% of septic neonates had a normal cord pH. Additionally, these data confirm the findings of Ramin et al. 15 that neither fetal nor maternal sepsis is an indication for cesarean section. The clinical significance of these findings lies in the increased incidence of fetal acidemia in septic neonates and in the association between increased duration of labor and fetal sepsis. From our study it is unclear whether fetal sepsis can contribute to a prolongation of labor or whether prolonged labor leads to fetal sepsis. However, our findings lend support to the concepts of active management of labor, which have been dem-
617
onstrated to reduce the duration of labor.16 In our institution, application of these principles has decreased the duration of labor without altering fetal acid-base status. 7 There are four major conclusions that can be drawn from our study: (1) Fetal sepsis with short duration of labor does not cause significant changes in acid-base status; (2) fetal sepsis alone does not cause an increase in the duration of labor; (3) there is, however, an association between the duration of labor, fetal sepsis and fetal acidemia at term; (4) preterm neonates have a significantly increased risk of fetal sepsis when compared with term neonates. Further study of the potential clinical effects of fetal sepsis is indicated. REFERENCES I. Andriole VT. Bacterial infection. In: Burrow GN, Ferris
2. 3. 4.
5. 6. 7.
8. 9. 10. 11. 12. 13. 14.
15. 16.
TF, eds. Medical complications during pregnancy. Philadelphia: WB Saunders, 1987:302-32. Peevey KJ, Chalhub EG. Occult group B streptococcal infection: an important cause of intrauterine asphyxia. AM J OBSTET GYNECOL 1983;146:989-90. Naeye RL, Peters Ee. Amniotic fluid infections with intact membranes leading to perinatal death: A prospective study. Pediatrics 1978;61: 171. Vintzileos AM, Campbell WA, Nochimson DJ, et al. The fetal biophysical profile in patients with premature rupture of the membranes-an early predictor of fetal infection. AM J OBSTET GYNECOL 1985; 152:510-6. Gilstrap LC, Leveno KJ, Cox SM, et al. Intrapartum treatment of acute chorioamnionitis: impact on neonatal sepsis. AM J OBSTET GYNECOL 1988; 159:579-83. Hauth JC, Gilstrap LC, Hankins GDV, et al. Term maternal and neonatal complications of acute chorioamnionitis. Obstet Gynecol 1985;66:59-62. Thorp .lA, Boylan PC, Parisi VM, et al. Effects of highdose oxytocin augmentation on umbilical cord blood gas values in primigravid women. AM J OBSTF.T GYNECOL 1988; 159:670-5. Freeman RK, Garite TJ. Fetal heart rate monitoring. ed I. Baltimore: Williams & Wilkins, 1981 :63-83. Schwartz RH, Fruiterman JP. Life-threatening infections in pregnancy. Clin Obstet Gynecol 1976; 19:561. MacVicar J. In: Charles D, Finland M, eds. Chorioamnionitis, obstetric and perinatal infections. Philadelphia: Lea & Febiger, 1973:491-5. Kubli FW. Influence of labor on fetal acid-base balance. Clin Obstet Gynecol 1968; 11: 168-91. St Geme JW, Polin RA. Neonatal sepsis. Progress in diagnosis and management. Drugs 1988;36:784-800. Siegel JD, McCracken GH. Sepsis neonatorum. N Engl J Med 1981;304:642-7. Boylan PC, Parisi VM. Fetal acid-base balance. In: Creasy RK, Resnick R, eds. Maternal-fetal medicine: principles and practice. ed 2. Philadelphia: WB Saunders, 1989:36274. Ramin S, Maberry M, Gilstrap L, et al. Fetal acidemia in pregnancies with chorioamnionitis. Obstet Gynecol 1990;76:351-4. O'Driscoll K, Foley M, MacDonald D. Active management of labor as an alternative to cesarean section for dystocia. Obstet Gynecol 1984;63:485-90.
